Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
Department of Mechanical and Industrial Engineering
ME TR JPL 952593 July 1970
AN APPROXIMATE THEORY OF IMPERFECT MODELING
WITH APPLICATION TO
THE THERMAL MODELING OF SPACECRAFTS
by
B T Chao and M N Huang
FINAL REPORT (
to
VJET PROPULSION LABORATORY
California Institute of Technology -
Contract No 952595
University of Illinois at Urbana-Champaign
Urbana Illinois 61801
Y [CES20NU)TIHR
O (PAGES) (ODE)
(NASA CR OR TMX OR AD NUMBER) (CATEGORY) (
N70-20517 -20548
SPACE PROCESSING AND MANUFACTURING
George C Marshall Space Flight Center Marshall Space Flight Center Alabama
5 February 1970
0
g 0
00-0- 001 0
Distributed to foster serve and promote the nations economic developmentand technological advancement
NATIONAL TECHNICAL INFORMATION SERVICE
US DEPARTMENT OF COMMERCE
0
This document has been approved for public release and sale
2N TECHNICAL REPORT STANOARD TITLE PAGE
O GN ZAT N t0-53993
RPRT 1y9N AS A TE CHN IC A L TLTE AND SUBTITLE ME O A D M4 PROCESSING AND MANUFACTURING February 5 1970MEMORANDUM SPACE ORGANIZATION6 PERFORMING C(DE
ORGANIZATION aNASA TM X-53993 7 AUTHOR(S) S PERFORMING REPoRr
9 PERFORMINGOROANIZATIONSAME AND ADDRESS I0 WORK UNIT NO
Manufacturing Engineering Laboratory 11 OR GRANT CONTRACT NO George C Marshall Space Flight Center
DMarshall Space Flight Center Alabama 35812 i TYPE OF REPORI amp PERIO COVERED
12 SPONSORINGAGENCYNlAMEAND ADDRESS Technical Memorandum
14 SPONSORING AGENCY CODE
IS SUPPLEMENTARY NOTES
I ABSTRACT
This publication contains the presentations from the Space Processing and ManufacturingSPACE PROCESSING AND MANUFACTURING necNg held at MISFC Huntsville Alabama October 21 and 22 1969
Contatas consist of the opening keynote address by Dr von Braun Director Marshall Manufactaring Engineering Laboratory Space Flight Center thirty-one presentations and the closing remarks by Dr M P Sebel
Manufacturing Engineering Laboratory MSFC In addition there is a complete listingFebruary 5 1970 of the speakers showing their affiliation mailing addresses phone numbers biographical
N70 51 ktles and alist of attendees
I O o~ 0t
NASA
George C Marshall Space Flight Center nama I7 K WORDS IDoOISTRIBUTION STATEMENTM atsball Space tlght
NA1IONatTECHN- IC 1I SECURITYCLASSIF(2I021 120 SECURITYCLASSIF0R1 1N F PAGES 22 PRICE
INIORMATION SERVICE ___________________________ _________
Department of Mechanical N7(-3q4 A and Industrial Engineering
ME TR JPL 952593 July 1970
AN APPROXIMATE THEORY OF IMPERFECT MODELING
WITH APPLICATION TO
THE THERMAL MODELING OF SPACECRAFTS
by
BT Chao and M N Huang
FINAL REPORT
to
JET PROPULSION LABORATORY
Contract No 952593
This work was performed for the Jet Propulsion Laboratory Califorshynia Institute of Technology as sponsored by the National Aeronautics and Space Administration under Contract NAS7-100
University of Illinois at Urbana-Champaign
Urbana Illinois 61801
REPRODUCEDBYNATIONAL TECHNICAL INFORMATION SERVICE
US DEPARTMENT -
OF COMMERCE_WRINGFIELD_VA22161
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED
FROM THE BEST COPY FURNISHED US BY
THE SPONSORING AGENCY ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE IT IS BEING RELEASED
IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE
TECHNICAL CONTENT STATEMENT
This report contains information prepared by the Department of
Mechanical and Industrial Engineering University of Illinois at Urbana-
Champaign under JPL subcontract Its-content is not necessarily enshy
dorsed by the Jet Propulsion Laboratory California Institute of Techshy
nology or the National Aeronautics and Space Administration
NEW TECHNOLOGY
No reportable items of new technology are identified
ii
ACKNOWLEDGMENTS
During the course of the investigation Mr W A Hagemeyer
JPL Technical Representative provided property data and assistance
in many ways without which this work would not have been possible
Professor J S de Paiva Netto of Instituto Tecnologico de Aeronaushy
tica Brazil and formerly of the Heat Transfer Laboratory of the
Department conducted preliminary studies of the theory To both
we offer our sincere thanks
The assistance of Miss Dianne Merridith in typing and Mr Don
Anderson in preparing the figures and of Mr George Morris and Mrs
June Kempka in supervising the overall preparation of this report
is greatly appreciated
iii
ABSTRACT
A theory has been proposed for conducring model experimentation
which permits the use of imperfect models or test conditions that
do not conform strictly to similitude requirements It is based on
a new concept of representing the error states in a multi-dimensional
Euclidean space when errors in the modeling parameters are small
A consideration of the error path in the hyperspace leads to expresshy
sions from which the global effect of these errors may be evaluated
approximately Conditions under which these expressions would yield
satisfactory results are given
To test the usefulness of the theory a computer experiment has
been carried out for the prediction of the transient and steady state
thermal behavior of a hypothetical spacecraft using perfect as well
as imperfect models The hypothetical spacecraft has major radiashy
tive and conductive heat flow paths that crudely simulate those of
the 64 Mariner family of space vehicles Errors in the conductance
and capacitance parameters of up to 25 percent and in surface emitshy
tances of up to 20 percent existed in the imperfect models Extenshy
sive temperature data were obtained for the various components of
the spacecraft when it was subjected to a sudden heating due to its
exposure to the suns radiation and the power dissipation within the
bus followed by its attaining the steady state and the subsequent
cooling as a result of removing the suns heat Data were also acshy
quired under the condition of a cyclically varying thermal environment
iv
Generally speaking errors of up to 10 20 and 300 R originally presshy
ent in the data of the imperfect modelswere reduced to less than
31R after correction according to the proposed theory
V
NOMENCLATURE
A = area ft
-aii --shy oefficient-matrix defined in--(32-15)
B = diffuse radiosity Btuhr-ft
b = inverse ofoaj - defined in (3217)
C = volumetric heat capacity Btuft -0R
C = weighting factor for 6 defined in (215)
d = thickness ft
E = exchange factor
F = shape factor
H = irradiation Btuhr-ft2
k = thermal conductivity Btuhr-ft-degR
Q = heat flow rate Btuhr
q = heat flux Btuhr-ft2
S = solar constant Btuhr-ft2
s = are length along error path
T = absolute temperature T o = absolute reference temperature
t = time hr
a = surface absorptance
= error ratio defined in (219)
= error ratio defined in (2114)
= error in modeling parameter
6 = Kronecker delta function
= surface emittance
vi
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
N70-20517 -20548
SPACE PROCESSING AND MANUFACTURING
George C Marshall Space Flight Center Marshall Space Flight Center Alabama
5 February 1970
0
g 0
00-0- 001 0
Distributed to foster serve and promote the nations economic developmentand technological advancement
NATIONAL TECHNICAL INFORMATION SERVICE
US DEPARTMENT OF COMMERCE
0
This document has been approved for public release and sale
2N TECHNICAL REPORT STANOARD TITLE PAGE
O GN ZAT N t0-53993
RPRT 1y9N AS A TE CHN IC A L TLTE AND SUBTITLE ME O A D M4 PROCESSING AND MANUFACTURING February 5 1970MEMORANDUM SPACE ORGANIZATION6 PERFORMING C(DE
ORGANIZATION aNASA TM X-53993 7 AUTHOR(S) S PERFORMING REPoRr
9 PERFORMINGOROANIZATIONSAME AND ADDRESS I0 WORK UNIT NO
Manufacturing Engineering Laboratory 11 OR GRANT CONTRACT NO George C Marshall Space Flight Center
DMarshall Space Flight Center Alabama 35812 i TYPE OF REPORI amp PERIO COVERED
12 SPONSORINGAGENCYNlAMEAND ADDRESS Technical Memorandum
14 SPONSORING AGENCY CODE
IS SUPPLEMENTARY NOTES
I ABSTRACT
This publication contains the presentations from the Space Processing and ManufacturingSPACE PROCESSING AND MANUFACTURING necNg held at MISFC Huntsville Alabama October 21 and 22 1969
Contatas consist of the opening keynote address by Dr von Braun Director Marshall Manufactaring Engineering Laboratory Space Flight Center thirty-one presentations and the closing remarks by Dr M P Sebel
Manufacturing Engineering Laboratory MSFC In addition there is a complete listingFebruary 5 1970 of the speakers showing their affiliation mailing addresses phone numbers biographical
N70 51 ktles and alist of attendees
I O o~ 0t
NASA
George C Marshall Space Flight Center nama I7 K WORDS IDoOISTRIBUTION STATEMENTM atsball Space tlght
NA1IONatTECHN- IC 1I SECURITYCLASSIF(2I021 120 SECURITYCLASSIF0R1 1N F PAGES 22 PRICE
INIORMATION SERVICE ___________________________ _________
Department of Mechanical N7(-3q4 A and Industrial Engineering
ME TR JPL 952593 July 1970
AN APPROXIMATE THEORY OF IMPERFECT MODELING
WITH APPLICATION TO
THE THERMAL MODELING OF SPACECRAFTS
by
BT Chao and M N Huang
FINAL REPORT
to
JET PROPULSION LABORATORY
Contract No 952593
This work was performed for the Jet Propulsion Laboratory Califorshynia Institute of Technology as sponsored by the National Aeronautics and Space Administration under Contract NAS7-100
University of Illinois at Urbana-Champaign
Urbana Illinois 61801
REPRODUCEDBYNATIONAL TECHNICAL INFORMATION SERVICE
US DEPARTMENT -
OF COMMERCE_WRINGFIELD_VA22161
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED
FROM THE BEST COPY FURNISHED US BY
THE SPONSORING AGENCY ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE IT IS BEING RELEASED
IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE
TECHNICAL CONTENT STATEMENT
This report contains information prepared by the Department of
Mechanical and Industrial Engineering University of Illinois at Urbana-
Champaign under JPL subcontract Its-content is not necessarily enshy
dorsed by the Jet Propulsion Laboratory California Institute of Techshy
nology or the National Aeronautics and Space Administration
NEW TECHNOLOGY
No reportable items of new technology are identified
ii
ACKNOWLEDGMENTS
During the course of the investigation Mr W A Hagemeyer
JPL Technical Representative provided property data and assistance
in many ways without which this work would not have been possible
Professor J S de Paiva Netto of Instituto Tecnologico de Aeronaushy
tica Brazil and formerly of the Heat Transfer Laboratory of the
Department conducted preliminary studies of the theory To both
we offer our sincere thanks
The assistance of Miss Dianne Merridith in typing and Mr Don
Anderson in preparing the figures and of Mr George Morris and Mrs
June Kempka in supervising the overall preparation of this report
is greatly appreciated
iii
ABSTRACT
A theory has been proposed for conducring model experimentation
which permits the use of imperfect models or test conditions that
do not conform strictly to similitude requirements It is based on
a new concept of representing the error states in a multi-dimensional
Euclidean space when errors in the modeling parameters are small
A consideration of the error path in the hyperspace leads to expresshy
sions from which the global effect of these errors may be evaluated
approximately Conditions under which these expressions would yield
satisfactory results are given
To test the usefulness of the theory a computer experiment has
been carried out for the prediction of the transient and steady state
thermal behavior of a hypothetical spacecraft using perfect as well
as imperfect models The hypothetical spacecraft has major radiashy
tive and conductive heat flow paths that crudely simulate those of
the 64 Mariner family of space vehicles Errors in the conductance
and capacitance parameters of up to 25 percent and in surface emitshy
tances of up to 20 percent existed in the imperfect models Extenshy
sive temperature data were obtained for the various components of
the spacecraft when it was subjected to a sudden heating due to its
exposure to the suns radiation and the power dissipation within the
bus followed by its attaining the steady state and the subsequent
cooling as a result of removing the suns heat Data were also acshy
quired under the condition of a cyclically varying thermal environment
iv
Generally speaking errors of up to 10 20 and 300 R originally presshy
ent in the data of the imperfect modelswere reduced to less than
31R after correction according to the proposed theory
V
NOMENCLATURE
A = area ft
-aii --shy oefficient-matrix defined in--(32-15)
B = diffuse radiosity Btuhr-ft
b = inverse ofoaj - defined in (3217)
C = volumetric heat capacity Btuft -0R
C = weighting factor for 6 defined in (215)
d = thickness ft
E = exchange factor
F = shape factor
H = irradiation Btuhr-ft2
k = thermal conductivity Btuhr-ft-degR
Q = heat flow rate Btuhr
q = heat flux Btuhr-ft2
S = solar constant Btuhr-ft2
s = are length along error path
T = absolute temperature T o = absolute reference temperature
t = time hr
a = surface absorptance
= error ratio defined in (219)
= error ratio defined in (2114)
= error in modeling parameter
6 = Kronecker delta function
= surface emittance
vi
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
2N TECHNICAL REPORT STANOARD TITLE PAGE
O GN ZAT N t0-53993
RPRT 1y9N AS A TE CHN IC A L TLTE AND SUBTITLE ME O A D M4 PROCESSING AND MANUFACTURING February 5 1970MEMORANDUM SPACE ORGANIZATION6 PERFORMING C(DE
ORGANIZATION aNASA TM X-53993 7 AUTHOR(S) S PERFORMING REPoRr
9 PERFORMINGOROANIZATIONSAME AND ADDRESS I0 WORK UNIT NO
Manufacturing Engineering Laboratory 11 OR GRANT CONTRACT NO George C Marshall Space Flight Center
DMarshall Space Flight Center Alabama 35812 i TYPE OF REPORI amp PERIO COVERED
12 SPONSORINGAGENCYNlAMEAND ADDRESS Technical Memorandum
14 SPONSORING AGENCY CODE
IS SUPPLEMENTARY NOTES
I ABSTRACT
This publication contains the presentations from the Space Processing and ManufacturingSPACE PROCESSING AND MANUFACTURING necNg held at MISFC Huntsville Alabama October 21 and 22 1969
Contatas consist of the opening keynote address by Dr von Braun Director Marshall Manufactaring Engineering Laboratory Space Flight Center thirty-one presentations and the closing remarks by Dr M P Sebel
Manufacturing Engineering Laboratory MSFC In addition there is a complete listingFebruary 5 1970 of the speakers showing their affiliation mailing addresses phone numbers biographical
N70 51 ktles and alist of attendees
I O o~ 0t
NASA
George C Marshall Space Flight Center nama I7 K WORDS IDoOISTRIBUTION STATEMENTM atsball Space tlght
NA1IONatTECHN- IC 1I SECURITYCLASSIF(2I021 120 SECURITYCLASSIF0R1 1N F PAGES 22 PRICE
INIORMATION SERVICE ___________________________ _________
Department of Mechanical N7(-3q4 A and Industrial Engineering
ME TR JPL 952593 July 1970
AN APPROXIMATE THEORY OF IMPERFECT MODELING
WITH APPLICATION TO
THE THERMAL MODELING OF SPACECRAFTS
by
BT Chao and M N Huang
FINAL REPORT
to
JET PROPULSION LABORATORY
Contract No 952593
This work was performed for the Jet Propulsion Laboratory Califorshynia Institute of Technology as sponsored by the National Aeronautics and Space Administration under Contract NAS7-100
University of Illinois at Urbana-Champaign
Urbana Illinois 61801
REPRODUCEDBYNATIONAL TECHNICAL INFORMATION SERVICE
US DEPARTMENT -
OF COMMERCE_WRINGFIELD_VA22161
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED
FROM THE BEST COPY FURNISHED US BY
THE SPONSORING AGENCY ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE IT IS BEING RELEASED
IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE
TECHNICAL CONTENT STATEMENT
This report contains information prepared by the Department of
Mechanical and Industrial Engineering University of Illinois at Urbana-
Champaign under JPL subcontract Its-content is not necessarily enshy
dorsed by the Jet Propulsion Laboratory California Institute of Techshy
nology or the National Aeronautics and Space Administration
NEW TECHNOLOGY
No reportable items of new technology are identified
ii
ACKNOWLEDGMENTS
During the course of the investigation Mr W A Hagemeyer
JPL Technical Representative provided property data and assistance
in many ways without which this work would not have been possible
Professor J S de Paiva Netto of Instituto Tecnologico de Aeronaushy
tica Brazil and formerly of the Heat Transfer Laboratory of the
Department conducted preliminary studies of the theory To both
we offer our sincere thanks
The assistance of Miss Dianne Merridith in typing and Mr Don
Anderson in preparing the figures and of Mr George Morris and Mrs
June Kempka in supervising the overall preparation of this report
is greatly appreciated
iii
ABSTRACT
A theory has been proposed for conducring model experimentation
which permits the use of imperfect models or test conditions that
do not conform strictly to similitude requirements It is based on
a new concept of representing the error states in a multi-dimensional
Euclidean space when errors in the modeling parameters are small
A consideration of the error path in the hyperspace leads to expresshy
sions from which the global effect of these errors may be evaluated
approximately Conditions under which these expressions would yield
satisfactory results are given
To test the usefulness of the theory a computer experiment has
been carried out for the prediction of the transient and steady state
thermal behavior of a hypothetical spacecraft using perfect as well
as imperfect models The hypothetical spacecraft has major radiashy
tive and conductive heat flow paths that crudely simulate those of
the 64 Mariner family of space vehicles Errors in the conductance
and capacitance parameters of up to 25 percent and in surface emitshy
tances of up to 20 percent existed in the imperfect models Extenshy
sive temperature data were obtained for the various components of
the spacecraft when it was subjected to a sudden heating due to its
exposure to the suns radiation and the power dissipation within the
bus followed by its attaining the steady state and the subsequent
cooling as a result of removing the suns heat Data were also acshy
quired under the condition of a cyclically varying thermal environment
iv
Generally speaking errors of up to 10 20 and 300 R originally presshy
ent in the data of the imperfect modelswere reduced to less than
31R after correction according to the proposed theory
V
NOMENCLATURE
A = area ft
-aii --shy oefficient-matrix defined in--(32-15)
B = diffuse radiosity Btuhr-ft
b = inverse ofoaj - defined in (3217)
C = volumetric heat capacity Btuft -0R
C = weighting factor for 6 defined in (215)
d = thickness ft
E = exchange factor
F = shape factor
H = irradiation Btuhr-ft2
k = thermal conductivity Btuhr-ft-degR
Q = heat flow rate Btuhr
q = heat flux Btuhr-ft2
S = solar constant Btuhr-ft2
s = are length along error path
T = absolute temperature T o = absolute reference temperature
t = time hr
a = surface absorptance
= error ratio defined in (219)
= error ratio defined in (2114)
= error in modeling parameter
6 = Kronecker delta function
= surface emittance
vi
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
Department of Mechanical N7(-3q4 A and Industrial Engineering
ME TR JPL 952593 July 1970
AN APPROXIMATE THEORY OF IMPERFECT MODELING
WITH APPLICATION TO
THE THERMAL MODELING OF SPACECRAFTS
by
BT Chao and M N Huang
FINAL REPORT
to
JET PROPULSION LABORATORY
Contract No 952593
This work was performed for the Jet Propulsion Laboratory Califorshynia Institute of Technology as sponsored by the National Aeronautics and Space Administration under Contract NAS7-100
University of Illinois at Urbana-Champaign
Urbana Illinois 61801
REPRODUCEDBYNATIONAL TECHNICAL INFORMATION SERVICE
US DEPARTMENT -
OF COMMERCE_WRINGFIELD_VA22161
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED
FROM THE BEST COPY FURNISHED US BY
THE SPONSORING AGENCY ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE IT IS BEING RELEASED
IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE
TECHNICAL CONTENT STATEMENT
This report contains information prepared by the Department of
Mechanical and Industrial Engineering University of Illinois at Urbana-
Champaign under JPL subcontract Its-content is not necessarily enshy
dorsed by the Jet Propulsion Laboratory California Institute of Techshy
nology or the National Aeronautics and Space Administration
NEW TECHNOLOGY
No reportable items of new technology are identified
ii
ACKNOWLEDGMENTS
During the course of the investigation Mr W A Hagemeyer
JPL Technical Representative provided property data and assistance
in many ways without which this work would not have been possible
Professor J S de Paiva Netto of Instituto Tecnologico de Aeronaushy
tica Brazil and formerly of the Heat Transfer Laboratory of the
Department conducted preliminary studies of the theory To both
we offer our sincere thanks
The assistance of Miss Dianne Merridith in typing and Mr Don
Anderson in preparing the figures and of Mr George Morris and Mrs
June Kempka in supervising the overall preparation of this report
is greatly appreciated
iii
ABSTRACT
A theory has been proposed for conducring model experimentation
which permits the use of imperfect models or test conditions that
do not conform strictly to similitude requirements It is based on
a new concept of representing the error states in a multi-dimensional
Euclidean space when errors in the modeling parameters are small
A consideration of the error path in the hyperspace leads to expresshy
sions from which the global effect of these errors may be evaluated
approximately Conditions under which these expressions would yield
satisfactory results are given
To test the usefulness of the theory a computer experiment has
been carried out for the prediction of the transient and steady state
thermal behavior of a hypothetical spacecraft using perfect as well
as imperfect models The hypothetical spacecraft has major radiashy
tive and conductive heat flow paths that crudely simulate those of
the 64 Mariner family of space vehicles Errors in the conductance
and capacitance parameters of up to 25 percent and in surface emitshy
tances of up to 20 percent existed in the imperfect models Extenshy
sive temperature data were obtained for the various components of
the spacecraft when it was subjected to a sudden heating due to its
exposure to the suns radiation and the power dissipation within the
bus followed by its attaining the steady state and the subsequent
cooling as a result of removing the suns heat Data were also acshy
quired under the condition of a cyclically varying thermal environment
iv
Generally speaking errors of up to 10 20 and 300 R originally presshy
ent in the data of the imperfect modelswere reduced to less than
31R after correction according to the proposed theory
V
NOMENCLATURE
A = area ft
-aii --shy oefficient-matrix defined in--(32-15)
B = diffuse radiosity Btuhr-ft
b = inverse ofoaj - defined in (3217)
C = volumetric heat capacity Btuft -0R
C = weighting factor for 6 defined in (215)
d = thickness ft
E = exchange factor
F = shape factor
H = irradiation Btuhr-ft2
k = thermal conductivity Btuhr-ft-degR
Q = heat flow rate Btuhr
q = heat flux Btuhr-ft2
S = solar constant Btuhr-ft2
s = are length along error path
T = absolute temperature T o = absolute reference temperature
t = time hr
a = surface absorptance
= error ratio defined in (219)
= error ratio defined in (2114)
= error in modeling parameter
6 = Kronecker delta function
= surface emittance
vi
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
NOTICE
THIS DOCUMENT HAS BEEN REPRODUCED
FROM THE BEST COPY FURNISHED US BY
THE SPONSORING AGENCY ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE IT IS BEING RELEASED
IN THE INTEREST OF MAKING AVAILABLE
AS MUCH INFORMATION AS POSSIBLE
TECHNICAL CONTENT STATEMENT
This report contains information prepared by the Department of
Mechanical and Industrial Engineering University of Illinois at Urbana-
Champaign under JPL subcontract Its-content is not necessarily enshy
dorsed by the Jet Propulsion Laboratory California Institute of Techshy
nology or the National Aeronautics and Space Administration
NEW TECHNOLOGY
No reportable items of new technology are identified
ii
ACKNOWLEDGMENTS
During the course of the investigation Mr W A Hagemeyer
JPL Technical Representative provided property data and assistance
in many ways without which this work would not have been possible
Professor J S de Paiva Netto of Instituto Tecnologico de Aeronaushy
tica Brazil and formerly of the Heat Transfer Laboratory of the
Department conducted preliminary studies of the theory To both
we offer our sincere thanks
The assistance of Miss Dianne Merridith in typing and Mr Don
Anderson in preparing the figures and of Mr George Morris and Mrs
June Kempka in supervising the overall preparation of this report
is greatly appreciated
iii
ABSTRACT
A theory has been proposed for conducring model experimentation
which permits the use of imperfect models or test conditions that
do not conform strictly to similitude requirements It is based on
a new concept of representing the error states in a multi-dimensional
Euclidean space when errors in the modeling parameters are small
A consideration of the error path in the hyperspace leads to expresshy
sions from which the global effect of these errors may be evaluated
approximately Conditions under which these expressions would yield
satisfactory results are given
To test the usefulness of the theory a computer experiment has
been carried out for the prediction of the transient and steady state
thermal behavior of a hypothetical spacecraft using perfect as well
as imperfect models The hypothetical spacecraft has major radiashy
tive and conductive heat flow paths that crudely simulate those of
the 64 Mariner family of space vehicles Errors in the conductance
and capacitance parameters of up to 25 percent and in surface emitshy
tances of up to 20 percent existed in the imperfect models Extenshy
sive temperature data were obtained for the various components of
the spacecraft when it was subjected to a sudden heating due to its
exposure to the suns radiation and the power dissipation within the
bus followed by its attaining the steady state and the subsequent
cooling as a result of removing the suns heat Data were also acshy
quired under the condition of a cyclically varying thermal environment
iv
Generally speaking errors of up to 10 20 and 300 R originally presshy
ent in the data of the imperfect modelswere reduced to less than
31R after correction according to the proposed theory
V
NOMENCLATURE
A = area ft
-aii --shy oefficient-matrix defined in--(32-15)
B = diffuse radiosity Btuhr-ft
b = inverse ofoaj - defined in (3217)
C = volumetric heat capacity Btuft -0R
C = weighting factor for 6 defined in (215)
d = thickness ft
E = exchange factor
F = shape factor
H = irradiation Btuhr-ft2
k = thermal conductivity Btuhr-ft-degR
Q = heat flow rate Btuhr
q = heat flux Btuhr-ft2
S = solar constant Btuhr-ft2
s = are length along error path
T = absolute temperature T o = absolute reference temperature
t = time hr
a = surface absorptance
= error ratio defined in (219)
= error ratio defined in (2114)
= error in modeling parameter
6 = Kronecker delta function
= surface emittance
vi
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
TECHNICAL CONTENT STATEMENT
This report contains information prepared by the Department of
Mechanical and Industrial Engineering University of Illinois at Urbana-
Champaign under JPL subcontract Its-content is not necessarily enshy
dorsed by the Jet Propulsion Laboratory California Institute of Techshy
nology or the National Aeronautics and Space Administration
NEW TECHNOLOGY
No reportable items of new technology are identified
ii
ACKNOWLEDGMENTS
During the course of the investigation Mr W A Hagemeyer
JPL Technical Representative provided property data and assistance
in many ways without which this work would not have been possible
Professor J S de Paiva Netto of Instituto Tecnologico de Aeronaushy
tica Brazil and formerly of the Heat Transfer Laboratory of the
Department conducted preliminary studies of the theory To both
we offer our sincere thanks
The assistance of Miss Dianne Merridith in typing and Mr Don
Anderson in preparing the figures and of Mr George Morris and Mrs
June Kempka in supervising the overall preparation of this report
is greatly appreciated
iii
ABSTRACT
A theory has been proposed for conducring model experimentation
which permits the use of imperfect models or test conditions that
do not conform strictly to similitude requirements It is based on
a new concept of representing the error states in a multi-dimensional
Euclidean space when errors in the modeling parameters are small
A consideration of the error path in the hyperspace leads to expresshy
sions from which the global effect of these errors may be evaluated
approximately Conditions under which these expressions would yield
satisfactory results are given
To test the usefulness of the theory a computer experiment has
been carried out for the prediction of the transient and steady state
thermal behavior of a hypothetical spacecraft using perfect as well
as imperfect models The hypothetical spacecraft has major radiashy
tive and conductive heat flow paths that crudely simulate those of
the 64 Mariner family of space vehicles Errors in the conductance
and capacitance parameters of up to 25 percent and in surface emitshy
tances of up to 20 percent existed in the imperfect models Extenshy
sive temperature data were obtained for the various components of
the spacecraft when it was subjected to a sudden heating due to its
exposure to the suns radiation and the power dissipation within the
bus followed by its attaining the steady state and the subsequent
cooling as a result of removing the suns heat Data were also acshy
quired under the condition of a cyclically varying thermal environment
iv
Generally speaking errors of up to 10 20 and 300 R originally presshy
ent in the data of the imperfect modelswere reduced to less than
31R after correction according to the proposed theory
V
NOMENCLATURE
A = area ft
-aii --shy oefficient-matrix defined in--(32-15)
B = diffuse radiosity Btuhr-ft
b = inverse ofoaj - defined in (3217)
C = volumetric heat capacity Btuft -0R
C = weighting factor for 6 defined in (215)
d = thickness ft
E = exchange factor
F = shape factor
H = irradiation Btuhr-ft2
k = thermal conductivity Btuhr-ft-degR
Q = heat flow rate Btuhr
q = heat flux Btuhr-ft2
S = solar constant Btuhr-ft2
s = are length along error path
T = absolute temperature T o = absolute reference temperature
t = time hr
a = surface absorptance
= error ratio defined in (219)
= error ratio defined in (2114)
= error in modeling parameter
6 = Kronecker delta function
= surface emittance
vi
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
ACKNOWLEDGMENTS
During the course of the investigation Mr W A Hagemeyer
JPL Technical Representative provided property data and assistance
in many ways without which this work would not have been possible
Professor J S de Paiva Netto of Instituto Tecnologico de Aeronaushy
tica Brazil and formerly of the Heat Transfer Laboratory of the
Department conducted preliminary studies of the theory To both
we offer our sincere thanks
The assistance of Miss Dianne Merridith in typing and Mr Don
Anderson in preparing the figures and of Mr George Morris and Mrs
June Kempka in supervising the overall preparation of this report
is greatly appreciated
iii
ABSTRACT
A theory has been proposed for conducring model experimentation
which permits the use of imperfect models or test conditions that
do not conform strictly to similitude requirements It is based on
a new concept of representing the error states in a multi-dimensional
Euclidean space when errors in the modeling parameters are small
A consideration of the error path in the hyperspace leads to expresshy
sions from which the global effect of these errors may be evaluated
approximately Conditions under which these expressions would yield
satisfactory results are given
To test the usefulness of the theory a computer experiment has
been carried out for the prediction of the transient and steady state
thermal behavior of a hypothetical spacecraft using perfect as well
as imperfect models The hypothetical spacecraft has major radiashy
tive and conductive heat flow paths that crudely simulate those of
the 64 Mariner family of space vehicles Errors in the conductance
and capacitance parameters of up to 25 percent and in surface emitshy
tances of up to 20 percent existed in the imperfect models Extenshy
sive temperature data were obtained for the various components of
the spacecraft when it was subjected to a sudden heating due to its
exposure to the suns radiation and the power dissipation within the
bus followed by its attaining the steady state and the subsequent
cooling as a result of removing the suns heat Data were also acshy
quired under the condition of a cyclically varying thermal environment
iv
Generally speaking errors of up to 10 20 and 300 R originally presshy
ent in the data of the imperfect modelswere reduced to less than
31R after correction according to the proposed theory
V
NOMENCLATURE
A = area ft
-aii --shy oefficient-matrix defined in--(32-15)
B = diffuse radiosity Btuhr-ft
b = inverse ofoaj - defined in (3217)
C = volumetric heat capacity Btuft -0R
C = weighting factor for 6 defined in (215)
d = thickness ft
E = exchange factor
F = shape factor
H = irradiation Btuhr-ft2
k = thermal conductivity Btuhr-ft-degR
Q = heat flow rate Btuhr
q = heat flux Btuhr-ft2
S = solar constant Btuhr-ft2
s = are length along error path
T = absolute temperature T o = absolute reference temperature
t = time hr
a = surface absorptance
= error ratio defined in (219)
= error ratio defined in (2114)
= error in modeling parameter
6 = Kronecker delta function
= surface emittance
vi
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
ABSTRACT
A theory has been proposed for conducring model experimentation
which permits the use of imperfect models or test conditions that
do not conform strictly to similitude requirements It is based on
a new concept of representing the error states in a multi-dimensional
Euclidean space when errors in the modeling parameters are small
A consideration of the error path in the hyperspace leads to expresshy
sions from which the global effect of these errors may be evaluated
approximately Conditions under which these expressions would yield
satisfactory results are given
To test the usefulness of the theory a computer experiment has
been carried out for the prediction of the transient and steady state
thermal behavior of a hypothetical spacecraft using perfect as well
as imperfect models The hypothetical spacecraft has major radiashy
tive and conductive heat flow paths that crudely simulate those of
the 64 Mariner family of space vehicles Errors in the conductance
and capacitance parameters of up to 25 percent and in surface emitshy
tances of up to 20 percent existed in the imperfect models Extenshy
sive temperature data were obtained for the various components of
the spacecraft when it was subjected to a sudden heating due to its
exposure to the suns radiation and the power dissipation within the
bus followed by its attaining the steady state and the subsequent
cooling as a result of removing the suns heat Data were also acshy
quired under the condition of a cyclically varying thermal environment
iv
Generally speaking errors of up to 10 20 and 300 R originally presshy
ent in the data of the imperfect modelswere reduced to less than
31R after correction according to the proposed theory
V
NOMENCLATURE
A = area ft
-aii --shy oefficient-matrix defined in--(32-15)
B = diffuse radiosity Btuhr-ft
b = inverse ofoaj - defined in (3217)
C = volumetric heat capacity Btuft -0R
C = weighting factor for 6 defined in (215)
d = thickness ft
E = exchange factor
F = shape factor
H = irradiation Btuhr-ft2
k = thermal conductivity Btuhr-ft-degR
Q = heat flow rate Btuhr
q = heat flux Btuhr-ft2
S = solar constant Btuhr-ft2
s = are length along error path
T = absolute temperature T o = absolute reference temperature
t = time hr
a = surface absorptance
= error ratio defined in (219)
= error ratio defined in (2114)
= error in modeling parameter
6 = Kronecker delta function
= surface emittance
vi
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
Generally speaking errors of up to 10 20 and 300 R originally presshy
ent in the data of the imperfect modelswere reduced to less than
31R after correction according to the proposed theory
V
NOMENCLATURE
A = area ft
-aii --shy oefficient-matrix defined in--(32-15)
B = diffuse radiosity Btuhr-ft
b = inverse ofoaj - defined in (3217)
C = volumetric heat capacity Btuft -0R
C = weighting factor for 6 defined in (215)
d = thickness ft
E = exchange factor
F = shape factor
H = irradiation Btuhr-ft2
k = thermal conductivity Btuhr-ft-degR
Q = heat flow rate Btuhr
q = heat flux Btuhr-ft2
S = solar constant Btuhr-ft2
s = are length along error path
T = absolute temperature T o = absolute reference temperature
t = time hr
a = surface absorptance
= error ratio defined in (219)
= error ratio defined in (2114)
= error in modeling parameter
6 = Kronecker delta function
= surface emittance
vi
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
NOMENCLATURE
A = area ft
-aii --shy oefficient-matrix defined in--(32-15)
B = diffuse radiosity Btuhr-ft
b = inverse ofoaj - defined in (3217)
C = volumetric heat capacity Btuft -0R
C = weighting factor for 6 defined in (215)
d = thickness ft
E = exchange factor
F = shape factor
H = irradiation Btuhr-ft2
k = thermal conductivity Btuhr-ft-degR
Q = heat flow rate Btuhr
q = heat flux Btuhr-ft2
S = solar constant Btuhr-ft2
s = are length along error path
T = absolute temperature T o = absolute reference temperature
t = time hr
a = surface absorptance
= error ratio defined in (219)
= error ratio defined in (2114)
= error in modeling parameter
6 = Kronecker delta function
= surface emittance
vi
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
p = surface reflectance p = diffuse component of surface reflecshytance p = specular component of surface reflectance
-8 2 o4 -= 01713 X 10 Btuhr-ft - R a = Stefan-Boltzmann radiation constant
= diffuse-specular overall radiant exchange factor
Subscripts
e = refers to empty space
ij = refers to surfaces A A - I ]
m = refers to model
p = refers to prototype
Superscri pts
refers to solar spectrum B E H Q q We a p I I ~ Py and P
also refers to modification experiments P(+) P(-) s( and s(-)
(+) = refers to subspace S(+)
--) = refers to subspace S(
(Others not found in the list are defined in the text)
vii
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145
TABLE OF CONTENTS
Page
1 MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING 1
2 AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS 3 21 BASIC IDEAS E 3 22 GEOMETRIC REPRESENTATION OF ERROR STATES 9 23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --
PARABOLIC ERROR PATH WORKING FORMULAE 15
3 APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT--PRELIMINARIES 26 31 THE HYPOTHETICAL SPACECRAFT 26 32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION
OF THE OVERALL EXCHANGE FACTOR 62 28 321 Radiant Flux in the Solar Spectrum 31 322 Radiation Heat Transfer in the Infrared 37
33 NODAL HEAT BALANCE EQUATIONS 43 34 MODELING REQUIREMENTS 53
4 SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT 55
5 NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS 65 51 A RECAPITULATION 65 52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM
AND THE METHOD OF SOLUTION 67 53 DETERMINATION OF OPTIMUM y AND a 71
6 RESULTS DISCUSSIONS AND CONCLUSIONS 76 61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING 76 62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING 96 63 SOME RELEVANT OBSERVATIONS 125
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability 125
632 Proper Control of Model Errors and Selection of Experimental Condition 126
64 CONCLUSIONS AND RECOMMENDATION 129
REFERENCES -1
APPENDIX A THERMAL RESISTANCE OF MULTI-LAYER INSULATION 132
APPENDIX B A NUMERICAL EXAMPLE 139
viii
Chapter 1
MOTIVATION OF A NEW APPROACH TO THERMAL SCALE MODELING
Theoretical requirements for the thermal scale modeling of unshy
manned spacecrafts are understood quite well at the present time
Difficulties arise in practice when the system to be modeled consists
of components fabricated from a number of materials and they all parshy
ticipate in influencing the systems thermal performance This is
particularly true when information on both the steady state and transhy
sient behavior is sought One possible means of satisfying modeling
requirements is to modify artificially the thermal conductance of
thin struts plates shells etc by electroplating with a high conshy
ductivity metal such as copper The main objective of a previous
research contract between the University of Illinois and the Jet Proshy
pulsion Laboratory JPL No 951660 was to develop such a possibility
Test results obtained with prototypes and models of simple configushy
rations have demonstrated the technical feasibility of the concept
either with or without solar simulation With the availability of
adhesive metal foil tapest in recent years it is conceivable that
they can be applied more conveniently than electroplating In this
connectionone also sees the possibility of modifying the materials
heat capacity by using plastic (teflon) tapes However the range
tOne type of copper and several aluminum foil tapes with pressure sensitive silicone adhesive manufactured by Mystik Tape Inc of Northfield Illinois may be suitable for the present purpose
of thickness of commercially available tapes is limited Consequentlyshy
the desired thermal conductance and capacitance of the various strucshy
tural members of the system may not be mpnufactured precisely in this
manner They can only be approximated Furthermore the application
of a metal foil tape to a surface would also introduce a minor dis-shy
turibance to its heat capacity Likewise the use of a plastic tape
---would produce a minor influence on its conductance It is also posshy
sible that the desired surface radiation properties may not be duplishy
cated accurately These considerations prompted us to explore and
examine another concept of model testing in which models that do not
satisfy the similarity criteria completely are used The theory
which is presented in the following chapter contains some heuristic
but plausible arguments and in this sense it is not mathematically
rigorous It was described first in its rudimentary form in a preshy
liminary proposal Thermal Scale Modeling with Imperfect Models
prepared by the senior author and submitted to JPL for technical evaluashy
tion in January 1969 The material contained in Chapter 2 is a genshy
eralization of the theory presented in that document It includes
additional information relating to its practical implementation
To test the validity of the theory it hs been applied to a
hypothetical spacecraft having a geometric configuration that crudely
simulates the global radiative and conductive paths of the Mariner
spacecraft The results of the present study firmly established the
usefulness of the proposed new concept and one sees the welcomed flexishy
bility in the use of model testing--a flexibility heretofore imposshy
sible to attain
2
Chapter 2
AN APPROXIMATE THEORY OF MODELING WITH IMPERFECT MODELS
To begin with it should be noted that while our concern here
is the thermal modeling of unmanned spacecrafts the concept to be
expounded herein has general applicability to the immense field of
the technology of model testing As such the analysis and discusshy
sion that follow are presented with this viewpoint
21 BASIC IDEAS
Consider a physical phenomenon y which depends on a number of
independent variables x1 x2 x Symbolically we may write p
y = y(x x2 x ) (211)
If the functional relationship has general validity it must be inshy
dependent of the units of measurement Thus a consideration of the
requirement of dimensional homogeneity leads to a dimensionless form
jof the relation
44( I12 11) (212)
Methods of dimensional analysis whether based on the governing dif-N
ferential equations of the problem or on the strict algebraic formalshy
ism or on physical intuition have been well documented and need no
further deliberation here Suffice it to state that for perfect
modeling the ]ts in the model system are made indentical to the corshy
responding ffs in the prototype system
3
In the modeling of the thermal performance of spacecrafts the
dependent variables of interest are usually the temperatures of their
various components Thus the number of s with which the experishy
menter has to deal could be quite large Moreover as has been noted
previously the precise satisfaction of all modeling requirements
may not be possible in practice In fact even if it were possible
it might not be desirable from the point of view of economy This
is then the fundamental reason for using imperfect models and some
deviation in the Hs of the model system from those of the prototype
will be tolerated To facilitate further discussion we rewrite (212)
as
( i ( (213)
whete i = 1 2 bull N N being the number of different components
of the spacecraft to which a characteristic temperature is assignashy
ble The subset of all relevant variables or parameters for which
the model system is free from error is denoted by 9 Clearly for
the problem under consideration all elements of Sare passive and
can be ignored In the transient thermal modeling of unmanned spaceshy
crafts n is usually several times the value of N as will become
clear later
If we designate the errors in H by 116 and define
n (1 + 6 ) s2 (214)= j [111 ( 1 + 61) 112(1 + 62)
then $ may represent the measured entity using the imperfect model
4
eg the temperature of the component surface A of the model spaceshy1
craft which does not satisfy the modeling requirements completelyThe 0)
The quantity sought is of course () When the Ss are sufficiently
small the right-hand side of (214) can he linearized to become
n
0) + C 6 j =
in which the coefficients C are not known It may be noted that
if Taylors series is used then
(0)() (216)C 0
However (216) does not necessarily give the best estimate of the
coefficient C In fact the validity of (215) does not require
the usual restriction that needs to be imposed on the function for
the existence of the Taylor expansion
To evaluate (0) from the measured 4 one would require theoshy
retically the knowledge of the n unknown coefficients Ci This
means that n experiments will have to be conducted with n different
sets of 6s The work may become formidable even when n is only modshy
erately large say 20 or so Furthermore the accuracy required
of the experimental measurements under such circumstance could be
so high that they become impractical We therefore seek alternashy
tive solutions with some loss of accuracy
5
We begin by noting that the 6 s associated with the imperfect
model may be either positive or negative For reasons to be seen
later they are grouped separately We designate that 6 6 6 shy
are positive and that 6r+ 6 6n are negative Furthershy
more we define
r
P) C (217a) j=l
n
p)3 c6 8(217b)
j =r+
+ ) Clearly and may be either positive or negative With the 1 1
foregoing (215) may be rewritten as
~ 0) (t+)+A(21)
If the imperfect model is modified or if the model test condition
is altered (in either case the extent of modification must be restricted
so as not to violate seriously the linear approximation used in the
theory) then 4 + ) and will change also A possible practical
means of effecting model modification is through the use of adhesive
metal and plastic tapes for altering the thermal conductance and heat
capacity as indicated in [lit Thus originally positive errors will
become more positive while originally negative errors will become
less negative or positive In what follows the errors following
tNumbers in brackets refer to entries in REFERENCES
6
modification will be designated with an asterisk
Let us consider that the modifications are carried out in two
steps In the first only the positive errors are altered leaving
the negative errors intact or essentially unchanged We denote the
error ratio by a Thus
6 = j = 1 2 r (219)
If Z-4+) represents the experimental data so obtained then
rn
+) + C (21) ]]] ] 21
j j = r +
Eliminating (0) from (215) and (2110) yields
Ir
j 1
If a weighted average of the error ratio is defined according to
r
6c
jl (2112)
rrthen C21ll) can be rewritten as
+)- = (a- 1) C 6( j=l
7
and consequently there is obtained from (217a)
4 - (2113)a-i
Obviously a cannot be evaluated without the prior knowledge of C
However in the event that all a s are identical say a then a =a
-and no knowledge of CI is needed While such special condition
is usually not met in practice the foregoing observation does sugshy
gest that in testing with imperfect models the experimenter shouid
seek small deviations in a s
In the second step of modification only the negative errors
are altered For clarity and for convenience of later discussion
the error ratio is to be denoted by Thus
= = j r + 1 r + 2 bull n (2114)
In this instance the measured data 4- can be written as
n1
)- + c06 (2115)
1j ++
It follows then
n
- C ( - 1) (2116)
j r+l
As in the first case if a weighted average of the error ratio is
- defined namely
n
j=r+l (l7)
E Q 6 =r+l 2
then
= 1 (2118)
The observation previously noted for a can likewise be made for
Thus without the knowledge of C s the precise evaluation of both
a and is not possible However they can be approximated It is
this desire of finding the appropriate mean values of the error rashy
tios that led to the creation of a new concept of model testing with
imperfect models
22 GEOMETRIC REPRESENTATION OF ERROR STATES
The central idea to be explored herein is the representation
of error states in a multidimensional Euclidean space whose elements
are 6 This will be referred to as the simple error space in con-I
trast to theweighted error space to be discussed later Such repshy
resentation is compatible with the lineal theory considered in the
present study
For j ranging from 1 to n there exist the following three points
in the n-dimensional error space
pound
(6 6 6 1 6)
S 2 r r+ n
pd- ( 6 6 6 6shy(6 1) 2 r r+1 n
V +) There are also two subspaces and ) with their respective orishy
0( + ) gin at and J-) The elements of subspace +) are 61 62 bull
and those of ) are 6 6 If the error ratio a
0 +) is a constant the three points P and P-(+) are colinear in
$ The straight line which originates at 0( + ) and passes through
the said points will be labeled M Similarly if is a constant
d-) P and P(-) are colinear in S(- and the straight line joinshy
ing them will be labeled hen n = 3 the foregoing can be repshy
resented pictorially as is illustrated in Fig 21 No such pictoshy
rial representation is possible when n gt 3
In (+)the lengths of the line segments 0(+)P and 0+) P ( + )
are
j 120T)P=LI1= 12
and
_______112 __
aL 12
0(+p(+ = (6D) = tj = a O+p1[jr-+
[j r 16)2
If we regard + as the abscissa of a plane plot with 4t as its ordinate
10
Origin of f 8 - Space 02010)0
-1 H
)--___l -- _ SubDspaceS
p__j 2) (l c2P
Subspace S
Figure 21 Illustration of subspace S(+ ) and S in a 3-dimensional simple error space (6 6 are + 6 is -is a
plane while S(- ) is a line)
12i
L+)then a point corresponding to P and another point 4A correspondshy
ing to P-d+) may be located as shown in Fig 22(a) A conjecture
+ )in toduced is that and varies lineshyi i depends uri-uely on
arly with it It follows then
+ =1 1 (221)J+) 4 +) -- 1a
0( + ) p
which is identical to (2113) when a = a as is the case here
An analysis based on analogous arguments can be made in subspace
S(-) Thus
= 1K 621
P 0(-shyand from the i vs -)plot as illustrated in Fig 22(b) it can
be shown readily that(- )0(-)P OPp 0shy
________ - 1 (222)
-1
which isprecisely (2118) when = B
12
--
0+ 1 p (+)
r-_l
(b)
-)p P(-) 0
Figure 22 Linear dependence of ti with 6 in S()and S(shy
13
3
The foregoing analysis plausible as it may seem is subject
to immediate criticism Referring to (215) one recognizes that
(0)the deviation of from 4 depends on the products C 6 instead
of 6 alone The coefficients C are in essence weighting fac-I 13
tars which prescribe the relative importance of the various errors
in influencing the dependent variable Thus a weighted error space
---whose elements--are 1C -16 -should-he used instead of the simple er-J 3
ror space and it is more logical to assume that the distance ascershy
tained in such error space from its origin is an appropriate measure
of the resultant error in i
- A-question which naturally arises is Why is it then that
for constant a and the analysis made in the simple error space
leads to correct results The reason for this may be seen easily
In S( + ) of the weighted error space the lengths of the line segments
0(+)P and 0 +) P (+ ) are
C ) P = (C )2 12
and
12
0($)P +) - (C 6)
j =
Hence when a is a constant (+)P - =a O(+)P and the three points
P and P + ) 0 M are colinear Since only the ratio of distances
14
is involved in the expression for L it is seen that (221) reshy
mains unaltered For the same reason (222) also holds It is
significant that the linear correction formulae (221) and (222)
valid for constant aand 0 and originally deduced from a conjecture
in the simple error space are formally identical to those based on
a consideration in the weighted error space It goes without saying
that the experimenter has normally no knowledge of C and hence
the error state-points like P P(+) and P(-) cannot actually be
plotted in the weighted error space
23 GLOBAL EFFECT OF POSITIVE AND NEGATIVE 6S ON --PARABOLIC ER-ROR PATH WORKING FORMULAE
We now proceed to examine the more realistic case for which the
error ratios u and are not constants We restrict ourselves to3 3
the consideration that the range of vaiiation among the q s and among
the f s is limited The analysis which follows closely parallels
that described in [] for the evaluation of the global effect of the
errors based on a consideration made in the simple error space
Referring to Fig 23(a1) the line joining 0(+) and P(+) is
designated as the -(+)-axisSince a is not a constant but varies
with j P does not lie in this line Hence the point P and the (+shy
axis together determine a plane in S(+ ) of the weighted error space
In this plane a line 0(1)T(+) is erected normal to the (+)-axs
as shown For convenience the T+)-axis is so oriented that P lies
on -the positive side It is pertinent that EM and imply
15
(a1) (a2) A
__(b1) ) ( _ _ C+P (b2) shy
_________
A
0)0
ampige (+d
) -Figure 23 Parabolic error paths in S( and S
and ( since the scale of the coordinates in the weighted error
spacedepends on i For simplicity the subscript i is omitted
Passing through the three points 0(+ ) P and P-+ a parabola
can be constructed Tracing the points along the parabola and toward
the origin 0( + ) may be interpreted as conducting a series of experishy
ments with controlled errors of successively diminishing magnitude
As the origin is approached all positive errors become vanishingly
small Thus the arc length along the curve measured from 0(+) is
an appropriate measure of the global effect of all positive errors
in t The choice of a parabola as the error path is to some extent
arbitrary but it is certainly reasonable Further argument in supshy
port of the choice will be given later
The equation of the parabola in the (+-7 plane is
T(+) a+) +b )2 (231)
where a gt 0 and b lt 0 The coordinates of P(+) are
r= 1142(C 6y) ]2j(a
(232b) = 0
In (232a) C implies C The coordinates of P are
r
J degO+)P (+) j=j I0 j
E - - j1n 2 (233a)
7+ shy+
17
in
(+) )
EVI[P)P12 - (4) I2 1~2 Ij=l1
+4~ 2 which the overscore arrow denotes vectors It
+)
a b +-
1-shy
2(96
can be shown that
(2s33b)
(234ab)
The arc length 0( + ) P-(+) is
s4 +) f 1
1 -
Ll+GV- i 2( + 12 C2i12
a( + a) + Zn [(1 + a2 + all (235a)
and the arc length 0( P is
1 =-0-[b- a(l + a2) 12 2a)+~~Y 2 2n2(i+2 a
where
- aX(l + a 2x2 12-- n [Ci-+ a 22 ) 2 + aX] (235b)
2b (+) = 2 P (-1 lt X lt 1) (235c) a P +(+)
18
_ _ _
The ratio of the two arc lengths is
2s( + )
2 (236a)(a
s(+) 1-f(aX)
where
2f(aAY = aX(l + a 2 ) + kn [(1 + a2X2) 1 2 + aX] (236b)a(l + a2 112 + Yn [(I + 2)12 + a]
and a is given by (234a) and X by (235c) In the latter two
equations we further note that
-12 12
C(2) Z 2(6) I_ __TIP I__-cent+ = p )2
(237a)
4+) Zc66 (237b)
4+ + shy2 -c2_ 2
R In the foregoing the summand implies Z Consequently within
j=1 the context of the present theory the correction required for all
positive errors is
+ - (238)s(+)
19
J(4) (4+) (and the appropriate mean ais seen to be given by the ratio s s
-For the practical application of these results the followfng
observatio-ns are made
(A) WFhen the range of o is restricted (237a) and (237b)3
are weak functions of C3 They may be replaced by the following apshy
proximate expressions
-12
(239b) (23___b)
j(+)
1(6)2
Alternatively if the physical nature of the problem is such that
the dominant C s are of similar magnitude our experience indicatesI
that these approximations are also valid
Following a detailed error analysis based on the same arguments
but made for the simpZeerror space it was found that (236ab)
(234a) and (235c) remained unaltered provided that in the latter
(+) (+)two expressions the ratios p P and were calculated
according to (239a) and (239b)
(B) If the experimental conditions are so chosen that in the
(+) = 1 then = f(a)= aweighted error space 2( and00
20
s + Is = 2 which is independent of the C s This condition is
of course not generally met in practice However if a is small
say less than 05 then f(aX) becomes somewhat insensitive to variashy
tions in a Furthermore if we require that jXI be kept small say
within 01 or 02 then f(aA) is small compared to 1 Under these
conditions the error in the calculated ratio of the are lengths reshy
sulting from inaccuracies in a and X will likewise be small Thus
the use of the parabolic error path provides guidance to the proper
selection of the experimental conditions for reducing errors due to
the lack of knowledge of the weighting factors Cj
The foregoing observations (A) and (B) taken together suggest
that in the application of the linear correction formula (238)
the ratio of the arc lengths (or its equivalent W) can be evaluated
from a consideration of the simpZe error-space when the range of cz
is restricted andor when the experimental conditions are chosen as
explained
If in the modified experiments the positive errors are made
Zess positive the relative location of P and P(+) with respect to
the origin of S(+) would be as shown in Fig 23(a2) The (+)-axis
is now chosen to pass through P Following a similar analysis it
was found that
4(2310a)
1
(+) s
21
in which
sd + ) i - f(ax) (2310b)
f +) 2 S
and f(aX) is given by (236b) with a and A redefined as follows
(+)
a i 1 (2310c)(+) (2310)qPP
1shy
(+)
P
(+)
wherein
gt (22gt 2o 2
2W-27 C2 6I L ((2310e)
and
(2310f)4+) c22=
Again if the range of a is restricted andor if the experimental
condition is such that lXl is small as compared to unity the ratios
and +4 +) can be calculated from a consideration of
22
the simpZe error space namely
12
(1gt2) 5 - 2 J(2311a)
and
4- (23lb)
Clearly in (2310ef) and (2311ab) the summand implies j l
The analysis of the global effect of the negative s on 4can
likewise be treated If the negative errors are Zess negative in the
modified experiments the error path in S(-) would be as shown in
Fig 23(bl) We shall refrain from presenting the details of the
calculation and instead shall summarize the results as follows
_ s 4--I ) (2312a)
1 - shy(- )
where the ratio of the arc lengths along the error path is
( -) s - 1 - f(aX) (2312b)
2
23
and f(aA) is again given by (236b) with a and X respectively
denoting
a = ------ (2312c)
---C-)
i 2 (2312d)
--wherein
12C 7Co2--v2s ) = (2312e)
and
=) (2312f)
n
In the above pound implies z The ratio s ( Is clearly denotes
j
the mean
The error path would be as that depjcted in Fig 23(b2) if the
negative errors become more negative after modification In this
case
24
-) _ I - - cent(2313a)
sC_ shy
where
-)_ 2 (2313b)
S-) 1 - f(aA)
and f(aA) is as defined earlier with a and X respectively given
by (234a) and (235c) when the superscript (+) is replaced by (-)
The ratios n -) F- and 8- ) are given by (237a) and (237b) n
with the summand r Z The discussionreinterpreted to mean j=r+l
on the validity of using approximations deducible from the simple
error space also holds
Finally the quantity sought is
25
Chapter 3
APPLICATION OF THE THEORY TO THE THERMAL MODELING OF A HYPOTHETICAL SPACECRAFT-PRELIMINARIES
31 THE HYPOTHETICAL SPACECRAFT
To ascertain the usefulness of the approximate correction theory
for imperfect modeling as expounded in the previous chapter we conshy
sider a hypothetical spacecraft shown in Fig 31 The geometric
configuration chosen is intended to simulate the global radiative
and conductive paths of the Mariner 64 spacecraft but with all secshy
ondary and minor details omitted The solar panels A1 -A2 and A-A4
are fabricated of an aggregate of aluminum sheets silicon crystals
for energy conversion and glass plate covers all mounted in an alumishy
num alloy celled structure While the local thermal properties are
of a highly anisotropic and inhomogeneous character it is quite
feasible to describe its overalZ behavior by equivalent values of
thermal conductivity and heat capacity Based on the information
transmitted to us from JPLt it appears that a reasonable value of
kd for the solar panel is 0058 Btuhr-F Also the mean specific
heat of the aggregate is reported to be 020 BtuIb-F and a typical
Mariner solar panel of 3 ft X 7 ft weighs about 30 Tbs From these
it is estimated that the Cd of the solar panel is approximately 029
2 Btuft -F
tThrough Mr William A Hagemeyer
26
---------- --
Collimated Sunlight (Either at Normal Incidence or 450)
Xii 450
A8
I 7 AG I
A7 )
i i i 5
A6 A4
Solar Panel A1 A2 A4 A5 -
Bus A A6 A7 A8 Aq Ao (A7 and Alo are covered wth superinsulation )
Antenna All All areas are idenlical and equal to 4 ft2
Figure 31 Hypothetical spacecraft
27
To keep the radiation geometry associated with the diffuse-specushy
lar surfaces as simple as possible the bus of the spacecraft is Taken
to be a cubical box instead of the usual polygonal configuration The
top and bottom faces of the bus are covered with multilayer superinsushy
lation The collimated suns rays are assumed to be either perpendicushy
lar to the solar panels or slanted at 450 to them but remain parallel
to the front and back faces of the bus TABLE 31 lists the materials
selected for the six component faces of the spacecraft bus The thickshy
nesses indicated afe equivaZent values they take into account the
electronic packaging subassemblies bolted to them Included is the
relevant information for the main directional antenna As will be
shown in Section 34 perfect modeling requires that the corresponding
surfaces of the model and the prototype have identical radiation propershy
ties Furthermore when the temperature fields are two-dimensional
the kd and Cd of all component surfaces of the model must bear a fixed
relationship to those of the corresponding surfaces of the prototype
TABLE 32 summarizes al pertinent surface properties and the equivashy
tlent (kd) s and (Cd)s for the hypothetical spacecraft
32 RADIATION HEAT TRANSFER ANALYSIS AND THE DETERMINATION OF THE OVERALL EXCHANGE FACTOR e9
When the surfaces of a spacecraft are under the irradiation of the
sun the absorbed and reflected radiation are in a wavelength region
totally different from that of the emitted radiation While it is posshy
sible to perform a general formulation of the problem there is neither
the need nor the justification of doing such analysis at the present
time It is well known that approximately 98 percent of suns radiation
28
TABLE 31
MATERIAL LIST FOR THE HYPOTHETICAL SPACECRAFT
(a) Solar Panel
Aggregate of silicon glass aluminum and its alloy
(b) Bus
Panel Material k C Equivalent Btuhr-ft-F Btuft -F Thickness
d in
A3 Magnesium Alloy 470 259 0125 AN-M-29
A6 Aluminum Alloy 802 371 0 1875 5086
A7 Titanium Alloy 460 350 010
A-110-AT
A Magnesium AlloyA781A(T4)
291 280 01875
A9 Magnesium Alloy 630 276 0125 ZK51-A
A1 0 Titanium Alloy 838 385 0025 -150-A(2)
(c) Antenna
Aluminum alloy 2011 (k = 823 Btuhr-ft-F and C- 393 Btuft3 -F)
Thickness - 00625 in
29
TABLE 32
SURFACE PROPERTIES AND OTHER DATA FOR THE HYPOTHETICAL SPACECRAFT
P
Solar
P or Pd
Infra-red
PS a(=E)
kd (equivalent)
Cd (equivalent)
Solar Panel
Sunlit Surfaces A1 A2 A4 As
Surface Away from Sun AiA 2A4Aw(painted with PV-100 white paint)
0
080
020
0
080
020
020
017
0
0
080
083 0058 029
A5 exterior (PV-lO0 paint)
interior (partially painted)
080 0 020
-- --
017 080
0 0
083 020
0489 0269
SA 6 exterior (PV-lO0 paint)
interior (partially painted)
080
--
0
--
020
--
017
080
0
0
083-125
020
0580
exterior (superinsulation)
interior (black paint) 0
--
085 --
015 --
050 015
0 0
050 085
00383 0292
Bus A8 exterior (PV-100 paint)
interior (black paint)
080
--
0
--
020
--
017 015
0 0
083 085
0455 0437
A9 exteior (polished aluminum) interior (black paint)
0 080 020 ----
0 015
095 0
005 085
0655 0287
A1 0 exterior (superinsulation)
interior (black paint)
0
--
085
--
015
--
050
015
0
0
050
085 00175 0080
Antenna A11 sunlit face (green paint)
face away from sun
025
--
005
--
070 018
095
0
0
082
005 0428 0204
2 All (kd)s are in Btuhr-F (Cd)s in Btuft -F
is contained within the 0-3 pm range On the other hand less than
001 percent of the radiant energy emitted by a black surface at
530 R (room temperature) is in that wavelength range Even at 750 Rt
less than 04 percent of the emitted radiation is of wavelengths beshy
low 3 Pm The semigray or two-band analysis suggested by Bobco [2]
is based on the foregoing facts The use of such a two-band model to
account for the spectral dependence of the surface properties of spaceshy
crafts has been examined by Plamondon and Landram [3] and found to
yield very good results
A simple and often realistic description of the directional propshy
erties of the surfaces is a subdivision of the hemispherical reflecshy
tance into diffuse and specular components ie p = pd + p In
the analysis that follows we shall adopt such an idealization which
together with the semigray approximation and the assumption of diffuse
emission makes possible an analytical treatment of the problem in
a reasonably straightforward manner The essence of such approach was
used in a recent paper by Hering E41
32] Radiant Flux in the Solar Spectrum
We first present a general formulation for an enclosure
of N surfaces each of which has a uniform temperature and is uniformly
irradiated The results will then be applied to the relatively simple
configuraticn selected for the hypothetical spacecraft All propershy
ties and radiant energies associated with the solar spectrum will
be designated with an asterisk
According to the two-band model theemitted radiation in the
tThis far exceeds the ordinary operating tempeiature range of present day spacecrafts
31
solar spectrum from any surface of the system is completely negligible
Hence the radiosity of a surface A is given by
B= p (SE + H-I) (321)i d i
in which E is a type of exchange factor denoting the fraction of
the solar flux incident on A directly and by all possible specular
reflections The irradiation H is given by
N
Hi= E (322)
j=l
where the exchange factor E denotes the fraction of the diffuse
radiation in the solar spectrum leaving A which arrives at A both1 3
direbtly and by a11 possible intervening specular reflections We
note that in writing (322) the reciprocity relation AE = E- ]1I 11-3
has been used Eliminating H from (321) and (322) yields
N
B BS= pa SE + p B (323)E 1 1 d i 1 3
j=1
which forms a system of N simultaneous linear algebraic equations
for the N unknown radiosities Once the Bi s are known the rate
of solar energy absorbed by Ai can be calculated from
32
+= S H)(st
A B (324)
Obviously B and consequently Ql are independent of the system
temperatures The decoupling of the solar fluxes from those of the
infrared radiation results in significant simplification of the analyshy
sis
If a surface k reflects purely specularly then pk = 0 and
accordingly B vanishes In this case
Q1 = Ad(sE + H ) (325)
We now proceed to apply the foregoing results to the hypothetishy
cal spacecraft when the suns rays are parallel to the front and rear
faces of the bus and are either inclined at 450 or normal to the
solar panels Figure 32 shows the relevant geometries involved
To evaluate Q Q and Q we need only to consider an enclosure
c6nsisting of A1 A2 A3 and an imaginary black surface A L at abshy
solute zero The latter comprises a rectangle with its long sides
shown as the dotted line in Fig 32 and two identical right triangles
both adjoining the rectangle Clearly BL 0 It is easy toeL
see that
= cos 45 B = 1 1 1 -2 1-3 1-3 (326a)
E=E = E =ED = 0 E =F 2 3 2 3-3 2-3 _3
33
Ao
Front -c--
A11 A1i0A
10_
A 2iComputer
__
logram)
_ __
2
Front
__
View
flA51 A4
Designates superinsulation
A A 2I
Lw
I A5 A 4
Figure 32
Top View
Radiation geometry associated with hypothetical spacecraft
Consequently (323) becomes
B1 - p 1F -B = p s cos 450
B - p 2F2 B3 = 0 (327a)d2 2-
p (F B + F3B -B = 0 d3 -1 -2 2 3
When the radiosities are evaluated Q Q and Q can be determined
from (324) or (325)
When the suns rays are normal to the solar panel
bull 1 B 1 = 0 E- F
E= 1 E = 0 E F (326b)
EB 0 E_3 0 I
Consequently
- P F Bwd 1 --3 3 = o SI -dI
B=p (327b)F S
P- (F B + B) - B = 0d 3- 1 F3-23-1 2- 3
To evaluate Q4 Q5 and Q we consider a similar enclosure conshy45 6
sisting of A4 A5 A6 and the associated imaginary black surface
A Rat absolute zero Again referring to Fig 32 we have
V cos 450 4o 4-6
E=(lp )cos 450 E =6 F1ES + P6 = B_ = F5 -6 (328Y
= ( + p ) 005 450 E- = 0 6 S5 6-6
35
and
B - pI F B z S cos 450
4 4 -66 cil 4 o 5
B - p F B = p ( p )S cos 450 (329)
4
d5 5-6 6 d 5 s6
p (F B- + F Bf) - B= -p- (U+ p )S cos 450 d 6 6-4 4 6-5 5 6 d 6 s 5
As before when the radiosities are evaluated Q- Qand Q can 4 56
be readily calculated
When the suns rays are normal to the solar panel the relevant
radiosities can be evaluated from (327b) provided that the subscripts
1 2 and 3 are replaced b y 4 5 and 6 respectively
By following the same procedure we may evaluate QP and Q
The results are
(a) For oblique solar irradiation
S cos 450sin 300 (3210a)
91 9 pw P F F
(b) For normal solar irradiation
It is only necessary to replace cos 450 in (3210ab) by unity
The top face A7 of the spacecraft bus is completely shielded from
the suns rays by a multilayer superinsulation The Appendix of this
report gives a detailed analysis of the thermal resistance of the
superinsulation and presents an expression from which the leakage
flux may be determined The rear and bottom face A8 and A10 of the
36
bus do not receive radiation in the solar spectrum either directly
or indirectly Hence Q = Q = 0- 810
322 Radiation Heat Transfer in the Infrared
Once again for generality we first consider an encloshy
sure of N surfaces each having a uniform temperature and being unishy
formly irradiated It is assumed that all emitted radiations are
diffuse and that the surface reflectances can be adequately described
by the pd-p model Our objective here is to deduce expressions
for the overall radiant exchange factor similar to Hottels diffuse
exchange factor 69 All surfaces are taken as gray within the specshy
tral range Clearly the results will be directly applicable to the
determination of similar exchange factors for internal radiation withshy
in the spacecraft bus
For any surface A of the enclosure
4
B = T- + P H (211)
and
S E B (3212)
j =3
37
Thus
N
B = T - P i EB (3213)
j=l
The system of simultaneous linear algebraic equations (3213) can
--be-rewritten as
N4
aBB i =1 2 N (3214)
jnl
where
a = 6j Pd iE (3215)
To evaluate the diffuse-specular overall exchange factor we-k
set aT = 1 and the temperatures of all remaining surfaces to zero
The heat flow rate at k is
AkakAB (3216)
where the presubscript i is a reminder of the original source of the
flux We recognize that B i(j = 1 2 N) is the solution
set of (3214) with the latters right-hand side set to zero except
for the ith equation for which it is S It is most convenient
to evaluate B by matrix inversion
38
Let [b] be the inverse of the coefficient matrix [a ] That
is
b1 1 - b1 2 bIN
a- [b 21 22b2
(3217)
bN bN bN
It can be readily demonstrated that
B -- bji (3218)
By definition i Qk = A it follows then
N Ak
-k 1gt (3219)
j=l
We now proceed to apply the foregoing general results to evalushy
ate V for the spacecrafts exterior surfaces We first consider
the enclosure formed by A1 A2 A3 and AeL the latter being black
and at absolute zero
1 A1 A2A are d- PS surfaces A A is black
A A A2 2(3) 2(3)
Figure 33 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A A2 A and AL
39
In Fig 33 the surfaces under consideration are identified with
letters or numerals irediately adjacent to them Thus AI refers
to the upper surface of the solar panel and A3 the exterior surface
of the left face of the spacecraft bus The image surfaces are desshy
ignated with two numerals the one in parentheses refers to the mirror
in which a surface in question is seen For instance 2(3) is the
image of A2 as seen in the mirror A3 With the aid of Fig 33
the various exchange factors can be readily written They are
ElB 1 = El_ 2 = 0 E1_ 3 = F_3 11_ = F _ p 3F1( )
E2- 2 z 0 E2_3 = F2_3 E2- = F2 -e + P 3 F2 ( 3 )-o (3220a)
EB_ 3 = 0 ES-e = F3 -e + PsI FS( 1) -e + Ps 2 FS( 2) e
and the elements of the corresponding coefficient matrix are
a1 1 =1 a1 2 = 0 a1 3 = -Pd IB-l 3 a -d IEl e
a2 1 = O a2 2 1 a2 3 = -Pd E2 3 a = - d 2 E2 -0 (322 b)
a 1 = _pd 3E3-I I a32 = -Pd 3ee-2 a3 3 = 1 a = -P se shy
ae1 =0 ae2 =0 a 3 =0 aCe =1
In (3220ab) the subscript e refers to A L Using the foregoing
information the diffuse-specular overall radiant exchange factors
- 1 -9 - -s and 4- can be evaluated immedishy1-2 1-3 1 -e 2-3 2-e 3 -e
ately from (3219) Exchange factors associated with A A5 6
and Ae R can likewise be determined Surface A8 sees only the empty
space thus the relevant 19 factor becomes its surface emittance
40
Next we consider the enclosure formed by A9 A (upper surshy
face of antenna) and the associated black surrounding Ae F Figure
34 shows the two relevant mirror images 9(i1) and 11(9) associated
with the system I A
Figure 34 Mirror Images of Interacting Surfaces in an Enclosure Consisting of A 9 A II and A F
It is easy to see from Fig 34 that the exchange factors are (with
subscript e eferring to A )
Eg 9-9 0 Eg _1I Fg-_1I E9 e = Fg-e + pS I Fg9(ii) -e (3221a)
EII I 0 E1l -C = I Il e + p 9 F11( 9) _-C
and the corresponding a Is are
ag _9 i a-19 1 -_d 9d9g-1 1 a9- = Pd99e I(
a 1 1 - 9 Pd111 1 1E 1 - 9 1 a 11 -l1 l a I- = Pd I IE 1 1-e
The remaining system of the spacecrafts exterior surfaces for
which the 1 factors are neededconsists of the dark side of the solar
141
panels (A A21 A4 and A5) the outermost plastic film A1 0 of the
superinsulation covering A1 0the shaded side of the antenna A
and the associated empty space Since none of the surfaces involved
has a specular component of reflectances in the infrared region the
evaluation of a and subsequently i- requires information on
shape factors but not on exchange factors The formulation presented
in this section and the expressions (3211) to (3219) remain valid
All one needs to do is to replace pdi by p and E -j by F1 It
goes without saying that the 6 factors for interior surfaces of the
bus can be likewise evaluated
42
33 NODAL HEAT BALANCE EQUATIONS
To formulate the nodal heat balance equations we need to evalushy
ate in addition to the overall radiant exchange factors the conshy
ductances of the various heat flow paths We shall assume that there
is no conductive link between the solar panels and the spacecraft bus
nor between the bus and the antenna Further the eleven sections
(22 surfaces) which make up the spacecraft are taken to be identical
squarest as illustrated in Fig 32 Each section may have a differshy
ent thickness however Under the foregoing conditions the conducshy
tance between two adjacent areas A and A having a common edge is 2 J
(kd) (kd) -) (k )- -2 + (331)
Clearly K - If A and A are fabricated from the same ma-K~ 1
terial and are of identical thickness like A and A (or A and A )1 2 4 5
of the solar panel then (331) reduces to
= (kd) = (kd) (331a)1 -3 1 3
It is pertinent to note that the crude subdivision used in the
analysis should in no way jeopardize our present objective since
the model is to be subdivided in precisely the same manner In this
regard we may further add that solely from the viewpoint of ascershy
taining the usefulness of the proposed theory it makes no difference
tA significant simplification of the arithmetic results from this choice It is not a restriction necessarily imposed on the analyshysis
43
whether a continuum model or a discrete model is used in the analyshy
sis
The top and bottom faes of the spacecraft bus are covered with
superinsulation The exposed layer of the superinsulation for the
bottom face interacts radiatively with Al1 and thenceforth with
- Air A2 t A4 and A5 An exact analysis of the thermal transients asshy
sociated with its many layers is too involved and is not warranted
for the problem under consideration If we ignore its heat capacity
then a reasonably simple analysis can be made It is demonstrated
in APPENDIX A that the outward leakage flux (from the base plate A10
toward the outer surface of the insulation) is
4 4 - a(T1 0 T10 1) (332)
AR
where ZR = (n - 1)R + R All quantities have been defined in
APPENDIX A
We now proceed to present the heat balance equations The dishy
mensional form will first be given Set (333)
dT1At (Cd) d- + )a (T4 T24 _ a (T 4)
I 9T -2 9 C 4
1-1 (1 T5 )-6tl -4 (T41 T42 4a T4
T4 K1 2IV 4 If-1_ 1aTl 1 1 A (T1 T2
44
dT2 4
44
- d a(T - ) - (T T2 _1 2 - T3 2
-2-I ( r ( T4 )a T112 11
44- TII) AA r(T 2 -T 14 4shy)
dT + q 4_(Cd)To 4T (T 4 -C (Tshy 4
A (T-T- T3 2
A (T$ - T8 ) A(T 3 6 -T T T
AT3 10T A 31- T 7 S-e 3 Y-T3 -
K 3 _1 0 A T3
2-1100AT10)
43i0
-- 5 )ar( A 3 4
A4cC 2 4- -5T 4 T5 or 4-6 ( 4 T +- 4-e4 4t-1 4 1
I4t-2
CT4
4 - T74)
2 - 27
4-101 T~4
4 T 4
10t
4 A K4 -5 14 11 4( 1)T
4~5
dT 54
A (Cd)- 4 T- - 10(T
+ S4 A (- T4 4
T q -4q( plusmn4-4(d )T -2 -(T -T)4(
6 66
6
T
o
-
4 -A
T ) 4
a
K (T-i r
T7
- T )
-~ OC(T7-T)676-8 6-7 aT 4
K6-8e
- A6-T14T
K
4-7 T
6-7
6- 4 6T 4
6-1
T
7
A-7
- -T 9)7_T
-lt76-e K7-66T 6--
A (T6 - T9 -42 OC OT (6C 10 At 6
- (
76)
(--T- 7
T
( 7 -T8
-T1 44 4 4
7-6 ( 7
45 (
Y7 --9 atr
A (-T7 467
- - 7
4 -4) - -
T K)-6
A7 - T6
8 CTT7
r(
(T
CT8
T8a 4 T
- ) 8
(T 7lc
-9 -- CT 7 T9
463
A8 (Cd) dT
- 9 - -7
4 -
(T 8
T7)7
4 4 a (sT - -
- 95 a (T8-9 8
9aC4
-
T49
4 T
-
- (T
-
-T
TI0) K8-7
)A--(T-7)
(TA
1(8-10
A (Cd)99
dT
9dt - 8-96plusmn q9
T4 T9-1 8-97K_0 - 4-
911 9 -e TIoT
9 T(T
A
Al0
(Cd)
d~lo 4 (Cd10d _ G (T o- T
- oC7~ -gt 9lto-3
A ((T 4- T
- 9-7 9 7
e2 GT4 T4
-10 9
(- T A (T9 - T
4 -- =( -7((T
10 10plusmn 0
41 T4 o) - - ( 1O
-9_ -b
K9o9-6
9 AC(T - T)
9-8 9 8
- 9lt93 (7T o10A 9 - T9
K- 7
A 9 7
4 4)4lT) 0 10-3 10
T4
9-_I0
A
4
3
0
- 6 10 6
4 OF8 a(T
10-8 0
K 03(T -T)-0-A 10 3
K~~ )- -T
A 10 8
10-7 10 7
CT(T - T4
10-90
(7 -T A 10
(T A 10 9
4F7
- 7 t4 )(T 4a1 41 0 o0 01 1
ZR
40-4 10 4
aC 4 4T4) OF(~T T4 4D510 o 5 lot-III 10t 11
10
T 4
A1 1 (Cd) dtT - T UT1111 dt 111-9 11 9 11-e 11
4 _TT4
c(T4 TT44 or (CT4- 74 4 11-1 11 4111-5 11 5
- -o 1 - lot - aT4 11 1 01 c T1)
Nodal Heat Balance Equations in Dimensionless Form
To reduce the system of heat balance equations to appropriate
dimensionless forms we introduce the following nondimensional varishy
ables
T2F (334a)
0
ki tS3
(33 4b)C A
3 cT A
= (334c)
-(kd)
where i 1 2 3 etc and T is a suitable reference temperature0
for the problem Its choice is to some extent arbitrary For
instance T may be taken as the initial uniform temperature of the
48
spacecraft or some representative average of its steady state temperashy
tures prior to the initiation of thermal transients Since the Physishy
cal time t must necessarily be the same for all is the dimensionshy
less time T2 for the various component parisA1
must be so related
that
kI k2 k3
TIT2T3 2
By using (334abc) we obtain the following non-dimensional form
of the nodal heat balance equations set (336)
deI qA
r - + - + z5 + z 7 + V A A(kd) - [1-- - e- 1 - 1
4+ + + + -I -St_ -1 t10 1 t 5 tI 1-2 1-~L~ 2~l3
(kd)-( ) (--
A 7 2 (kd)2To 2 It 1 t 23 +22+Y 4t
4 4
2L-1 -5 amp 21-10 tO
457 1 K21 2tL1T 1J (kd)2
4~9
t
dt3
~
(q + q3 )A
( d T-~+c
[
+ +
04
S 73-10 10]9 4 4
-7+plusmn 2 4
- z- -4
4
Y3 -7K-_43 9K3_
4 d4= 4 -
(kd)4-( 4
53-2 2 -o3-_6 +63_-7
-k- 97- 2s 33 3)
qA [T- $ + (d) T 4464+
+ 5 44T411plusmn47plusmn4+47 4 7
A14
4
8
3
+
3 _9
_7 9 _
1-
L1 L2
+4
A + 4)
4
4- d It]-21 2shy
-V
K4-5 ( k)4
4 10 A0
5
is2
55Aq6A
-Rd)5sT +T4 e) ) 4 plusmn5 tTA4
5 -6 ~6 75-llI Z5- 2t - L1tl
4
OloI
ii kd) 5 ( 4
50
+ 66-7+lt(4 t + 6-(gj ~d~~ )A
66A6
44-
-)069 +-F + 6 _8c 66 _(K6
+ 61091(6 (kdKS IVplusmn 6
+ -3 4
Td6
0 +
-kshy-- +
--
d c1-76~l 9 2
-d 7 -425 - 7-
A ( 7Tk70 t0 )R
_ +74987-33 +Th71 e7 7
4 y 7 - + r 7
9 (k 7 -1 1 Tkd) 7
-81 plusmn K7 -6~+ -7 plusmn+5 7 -9
soAa _4 44qSA4 - raaS-6+9
41
8 7tY-ss +1S-10 5 K8 7
I K
dcent (q9 + +
A 9 7 kd 9+ (k9+09-7 - 0
K1 -6
1 (lcd)9
( 6
(K9
- [e 1_plusmn 6_plusmn4o_I 0 )(4k c(1d) 99 -8 -1 0 91 1
1 - 9 8 81 - 0
- + K(9-6 + Y1(-7 + K9 -10 )09
+ k--79 _39-R
1(9_3 + (lcd) 9
4 3
0K9 -9-7
+ (kld) 9 6 + 4(kd) 9
_ _Z7 40Y9 -0 (ld)9 10
A1 0
do 10 d1 0
i __I5103+ + 10-6 10 -7+_
o4-_ 0 -8
+ 4 2
0-
(P4
10
421
R
4
lt
4qi
10-3
44 4
4)04 0-6 10-7 7
Z42 18
A)44474 8
44
lO- 9 j
A
-(lcd)
K 10-6
(lcd) 1 0
101 ++
RL 4)4o
4 10P
R0-1 11
0 -3
1 OK0-8
6
0-
i tI
So06 + 0-8+
K 0-9
Sl (lcd)1
10
10-2 02
+ 42Z p +
Kl0 -9010
9
1 0 14 T 4)4
47O21 42
II0
+ (lcd)
1015
1 -
3
4)4
+ OW5 1O 11 0e)io _e
52
(kd)q T ii+2 AI1 64)1 = fd 1 T 118 11-e- 111-11 11-2
11111-11 e 11shy
-42 1 1 it2 Th 11-4 4 1 t5
Z11i-lot 10J
In numerical computations if we arbitrarily select T3 as the stanshy
dard dimensionless time then
) (k3 C1 dTl k~I C3)dT3
d642 k C 6)d2
dT etcd2
34 MODELING REQUIREMENTS
An inspection of the dimensionless heat balance equations reveals
that for perfect modeling the following scaling requirements must
be met
1 Model and prototype are geometrically similar in order that all configuration factors among corresponding surfaces are identical (Thickness distortion would have very minor or little influence on radiation exchange and is thus permitted)
2 Radiation properties of all corresponding surfaces are idenshytical
The satisfaction of (1) and (2) ensures that the model and the protoshy
type have the same e2 for all corresponding surfaces
3 The simulated solar radiation used in model testing has a spectral distribution intensity and direction the same as those of the local suns rays incident on the spacecraft
53
4 The (kd)s and (Cd)s of all corresponding surfaces must satisfy the following relationships
(kd) A L 2
(k) = - a fixed quantity (337a) P P P
(Cd) t
(Cd)p
- p
a fixed quantity (337b)
The first is associated with the geometric scale ratio and the second the time scale ratio
5 The internal power dissipation when expressed in terms of the equivalent surface flux is the same for the prototype and the model ie
S337c)
for all corresponding surfaces of the spacecraft bus
Finally we note that for complete similitude the initial
temperature field of the model and of the prototype must either be
uniform or have identical distributions The foregoing modeling
requirements have previously been stated in the literature eg
[51 and experimental verifications are available ESabcJ
54
Chapter 4
SOLAR FLUXES DIFFUSE-SPECULAR OVERALL EXCHANGE
FACTORS AND CONDUCTANCES OF HEAT FLOW PATHS
-ASSOCIATED WITH THE HYPOTHETICAL SPACECRAFT
As a prerequisite for the numerical solution of the nodal heat
balance equations the solar fluxes QA the overall radiant exchange
--- factors and the conductance of the heat flow paths K - associated
with the various surfaces of the hypothetical spacecraft must be
evaluated All relevant configuration factors can be readily detershy
mined since all surfaces or their subdivisions are chosen to be idenshy
tical squares as illustrated in-Fig 31 The results are listed
in TABLES 43 and 48 for surfaces facing the sun and in TABLE 413
for surfaces away from the sun The data were taken either directly
from an NACA Technical Note by Hamilton and Morgan [ 7] or indirectly
evaluated from the information given therein The solar exchange
factors were calculated from (326a or b) and (328) and the solar
radiosities from (327a or b) and (329) The irradiation H1 is
then given by (322) and the required Q follows from (324) or
(325) For surfaces A and ill the solar fluxes can be directly
determined from (3210a or b) In these calculations-the solar
- constant S was arbitrarily assumed to be 442 Btuhr-ft2 and the releshy
vant reflectances and absorptances were taken from TABLE 32 The
solar fluxes so calculated are listed in IABLES 41 and 42 respecshy
tively for the normal and oblique incidence of suns rays
55
The procedure used in the determination of the diffuse-specular
overall exchange factors for the spacecrafts exterior surfaces has
been explained in detail in Chapter 3 Following the evaluation
of the exchange factor E the elements of the coefficient matrix
a i were calculated according to (3215) and those of its inverse
b -j were determined by using an available subroutine in our compushy
ter laboratory For the convenience of the reader these intermedishy
ate data are listed in TABLES 44 and 49 for E_ in TABLES 45
410 and 414 for aa -j and in TABLES 46 411 and 415 for bl j
The desired diffuse-specular overall exchange factors e_- were calshy
culated from (3219) and the results are listed in TABLES 47 and
412 for surfaces facing the sun and in TABLE 416 for surfaces away
from the sun It is pertinent to note that for an enclosure of
N surfaces the overall exchange factors satisfy the following simple
relation
N
j =zs (41)
j=l
which is the consequence of the requirement of energy conservation
Equation (41) has been used to check the numerical accuracy of the
data list in the three tables last mentioned
For the interior of the spacecraft bus the radiation was asshy
sumed to be devoid of specular component and hence only the difshy
fuse overall exchange factor is involved TABLES 417 418 419
56
and 420 list respectively the data for the configuration factors
F- the elements of the coefficient matrix a the elements of
the inverted matrix b and the diffuse overall exchange factors
Finally the conductances K are listed in TABLE 421 They
were evaluated from (331) using information given in TABLE 32
It may be noted that there is no conductive link between the solar
panels and the spacecraft bus nor between the bus and the antenna
57
58
NOTE EXTERIOR OF SURFACE 10 1S DESIGNATED AS SURFACE 12
TABLE 41 SOLAR FLUX AT NORVAL INCIDENCE QA BTUHRSQFT
1
3535999
2
3535999
3
00
4
3535999
5
3535999
6
00
7
00
8
00
9
02376
10
00
11
1547000
12
00
8 TABLE 42 SOLAR FLUX AT 45 NEINCIDENCE OAA BTUHRSOET
1
2500340
2
00
3
00
4
2579092
5
2980498
6
750102
7
00
8
00
9
01680
10
00
11
1093899
12
00
~~4t EXTEIOR SURFACES ct
TABLE 43 CONFIGURATION FACTORS FA(IJ)
I
1 2 3 1(3) 2(3)
3(l) 3(2)
J 1
00 00 0032810
0 03 28 10
2
00 00 0200040
0-200040
3
0032810 0200040 00 0032810 0200040
E
0967190 0799960 0767150 0032810 0200040
0032810 0200040
__ - FOLOUT FRAMME -___ ____--- - IFOLDOUT FRAM
TABLE 43 CONTINUED 59
1 3 4 5 6 E
4 00 00 0032810 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767150 4(6) 0032810 0032810 5(6) 0200040 0200040 6(6) 0032810 0032810 6(54 0200040 0200040
__ _ TABLE 44 EXCHANGE FACTORS E(IJ)
I 1 1 2 3 E
1 00 00 0032810 0967190 2 00 00 0200040 0799960 3 0032810 0200040 00 0767190
__ TABLE 44 CONTINUED
I 4 5 6 E
4 00 00 OO328lO 0967190 5 00 00 0200040 0799960 6 0032810 0200040 00 0767190
TABLE 45 COEFFICIENT MATRIX AA(IJ)
Il J 1 2 3 E
1 1000000 00 -0006562 -0193438 2 00 1000000 -0040008 -0159992 3 -0005578 -0034007 1000000 -0130422 E 00 00 00 1000000
444 TABLE 45 CONTINUED 44
I 4 5 6 E
4 1000000 00 -0006562 -0193438 5 00 1000000 -0040008 -0159992 amp RME2
FOLDOUT FRAME
60 6 -0005578 -0034007 1000000 -0130422
F 00 00 00 1000000
TABLE 46 INVERSE OF COEFFICIENT MATRIX BBIJ)-
I
1 2 3
E
1
1000036 0000223 0005585
00
2
0000223 1001362 0034054 00
3
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
TABLE 46 CONTINUED
I
4 5
6 E
I J 4
1000036 0000223 0005585 00
5
0000223 1001362 0034054 00
6
0006571 0040064 1001399 00
E
0194338 0165478 0137133 1000000
44 TABLE 47 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FF(lJI 4
1
1 2 3
J 1
0000117 0000715 0021816
2
0000715 0004360 0133012
3
0021816 0133012 0005670
E
0777351 0661913 0669534
TABLE 47 CONTINUED -
I
4 5 6
I 3 4
0000117 0000715 0021816
5
0000715 0004360 0133012
6
0021816 0133012 0005670
E
0777351 066113 0669534
FOLDOUT FRAMEFOLDOUT FRAME 0
61 ___ TABLE 48 CONFIGURATION FACTORS FA(IJ) a
I J 9 11 E
9 00 0021500 0978500 11 0021500 00 0978500 9(11) 0489250 0489250 00 11(9) 0021500 0021500
TABLE 49 EXCHANGE FACTORS E(IJ)
I I J 9 II E
9 00 0021500 0978500 11 0021500 00 0998925
bull TABLE 410 COEFFICIENT MATRIX AA(IJ)
I J 9 II E
9 1000000 00 00 11 -0003870 1000000 -0179806 E 00 00 1000000
__ _ TABLE 411 INVERSE OF COEFFICIENT MATRIX BB(TJ1
J 9 11 E
9 1000000 00 00 11 0003870 1000000 0179806 E 00 00 1000000
TABLE 412 DIFFUSE-SPECULAR OVER-ALL EXCHANGE FACTORS FFIJ)
I I J 9 11 E
9 0000000 0000881 004911811 0000881 00 0819118
F__O__ _TLE4 COFGOLDOUAI FACORL =I7
FOL~D1JLEA -a -TABLE413 CONFIGURATION FACTORS FA(IJ) ___
62
I J 1 2 10 5 4 11 E
1 00 00 00 00 00 0002700 0997300 2 00 00 00 00 0019500 0980500 10 00 00 00 00 00 0087000 0913000 5 00 00 00 00 00 0002700 0997300 4 00 00 00 00 00 0019500 0980500 11 0002700 0019500 0087000 0002700 0019500 00 0868600
TABLE 414 COEFFICIENT MATRIX AA IJ)
I I 2 10 5 4 Ii E
i1 1000000 00 00 00 00 -0000459 -0169541 2 00 1000000 00 00 00 -0003315 -0166685 10 00 00 1000000 00 00 -0043500 -0456500 5 00 00 00 1000000 00 -0000459 -0169541 4 00 00 00 00 1000000 -0003315 -0166685 11 -0002565 -0018525 -0082650 -0002565 -0018525 1000000 -0825170
_4 TABLE 415 INVERSE OF COEFFICIENT MATRIX 58IJ) R
I J 1 2 10 5 4 i1
1 1000001 0000009 0000038 0000001 0000009 0000461 0169942 2 0000009 1000061 0000275 0000009 0000062 0003327 0169580 10 0000112 0000809 1003608 0000112 0000809 0043662 0494484 51 0000001 0000009 0000038 1000001 0000009 0000461 0169942 4 0000009 0000062 0000275 0000009 1000061 0003327 0169580 II 0002575 0018594 0082959 0002575 0018594 1003734 0873194
TABLE 416 DIFFUSE OVER-ALL EXCHANGE FACTORS F(IJ)
I I J I 2 10 5 4 1I
1 0000005 0000035 0000093 0000005 0000035 0000112 0829715 2 0000035 0000250 0000671 0000035 0000250 0000812 0827947 10 0000093 0000671 0001804 0000093 0000671 0002183 0494483 5 0000005 0000035 0000093 0000005 0000035 0000112 0829715 4 0000035 0000250 0000671 0000035 0000250 0000812 0827947 11 0000112 0000812 0002183 0000112 0000812 0OOOOTO o0455
FOLDOUT FRAME -FOLQOLER fr
INTERIOR SURFACES OF BUS 63
TARLE 417 CONFLIGURATION FACTORS FA(IJ)I
I 3 6 7 10 8
3 00 0199820 0200040 0200040 0200040 6 0199820 00 0200040 0200040 0200040 7 0200040 0200040 00 0199820 0200040 10 0200040 0200040 0199820 00 0200040 8 0200040 0200040 0200040 0200040 00 9 0200040 0200040 0200040 0200040 0199820
TABLE 418 COEFFICIENT MATRIX AA(IJ)
I 3 6 7 t0 8
3 1000000 -0159856 -0160032 -0160032 -0160032 6 -0159856 1000000 -0160032 -0160032 -0160032 7 -0030006 -0030006 -1000000 -0029973 -0030006 10 -0030006 -0030006 -0029973 1000000 -0030006 8 -003000f -0030006 -0030006 -0030006 1000000 9 -0030006 -0030006 -0030006 -0030006 -0029973
TABLE 419 INVERSE OF COEFFICIENT MATRIX BB(IJ)
I I J 3 6 7 10 8
3 1057707 0195533 0220391 0220391 0220391 6 0195533 1057709 0220391 0220391 0220301 7 0041323 0041323 1017-411 0046514 0046545 10 0041323 0041323 0046514 1017411 0046545 8 0041323 0041323 0046545 0046545 1017412 9 0041323 0041323 0046545 0046545 0046514
__4 TABLE 420 DIFFUSE OVER-ALL EXCHANGE FACTORS FFIJl
I 3 6 7 10 8
3 0002885 0009777 0046833 0046833 0046833 6 0009777 0002885 0046833 0046833 0046833 7 0046833 0046833 0083875 0224043 0224193
FO-L1U I0 0046833 0046833 0224043 0083875 0224193 8 0046833 0046833 0224193 0224193 0083875
9
0200040 0200040 0200040 0200040 0199820 00
9
-0160032 -0160032 -0030006 -0030006 -0029973 1000000
9
0220391 0220391 0046545 0046545 0046514 1017412
_
9
0046833 0046833 0224193 0224193 0224043 rfltlYO
T5_ 64 9 0046833 0046833 0224193 0224193 0-224043 0083875
tt TABLE 421 DIMENSIONAL CONDUCTION LINKS AKIIJ) BTUHRDEG t _
I 11 2 3 4 5 6
1 2 3 4 5 6 7 8 9
10 11 12
00 0058000 00 00 00 00 00 000 00 00 00 00
0058000 00 00 00 00 00 00
0 00 00 00 00
00 00 00 00 00 00 0071100 0471490 0560795 0033714 00 00
00 00 00 00 0058000 00 00
00 00 00 00
00 00 00 0058000 00 00 00
7 760 00 00 00 00
00 00 00 00 00 00 0074391 0667263 0861395 0034437 00 00
S TABLE 421 CONTINUED
I J 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9
10 11 12
00 00 0071100 00 00 0074391 00 0070706 0072435 00 00 00
00 00 0471490 00 00 0667263 0070706 00 00 0033626 00 00
00 00 0560795 00 00 0861395 0072435 00 00 0034012 00 00
00 00 0033714 00 00 0034437 00 0033626 0034012 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00
FnIFODOT Mfl ERAMEamp (-FOLDOUT FRAME
CHAPTER 5
NUMERICAL SOLUTION OF NODAL HEAT BALANCE EQUATIONS
51 A RECAPITULATION
The dimensionless nodal heat balance equations for the hyposhy
thetical spacecraft presented in Chapter 3 and designated by the
set (336) are simultaneous non-linear first order ordinary
differential equations They can be compactly described by (511)
which is written for an arbitrary node i
MjT4 Q - Fgtkc _ + E s _ + 7 9Ljj
1 j1a
K K2j)-
ji ji
where
k C M __0 subscript o refers to a convenient reference
0kC1 0 surface for which the dimensionless time t
is related to the physical time t according
to T = (k C A)t One of the surfaces in 0 0
the system may suitably serve for this purshy
pose
65
Cq + q)A Q (d) Texcept Q7 which is
1 0
a S A cos 9S ( T Also Q10 = Q 01 2
ER (kd )7 T0
-C = conductance of multi-layer insulation (R) 1
The three nonvanishing conductances are C7-e
C012 and C12-1 0 bull
i- overall radiant exchange factor between A and A
The indicated summand refers to all A s which have
direct andor indirect radiant exchange with A
W-hen approriate A includes the empty space A
similar interpretation should be given toZ Xj 0
K conductance between A and A Here the indicated
summand refers only to those surfaces which have physishy
cal contact with A
All other quantities have previously been defined in Chapter 3
The introduotibn of the dimensionless parameter MI is to obtain
a Iiniform time scale for all component-surfaces in the system An
alternative is to replace the left hand side of (511) by
Ci A d4
k dt I
and the integration is to be carried out in the physical time domain
The latter is indeed adopted in our computer program for which separate
66
documents were prepared and submitted to JPL
52 NODAL HEAT BALANCE EQUATIONS IN FINITE DIFFERENCE FORM AND THE METHOD OF SOLUTION
Equation (511) may be integrated from T to T + AT AT being
deg-small indrement to give
M (T+ AT) - (T)] = Q AT- (c + (+Y4
j
+ i( fit a + _ ft dTplusmnT I
_j f dT) - d) K_ dT
J ji
iK T (521)
j - TI-AT
In (521) all integrals J imply f They will now be replaced by the T shy
the approximate two point integration formula namely
T+ AT
J )dt A-rf(l - a)4)Cr) + d4Kt + AT)] (522a)
T AT
f aT AT[(l - cd4f(t) + a44(T + AT)] (522b) T
in which a is the so-called integration parameter and is a pure numeral
between 0 and 1 This scheme of approximation has been used by Strong
67
and Emslie [ 8 ] in their treatment of relatively simple systems inshy
yolving combined radiation and conduction heat transport Upon
substituting (522ab) into (521) followed by some rearrangeshy
ment one obtains
4
f4AT) + fl(T+ AT) +f ( 0
where
= (c gt+ - 7 _ )
fl (kd) K AT
ji
M f -Qt _t (T) +l _a-k K i (T)
Ta AT (d) _12 ji
(523)
(523a)
(523b)
1-1
(lcd) 1
5
4t
-) (1cfS d(T) + (T)
+
j68
4t1 ar)- i [CI 4tct + AT)
68
+ ~ _ 57(T + AT) + Zd~ Cplusmn rl(23c0j 32
Here again ) E 4x (T) and 4j (T + Ai) Ht (T + AT)
The equation set (523) was solved by a modified Gauss-Seidel
iterative procedure Current values of (T)were used as starting
approximations for (T + AT) and the Newton-Raphson method was used
for finding the physically meaningful-roots If we denote the current
root by (T + AT) then the new value of 0i (T + AT) is calculated from
Ti+ AT) e=w (T + AT) old plusmn Y T + AT) - (T plusmn AT) old] (524)
where y is the acceleration factor The next equation in the set which
is for A +I is solved in precisely the same manner using improved values
of t + AT) as soon as they are generated It has been found that in
general a and y taken together have a decisive influence on the convershy
gence rate of the iterative process
The steady state solution was obtained in a similar manner except
of course that it does not involve a By setting the left hand side
of (511) to zero and rearranging we obtain
g4i -i-g+i + go 0 (525)
where
(525a)4 1 -1 ++ plusmnC
69
g 1 (d) K _ (52 5b)
jjigo k )-d)( shy
go - K
j
The iteration began with a suitable set of guessed values of f For
bach node the Newton-Raphson procedure of successive approximations
was again used for determining the root of (525) The new value of
P is calculated according to
+ Y ( Snew 0i old i ol d
Thus for either the transient or steady state computations there are
two distinct iterations involved One is associated with the Newton-
Raphson method of evaluating the roots of (523) or (525) and another
is associated with the Gauss-Seidel procedure In the former when
the successive 4k is differ by less than 001 percent the result is
considered satisfactory and the calculation will proceed to the next
node The Gauss-Seidel iteration will terminate when the residual of
each nodal equation is less than a specified value This residual is
a measure of the deviation from the perfect heat balance required for
70
each node Thus we may write
g g (527a)i O i i i g4 i 4
for the steady state problem and
A t1 + f 1 4) + f4 1 4
(52 7b)
for the transient case For the hypothetical spacecraft described in
Chapter 3 go i or f0 i is of the order 102 for all sunlit surfaces and
hence we have arbitrarily set HB lt 10 as the condition to stop the
iteration This value was chosen essentially from a consideration of
the computer time on one hand and the desired accuracy of computed reshy
sults on the other hand
53 DETERMINATION OF OPTIMUM y AND a
To ascertain the optimum values of y and a for rapid convergence
of the iterative process a series of preliminary computer experishy
ments was conducted The results for the steady case with normal inshy
cidence from the sun and with a distribution of electronic package heating
similar to Mariner 64 spacecraftt are shown in TABLE 51 A redistrishy
bution of the heating source inside the bus or a change in the relative
importance of the radiative and conductive mode of heat transfer would
alter the number of iterations required to achieve the same heat balance
tThis calls for the following distribution A-29 percent A6-24 percent A8-36 percent and A9 -II percent (See Figs 31 and 32 for identifishycation)
71
TABLE 51
INFLUENCE OF ACCELERATION FACTOR ON THE CONVERGENCE RATE OF THE ITERATIVE PROCEDURE-STEADY STATE
Acceleration No of iterations factor y required to achieve
HB 10 - 3 1
03 gt80
05 65
07 38
09 24
11 14
13 10
15 17
17 33
19 gt80
Note Normal incidence from the sun distribution of heat generation
inside the bus similar to Mariner 64 spacecraft
72
ie HBi lt10 -3I -
However these changes if kept within reasonable
limits have only a minor effect on the optimum value of ywhich is
around 13 This finding is at variance with the suggestion made by
Strong and Emslie [8]
The transient results are presented in TABLE 52 They were obshy
tained for the assumed condition that the spacecraft had initially a
uniform temperature of 400degR and at an arbitrarily chosen instant
t = 0 the internal heating over bus surfaces A3 A6 A8 and A9 was
suddenly raised to their full power and simultaneously the spaceshy
icrafts exterior surfaces were exposedto suns irradiation To a
large extent the table is self-explanatory It is seen that the
acceleration factor has a dominant influence on the convergence rate
while the integration parameter plays only a minor role The optishy
mum vAlue of the acceleration factor has been found to be 11 under
the stated conditions
The computed temperature data for the 12 nodal surfaces of the
hypothetical spacecraft at 10 100 and 300 minutes after the thershy
mal disturbance are listed in TABLE 53 Included are the data for
the steady state temperatures As one might expect the solar
panels antenna and the exterior surface of the multi-layer inshy
sulation (A1) respond very rapidly to the changing environmental
condition
73
TABLE 52
INFLUENCE OF INTEGRATION PARAMETER AND ACCELERATION FACTOR ON THE CON-VERGENCE RATE OF THE ITERATIVE PROCEDURE--TRANSIENT CASE
No of iterations required to achieve Integration Acceleration HB 10-3
parameter factor Prototype time interval Min y 0-10 90-100 290-300
02 09 5 5 4
04 09 5 5 5
06 09 6 6 5
08 09 6 6 5
10 09 6 7 6
02 12 5 4 3
04 11 5 4 3
06 11 5 4 3
08 11 5 4 3
10 11 5 4 3
02 13 9 7 6
04 13 9 8 6
06 13 9 8 6
08 13 9 7 6
10 1 9 7 6
02 15 16 13 10
04 15 16 12 10
06 15 16 11 9
08 15 16 1 9
10 15 16 11 9
02 17 31 24 18
04 17 31 21 17
06 17 31 26 19
08 17 31 24 18
1810 17 31 23
Note 1 Prototype At = 10 min
2 Normal incidence from the sun distribution of heat
generation inside the bus similar to Mlariner 6i spacecraft
74
TABLE 53
THERMAL RESPONSE BEHAVIOR OF THE
HYPOTHETICAL SPACECRAFT
Uniform Initial Temperature = 4000R
Temperature OR
t = 10 min t =100 min t = 300 min Steady State
A1 5181 5978 5979 5980
A2 5205 6038 6047 6048
A 4294 5191 5339 5344
A 5181 5976 5979 59794
A 5202 6028 6042 6043
A 4122 4988 5244 52556
A 4015 4796 5277 5295
A8 4207 4995 5240 5250
A9 -4159 5116 5462 5475
A 0 4030 4908 5277 5290
All 4763 5680 5681 5681
A 2(= A 0) 1528 1811 1848 1850
Note 1 Normal incidence from the sun solar constant taken to be 442 Btuhr-ft2
2 Cubical bus consists ofsix 4 ft2 surfaces total electronic
package dissipation 1100 Btuhr distributed over surfaces A3 (29 percent) A6 (24 percent) A8 (36 percent) and Ag (ii pershy
cent)
75
Chapter 6
RESULTS DISCUSSIONS AND CONCLUSIONS
The similitude requirements for the thermal modeling of the hyshy
pothetical spacecraft have been given in Section 34 In the present
study it was assumed that the models entail errors in the desired
values of kd Cd e and E (the prime refers to the interior of the
surfaces) There was no error arising from the requirement of geoshy
metric similarity nor was there error in the heat fluxes q and q
As has been noted previously this choice of error distribution is
based primarily on a consideration of practical model fabrication
with the present-day technology It has no bearing on the theory
itself
The general theory of imperfect modeling presented in Chapter
2 is described logically in terms of dimensionless quantities Howshy
ever in the present instance it is possible to speak of errors in
the dimensional quantities like (kd) and (Cd) since not only are they
the only quantities which may be varied in their respective dimensionshy
less groups but also they are physically apprehended more readily
by the designer
61 SPACECRAFT SUBJECTED TO SIMPLE HEATING AND COOLING
To ascertain the usefulness and limitations of the proposed theor
several sets of errors in the modeling parameters have been assigned
TABLE 61 lists the values for the first of such sets In the table
76
TABLE 61 Errors inmodeling parameters
Percentage Error in Modeling Parameters
Model MI Model MII Model MIII Surface Okd Cd 06 6E 6kd 6 Cd r GEt 6kd Cd 66 6St
K 100 X 00 X100 X 100 X 100 lt 100 00 X 100 X 100 100 X 1OOO00
A 0 0 0 0 0 0 0 0 0 0 0 0
A2 0 0 0 0 0 0 0 0 0 0 0 0
A3 250 250 -180 100 550 525 -180 200 250 250 -288 100
A4 0 0 0 0 0 0 0 0 0 0 0 0
As 0 0 0 0 0 0 0 0 0 0 0 0
A6 250 200 -200 100 450 500 -200 200 250 200 -320 100
A 200 250 150 -80 400 550 270 -80 200 250 150 -147
A8 250 250 -180 -80 575 575 -180 -80 250 250 -259 -115
A9 250 250 250 -80 450 450 450 -60 250 250 250 -160
A10 200 250 150 -80 440 600 330 -80 200 250 150 -128
A 0 0 0 0 0 0 0 0 0 0 0
A12 (=A0 ) 0 0 150 0 0 0 330 0 0 0 150 0
6 refers to the exterior of the surface in question and 6 refers to its interior
and others which follow modei MI refers to the original imperfect
model MII refers to that for which the positive errors are modified
(in this case they become more positivp) while the negative errors
remain unaltered and MIII denotes that for which the negative errors
are modified (in this case they become more negative) while the posishy
tive errors are unchanged Thus the error state of model MI would
be represented by the point P in the multidimensional hyperspace reshy
ferred to in Chapter 2 Likewise the-error states of model MII and
- )MIII are represented respectively by Pz(+) and P
Initially the spacecraft is assumed to have a uniform temperashy
ture of 4000R At some arbitrarily selected instant t = 0 intershy
nal heating within the spacecrafts bus and of a distribution as preshy
scribed in TABLE 53 is assumed to begin instantaneously and at the
same instant its exterior surfaces are exposed to suns irradiation
Subsequent to the establishment of a steady thermal condition the
spacecraft is conceived to move into the shade of a planet thus cutshy
ting off the suns irradiation In this manner a cooling transient
is created
Using the equation set (333) and the computational procedure
explained in Chapter 5 temperature data for the heating transient
the steady state and the cooling transient are obtained for the proshy
totype (or equivalently the perfect model) and for the three models
The results are summarized in a table set 62 for normal incidence
of the solar irradiation and in table set 63 for 450 oblique incishy
dence The prototype temperatures are denoted by Tp and those of
the models by J Ji and The errors in the latter temperatures
78
TABLE 62(a) Model temperature before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A A( A Temperature 1t
TJ_ I I T Iv TMc
A 5181 5181 0 5180 -0i 5180 -01 -0-01 -005 006 5181 0
A2 5205 5200 -05 $200 -05 5198 -07 -005 -032 7037 5204 -01
A_ 4294 4262 -32 4219 -75 4275 -19 -389 188 -201 4282 -12
A4 5181 51810 -01 5180 -01 5180 -01 ~0 -004 -004 5181 0
A 5202 5200 -02 5199 -03 5197 -05 -02 -025 027 5202 0
6 4122 4114 -08 4092 -30 4120 -02 -195 098 -097 4123 01
A7 4015 4009 -06 4006 -09 4009 -06 -029 -011 -040 4013 -02
As 4207 4190 -17 4154 -53 4200 -07 -323 137 -186 4209 02 A9 4159 4129 -30 4109 -50 4130 -29 -176 023 -153 4144 -15
A1 0 4030 4018 -12 4010 -20 4016 -14 -066 -029 -095 4027 -03
A1 47E3 4763 0 4763 0 4763 0 -0 -0 -0 4763 0
12(=A01 )1A (-A
1528 OR
1533 05 32 04 154- 13 007 118 1n8
111
1522 1522 ~06
All temperatures in 0R
TABLE 62(b) Model temperatures before and after cortection
Heating Transient 30 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Surface
Prototype
or Perfect Model
Temperature
Temper atures of Imperfect Models
Model MI Model MII Model MIII
A ()
At
Model Teimperature
after Correction
A1 5907 5906 -01 5906 -01 5905 -02 -004 -007 -011 5907 0
A2 5949 5942 -07 5939 -10 5939 -10 -022 -043 -065 5948 -01
A
A4
A
A
4745
5905
5939
4403
4690 -55
5905 I 0
5935 0
4368 -35
4594
5905
5934
4303
-151
0
-01
-00
4720
5905
59341
4383
-25
0
01
20
-861
-002
-011
-583
447
-004
-020
224
-414
-006
-031
-359
4731
5905
5939
4404
-14
0
0
01
A7 4131 4093 -38 4065 -66 4091 -40 -252 -030 -282 4121 -10
A8 4506 4489 -17 4410 -96 4516 10 -7j10 409 -301 4519 13
A9 4454 4387 -67 4328 -126 4396 -58 -524 137 -387 4426 -28
A1 0 4243 4183 -60 4129 -114 4183 -60 -479 -001 -480 4231 -12
A1 1 5476 5476 0 5476 0 5476 0 001 -0 001 5476 0
A (=A1o0 ) 1722 1723 01 1719 -03 1729 07 -036 087 051 1717 -05
All temperatures in OR
TABLE 62(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Prototype Model Temperature
Surface or Perfect
Model Model MI Model MIT Model MIII
At ) 4( A after Correction
Temperature A t
A1 5978 5977 -01 5976 -02 5976 -02 -006 -008 -014 5978 0
A2 6038 6031 -07 6026 -12 6027 -11 -036 -047 -083 6039 01
A3 5191 5243 52 5150 -41 5325 134 -835 -1224 389 5204 13
A4 5976 5975 -01 5974 -02 5975 -01 -006 006 -012 5976 0
A5 6028 6021 -07 6017 -11 6019 -09 -037 -033 -070 6028 0
A6 4988 4998 10 4880 -i08 5059 71 -1052 913 -139 5011 23
A7 4796 4694 -102 4541 -255 4711 -85 -1368 255 -1113 4805 09
A8 4995 5050 55 4948 -47 5123 128 -914 1081 167 5034 39
A9 5116 5069 -47 4942 -174 5119 03 - 1 1 36 751 -385 5108 -08
A 0 4908 4862 -46 4725 -183 4900 -08 -1225 579 -646 4926 18
A 5680 5681 01 5681 01 5681 01 ~0 -0 0 5680 0
A2 (=Alot1) 1811 1816 05 1802 -09 1830 19 -118 222 104 1805 -06
All temperatures in 0R
TABLE 62(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
II
-
Surface
Prototyp= or Perfect
Model Temperature
Temperatures of Imperfect Models
Model 1I Model MI Model MIII
iv TbUII E1 I I A AA
t
Model Temperatureafter Correction
A 5980 5979 -01 5979 -01 5979 -01 -003 -005 -008 5980 0
A2 6047 6044 -03 6042 -05 6042 -05 -015 -030 -045 6049 02
00 A
A4
5339
5979
5507
5978
168
-01
5476
5978
137
-01
5636
5978
297
-01
-285
003
1922
-003
1637
-006
5344
5979
05
0
A 6042 6041 -0 6039 -03 6039 -03 -017 -019 -036 6044 02
5244 5421 177 5383 139 5554 310 -340 1989 1649 5256 12
A7 5277 5406 129 5338 61 5505 228 -614 1468 854 5321 44
A3 5240 5430 190 5392 152 5553 313 -346 1832 1486 5282 42
A9
A10
11 5462
5277
5583
5412
121
135
5517
5355
55
78
5696
5513
234
236
-593
-511
1684
1504
1091
993
5474
5313
12
36
A 568 01 5682 01 5682 01 002 002 004 5681 0
A12 (=A 1 ) 1849 1877 28 1869 20 1906 57 -063 442 379 1839 -10
All temperatures in OR
TABLE 62(e) Model temperatures before and after correction
HeatingTransient 500 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model 1I Model MIl Model MIII after Correction
Surface Model A( ( Temperature ht
A1 5980 5979 -01 5979 -01 5979 -01 -001 -005 -006 5980 0
A2 6048 6045 -03 6045 -03 6044 -04 -005 -027 -032 6048 0
A3 5344 5528 184 5517 173 5663 319 -100 2011 1911 5327 -17
A 5979 5979 0 5979 0 5978 -01 -001 -003 -004 5979 0
A 6043 6042 -01 6042 -01 6041 -02 -004 -015 -019 6044 01
A6 5253 5453 200 5444 191 5595 342 -077 2123 2046 5235 J8
7 5294 5459 16 548 144 5559 275 -189 1637 1448 5301 07
A3 5250 5461 211 5449 199 5591 341 -104 1940 1836 5269 19
A 5475 5623 148 5591 116 5745 270 -293 1819 1526 5468 -07
A10 5290 5455 165 5435 145 5564 274 -178 1638 1460 5297 07 A1 5681 5632 01 5682 01 5682 01 002 002 004 5681 0
A12(=A 0 ) 1850 1832 32 1879 29 1913 63 -023 470 447 1844 J06 All temperatures in OR
TABLE 62(f) Model temperatures before and after correction
Steady State
Temperatures of Imperfect Models
Prototype or Perfect Model MI Model MI Model MIII
Model Temperature after Correction
Surface Model
Temperature
TP I
I (+)
A
A(-)
At
A1 5979 5979 0 5979 0 5979 0 -001 -005 -006 5980 01
A2 6048 6045 -03 6045 -03 6044 -04 -003 -026 -029 6048 0
A3 5344 5530 186 5522 178 5665 321 -068 2020 1952 5335 -09
A4 5979 597910 5979 0 5979 O -0 -002 -0-03 5979 0
A 6043 60421 6042 -01 6042 -01 -001 L014 -015 6044 01
A 5253 54551 202 5452 199 5598 345 -031 2136 2105 5244 -09
A 5295 5463 168 5450 155 5574 279 -115 1654 1539 5309 14
A8 5250 5463 213 5456 206 5594 344 -Ot60 1951 1891 5274 24
A9 5475 5626 151 5600 125 5749 274 -238 1833 1595 5467 -08
A10 5290 5458 168 5445 155 5568 278 -119 1652 1533 5304 14
Al 5681 5682 01 5682 01 5682 01 002 002 004 5681 0
A12(=AID)0 1850 1882 32 1880 30 1914 64 -016 473 457 1836 -14
All temperatures in OR
TABLE 62(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Temperatures of Imperfect Models
Prototype Model Temperature or Perfect Model MI Model MIX Model MIII after Corre6tion
Surface Model Temperature T Fir 1
A() A A-) jt__
A1 3765 3765 0 3765 0 3764 -01 001 -012 -011 3766 01
A2 3905 3901 -04 3902 -03 3897 -08 003 -066 -063 3908 03
A 5162 5396 234 5404 242 5547 385 078 2257 2335 5162 0
I A44 A5 j
3764
899
3763
3895
-01
-04
3764
3895
0
-04
3762
3889
-02
0
001
005
-014
-080
-013
-075
3765
3902
01
03
A6 5130 5368 238 5379 249 5523 393 094 2314 2408 5127 -03
A7 5270 5445 175 5436 166 5557 287 -081 1673 1592 5286 16
A8 5226 5445 219 5442 216 5578 352 -034 1978 1944 5251 25
A9 5434 5596 162 5574 i40 5722 288 -202 1880 1678 5428 -06
A10 5255 5432 177 5424 169 5545 290 -075 1686 1611 5271 16
All Al2 (=AI0 )
3952 1504
3953 1565
01 61
3953 1563
01 59
3953 1621
01 17
004 -018
004 837
008 819
3952 1483I
0
All temperatures in OR
TABLE 62(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
1 Temperatures of Imperfect Models
Prototype - M Model Temperature
or Perfect M M after Correction Surface Model
Temperature Tp Tt
I E
M1 TM I TDI Ic E
A( + ) A(- )
TIc
_
F _
A1 2298 2297 -01 2298 - 2295 -03 007 -036 -029 2300 02
A2 2757 2754 -03 2757 0 2745 -02 025 -133 -108 2765 08
A3 5003 5246 243- 5251 247 5409 406 045 2446 2491 4997 -06
A4 2290 2289 -01 2291 01 2285 -05 014 -058 -044 2294 04
A5 2727 2723 -04 2729 02 2709 0 050 -214 -164 2740 13
AA6 4918 5184 266 5196 278 5356 436 111 2580 2691 4914 -04
A7 5088 5286 198 5283 195 5413 325 -024 1901 1877 5098 10
A3 51041 5332 228 532-8 224 5474 370 -040 2123 2083 5124 20
A9 5256 5436 180 5416 160 5577 321 -183 2113 1930 5243 -13
A10 5083 5278 195 5272 189 5406 323 -054 1909 1855 5093 10
A11
A12 (=A10 )
2356
1348
2360 04
1427 79
2361
1426
05
78
2361
1498
05
150
016
-015
020
1051
036
1036
2356
1324
0
-24
All temperatures iii OR
TABLE 63 (a) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Teoeaures of imperfect Models
Prototype Model Temperature
or Perfect Model M11I after Correction Surface Model 7 0l MI A__ A ) A[Temperature T
Tp T I TM TV1 vM 4 T
4774 4773 4773 -01 4773 -01 -001 -007 -008 4774 0I fI -
2 3701 3695 -06 3695 -06 3693 -08 -006 -044-05 3701 0A 4257 4237 -20 4198 -59 4253 -04 -35 6 - 74249 I -
A4 4810 4810 0 4810 0 4809 -01 -001 -008 -007 4810 0
As 4993 14988 -05 4988 -05 4986 -07 -006 -033 -039 4992 -01
4304 4269 -35 4218 -86 4277 7 -457 4303 -01 A7 4017 4011 - 06 4007 -10 4010 -07 -034 -011 -045 4016 -01I
A8 4213 4195 18 4158 -55 4205 -08 -33 140 -193 4215 02
A 4169 4138 -31 4117 -52 4140 -29 -193 027 -166 4155 -14
AI 4036 4922 -14 4014 -22 4021 -15 -078 -026 -104 4033 -03
All 4476 4476 0 4476 0 4476 9 001 0 Q01 4476 0
A 1421 1429 -08 1427 -06 1439 -18 -010 160 150 1414 -07
All temperatures in OR
TABLE 63(b) Model temperatures before and after correction
Heating Transient 30 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
A2
V Proottype PoPerfect
Mode Temperature
5355
3353
Temperatures of Imperfect Models
Model MI ModelM-ll Model MIlI
T P -- Ilz~ v L ~I I
AI
5354 5354 -02 53541 -02 3336 -17 3331 -22 3328 -25
f
At T
-0041-01
-037 -416
I
-016
-153
Model Temperatur after Correction
F6
5356 3351 -02
A 577 459 -08 4488 -89 4614 37 -718 -679 -039 4572
A
A
A
5419
5683
4806
51-01
8 3 4732
09
-74
5417
56 4608
-02
-12
-198
5417
5671
4757
-02
-12
-49
-006
030
-1118
-009
-048
364
-015
-078
-754
5419
5682
1 6 -4808
0
-0
02
A7 4149 410 7 -42 4075 -74 4106 -43 -292 -023 -315 4139 -10
A 4535 4517 -18 4437 -102 4546 11 -757 438 -319 4549 14
A~i I 14505 4436 -69 4368 -137 4448 57 - 8 182 -426 4478 -27
A1 4276 4213 1 -63 4151 -125 4214 -62 -554 029 525 4265
A I 4980 4980 0 4980 0 80 0 001 -0 001 -4980 0
Aj(tAoI 1555 1556 01 1550 05 1566 11 -057 147 090 1547 -08
All temperatures in OR
TABLE 63(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temoeratures of Imperfect Models
Prototyne Model Temperature or Perfect Model MI Model MI Model MIII after C6rrecLion
Surface Model A(+) A-Temperature A A
A1 5463 5461 -02 5460 -03 5460 -03 -008 -013 -021 5463 0
A2 2962 2923 -39 2903 -59 2906 -56 -173 -251 -424 2965 03
A 4981 5089 108 5002 21 5200 21-9 -781 1667 886 5000 19
A4 5534 5531 -03 5530 -04 5530 -04 -015 -013 -028 5534 0
A 5816 5802 -14 5794 -22 5798 -18 -078 -068 -146 5817 01
A6 5528 5560 32 5403 -125- 5657 129 -1410 1445 035 5557 29
A7 4906 4811 -95 4632 -274 4842 -64 -1604 455 -1149 4926 20
A8 5079 5151 72 5040 -39 5237 158 -999 1283 284 5123 44
A9 5249 5227 -22 5083 -166 5297 48 -1284 1049 -235 5250 01
A10 5026 4997 -29 4841 -185 5054 28 -1400 8L45 -555 5053 27
A 5206 5207 01 5207 01 5207 01 -0 001 001 5207 01
A (=A )0 1672 1681 09 1660 -12 1706 34 -185 369 184 1663 -09
All temperatures in OR
TABLE 63(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal beating Temperatures of Imperfect Models -
Model TemperaturePrtotypeor Perfect Model MI Model MIl Model MII after Correction Surface Model _ _ _ _ __t (_+ )
Temperature j T At
^ 5465 5464 -01 5464 -01 5464 -01 -002 -Q08 -0 i r0
A2 2908 2898 -10 2886 -22 2887 -21 -108 9-a i 1-17
(0 A 5152 5399 247 5387 235 5564 412 -104 2472 10__2o
A4 5538 5537 -01 5536 -02 5336 -02 -007 -008 -051 5538 0
A5 5834 5829 -05 5825 -09 5826 -08 -033 -038 -071 58 02
A6 5734 59 6 202 5884 150 6102 368 -462 2485 2023 5734 0 A7 5401 5576 175 5515 114 5700 299 -542 1860 1318 5444 43
A8 5331 5556 225 5525 194 5699 368 -277 2138 1861 5370 39
AS9 U96 5763 167 570-3 107 5902 306 -534 2076 1542 5609 13
A10 5400 5580 180 5530 130 5706 306 -446 1887 1441 5436
Aia 5209 5210 01 5210 01 5210 01 002 004 006 5209 0
A1(=A) 1722 1767 45 1759 37 1812 90 073 671 598 1707 -15
All temperatures in OR
TABLE 63(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype dModel Temperature or Perfect Model MI Model Ml Model IIl after Correction
Surface Model A+_A( ) At Temperature I j
A_ 5455 5465 0 5465 0 5464 -01 001 -006 -005 - 5465 0
A2 2914 12917 03 2919 05 2910 -04 019 -105 -086 2926 12
A3 5157 5417 260 5427 270 5587 430 085 2541 2626 5155 -02 A4 5538 5537 -01 5537 -01 5537 -01 -003 -007 -010 5538 0
A5 5834 15831 -03 5829 -05 5828 -06 -018 -034 -052 5836 02
TI
A___ 573 595 218 5929 1906128 389 -250 2566 2316 5725 -14
A7 5413 5618 205 5602 189 5750 333 -140 1979 1839 5434 21
A3 5337 5581 244 5575 238 5729 392 -047 2216 2169 5364 27
A9 5605 15794 189 5766 161 5939 334 -253 2169 1916 5603 -02
A1 0 5410 95613 203 5598 188 5746 336 -i33 1981 1848 5428 l-8
A11 5209 15210 01 5210 01 5210 01 003 004 007 5209 O
o723 j1772 49 1770 47 1819 96 -023 696 673 1705 -18
All temperatures in OR
TABLE 63(f) Model temperatures before and aftercorrection
Steady State
Temperatures of Imperfect Models
Prototype Model MI Model MIl Model MIll Model Temperatureafter Correction
or Perfect Surface Model
Temperature1 Ej T1 4
r I______
T 5
) I
IMC IA A(-)
71Cz
5465 5465 0 5465 0 5464 -01 002 1-006 -004 5465 0
A2 2915 2919 04 2924 09 2913 702 045 -098 -053 2925 0
5158 5419 261 5431 273 5589 1 431 110 2546 2656 5153 j -05
A4 5538 5537 -01 5537 -01 5537 -01 -003 007 - 10 5538 0
A 5834 5831 -03 5829 -05 5828 -06 -015 -034 -o49 5836 02
A 5739 5958 219 5933 194 6130 391 -223 2572 2349 5723 -1
A 5414 5620 206 5610 196 5753 339 -088 1987 1899 5430 16
A8 5338 5582 244 5580 242 5731 393 -017- 2221 2204 5362 24
A0 5606 5796 190 5772 166 5942 336 -21Z 2175 1960 5600 -06
A1 o 5410 5615I _ 205
_ _ 5605
_ 195
_ 5748
_ 338
_ _ -092
_ 1987
_ 1895
_ 5425 15
A11 5209 5210 5210 01 5210 01 003 004 007 5209 0
A2(=f 0 1723 1773 50 1771 48 1819 96 7016 698 682 1704 -19
All temperatures in OR
TABLE 63(g) Model temperatures before and after correction
Cooling Transient 30 minutes after spacecraft moving into shade
Prototyv T mo eatures of Imperfect Models F SPrototyp Model Temperature
or Perfect Model MI Model MI I Model MIII after Correction ModelSurface FTemperature A A At )
__ _ I - T1 N Tz 1j _ _ Tc _I_ I__ _1 A 3665 3665 0 3665 0 36 01 03 -010 -007 3665 0
A2 J 2886 2890 04 2895 10 2883 -03 -104 -056 2896 10j0248 A 5125 5391 266 5408 283 5562 437 146 255 2701 5121 -04
A4 3705 3704 -01 3705 0 3703 -02 002 -015 -013 3706 01 AS 3904 3900 04 3902 -02 3895 -09 016 03081 -065 3907
A6 5295 5577 282 5607 312 5754 459 272 2651 2923 5285 -10
A7 5367 5582 215 5581 214 5714 347 -006 1975 1981 538-5 18
A8 5290 5540 250 554-5 255 5688 398 049 2211 2260 5314 24
A0 5526 5726 200 5713 187 5872 346 -119 2175 2056 5521 -05
A 5343 5557 214 5558 215 5688 345 013 1966 1979 5359 16
A 3831 3833 02 3833 02 3833 02 005 006 011 3831 0
A1 2 (=A1)0 1497 1571 74 1571 74 1635 138 001 964 965 1474 -23 All in OR___
Al] temperatures in oR
TABLE 63(h) Model temperatures beforeand after correcti6n
Cooling Transient 200 minutes after spacecraft moving into shad6
Surface
Prototype or Perfect
Model
TemperatureTp
Temperatures of Imperfect Models
Model MI Model Nil Model MIII
_ _(
T ENpound ~~Eua Tv IZ ~ I i~vI )
Model Temperatureafter Correction
____t___
1T
A2
2278
2699
2279
270 0J 01
01
2281
2707
03
08
2277 1-0i
2692 -07
016
060
-033 1 -017
-131 -071
2280
2707
02
08
A 5005 5253 248 5264 259 5419 414 097 2480 2577 4996 -09 1
A4
A
2288
2734
12288
2735
0
01
2291 1
2745
03
01
2285
2722
-03
-02
026
090
-052
-190
-026
-10
2291
2745
03
11
A6
7
A
L4924
5096
5109
5197
5306
5344
273
210
235
5217
5314
5346
293
218
237
5372
5437
5488
44B
34
379
185
078
023
2626
1958
2160
2811
2036
2183
4916 f-08
5102 06
5125 16
A9
A10
5262
5090
5451
5294 i
189
204
544 0
I 5298
178
208
5596
5425
334
335
r-l06
030
2160
1955
2054
1985
5246
5096
-16
06
JA
All
(=A )i
2340
1348
2344
1431
04
83
2346
1432
06
84
2346
1502
06
154
019
006 1067
041
1073
240
1323
0
-25
All temperatures in OR
are designated by t E-ii and Irii - Equation (238) is used
to evaluate A plusmn) which is the correction for all positive errors
and A the correction for all negative errors is evaluated from
(2313) The combined correction At is simply the algebraic sum
of A and A Finally the corrected model temperatures calcushy
lated according to (2314) are denoted by 9c and their errors are
c In either set of the foregoing tables data for the heating
transients are listed in tables (a) to (e) the steady state in tashy
ble (f) and the cooling transients in tables (g) and (h)
An inspection of these tables shows that the proposed correction
theory is indeed capable of providing good results when errors in
kd and in Cd of the original model MI are up to 25 percent and ershy
rors in surface emittance up to 20 percent (corresponding errors in
models MII and MIII are much larger) As an example take the case
of heating transient at t = 100 minutes after the sudden change of
the thermal condition Uncorrected model temperatures of errors up
to approximately 10degR in MI 260R in MII and 130R in MIII are brought
to within 4OR of the prototype value after correction [see TABLE 62(c)]
For the steady state corresponding errors of up to 213 206 and
345 0R are reduced to less than 250R [see TABLE 62(f)] At a time
30 minutes after cooling begins uncorrected model temperatures exshy
hibit errors of up to 238 249 and 3930R respectively for the
three models After correction the maximum error is 250R [see TA-
BLE 62(g)] Data for the 450 solar incidence show similar results
95
To test further the performance of the theory a second set of
errors in kd Cd s and c was studied They ate listed in TABLE
64 The magnitudes of the various errors are approximately the same
as before However both positive and negative errors now occur in
kd and Cd Moreover the modifications for model MIII are carried
out in such a manner that the negative errors become less negative
The results of this study are summarized in TABLES 65(a)-(h) when
the spacecraft receives the suns radiation at normal incidence and
in TABLES 66(a)-(h) for the 450 oblique incidence The spacecrafts
initial thermal condition the manner of the initiation of the heatshy
ing and cooling transients the relative distribution of internal
power dissipation within the bus etc are identical to those conshy
sidered in the earlier study All data again lend support to the
effectiveness of the proposed correction technique The errors lc
which remain in the corrected model temperatures are in-general
smaller than those shown in TABLES 62 and 63 Our limited experishy
ence seems to indicate that the theory would yield better results
when positive and negative errors coexist in each set of the modelshy
ing parameters
62 SPACECRAFT SUBJECTED TO CYCLIC HEATING AND COOLING
For an orbiting spacecraft the temperatures of its various comshy
ponents undergo continuous cyclic changes Under such condition
there was the concern that the error in the model temperatures might
accumulate and it was not certain that the proposed correction technique
96
TABLE 64 Errors in modeling parameters
Percentage error in modeling arameters
Surface
A1
x 66
Ao 0
Model MI
6 6 6Cd
00O o K A o 0 0 0
6lc6d
1
0
Model MII
6Cd 6Cd
iMo
0 0
6J00X6Id lcX66d
x 0o xAOO 0 0
Model MIII
X610K6l 6Cd
x 1000o 0 0
6lti o
0
A2
As
0
-200
0
200
0
-288
0
i00
0
-200
0
420
0
-288
0 0
200 -110
0
200
0
-180
0
100
A4 0 0 0 0 0 0 0 0 0 0 0 0
A5 0 0 0 0 0 0 0 0 0 0 0 0
A6 180 -150 -320 100 396 -150 -320 200 180 -82 -200 100
A7 200 150 150 -147 420 330 270 -147 200 150 150 -80
A8 80 -150 -259 -115 192 -150 -259 -115 80 -90 -180 -80
A9 -100 -150 250 -160 -100 -150 450 -160 -65 -82 250 -80
A10 -180 150 150 -128 -180 360 330 -128 -108 150 150 -80
A11
A12(=A )
0
0
0
0
0
15 0
0
0
0 0
0
0
330
0
0
0
0
0 0
150
0
0
6 refers to the exterior of the surface in question and 6 refers to its interior
TABLE 65(a) Model temperatures before and after correctibn
HeatingTransient 10 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models 1 Prototype IModel Temperature or Perfect Model MI Model MII Model MIIN after Correction
Surface Model _(+)__
TemperatureTp T E Tw Ei Iz ms I I ElI I I
A Tr117 FMIC
A 5181 5180 -1 5180 -01 5181 0 -001 -009 - 010 5181 0
A2 5205 5199 -06 5198 -07 f 520 1 -04 -004 -054 -058 5204 -01
A3 4294 42927 -02 4253 -41 4277 -17 -359 378 019 4290 -04
A4 5181 5180 rOl 5180 -01 5180 -01 -0 -007 -007 5181 0
A5 5202 5198 -04 5198 -04 5199 -03 -0 -041 -041 5202 0
A6 4122 4167 45 4167 45 4147 25 -001 517 516 4116 -06
A7 4015 4015 0 4013 -02 4014 -01 -016 015 -001 4015 0
A3 4207 4278 71 4276 69 4250 43 -019 709 690 4209 02
A 4159 4183 24 4178 19 4169 10 -043 355 312 4152 -07
A)o 4030 4030 0 4025 -05 4029 -01 -044 040 -004 401 01
All 4763 4763 0 4763 0 4763 0 -0 0 0 4763 0
A(=AO) 1528 j1534 06 1534 06 1531 03 -005 090 085 1526 -02
All temperatures in OR
TABLE 65(b) Model temperaturesbefore and after correction
Heating Transient 30 minutes after spacecraft subject to solarirradiation (noAmaln nemo) and internal heating
Temperatures of Imperfect Models
Protot pe Model MI Model M11 Model MIII T after Correction or PerfectI
Surface Model I _ -- A A A Temperature T E T 1ET
0 F
T ~5 x ~ I0 - -01 I0 -0 0 A1 5907 5906 -01 5905 -02 1 003 -012 -015 5907 0
A 5949 5940 - 09 5938 -11 5943 -06 -017 -070 -087 5949 0
A 4745 4780 35 4702 -43 4739 -0-6 -719 1052 333 4747 02
A4 5905 5905 10 5905 01 5905 0 ~0 -004 -004 5905 0
A5939 5936 -03 5936 -03 5937 -02 0 024 -024 59391 0
A 4403 4516 113 45114 11 4464 61 -024 1327 1303 4386 -17
4131A 4134 03 4118 13 4128 -03 -149 160 011 4133 02
Aa 4506 4662 156 465-1 145 4599 93 -0-99 1588 1489 4513 07
A0 4454 4525 71 -450S6 52 4484 30 - 3 1025 852 4439 -15
AI 4243 41261 18 4228 -15 4243 0 -300 443 143 4246 03
A1 5476 5476 0 5476 0 5476 10 001 -0 001 54760
A (=A i) 1722 1728 06 1726 041 1725 03 -023 098 075 1721 -01A01 temperatures in OR
All temperatures in degR
TABLE 65(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (normal incidence) and internal beating
Surface
Al
Prototypeor Perfect
Model Temperature
Tp
5978
I
I
Temneoatures of Imperfect Models
Model MI Model MIX Model MIII
__ _) _4 T r p ST I EplusmnIu TVU I I Ei I I
5977 -01 5976 -02 5977 -01 -003
----shy
-010 -013
Model Temperatureafter Correction
__ _t
T c I
5978 0
A2 6038 6031 -0 6029 -09 6034 -04 -019 -066 -079 6039 01
A 5191 5976
5420 5976
229 0
5374 5976
183 0
5313 5976
122 0
-426 -001
2724 -004
2298 -005
5190 5976
-01
A5
A6
6028
4988
6025
5237
-03
249
6025
5216
-03
228
6026
5124
-02
136
-007
-196
023
2872
-00
2676
6028
4969
0
-19
A 4796 4901 105 4825 289 4840 44 -700 1538 838 4817 21
A3 4995 5261 266 5235 240 5154 1519 -2t45 2720 2475 5014 19
A5116 5306 200 52t8 142 5209 93 -441 2456 2015 5105
A0 4908 5076 168 5019 lli 4993 85 -521 2097 1576 4918 10
A A
(=A1 5680 1811
5681 1837
01 26
5681 1831
01 20
5681 1824
0l 13
001 -056
002 346
003 290
5680 1808
0 -03
All temperatures in OR
TABLE 65(d) Model temperatures before and aftercorrection
Heating Transient 300 minutes after spacecraft subject to solar irradiation (normal incidence) and internal heating
Temperatures of Imperfect Models
Surface
AI
rPerfectPrototSe
Model
Temperature
I bull
5980
11
5979
I Tj I
-01
Model Nil
I TI
5979 -01
Model MIll
II iiD
5979 -01
A(+)
-001
-
A(-)
-007
Po p
I
bull -008
Model Temperature
after Correction T
7C
5g8 o 0
A047 6043 -04 6043 -04 6045 -02 -006 -043 -049 604R 0
A A
5339 5979
5659 5978
320 -01
5646 5978
307 -61
5522 5979
183 0
-113 -shy001
3462 -004
3349
-004
5
5
shy 1
0
5
A1 A5
A6
6042
8244
tI 604 0
5584
-0 2
340
604 0
5573
-02
329
604 2
5439
0
195
-005
-103
-021
3666
-0 25
13563
16043604 8
5228
0110
-16
A7 5277 5565 288 5542 265 5447 170 -217 2987 2770 5288 11
A8
A0
A0
5240
5462
5277
5585
5749
5566
345
287
289
5572
5718
5544
332
256
bull 267
5452
5622
5448
212
160
I 171
-119
-286
-198
3360
3209
3000
j 3241
2923
2802
5261
5457
5286 I
21
-05
09
All A 2(=A1o )
5681 1849
5682 1895
01 46
5682 189
01 5682 44 1874 01 25
002-026 004 553
006 527
5681 1843
0 -06
All temperatures in OR
TABLE 65(e) Model temperatures before and after correction
Heatinz Trim n nt 500 minutes alter spacecraft subject o solar irradiation (normal incidence) and internal heating
TemPoratures of imperfect Models
Surface
Prototypeor Perfect
Model Temperature
5980
Model MI or
z I Era
5979 -01
Model M11 Model MIII Pefectafter
T1TIM I I
5979 -01 5979 -01
A(+)
-001
I
(
-007
At
-008
Model Temperature Correction
T
T c
5980 0
2
A3
6048
5344
6044
5669
-04
325
6043
5662
-05
318
6046
5532
-02
188
-003
-072
-042
3491
-045
3419
6049
5337
01
-07
A
A6
5979
6043
5253
5979
[6042
5599
O
-01
346
5978
6041
5592
-01
-02
339
5979
6042
5453
0
-01
200
-0
-003
-060
-004
-021
3695
-004
-024
3635
5979
6044
5248
0
-01
-05
A7 5294 5592 298 5579 285 5473 179 -120 3031 2911 5315 21
A3 5250 5600 350 5591 341 5466 216 -075 3383 3308 5277 27
A1o
All
50
5290
5681
5768
5587
5682
29
297
01
5573
5689
269
283
01
5641
5467
5682
166
177
01
-228
-128
002
3234
3029
004
3006
2901
0106
5471
5309
5681
-04
19
0
__ 0)_ __1850 1898 48 1896 46 1876 26 -018 5 544 1837 -13
All temperatures in OR
TABLE 65(f) Model temperatures before and after correction
Steady State
SjTemperatures of Imperfect Models
Prototype Model MI Model MII Model MIII or Perfect - after Correction
Surface I Model (+) () I oeTemperaturebullT- rz t TM MUi I TFIZZI vr T
At T11c
A 5979 5979 0 5979 0 5979 0 -001 -007 -008 5980 01
6048 6044 -04 6044 -04 6046 -02 -003 -042 -045 6048 0
A_ 5344 5670 326 5662 318 5532 188 -069 3491 3422 5328 -16
A3j 5979 597940 0 5978 -01 5979 0 -0 -0044 -004O 59799 o
A5 6043 6042 -01 6041 -02 6042 -01 -003 -021 -025 6044 01
A 5253 5599 346 5593 340 5453 200 -057 3696 3639 5235 -18
A7 5295 5594 299 5581 286 5474 179 -113 3032 2919 5302 07
-A 5250 5600 350 5592 342 5467 217 -071 3383 3312 5269 19
A 5475 5769 294 5745 270 5642 167 -223 3234 3011 5468 -07
A10 5290 5588 f298 5575 285 5468 178 -122 3030 2908 5297 07
All 5681 5682 01 5682 01 5682 01 002 004 006 5681 0
A (=Ao) 1850 1898 48 1896 46 1876 26 -017 562 545 1844 06
All temperatures in OR
TABLE 65(g) Model temperatures before and after correction
Cooling Transient 30 minutes after sp-acecraft moving into shade
Temperatures of Imerfeoct Models
Prototype odel MI Model MII Model MIII Model Temperatureafter Correction
or Perfect
Surface Model A() A1 Temperature 1 T AtT
A 3765 3764 -01 3764 -01 3765 0 0 -019 -019 3766 01
A 3905 3897 -08 13897 -08 1 3901 -04 -0 -104 -104 3907 02
A 5162 5544 382 5547 385 5390 22-8 033 3889 3921 15151 -09
A4 3764 3762 -02 3762 0 3763 -01 -001 -023 -024 3765 01
A I 3899 3888 -01 3888 -01 3893 -06 -005 -129 -134 3902 03
A6 5130 5504 374 5499 369 5348 218 -043 3944 3901 5114 -16
7
A
52705226 5574
5579
304
353
5563
5572
293
346
5453
5445
183
219
-095
-q69 j 3061
3411
2966
3342
5277
5245
07
19
A9 5 34 5736 302 5713 279 5606 172 -214 3292 30 78 5428 -06 A I
A0 5255 5560 305 5549 294 5438 183 -102 3074 2972 5263 08
A 3952
I__1504
3953
1 _592
01
88 3954
1589 _
02
85
3953
1553 01
4 9 -004
0249 -007
96 -011 972
3952I1 9 5 0 -0
A 24 996 972 1495 -09
All temperatures in OR
TABLE 65(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft movinginto shade
Tomoeatures of Imperfect Models
Surface
Prototype or PerfectjModel
Model MI I Model Mll Model MIII A(+) K-A(-)
Model TempIrature after Correction
T e m p e r a t u r v TP
a T
1 I T t TIITNII
Al 1 2298 2294 -04 2294 -04 2296 -02 -001 -057 -058 2300 02
A2 2757 2742 -15 2741 -16 2751 -06 -008 -208 -216 2764 07
A 5003 5403 400 5398 395 5237 234 -039 4184 4145 4988 15
A4 2290 2284 -06 2284 -06 2287 -03 -001 -092 -093 2293 03
A5
A6
2727
4918 2703
5345
-24
427
2703
5343
-24
425
2717
5170
-10
252
-005
-016
-338
4434
-343
4418
2738
4903
-11
l5
A 5088 5426 338 5414 326 5290 202 -101 3430 3329 5693 05
A8 5104 5478 374 5470 365 5334 230 j - 81 3653 3572 5121 17
A0 5256 5596 340 5572 316 5450 194 -215 3692 3477 5248 -08
A10 5083 5420 337 5407 324 5285 202 120 3440 3320 5088 05
A 1 2356 2361 05 2363 07 2360 04 019 034 0153 2156 0
121348 1463 15 1459 111 1412 64 -033 1290 1257 1337 i1
All temperatures in OR
TABLE 66(a) Model temperaturesbefore and after correction
Heating Transient 10 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imperfect Models
Prototype I Model MIX Model MIII ModelModel Temperature or Perfect after Correction
Surface Model _(+) -
Temperature Tix Tm 1 Tua I I
A j 4774 4773 -01 4773 -01 4773 -01 -001 -012 -013 4774 0
A2 3701 3693 -08 3692 -09 3696 -05 -005 -074 -079 3700 -01
A 4257 4270 13 4234 -23 4251 -06 -330 463 133 4256 -01
A4 A 4810 4809 -01 4809 -01 4810 0 -0 -009 -009 4810 0
A5 4993 4987 -06 4987 -06 4988 -05 -0 -050 -050 4992 O1
A6 4304 4381 77 4380 76 4345 41 -015 922 907 4291 -13
A7 4017 4018 01 4016 -01 4017 0 -016 023 007 4017 0
A8 4213 4287 74 4286 73 4258 45 -009 745 736 4213 0
A9 4169 4198 29 4194 25 4182 13 -036 407 371 4151 -08
A1 0 4036 4038 02 4032 -04 4035 -01 -048 061 013 4036 0
A 4476 4476 0 4476 0 4476 0 001 -0 001 4476 0
A12 (=A1 0 1) 1421 1430 09 1430 1 09 1425 04 -006 125 119 1418 -03
All temperatures in OR
TABLE 66(b) Model temperatures before and after correction
Heating Transient 10 minutes after spacecraft subjectto solar irradiation (450 incidence) and internal heating
Temperatures of Imoerfect Models y -Model Temperature
Surface Prototypeor Perfect
Model bullModel MI Model NIX Model MII
A)(-) after Correction
Temperature V E T I T I EI1 x TMC
A 5356 5354 -02 5353 -03 5354 -02 -003 -021 -024 5356 0
A2 3353 3330 -23 3327 -26 3338 -15 -030 -190 -220 3352 -01
A5 4577 4670 93 4504 27 4612 35 -606 1468 862 4584 07
A4 5419 5418 -01 5418 -01 5419 0 -0 -01 600 549 0
As 5683 5677 -06 5677 -06 5679 -04 -002 -054 -056 5683 0
A6 4806 4992 186 4981 175- 4904 98 -098 2226 2128 4779 -27
A 4149 4160 11 4144 -05i 4151 02 -155 246 091 4151 02
A
A9
A
4535
4505
4276
471
4604
4312
176 4705 170
99 4589 84
36 4278 02 j
4641
4552
4287
106
47
11
-054
2142
-311
1794
1325
635
1740
1183
324
4537
4486
4279
02
-19
03
All 4980 4980 0 4980 0 4980 0 001 P02 4980 0 1A(=A 1555 1567 12 1563 08 1560 05 -033 17 143 1552 -03
All temperature in OR
TABLE 66(c) Model temperatures before and after correction
Heating Transient 100 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperatures of Imverfect Models
Prototype or Perfect
Model MI Model MI I iI
Model MIII
)
Model Temperature after Correction
Surface Model Temera tu
(I4 AA A t
Tf T I IT1 3S- T 7 I1 flvnI I
54 5451 I 2 541 -02 42 O-004 -017 -021 5463 0
A2 2962 2924 -38 2913 -49 2937 -25 -105 -319 -424 2967 05
A A01
41 5534
523 5533
I
8s2 -01 0
5250 269 5532 j-02 5153
5533 35534
172 -01
-388 -003 - 3
3530 -011 01
314 -014 01
4979 5534
-02 0
1
A5 5816 5809 -07 5807 -09 5811 -05 -016 -055 -073 5817 0i
A 6 552 -I8581 33 5825 297 5708 180 -331 3888 3557 5506 -22
A7 4906 5075 169 4996 90 4987 81 -729 2217 1488 4926 20 As 5079 5398 37
353 2
297 5270 191 052
f915093
14
A9 249 5508 259 546- 212 5383 134 -427 3166 2739 5234 - 5
A 5026 5251 235 5206 180 5150 124T j -505 2801 2296 5031I 05
A11 (A )[
5206 1672
5207 1718
01 46
5207 1710
j02 38
5207 1694
01 22
j002
074 004-
595 006 52060
1666 -06 A12 ltemperatues-0ndeg
All temperatures in OR
TABLE 66(d) Model temperatures before and after correction
Heating Transient 300 minutes after spacecraft subject to solar irradiation (450 incidence) and internal heating
Temperxatures of Imperfect Models
Prototype ior Perfect
bullModel NI Model MIX Model MIII
bull bull
Model Temperature after Correction
Surface i Model A(+) A(-) TemperatuTe i 1 TT Fi
T 9 0 I - m1 IvtI
A 5465 5464 -01 5464 -01 5464 -01 -0Oi -012 -012 0
A2 2908 2894J -14 2890 -18 2902 -06 -037 -179 -216 2916 08
A 5152 5564 412 5561 409 53914 bull24-2 -031 4310 4279 5137 -15
A4 5538 5537 -01 5536 -02 5537 -01 -03 -OlO_ -013 5538 0
5834 5829 -05 5828 -06 5831 -03 -o13 1-050 --063 5835 01
A6 5734 6140 406 6118 384 5964 230 -2p4 4473 4269 5713 -21 A 5401 5753 352 5735 334 5610 209 -166 3608 3442 5409 08
A8 5331 5725 395 5719 388 5574 243 -069 3855 3786 5348 17
A 5596 5945 349 5918 1322 5794 198 -250 3820 3570 5588 -08
A 5400 5752 352 5735 335 5609 209 -154 3612 3458 5406 06
A 1 5209 5210 01 5211 02 5210 01 003 007 010 -5209 0
il 2 (=Ai 1722 1794 72 1791 69 1761 39 -028 844 816 1713 -09
All temperatures in OR
TABLE 66(e) Model temperatures before and after correction
Heating Transient 500 minutes after spacecraft subject to solar irradiation (4 50 incidence) and internal heating
Tcmeatures of Imoerfect Models
Model TemperaturePrototype I or Perfect Model MI Model MIX Model MIII after Correction
Surface Model A___ A__ I Temperature A A A
1 5465 5464 -01 5464 -01 5464 -01 0-00 -010 5465
2 214 29 08 29 08 2912 -02 ~0 -157 -157 2921 07
A 5157 5573 416 5574 417 5402 245 009 4323 4332 5140 -17
4 58 8 53 - 1 55 7 -01 5537 -01 -002 - 0 -012 55380
ls 5834 1 5830 -04 5828 -06 5832 -02 -O1 -051 -032 5836 02
bull A579 61 409 6130 391 5971 232 T -1 7 3 4482 4309 5717 -
bullA 7 5413 5772 359 5762 349 5629 216 -087 3618 3531 5419 06
A 5337 5737 400 5733 396 5584 247 -cent32 3860 3828 5354 17
A9 $5805 5959 354 5937 332 5808 203 -202 3823 3621 5597 -08
Ao 5410 5766 356 5756 346 5624 21T 4 -097 3616 3519 5414 04
S All 5209 I52i0 01 5211 02 ]5210 01 004 007 0ii1 5209 0
IA12(=A i ) 1723 1797 74 179-S 72 17683 40 - 018 848 83 1714 -09
All temperatures in OR
TABLE 66(f) Model temperatures before and after correction
Steady State
Temoeratures of Imperfect Models
Surface
Prototype or Perfect Model
Temperature
Model
T
M Model M1 Model MIII t A(-)
Model Temperature after Correction
M j A I 46 5464 546 5464 -0546001 - -009 5465 0
2 2915 2907 -08 2907 -08 2913 -02 064 -156 -152 2922 07
AA 5158 5573 415 5575 417 5403 245 011 4323 4334 5140 -18
5538 5537 101 55371 -01 10
5 53
-021 -002 010 -062 5538 p2 As 5834 5830 -04 5828 06 5832 -02 -011 -051 -062 5836 02
6 5739 6149 410 6136 391 5972 233 -171 4482 4311 5718 -21
A7 5414 5772 358 5763 349 5629 215 -083 3618 335 5419 05 A8 5338 5737 399 5734 396 5585 247 -630 3860 3830 5354 16
A 5606 5959 353 5937 331 5808 202 -199 3822 3623 5597 -09
A 5410 5767 57 5757 347 5624 214 -093 3616 3523 5415 05
A 5209 5210 01 5211 02 5210 01 004j 007 011 5209 0
-2723 1797 74 1795 72 1764 41 -017 848 1714
All temperatures in OR
TABLE 66(g) Model temperatures before and after correctioh
Cooling Transient 30 minutes after spacecraft moving into shade
Surface
A1
Prototype or Perfect
ModelIT aure Tempert e
36
3665
Temeratures of Imperfect
Model MI Model MIT
Ei v__ __
0 36
3664 -01
Models
Model Mill
E___ _ _ _
3664 -01
oeMood M
A _ _ _ _ __ _ _
~0 -016
_ _ _
1
__
Model Temperature after Correction
_ _
36T
3665 0
A2 2886 2877 -09 2877 -09 883 -03 005 -168 -163 2893 0
2 5 5542 417 5544 419 5372 247 025 4303 4328 5109 -16
A
11 3705 3904
3702
3890
-03
-14
3702
3888
-00
16
3703
3896
-02
-08
- 03
-01 6
-030
-156
-b33
172
3706
3907
01
03
A6 5295 5687 392 5676 381 5523 228 -102 4162 4060 5281 -14
A7 5367 5718 351 5712 345 5580 213 -057 3516 3459 5372 05
A8 5290 5674 384 5668 378 5528 238 j056 3709 3653 5309 19
A
A
5526
5344
5864
5687
338
344
5841
5679
315
335
5720
15551
194
207
-212
-075
3651
3454
3439
3379
5521
5349
-05
05
A
A 2(=A-0
3831
1497
3833
1600
02
103
3834
+598
03
1101 3833
1554
02
57
006
-019
011
1148
017
1129
3831
1487
0
-10
All temperatures in OR
TABLE 66(h) Model temperatures before and after correction
Cooling Transient 200 minutes after spacecraft moving into shade
Surface
A2
Prototype or Perfect
j Model Temperature
S2
1 2278
2699
5005
I
Tompexatures of imperfect Models Model MI ModelMII Model M111
o MIafter
M I E I- TII tIIM
2274 -04 2274 -04 2277 -01
2684 -15 2683 -16 2693 -06
5407 402 5403 398 524 26
-o
-003
-029
A(-))
-057
-227
4200
-057
-230
4171
Model Temperature
Correction
_ _
ml
2280 1 02
2707 08
4989 -16
A 2288 2281 -07 2281 -07 2285 -03 -001 094 -095 2291 03
A 2734 2710 -24 2710 -24 2724 -10 -003 -836 -339 2744 10
A 4924 5352 428 5351 427- 5176 252 -008 4446 4438 4908 -16
A7 5096 5437 341 5428 332 5301 205 -082 3450 3368 5100 04
A3
A9
5109
5262
5485
560
376
342
5477
5582
368
320
5340
J5458
231
196
-073
-204
3665
3704
3592
3500
5125
5254
16
-08
A 10 5090 5429 339 5418 32 8 1 5293 20 3 -06 3454 3348 5095 5
A1 2340 2346 06 2348 08 2344 04 020 036 056 2340 0
A 1348 1465 7 1461 113 1413 65 -030 1f97 1267 1338 10
All temperatures in OR
would still be capable of providing satisfactory results To obtain
some answers to these questions additional computer experimentation
was carried out in which the hypothetical spacecraft was subjected
to direct suns irradiation at normal incidence for 200 minutes folshy
lowed by its -ntry into the shade for 100 minutes and the cycle re-shy
peated itself thereafter The internal power dissipation was left
unaltered at all times At the very beginning the spacecraft had
a uniform temperature of 4000R throughout and the errors in the modshy
eling parameters were taken as -those listed in TABLE64 Other conshy
ditions were the same as those used in the earlierstudies Thecomshy
putation was continued for five complete cycles Because of the masshy
sive data generated only temperatures at 30 and 200 minutes from
the start of each cycle and those at the end of each cycle are preshy
sented herein In addition data for the third and fourth cycle are
omitted in the interest of conserving space This is illustrated
in Fig 61 Details of the temperature data are given in TABLES
67(a)-(i) It goes without saying that -the results are very encour-
Aging indeed A second experiment of the same kind was also conducted
in which the heating and cooling time were doubled so was the period
of-the cycle The results were equally good Details may be supplied
upon request
The foregoing studies have established beyond doubt the useshy
fulness of the proposed concept of imperfect modeling when the errors
in the modeling parameters are restricted such that the linear theory
is applicable and when they are properly controlled as pointed out in
114
700
A7 (Upper Surface of But
600f - t
o_400
H i HeaQampgIlt
Al (Solar Panel) 20o 00 mi L10in
Cooling Time m n
Pigure 61 Cyclic heating and cooling of the hypothetical spacecraft (open circle o denotes instant for which data are presented in TABLE 67)
TABLE 67(a) Model temperatures before and after correction (Spacecraft subject to cyclic heating and cooling)
30 minutes from start of the first cycle
Surface PrototypeIOr Perfect odPelfecModelA
Temzeratures of Imperfect Models
Model MI Model M11 Model M111 C-)_____A______
Model Temperature after Correction
A
A2
Temperature
5907
5949
rT
5906
5940
E
-01
-09
5905
5938
z I
02 5906
-11 5943
-01
-06
shy003
-017
-012
-070
-015
-087
T
5907
5949
FC
0
0
A
A4
4745
5905
4780
5905
35
0
4702
5905
-43
0
4739
5905
-016
0 -
-719
0
1052
-004
333
-004
4747
5905
02
0
A 5939 5936 -03 5936 -03 5937 -02 -0 -024 -014 5939 0
A6 4403 4516 113 4514 111 44641 61 -024 1327 1303 4386 -17
4131 4134 03 4118 -13 4128 -03 -149 160 011 4133 02
A1
A
4506
4454
4662
4525
156
71
4651
4506
145
52
4599
4484
93
30
-099
-173
1588
1025
1489
853
4513
4439
07
-5
A10 4243
Al 547 6
4261
5476
18
0
4228
547 6
-15
0
4243
547 6
0
0
-300
001
443
-0
143
001
4246
547 6
03
0
A 2(=A I 1722 1728 06 726 04 172 -024 098 07- 1721 -01
All temperatures in OR
TABLE 67(b) Model temperatures before and after correction
200 minutes from start of the first cycle
Surface
A
PrototypePrttp or PerfectMode1
Temperature
5979
Toriratures of Imperfect
Model I Model M11
T I
5978 1-01 5978 -01
Models
Model MIll
I
5979 0
AAAt
-001
A
-008 1-009
I
Model Temperatureafe oreto after Correction
y
5979
A 06046 6041 5604 -06 6043 -03 -01o -046 -056 6047 01
A
A
As
5313
5978
6040
5615
5978
6038
302
0
-02
5592
5978
6037
279
0
-03
5483J 170
5978 0
6039 101
-207
-Q01
-008
3331
-004
-022
3124
- 005
-030
15302
5978
1041
-1
01
A6 5203 5523 3197 5503 300 5383 180 -178 3522 3344 5188 -15
A
A
A9
5199
5198
5404
5448
5524
5668
249
326
264
5400
5502
5625
201
304
221
5340
5396
5546
141
198
14
-442
-21
I3 94
2753
3243
3077
2311
3042
2683
15217 5220
5399
18
22
-5
A10 5215 5477 262 5440 225 5365 150 -346 2847 2501 5227 12
IA
A
(=A 0 )i
5681
1842
5682
1884
01
42
5682
1879
01
37
5682
1864
01
22
002
044
003
514
005
470
5681
1837
0
-05
All temperatures in OR
TABLE 67(c) Model temperatures before and hfter cortection
End of first cycle I
Surface
Prototype or Perfect
Model
Toecratures oil imperfect Models I SPrototypeI I oModel-odcl 141 Od l Tfe-Cme
Md MI MModl M111
_ _ _ _ _ __) i
A Temperiature
after Correction
Temperature C
A
A0
2747
3042
27747
2744
3828
-03
-14
2744
3026i
-03
-16
2745
3035
-02
-07
-003
-014
j-0381J-175
-041
-189
2748
3047
01
05
A 5033 5423 390 54171 384 5262 229 -056 4085 4029 5020 -13 A4 2744 2739 05 2738 -06 2741 -03 -004 058 -062 2745 01
A- 3026 13004 -22 3002 -24 3015 -1i -020 -2 6 8 288 3033 07
A 4964 5371 407 536 5400 5202 238 -061 4267 4206 shy 4950 -14
A7 5141 5455 1314 54f6 295 5328 187 -175 3213 3038 5151 10
A 55136 3_ __
5497 _ _
351 _ _
5484 _ _ _
1348 _ _ _
5358 _ _
222 _ _
Z118 _ _
13529 3411 5156 _ _
20
A9 5102 5621 319 5592 290 5483 181 -266 6509 3243 5297 -05
A 5130 5q46 316 5427 (297 5318 188 -173 b254 3081 5138 D8
A11 2897 2900 03 2902 04 2899 02 009 018 027 2897 0
A (=A- 1488 105 1484 101 1441 58 -045 1195 1150 1373 -10
All temperatures in OR
TABLE 67(d) Model temperatures before and after correction
30 minutes from start of the second cycle
Tcmno-ratures of Imperfect Models
Surface
PrototypeSor Perfect
Model Temperature
Model MI
T7 E
Model
T1
MIX Model
NtlTI
MIII A(
_ __
A(- )
00
^
04
after Correction
580c 0
A- 5806 5805 501 5 805 1 -001 -010 5806 0
-A 5905 5901 1-04 5900 -05 5903 -02 -008 -04d -048 5906 21
A 5183 5523 339 5505 32r2 5374 191 -154 3756 3602 5162 -21
A 5804 5804 0 5803 -01 5804 0 -01 -009 -010 5804 0
A
A6
A7
] 5896
5045
5135
5895
15437
545 4
-01
392
320
5895
5431
5436
-01
386
302
5896
5274
5325
0
229
190
-005
-056
-166
-030
4128
3284
-03
4072
3118
5899
5030
514 3
03
-15
08
A
A 9
A0
A
5139 a
5$11
5134
5212
5502
5633
5453
5214
1363
322
1319
02
5490
5605
5434
5214
351 I
294
300
02
5362
5493
5323
5213
22-3
182
19
01
-108
-258
-175
005
3561
3544
3301
011
3453
3286
3126
916
5157
5304
5140
5212
18
-07
06
0
A(=A) 1764 1818 54 1815 51 1793 29 -025 627 602 1758 -06
All temperature in OR
TABLE 67(e) Mcdel temperatures before and after correctioh
200 minutes from start of the second cycle
Prototype or Perfect
Model
Tod odel MI Mod 2
IM _
1 Model MIII[ _ A
lt - AC-P A
Model Temperature after Correction
Surface Temperature I t
A 5980 2gt
5979 rt
-01 va I
5979 )V IvX TIM I
i-01 597 9 b1 -O 0 -00771-008 j C
598 0 --C
SI -01 -0 A2 6047 6044 -03 6043H -04 6045 -02 -004 -041 -045 6048 01 A 5340 15665 325 5656 316 5527 18-7 -083 35W0 34 pound7 5323 -17
A
A5
A ____ ___
A
5979
6042
5246 __28 __ 5281
__
5978
6041
5592
15581
-01 5978159 3I_ _
-01 60411 I I1
346 5585130o___ __ 5565j 300 5565i
-01 _
-01
339284 2
_I_ 5979
_
6042
5446
0 _
0
200
178
_ -001
_ _
2003
-070
-144
_ -003 -004 5979 0I _ _ __ _ 0_ _
-019 -022 6043 J01
I - - 7j3714 3644 5228 -18I_5459_ _3 ]_52I 3071 2927 5288 07
A9 I
5242
5465
5593
5760 3-51
295II
5584
5734
342
26L9
5459
5631
2i7
166
-085
-241
3405
3266
3320
3025
5261
5457I
19
-08
A0 5279 5578 299 5562 283 5457 178 -144 3063 2919 5286 07
A 568568 15682 01 5682 01 5682 1 01 00
2004 o4 6 _____ 5681 0o6
AC=Al0 1849 8 8 1895 46 1875 26 -020 563 53 1843 -06
All temperatures in OR
TABLE 67(f) Model temperatures before and after correction
End of second cycle
Tcmopcraturcs of Imperfect Models iModel Temperature-
Prototype or Perfect
Model MI Model Mil _
Model MI ( A
odeafter Correction aeC ct
Surface Model Temperature T E TL
-Tj 1 E4
A A At
A1 2748 2745 -03 2745 -03 2746 -02 shy0 -036 -036 2749 01
A2 3044 3032 -12 3033 -11 3039 -05 001 -166 -165 3049 05 50 54 396 5433 394 25272 233 -011 4116 1 04105 I 0 5024 -15
A4 2745 2740 -05 2740 -05 2742 -03 -001 -057 -058 2746 01
As
A5157
3030
4973
3010
543
-20
325
3009
5384 5473
-21
4L1 316
3020
5216 5353
-10
243 196
-006
-016 -078
-262
4300 3266
-268
42184 3188
3037
4958
5163
07
[-15 06
A8 5145 5511 366I 5503 358 5371 226II
-073 3557 3484 5163 18
A 5314 5641 _1_
3271 5618 _
3041 1 5501
1 187 -207 3541 3334
_ _ II 5307
J_
-07
A0 5142 5467 325 5456 314 5337 195 -102 3290 3188 5148 05
Ai 2898 2900 02 290l[ 03 2900 02 0
010 -02
018 028 1179
2898 1^7I
0
1 1I(=A I ) 1386 1494 108 1491 105 1446 60 -028 1207 1376 -10
All temperatures in OR
TABLE 67(g) Model temperatures before and after correction
30 minutes from start of the fifth cycle 1 I
Tomperratures of Imperfect
frototype Model MI Modl MiSor Perfect Surface T~emperature WI
I
Models 1
Model MTI T Z A(
bullModel
AI At
ot Temperature
after Correction C Fv
A
2
A
A
5806
5905
5186
5804
5805
5902
15530 I5804
-01
-03
344
0
580
590i
5516
D1
D4
330
0
5805
5903
5381
5804
-01
-02
195
0
-001
-004
-122
-0
-008
-038
3774
-008
-009
-042
3652
-008
5806
5906
5164
5805
0
01
-22
01
_
A
_ _
A-
_
5897 5051
5145
1 5896
5447I 5472
396
327
58g4 5444
5461
-0101 393
616
5898 5283 __ 1
5341
01 232
196
-001
-023
-09
308 4148 3 1 3315
-031
4125 2 1
5899
5034
5150
02I-17
05 I__
A
5144
518
5512
5645
368
327
5503
5622
359
304
5370
5505
226
I187 -076
3561
3577 3501
3347
5161
5311
17
-07
A 5142 5467 25 5453 311 5335 193 -123 3322 3199 5147 05
A 2 (=Ao )
5212
1765
15214
11820 1
02
55
S02
1818
5213
53 1795 30
6
-018
1
633 017j 5212
615 1758
jFo 7
All temperatures in OR
TABLE 67(h) Model temperatures before and after correction
200 minutes from start of the fifth cycle
Temopratures of Imoerfect Models bull II Model Temperature
Prototype odel MI Model MIX Model after ComectionM o or Perfect aeC c I
Surface Model AA A Temperature E poundUZ-00
1 1 5980 5979 -01 5979 -01 5979 -01 -001 -0 -008 5980 o
6047 6044 -03 604 -04 6045 -02 -004 -041 - 5 6048 j 01
A 5340 5665 325 5657 317 bull 5527 187 -079 3501 3422- 5323 -17
A4 5979 5978 -01 5978 -O1 5579 0 0 -003 -003 5979 0
A5 604 041 -01 6042 0 -019 -022 6049 01i003 A 5246 347 5586 340 5446 200 -066 3715 3649 5228 -18
5281 5582 301 5567 286 50 136 3071 2935 5288 07 30582 567 5461 18 -1360
5243 551 5585 342 5459 216 -091 3405 3324 5261 18
A 5465 5761 296 5735 270 5632 167 -235 3266 3031 5458 -07
A0 5280 5579 299 5564 284 5458 178 -138 13063 2925 5286 06
A 5631 5682 01 5682 01 5682 01 002 004 006 5681 0 (=A 0 1 1849 1897 48 1895 46 1875 26 -066 3715 3649 1843 -06
All temperatures in CR
TABLE 67(i) Model temperatures bef6re and after correction
End of the fifth cycle
Temrocratures of I -mperfectModels
IModelI TemperaturePrototype Model MI Model MIX Model M1I after Correbtion or Perfect t A ) A-) ASur f a c e M o d e l t
Temperature r P E Tc A I 2748 2745 -03 2745 -03 -82 -036 -036 2749 01
A 3044 3032 -12 3033 -11 3039 -65 001 j 01 -016 3049 05
A3 5039 5435 396 5434 395 5272 2 3 -010 4107 5024 15 I A 2745 2740 -05 2740 -05 2742
I--63 -001 -057 -058 2746 I 012 2 II - 3030 3010 -20 3009 -2-1 3020 -10 -006j -268 3037 07
A4973 5386 4 53843 85 412 5216 243 -015 4306 4285 495 -A I_o 7 i _____________366 5474 1 5353 -6 -077 j3189 5163 06__
A3 5145 5511 366 5504
A7 5157 5482 325 i535 317 526[ 3266 3189 151631 06
359 5371 226 -072 3557 3485 5163 )8
A 5314 5641 327 5618 [304 5501 187 -206 354d 3334 5307 -07
A 5142 5467 325 5456 314 5337 195 -101 3290 3189 5148 06 A1 2898 2900 02 2901 03 2900 62 0 oao 018 028 o8986
[2(1A 1494 108 1491 105 1446 60 -0 27 1207 1180 1376 10 All i [R -ir
All temperatures in 0R
Chapter 2 During the course of the present investigation additional
observations some of which are quite essential for the successful
application of the theory have been made They are discussed briefly
in the section which follows
63 SOME RELEVANT OBSERVATIONS
631 Influence of Decreasing and Increasing Model Errors on Prediction Reliability
The errors in the various modeling parameters as listed
in TABLES 6] and 64 are of the order of 20 to 25 percent in kd and
Cd and of 10 to 20 percent in c and e for the original model MI
It would be of interest to know how the reliability of the final corshy
rected temperature data be altered by increasing and by decreasing
such errors To this end the computations described in Section 61
were repeated first with all errors reduced to one-half of their
respective values as shown in the tables It was found that the imshy
provement in the accuracy of the corrected model temperatures of c
was totally insignificant In no case did it exceed 10R Thus if
the present theory is used for the correction of data gathered with
imperfect models there is no real need of striving for the undue
reduction of modeling errors provided that they are kept within
limits This welcoming fact allows additional flexibility in model
fabrication On the other hand when errors in the modeling parameshy
ters are doubled the reliability of the correction procedure deteshy
riorates rapidly Errors in Thc as large as 20OR have been noted
125
This points to the need of a development of a more general theory
-whichcould accommodate greater errors in the modeling parameters
632 Proper Control of Model Errors and Selection of Experishymental Conditions
-TIt-has been-demonstrated tin-Chapter 2 that the best reshy
sults obtainable from the present theory arise when the ratios
and -) are close to 12 [see Fig 23(al) and (b2)]
or 2 [see Fig 23(a2) and (bl)] depending on whether in the modishy
fication experiments the magnitude of the 6s increases or decreases
These ratios evaluated for the simple error space are respectively
0474 and 0532 for models prescribed in TABLE 61 and are 0474 and
164 for models prescribed in TABLE 64 The ranges of errors shown
in Jhese tables are in fact selected with the guidance of the said
theoretical requirements- The question which naturally arises is
In the modification experiments with models MII and MIII would it
Tnot be desirable to strive for smaller errors That is to make the
positive errors become less positive and negative errors less negashy
tivet To see if this were the case the computations described unshy
der Section 61 and pertaining to models prescribed in TABLE 64 were
repeated but with both positive errors in model MII and negative
errors in MII kept less than 5 percent The change gave rise to
rather unfavorable ratios in namely F(j42 - 031 and
4-) ) - 022 It was found that the computed data for Te showed
tSuch modification scheme may be unrealistic in practice However this is NOT the point in question here
126
little or no improvement over those given in TABLES 65 and 66
In the discussion of error paths in the multi-dimensional error
space four possibilities are considered as illustrated in Fig 23
In the first modification experiment all positive errors remain posishy
tive after modification whether they become smaller or larger Likeshy
wise in the second modification experiment all negative errors reshy
-nmainnegative -Clearly such are -not-the-only-possibilities -it-is
physically feasible to have mixed errors (positive and negative) afshy
ter modification although they should always be avoided for reasons
to be seen shortly Consider for instance the experimentation with
MIII Some of ther~piginally negative errors may become positive
after modification while others remain negative Under such circumshy
stance the parabola joining the error states P P(-) and 0-) in
-subspace S( ) would have a relatively sharp curvature thus creating
a condition that is in direct conflict with the fundamental assumpshy
tion upon which the theory is built
To provide a pictorial illustration we consider a three-dimensional
simple error space shown in Fig 62 The coordinates of the point
P which is associated with the original model MI are all negative
Suppose that after modification 6 and remain-negative but P 1 2 3
become positive Thus the point P(-) would be a8 shown in Fig 62(a)
The error path joining P P(-) and 0(-) makes a sharp bend as illusshy
trated and it would totally invalidate the linear theory developed in
[1 and in this report On the other hand if P remains negative3
as shown in Fig 62(b) the error path would have a gentle curvature
127
F-
I
I
7D
(a)
Figure 62
-83 -83
(6) (a)
Schematic representationof error path as affected by model design
which is small everywhere along the path If all s are made posishy
tive it is again possible to have an error path which has a uniformly
small curvature This is illustrated in Fig 62(c) The three points
through which the parabola passes must now be of the sequence P
and P(-) The extension of the foregoing consideration to the design
of model MII experimentation is obvious
- Extensive numerical calculations have been carried out not only
for the hypothetical spacecraft considered in the present report
but also-for a totally different problem namely the correlation
of boiling heat transfer data and the results conclusively demonstrate
the validity of the arguments just given The nature of the error
paths has a controlling influence on the reliability of the theory
Error paths exhibiting small and sharply changing radius of curvature
such as that of Fig 62(a) should not be considered in any circumshy
stance
64 CONCLUSIONS AND RECOMMENDATION
Based on the evidence provided by a computer experimentation
of the application of the proposed new theory of imperfect modeling
to the study of the thermal performance of a hypothetical spacecraft
the following conclusions may be drawn The hypothetical spacecraft
has major radiative and conductive heat flow paths that crudely simushy
late those of the 64 Mariner family of space vehicles
(1) When errors in kd and Cd of up to 25 percent and in surface emittance of up to 20 percent exist in the model and when the spacecraft is subject to either a simple heating and
129
cooling transient with an intervening steady condition or whenit experiences a continuous cyclic transient the inshydicated model temperatures would entail errors of 10 20
and 30Rand higher These errors were reduced to less than 30R after a correction according to the theory There is no evidence of error accumulation under the cyclic conshydition investigated
(2) While the theory is deduced under the basic assumption that the errors in the modeling parameters are small and that they have independent effects on the dependent variable or variables there is no need to strive for the undue costly reduction of these errors Rather in the modifishycation experiments the model design andor the test conshyditions should be properly controlled in order that the ratios p+) 4Q) and 4p-) 4_) do not deviate excessively from 12 when the magnitudes of the 6s generally increase or 2 when they generally decrease
(3) The reliability of the theory deteriorates rapidly when errors in the modeling parameters become large At the present time it is not feasible to describe quantitativeZy what is meant by large or small error
In view of the success as well as the limitations found in connecshy
tion with the performance of the theory it is recommended that furshy
ther investigation be launched with emphasis on developing a means
of treating large errors
130
REFERENCES
[1] Chao B T A Study of Thermal Scale Modeling with Imperfect Models Proposal Gr-2-2404 submitted to Jet Propulsion LaborashytoryPasadena California May 28 19
[21 Bobco R P Radiation Heat Transfer in Semigray Enclosures with Specularly and Diffusely Reflecting Surfaces J Heat Transfer Trans ASME Series C 86 123-130 (1964)
[3] Plamondon J A and Landram C S Radiation Heat Transfer from Nongray Surfaces with External Radiation in Thermophysics and Temperature Control of Spacecraft and Entry Vehicles edited by G B Heller Progress in Astronautics and Aeronautics v 18 173-197 (1966)
[41 Hering R G Radiative Heat Exchange and Equilibrium Surface Temperature in a Space Environment J Spacecraft and Rockets 5 47-54 (1968)
[51 Chao B T and Wedekind G L Similarity Criteria for Thermal Modeling of Spacecraft J Spacecraft and Rockets 2 146-152 (1965)
[6a] Chao B T DePaiva J S and Huang M N Transient Thermal Behavior of Simple Structures in a Simulated Space Environment by Model Testing ME-TR-NGR-048 University of Illinois Urbana July 1967 78 pp
[6b] Chao B T DePaiva J S and Huang M N Results of Transhysient Thermal Modeling of Simple Structures in a Simulated Space Environment ME-TR-JPL-951660-1 University of Illinois Urbana November 1967 51 pp
[6c] Chao B T and Huang M N Transient Thermal Modeling with Simulated Solar Radiation NE-TR-JPL-951660-2 University of Illinois at Urbana-Champaign March 1969 55 pp
[7] Hamilton D C and Morgan W R Radiant-Interchange Conshyfiguration Factors NACA TN 2836 1952
[8] Strong P F and Emslie A G The Method of Zones for the Calculation of Temperature Distribution report prepared by Arthur D Little Inc to Jet Propulsion Laboratory Pasadena California No 1 July 1963
131
APPENDIX A
THERMAL RESISTANCE OF MULTI-LAYER INSULATION
The superinsulation used in todays spacecraft is a multi-layer
aggregate of aluminized mylar or teflon In the ideal condition
the layers form parallel planes and have no physical contact with
each other For the purpose of our present analysis such condition
is assumed Here again we adopt a semi-gray two-region spectral
subdivision of properties namely the solar and the infrared region
The directional properties of the surface are taken into account
by the simple diffuse-specular model ie p = pd + p All emitshy
ted radiant energies are assumed diffuse
Aluminized Surface Q-Exposed Surface (Plastic)
P T (=Txp Temp of Exposed Layer) C P
sPdbPspI T-iT
0
itrPd bPsb
yb (Temp of the Base Plate)
Fig Al A Base Plate Covered with Superinsulation
Figure A1 depicts a base plate at temperature Tb covered with
a superinsulation of n-layers The exposed surface is plastic and
132
is absorbing suns radiation at a rate
Q = S Cos 6 (A1)
where a is the solar absorptance of the plastic layer and S is the Zocal PI
solar constant (at the outer fringes of earths atmosphere S = 442
Btuhr-ft2) All properties shown in Fig A1 without an asterisk
are for the infrared range and the subscript a refers to aluminum
and p refers to plastic Appropriate values of reflectances and
absorptances of aluminized mylar or teflon currently used in spaceshy
craft applications are listed in Table Al
TABLE A1
REFLECTANCES AND ABSORPTANCES OF ALUMINIZED MYLARt
Solar Range (0251-25p) Infrared Range (gt 5p)
Aluminized side 085 0 015 095 000 005
Plastic side 085 0 015 000 050 050
Consider a typical section of two adjacent layers of the supershy
insulation eg A2 and A3 of Fig A1 We assume that the temperashy
ture of each layer is uniform and that the edge effects are neglishy
gible Figure A2 shows the equivalent resistance network when steady
condition prevails in the figure Eb is the black body emissive
power B denotes the diffuse radiosity and E23 is the exchange factor
tData transmitted to the author by Mr W A Hagemeyer of JPL
133
The subscripts 2 and 3 have the usual meaning
B2 B3
Eb 2 1 -P 1-pp Eb3
p pdo di p
E A (1-P A 1-p)
1
A2 E2 3 (l-p )(1-P
Fig A2 Radiation Network for Infinite Parallel Planes
Since A2 = A = A the thermal resistance per unit area is
R = b - Eb 2 Pd a Pd pQA (i -P ) + Pa sxa p sp
++ Es~(P - )(1 P (A2)2 l sa)(l - P)
The exchange factor E2 3 can he easily evaluated using the image techshy
nique Recognizing that the separation distance between layers is
very small as compared with the lateral extent of the superinsulashy
tion one finds
+ + 2 ++ (A3) 2 l+PSa sIp s~a sp -P P
sa SIp
Hence
Pd ad p iP P (1- ) (1- P ) (l1 s )(l - P)Sa S+p (A4)
a sa p sp sa Sp
134
If the surfaces are diffuse p = Pp = 0 and Eq (A4) simplishy
fies to
1 1 R=-plusmn+---1
a Ea p
---which is a well-known result
The resistance per unit area R I between the aluminum surface
of the last layer of the superinsulation and the protected surface
of the base plate can also be evaluated from Eq (A4) provided
That the three properties associated with the plastic namely E
pd P and p P and replaced by the corresponding properties of the
base plate Sb Pd b and Pb
The inward leakage heat flux through the superinsulation
can be determined from a consideration of the heat balance at its
outermost exposed surface It is
oT4 -aT
kt SCOS aO - T4 Tep (A5a) A p p exp
with
ER = (n - 1)R t RV C A 5 b )
n being the total number of layers in the superinsulation In pracshy
tice for the sunlit surfaces of the spacecraft
135
Sltlt A
COI S Cos e
or E aTgT
4
p exp
(A6)
and hence a valid approximation for Eq (A5a) is
A R(EPS Scos - a1T)
b (A7)
which may be either positive or negative When the superinsulation
sees only the empty space any possible heat leak is away from the
surface A straightforward calculation leads to
Q A
-1
P
4 Tb
+ R
(A8)
where E R is again given by Eq (ASb)
While the foregoing results were obtained under the assumption
of steady heat flow they may be used for transient analysis withshy
out entailing significant errors since the heat capacity of multishy
layer insulation is usually quite small and in spacecraft applicashy
tion the transients are seldom very rapid
To ascertain the effectiveness of of superinsulation in reducshy
ing surface heat flux we consider an aluminum base plate with and
without the superinsulation when it is exposed to
a sunts irradiation of a local intensity of 300 Btuhr-ft2
and
136
b the black empty space
The plate has a temperature of 528 0R and is either highly polished
or is sprayed with PV-100 white paint By using Eq (A4) and the
relevant data in TABLE AM one finds
R = 21 for aluminized mylar or teflon plastic side out
388 when the aluminum base plate is highly polished Rt =
202 when the base plate is covered with PV-100 white paint
Then it follows from Eq (A5b) that for a five-layer superinsushy
lation R = 1228 and 1042 respectively for the highly polished
and painted base plate The corresponding valuesfor a-ten-layer
insulation are 2278 and 2092 and for a twenty-layer insulation
4378 and 4192
With the foregoing information the leakage surface flux can
be readily calculated from Eq (A7) and Eq (A8) The results
are shown in Table A2 In practice the effectiveness of insulashy
tion is somewhat less due to the unavoidable conduction leaks
137
TABLE A2
Effect of Multi-Layer Insulation on Surface Heat Flux Btuhr-ft2
of an Aluminum Base Plate at 528degR
Based Plate Highly Polished Base Plate Covered with PV-100 White Paint
Without With Superinsulation Without With Superinsulation Superinsulation Superinsulation
(number of layers) (number of layers)
5 10 20 5 10 20
Surface Exposed 533 -035 -019 -010 - 506 -041 -021 -010 to Suns Irradiation
Surface Exposed - 67 -107 -058 -030 -111 -125 -063 -032 to Empty Space at 00R
The positive value is for heat flow into the base plate all negative values are for heat flow away from the base
plate
APPENDIX B
A NUMERICAL EXAMPLE
As an example we consider the case in which errors in the modshy
eling parameters are as shown in TABLE 61 and the suns rays strike
the solar panels at normal incidence Specifically we refer to data
listed in TABLE 62(d) which is for a heating transient 300 minutes
after the spacecraft is subjected to solar radiation and internal
heating The choice of TABLE 62(d) for illustration is totally arshy
bitrary any other tabulated data given in the report may serve the
purpose equally well
First we note that the perfect model temperatures listed under
T are calculated from the equation set (336) using the theoreti-P cally correct values of the modeling parameters This is achieved
by the combined Newton-Raphson and Gauss-Seidel procedure as explained
in Chapter 5 The various temperatures of-the three imperfect models
II II and Tlll are computed from the same equation set using
the same procedure but with modeling parameters entailing errors
as listed in TABLE 61
To illustrate how the various corrections are computed-it is
sufficient to work out the details for one of the surfaces say A6
At t= 300 minutes after the start of the heating transient
TI = 542JPR TII = 5383degR and M1 II 55540 R These may be MI
compared with the theoretically correct temperature of 52440R and
hence their respective errors are 177 139 and 3100 R All
139
temperatures cited here may be read from TABLE 62(d) in entries folshy
lowing A In passing we note that the three temperatures associshy
ated with the three imperfect models correspond to the function
of Chapter 2 as follows
T~
T TLIII i
T
4 ) DETERMINATION OF CORRECTION FOR POSITIVE ERRORS A
Prom TABLE 61 it is seen that the various positive errors are
(a) for Model MI
A A A4A 5 A A A A0 A A1
kd -- 025 025 020 025 025 020
6 025 020 025 025 025 025 Cd
6 015 -- 025 015 -- 015
6 010 010
(b) for Model MII
A A A A4 A5 A6 A7 A8 A9 A 0 A11 A12
6 - 055 045 040 0575 045 044 kd
6 0525 050 055 0575 045 060 -shycd
6 -- -- 027 -- 045 033 -- 033
6 -- 020 - 020
140
The asterisk designates errors after modification In the present
instance all errors become more positive after modification Hence
the error path is as illustrated in Fig 23(a1) and (238) is the
appropriate formula for determining A + )
From the foregoing data we may readily calculate
2 2 2 2 2 2 2 = 025 + 025 + 010 + 025 + 020 + 010I
+ 0202 + 025 + 0152 + 0252 + 025
2 2 2 2 2 + 025 + 025 + 020 + 0252 + 0152 + 015
= 0833
lt 052 2 2 2 5= + 0525 + 020 + 045 + 050 + 020I
+ 0552 + 0272 + 05752 + 05752 + 0452 + 0402
2 2 2 2 2 + 045 + 045 + 044 + 060 + 033 + 033
- 3687
6 6 = 025 X 055 + 025 X 0525 + 010 X 020 + 025
X-O45 + 020 X 050 + 010 X 020 + 020
X 040 + 025 X 055 + 015 X 027 + 025 X 0575
+ 025 X 0575 + 025 X 045 + 025 X 045 + 025
X 045 + 020 X 044 + 025X 060 + 015 X 033
+ 015 X 033 = 1740
Using (239a) we find
141
p F0833 X 3687 11011i(174 0= 0112
5 1740 = 0472
3687
Also ( = 3687 192 Hence from (234ab) we determine p
1 a = 0112 X = 02121 - 0472
0212 b = - 021 = -0110
192
and from (235c)
A = 1 - 2 X0472 0056
Substituting the foregoing values of a and X in (236b) gives
f(aA) = 00549
and hence the ratio of arc lengths is
- 2 = 2116
I 00549S(+)
as according to (236a) Finally the required correction for all
positive errors is given by (238)
+ 5383 - 5421 21=- 340oR2116 - 1
which is listed in TABLE 62(d)
142
-DETERMINATION OF CORRECTION FOR NEGATIVE ERRORS A(
From TABLE 61 it is seen that the various negative errors are
(a) for Model MI
A7A A3 A A A A1 -A AAAIA
kd
Cd
-- 018 -020 -- -018 -- -shy
6 - -008 -008 -008 -008
and 6 = -0102 for superinsulation covering A and A 0
(b) for Model MIII
A A A A A A A6 A7 A8 A Ao A1A1
kd
Cd
6--- -- 0288 -032 -- -0259 -
-- -0147 -0115 -016 -0128
and 6 = -0205 for superinsulation covering A and A0 Again the ER7
asterisk denotes errors after modification Since all errors become
more negative after modification the error path is as illustrated
in Fig 23(b2) From the data we find
+ (-008)2262 = (-018)2 + (-020)2 + (_008)2 + (_018)2
+ (-008)2+ (-008)2 + 2x(-0102 )2 = 01512
143
Z (602 = (-0288)2 + (-032)2 + (-0147)2 + (-0259)23
+ (-036)2 + (-D128)2+ (-0115)
2 + 2x(-0205) = 0412
661 = (-018)(-0288) + (-020)(-032)
+ (-008)(-0147) + (-018)(-0259)
+ (-008)(-0115) + (-008)(-016)
+ (-008)(-0128) + 2(-0102)(-0205) = 0247
Hence
n) _ -01512 X 0412 112 = 0158
(-) 02472
p _ 02470412- 0601
= 412= 0642
and
r1
a = 0158 1 03951 0601
0395b = 064 =0 6150 642
F 1 - 2 X 0601 = -0202
f-(aX) -0198 (from 236b)
s ( -) ( -) _ 2 s= 1+0198 = 1669 (from 2313b)
144
and
= 5554 - 5421 _ 19890R (from 2313a)1669 - 1
which is given in TABLE 62(d) The total required correction is
At = P + - -340 + 1989 r 1649degR
and hence the corrected model temperature is
1649 = 52560Rc = 5421 -
Upon comparing with the theoretically correct value of 52441R one
sees that the error in TC is 120R
145