PERFORMANCE RANKING OF PLATE-FINNEDHEAT EXCHANGER SURFACES

James Gi Ibert Soland

3AUUATE SCHOOiILIFORNIA «3<Uf»

PERFORMANCE RANKING OF PLATE-FINNED HEAT EXCHANGER SURFACES

by

JAMES GILBERT SOLAND

SoBo, United States Coast Guard Academy(1968)

SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THEDEGREE OF OCEAN ENGINEER

AND MASTER OF SCIENCEIN MECHANICAL ENGINEERING

at the

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

June, 1975

OUDLEY KNOX UCSAR*MAVAL POSTCniJUATE SCHOOLMONTEREY. CALIFO:.;

PERFORMANCE RANKING OF PLATE-FINNED HEAT EXCHANGER SURFACES

by

JAMES GILBERT SOLAND

Submitted to the Department of Ocean Engineering on May 9, 1975

in partial fulfillment of the requirements for the degrees of

Ocean Engineer and Master of Science in Mechanical Engineering.

ABSTRACT

The Kays-London (4) way of presenting heat exchanger performanceis modified to use the data in the simplified analysis of La Haye et al.

(5) which was extended to permit the ready comparison of the performanceof the various plate-finned surfaces. Based on the four comparisoncriteria considered, the wavy finned 17.8 - 3/8 W surface of Kays-London is the best.

THESIS SUPERVISOR: Professor Warren M. RohsenowTITLE: Professor of Mechanical Engineering

ACKNOWLEDGEMENTS

I would like to thank Dr. William M Mack for the use of his

thesis which served as the basis for my work, and to Professor

Warren M. Rohsenow for his untiring help and guidance.

I would expecially like to thank my wife, Jan, whose under-

standing, constant faith, and abiding love kept my goals in their

proper perspective.

TABLE OF CONTENTS

PAGE

TITLE PAGE 1

ABSTRACT 2

ACKNOWLEDGEMENT 3

NOMENCLATURE 5

LIST OF FIGURES 9

I„ INTRODUCTION 11

II. COMPARISON METHOD 12

III. COMPARISON OF KAYS AND LONDON PLATE-FINNED SURFACES 28

IV. COMPARISON WITH SAND GRAINED ROUGHENED SURFACES 45

V. RESULTS OF COMPARISONS 47

VI. CONCLUSIONS 48

REFERENCES 49

NOMENCLATURE

Symbol

*b

Definition

heat exchanger width

heat transfer area of base surfaceignoring any enhancement; equalslength times heated perimeter

minimum free flow area

frontal area of heat exchanger core

Units

ft

ft'

ft'

ft'

flow area ignoring any enhancingsurfaces

ft'

total heat transfer area ft'

plate spacing

specific heat

ft

Btu/lbm- F

D nominal diameter; defined by (lb) ft

n

roughness height for a granular surface ft

friction factor based on total area (A )

defined by (4a)

friction factor based on base area (A,) ;

defined by (4b)

friction factor for a smooth surface;defined by (29)

conversion factor (= 32.174 lbm-ft/lbf-sec )

mass flux based on minimum free flowarea; defined by (2a)

lbm/hr-ft'

2G mass f lux based on free flow area (A ) ; lbm/hr-f

t

ndefined by (2b)

b heat transfer coefficient I a" m tr*-al Btu/hr-ft - F

m

area (A ) ; defined by (5a)

h heat transfer coefficient based on base Btu/hr-ft - F

area (A, ) ; defined by (5b)

j Colburn j-Factor based on total area (A ) ; - -

defined by (7a)

j Colburn j-factor based on base area (A, ) ; - -

defined by (7b)

j Colburn j-factor for smooth surface; - -

defined by (27)

k thermal conductivity Btu/hr-ft- F

£ fin length from root to center (=b/2) ft

L heat exchanger length ft

component of fin efficiency (r| f ) ;

defined by (10)

NTU number of transfer units; defined by - -

(21)

2p pressure lbf/ft

P pumping power hp

q heat transfer rate Btu/hr

2q/A heat flux Btu/hr-ft

r hydraulic radius; defined by (la) ft

T temperatureo,

U overall heat transfer coefficient Btu/hr-ft - F

3V heat exchanger volume on one side ft

DIMENSIONLESS GROUPS

Nu ITuttiBidt: tnsnber; defined by (6a)

Nu Nusselt number; defined by (6b)

Pr Prandtl number

Re Reynolds number based on minimum freeflow area (A ) ; defined by (3a)

c

Re Reynolds number based on free flow arean

(A!); defined by (3b)

SUBSCRIPTS

a case a parameter (Shape, V = const.)

b case b parameter (P, V = consto)

c case c parameter (NTU, P = const.)

d case d parameter (NTU, V = consto)

e enhanced surface

m heat exchanger metal

s smooth surface

MISCELLANEOUS

3

AP]

-l

p

e

6

ratio of total heat transfer area (A ) to ft

volume (V)

friction pressure drop lbf/ft'

mass flow rate lbm/hr

fin efficiency; defined by (9) - -

total surface temperature effectiveness; - -

defined by (8)

viscosity

density

heat exchanger effectiveness - -

fin thickness ft

lbm/hr-ft

lbm/ft3

* pin diameter ft

LIST OF FIGURES

FIGURE TITLE PAGE

1 Sample Calculation of Nominal Diameter and Mass 16

Flux for Rectangular Flow Passages

2 Comparison of Colburn j Factors (j and j ) , and 18

Comparison of Friction Factors (f , and f ) c

3 Typical Plot of Heat Exchanger Effectiveness 20

( £ ), versus Number of Transfer Units (NTU)

.

4 Performance Parameter Curves for Two Surfaces 25

Showing Points Used in Sample Comparisons.

5 Relative Heat Exchanger Shape for Sample 26Comparisons Assuming Unit Height.

6 Typical Performance Comparison Results 27

7 Performance Parameter Curve for Plate-Fin 29

Surfaces.

8 Performance Comparison Results (NTU /NTU ) for 33e s

Case a (i c e. Same shape and V = constant).

D = o 50" for both the enhanced and plainn r

surface. •*. Re = constantn

9 Performance Comparison Results (P /P ) for 34e s

Case a (i e same shape and V = constant)

„

D = 0.50" for both the enhanced and plainnsurface. ••• Re = constant,

n

10 Performance Comparison Results (NTU /NTU ) for 35

Case b (i c e. V = constant, P = constant).D = o 50" for both enhanced and plain surface,n

11 Performance Comparison Results (V /V ) for 36Case c (i u e. NTU = constant, P = constant).D = 0.50" for both enhanced and plain surface,n

10

Figure Title Page

12 Performance Comparison Results (P /P ) for 37

Case d (i.e. NTU = constant, V = constant)D = 0.50" for both enhanced and plain surface,n

13 Performance Comparison Results (NTU /NTU ) for 40

Case a (i e same shape and V = constant) to

a Smooth Surface with D = o50" o

n

14 Performance Comparison Results (P /P ) for 41Case a (i e same shape and V = constant) to

a Smooth Surface with D = 0.50" on

15 Performance Comparison Results (NTU /NTU ) for 42

Case b (i.e. P = constant, V = constant) to a

Smooth Surface with D = o50" o

n

16 Performance Comparison Results (V /V ) for 43S 6

Case c (i e NTU = constant, P = constant) to a

Smooth Surface with D = 0.50".n

17 Performance Comparison Results (P /P ) for 44Case d (i.e. NTU = constant, V = constant) to

a Smooth Surface with D = 0.50".n

18 Performance Parameter Curves for Best Plate-fin 46Surfaces and Dipprey & Saberski (3) e/D = o 049Surface.

11

I. INTRODUCTION

The goal of any heat exchanger -omparisor Method is to enable

the designer to select from among the various enhanced surfaces that

surface which is most beneficial. The comparison method should allow

this selection to be made as easily and as accurately as possible.

Enhanced surfaces, particulary in the form of plate-fins, have

been used in heat exchangers for many years. Attempts at comparing

performance of various surfaces have also been made for many years.

The more recent of these are by Bergles et al. (1), (2).

La Haye et al. (5) developed a method of comparing surfaces and

used it to show how an effective uninterrupted flow length to diameter

ratio could parametrically order the performance of surfaces. Here

their comparison method is modified and used to compare the performance

of all of the Kays-London (4) plate finned surfaces, unfinned surfaces

and sand roughened surfaces (3) . The method is universally applicable

to any other heat exchanger surface as well.

12

II. COMPARISON METHOD

To simplify the comparison method, consider the performance of

only one side of a plate-finned heat exchanger. This is equivalent

to considering a heat exchanger with the controlling heat transfer

resistance on one side - e.g., gas flow on the side of interest and

condensing or boiling fluid on the other. We compare the performance

of various finned and unfinned surfaces for the following quantities

being the same:

1. 0) , flow rate

2. T, , hot fluid inlet temperatureh, in

3. T . , cold fluid inlet temperaturec , in

Note also that the heat transfer resistance of the plate separating

the two sides of the heat exchanger shall be considered to be negligible,

Kays and London (4), hereafter referred to as K-L, present data

for many plate-finned surfaces in terms of heat transfer coefficients,

h, and friction factors, f, referred to the exposed area, A , as a

function of Reynolds number, Re, based on the minimum free flow area,

A .

c

The proposed comparison method converts these h and f magnitudes

to the base plate area, A^; hence, the effect of the fins is included

in the new h and f based on A, . Further, the new Reynolds number,

Re , will -be based on the open flow, A^, as though the fins were not

present. This requires that the metal conductivity of the fins be

13

specified in incorporating the effect of the fins into the h .r n

The comparison of plate-finned performance is then converted to

the same base that is currently used in comparing enhanced surfaces

(such as those utilizing roughness, turbulence promoters, etc.) with

plain, smooth surfaces.

Table I shows the proposed new definitions of the various quanti-

ties compared with the definitions used by K-L (A) . Note from Figure

1, D = 2b for either a finned or unfinned parallel plate passage.

To convert data of K-L to the new basis, the following ratios

are obtained from the definitions, eqs. (1) through (10) , and Figure 1,

\ _ 2a L = _2_Kj, 3 V 3 b

(ID

where 3 = A /V

^F Lab 1

Ac \ r

h *rh

(12)

Gn

A

Ap hGc

Ren

D Gn n 6 b

Re 4rh

Gc

2

fn *F

AT

Gc

2

A A, G2

c Id n

b

f 2 32B r

h

3

(13)

(14)

(15)

14

CD

CO

ocHoy<

coX)coJ<&

CO

>

J

J*

<J°

<r

III

Q

cO

.J

III

J-i

CD

4J

cu

ecO CO

•H 3X) •H

O Cli

>. •H M4J iH•H 3 !-!

P cO OC }-i

cfl T33 >.C x:

CM

<^III

O

CO

CM

III

o

•HOO

0)

>

co

CO

cO

X>CO

OIII

0)

Pi

CO

3-

III

0)

0)

CO

13

XI

<4-l

a,<

o

bO

dCM

III

cO

In(X<

CM

O

00

Q.

CM

i-J m

III

14-1

uo4-J

uCO

mCo•H•p

o

^cw

<f°

in

CO

o

cr

in

<U•rl

O•H4-1

4-J

oo

J-i

V-i CD

CD rQIWCO 3ti SCO

U 4J4J tH

CD4J CO

CO CO

0) 3.d ^

Q

III

35S

CO

UM

III

15

<u

a•H

aoo

o

wp

wJ

ai

CO

oCMonP-.

co

o

CO

H

CM

a.

oin

to

ro

CN

a.

o

in

00

En

CT>

aCO •Ha U-l

•Hm pi

•H•p P.<D

a) U.fl rt

w t-H

a O•H l-i

^ •H•U O

/~\ »»-%

aJ J=>o o

o*e

o*42 6 43Cnj

4-1 ™|

III

ccr

4-4

in

e

:>>

u PJ

•H MU 3a 43cfl H3 OO* O

16

a) Smooth Surface Passage

*F= ab

*b== 2aL

V -- abL

Dn

4V

G =n

03 =

abL2aL

2b

03

ab

b) Plain Plate Fin Surface 6.2

» * 1 i

*F- ab

A, = 2aL

V = abL

a

/ \ 7i

i4V_ = 4

abL2aLh

ab

03 03G = =- =n

*Fab

Figure 1. Sample Calculation of Nonimal Diameter and Mass Flux for

Rectangular Flow Passages

17

j h G. n b

(16)j h G 2 r,J n h

These ratios were used to convert K-L data to the proposed fn

and i vs. Re . Two assumptions were required in order to solve forn n

the proper fin efficiency, r\ , required in the conversion. The heat

exchanger used was constructed from aluminum (k - 100 Btu/ft-hr- F)

and the gas used was air at 90 F. Figure 2 shows the two sets of

curves on both basis for one of the K-L surfaces, namely their 6.2

plain plate-fin surface, Figure 61 of reference 4.

For any heat exchanger, the power per unit volume on one

side is given by:

(17)p

w Apf

V P v

or from Table I,

P _ n3

2

So P

f Re3

n n

V

n

(18)

For the same fluid at the same temperature level, y and p are constant

Therefore:

P 3 4- « f ReJ/D (19)V n n n v

'

The heat transfer for any heat exchanger is given by:

q = e (T, . - T ) toe (20)M h, in c, in p

18

T-i—i

r 1 1 i I i i r o

J= -

o o

ii i i i i I I I nn i I i i

oo

ulV uf'f

COo

oroo

o

aoW•H5-i

CO

ou•x)

GfO

TJcctj

•i-i

to

uo4J

u •

C(J /~NtM c

M-l

•I-

) T3acd

Cn IHd **S

,oHo to

O 5-1

oIH J-l

O oCO

a fJ4

oen G•H OU •Hctj •Pa Oe •Ho Mu ^

oCM

a)

s^

360

Fn

19

For any given flow arrangement, there exists a curve similar to that in

Figure 3 which relates C to NTU, where;

A h

NTU = — (21)wc

P

This relationship between e and NTU is always monotonically

increasing. If fluid properties and flow rate are held constant,

an increase in A h results in an increase in NTU which results inn

an increase in £ and a higher heat transfer rate, q. Therefore,

knowledge of either Ah or NTU will allow determination of the heatn

transfer rate. Thus the ratios A b /V and NTU/V are quantities ofn

interest from (21)

;

A hNTUV V wc

P

Substituting from Table I yields:

(22)

NTU 4 u ^nRe

n

or.

V iHT 2Pr '

03 Dn

Ah 4 c u j Ren p " J n n_

V=

pr2/3 2

n

(23)

(24)

For the same fluid at the same temperature levels, c , y, and Pr areP

constant, therefore:

Ah j Re

-V2- - -JLTJL

C25)Dn

20

Figure 3 Q Typical Plot of Heat Exchanger Effectiveness,e , vs. Number of Transfer Units, NTU.

21

and, here the flow rate, oo , is also constant

m « V!^_(26)

VDn

With enhanced surface data in the form f vs Q Re and i

n n n

vs. Re , it is a simple matter to construct the performance parametersn

f Re /D and j Re /D '

, and to plot them as in Figure 4. Differentnnn Jnnn' v b

curves will result for different surfaces. Two surfaces shown by

Figure 4 will be used to demonstrate the use of these curves in deter-

mining heat exchanger relative performance. All comparisons will be

made for the same w , T , T, This implies that anyc, in h, in

comparison which results in the same value for NTU/V will also have

the same q/V.

Point o has been selected on surface 1 of Figure 4. It may

lie anywhere on the surface 1 performance curve and represents a

reference heat exchanger having the following specifications: P ,

o

NTU , q , L , A_ , V , where the subscript zero refers too o o F, o o

surface 1. Diagramatically, the shape of the reference exchanger is

labelled o in Figure 5.

Four different performance comparisons are immediately available

from Figure 4 and are indicated by points a , b , c and d on surface

2. Here the curve for surface 2 lies above the curve for surface 1

in Figure 4.

22

Case a;

Same heat exchanger shape and volume (L = L , V = V ,

a o a o

*F. a" AF, o>-

This case represents a comparison of points o and a of

Figure A where:

D

Re = Re x'na no D

no

Since the flow rate (w) and the frontal area (A-) are fixed. The

results of this comparison are easily obtained as the ratios of the

ordinate values and abscissa values and are shown in Figure 6a.

Figure 5 shows the same heat exchanger shape for this case. Figure

6a shows the magnitude of the increase in pumping power and the heat

transfer when using surface 2 instead of surface 1 in the same shape

heat exchanger.

Case b

Same heat exchanger volume and pumping power. (V, = V ,

b o

P - P ).b o

This case represents a comparison of points o and b in

Figure 4 since a vertical line in the performance plot has a fixed

value of power per unit volume. This comparison yields the NTU ratio

of the two heat exchangers and is easily obtained as the ratio of

the ordinate values at o and b. Results are shown in Figure 6b and

23

shows the magnitude of the increase in the NTU for the same pumping

power for surface 2 compared with surface 1. The relative shape for

this case is shown in Figure 5 Since surface 2 has higher friction,

the same pumping power will be obtained for a smaller Reynolds number,

i.e. Re, < Re This causes the frontal area to increase, A_ .>bo T, b

A^, , while the same volume requires the length to decrease, L. < LF, o bo

Case c:

Same pumping power and number of transfer units. (P = P ,

NTU = NTU (or q = q )).c o c o

This case represents a comparison of the heat exchanger size

required for the same "job" - - a job being defined as the same pumping

power resulting in the same heat transfer rate. This comparison is

a line having a slope of 1 in Figure 4 (i.e points o and c)

,

since each axis is inversely proportional to the volume (NTU and P

are constant) . This results in the ratio of the heat exchanger volume

required for the two surfaces obtained as the ratio of either the

abscissas or the ordinates at points o and c. As long as surface

2 lies above surface 1, the result will be a smaller volume required

to do the same "job". The resulting reduction in volume for surface 2

is shown in Figure 6c and the relative shape is shown in Figure 5.

24

Case d:

Same volume and number of transfer units (V, = V , NTU,d o d

= NTU (or q, = q ))o doThis case compares the pumping power required by surface 2 to

that required by surface 1 for the case when both heat exchangers yield

the same overall heat transfer performance. It is obtained as the

ratio of the abscissas of points o and d in Figure 4, since a

horizontal line on the performance plot has a constant value of

NTU/V. The results are illustrated in Figure 6d; they show the decreased

pumping power required for surface 2. The surface 2 Reynolds number

is the smallest of all four cases and consequently, the flow area

is the largest. Since the volume is the same as the surface 1 exchanger

volume, the length must decrease and the general shape is shown as

d in Figure 5

2Thus for a very simply constructed plot of j Re /D vs.3 r J r J n n n3 4

f Re /D , it is possible to obtain useful performance comparisonsn n n ' r v v

between two heat exchangers for four different criteria.

25

inRen

Dn2

f nncn3

D 4

Figure 4. Performance Parameter Curves tor Two Surfaces ShowingPoints Used in Sample Comparisons.

26

fEow

r

^

J

J b

d

o,a

Figure 5 Relative Heat Exchanger Shape for SampleComparisons Assuming Unit Height.

27

>ii

A

ii

cr

o

Pm

II

Pm

o

ccXI ii

H£5

-O

o0)

DC

CO

o

^

co

o

II

JII

<u

CO

c_>

o' -

" 1

o

DC

O

cr

ii

HS3

HS5

Pw

II

c

Pm

oQ?l£

o

DC

Of o

CO

•ui-H

CO

S!p<3

aoCO

•H

IoCD

odcd

6

oM-l

M0)

Pm

o•H

}-l

00•HPm

28

III. COMPARISON OF KAYS AND LONDON PLATE-FINNED SURFACES

All of the plate-fin surfaces were plotted as performance curves

2 3 4of i Re /D vs f Re /D , for the same w , T_ . and T, . ,J n n n n n n ' ' c

}in h, in

This is a plot equivalent to NTU/V vs. P/V.

K-L data are available for various finned surfaces having about

nine different nominal diameters between 0.50 inches and 1 646 inches

(b = o 25 inches to o 823 inches). For each nominal diameter, all of

the finned surfaces were plotted as shown in Figure 7. However, in

Figure 7, only the highest performance curve at each plate spacing

is shown.

Table II lists all of the K-L plate-fin surfaces by type and

surface designation Due to the large number of surfaces used in the

calculations, all Figures 7 through 17 will use the surface numbering

system which is listed in the right hand column of Table II.

All of the plate fin surfaces were then compared to smooth

surfaces having no enhancements and having the same nominal diameter

(D ) , (or plate spacing (b)), as the plate-fin surface being compared.

This was accomplished using all plate-fin surfaces having a common

nominal diameter of o 50 inches (b - .25 inches) This was chosen

due to the quantity of data available. There are 17 surfaces having

this nominal diameter, whereas, the next nominal diameter (D =0.82' '

v n

inches) has data for only 3 different surfaces The subscript e

will be used to denote the enhanced surface and the subscript s

will be used to denote the smooth surface.

29

pnn i

ii i i i i i

o

co

CU

Oco

p.

CO

PnI

CD

co

rHPM

o

CD

>

CO

CD

OCO4-1

CO

a•Hiw

I

a•H

a1

43•p

H3

cO

a•HO03

CO

CD

4J

a)

a.

43oCO

CD

>-l

o

ri Ou 43

CO

5-i

CI) CO

•u •H<d

CD

cO au co

CO IHP4 5-1

3CD CO

o

uoCH*H

CD

Pm

a)

j-i

•H

CO

CU

43

CD

43•P

ao

wHo!3

(!)

ccCMCQ

CM

30

TABLE II. KAYS AND LONDON PLATE FIN SURFACES.

GENERAL PLATE SPACING SURFACE SURFACE NUMBEREDSURFACE TYPE (b) (inches) DESIGNATION IN FIGURES AS:

Plain plate-fin .47 5.3 1

.41 6,2 2

.82 9.03 3

.25 11.1 4

.48 ll„l(a) 5

.33 14.77 6

.42 15.08 7

25 19.86 8

Louvered plate- .25 3/8 - 6.06 9

fin.25 3/8(a) - 6.06 10

.25 1/2 - 6.06 11

25 l/2(a) - 6.06 12

.25 3/8 - 8 7 13

25 3/8(a) - 8.7 14

.25 3/16 - 11.1 15

.25 1/4 - 11.1 16

«25 l/4(b) - 11.1 17

.25 3/8 - 11.1 18

.25 3/8(b) - 11.1 19

.25 1/2 - 11.1 20

.25 3/4 - 11.1 21

.25 3/4(b) - 11.1 22

Strip-fin plate- .25 l/4(s) - 11.1 23

fin.49 3/32 - 12.22 24

41 1/8 - 15.2 25

wavy-fin plate- o41 11.44 - 3/8W 26

fin.41 17„8 - 3/8W 27

pin-fin plate- 24 AP-1 28

fin.40 AP-2 29

c 75 PF-3 30

.50 PF-4(F) 31

.51 PF-9(F) 32

31

Smooth surface i Re /D calculations were completedJ n ns ns r

utilizing the Colburn correlation for forced-convection, turbulent

flow in tubes with the appropriate nominal diameter (D ) from

reference 8. This reduces to:

j 5 o 023 Res

-0 o 2(27)

For a smooth surface h«=h , G = G , Re = Ren n ' n

Then:

j Re = 0o023 Re°°8

J s n n (28)

3 4Smooth surface f Re /D was calculated using the linear

n n n °

approximation:

-0.25f = O 0791 Res n (29)

For a smooth surface f = f .

n

Thus:

3 2 75f Re = .0791 Res n n

(30)

Combining (28 and 30) to obtain the equation for the smooth

surface leads to:

j Re = O 0481 (f Re3

)

' 291Js n s n

(31)

or:

j ReJs n

D2

0481.836

n

-, 0.291f Re

*

s n

1)

n

(32)

32

These lines were calculated for all different nominal diameters. How-

ever, only three separate nominal diameters (0 o 50" , 0.94" , 1.646")

which span the range of values are shown on Figure 7 as the three

straight lines at the bottom of the figure.

Cases a through d were then calculated for all plate-fin

surfaces having a nominal diameter of o 50 inches and were compared

with a smooth surface having a nominal diameter (D ) equal to 0.50

inches. Figures 8 and 9 show the case for same Re since the nominaln

diameter is fixed. Figure 8 shows the ratio of ordinates (NTU /NTU )

vs. Re and Figure 9 shows the ratio of abscissas (P /P ) vs. Ren e s n

Note that a larger ratio on Figure 8 is preferred and a smaller ratio

on Figure 9 is preferred

Figure 10 shows the case for P = constant, V = constant or case

b (i.e. NTU /NTU vs c Re ) .

e s n

Figure 11 shows the case NTU = constant, P = constant or case

c (i.e. V /V vs. Re )

.

s e n

Figure 12 shows the case NTU = constant, V = constant or case

d (i.e. P /P vs. Re )

.

s e n

For Figures 10 through 1? higher ratios are preferred.

With the completion of the performance comparison for a nominal

diameter of 0.50 inches all of the plate fin surfaces of K-L (4) of

all diameters will now be compared to a smooth surface having a nominal

33

CO *

CO

incu

o

o a)

cl

CO

4-1

>-<

CO

cu

dCO •

rHd CN

• d CU

H •H cu uv_^ CO

rH£ cu

CO P* CU g^ d(D 13 dCO d cu

CO cO •H cu

CJ>

13<-\ o

CO

n 0) rH 4-1

o o rH U44 d

cO

co dCO

,—

v

^c CMtod CN CO

Pcu

13 O£5 cu

J3d rHCO rH

CU4-J OP> CT\ 4-1

H ,d rH53 •u 1 CNv_^ O \£> CN

rHCO >-l

o ®4->

rH3CO

O4-1 d

CO I

as

umbe

0)~

4-1 m drt o CO

LT> d w <u

d • o J ooCO

o CJ pq co

<! 4-<

•H II II H ^}-i

rtd

CO d g coPuQ d Oe cu Mo Pi 4-1 o

u • o•—

s

• X) CNcu > • CD

o • M 13d 13 CU dCO d • ,Q cO

cO cu

o d mo cu cfl d rH4-1 a. 4-j

U cO u CO 13CU ,d d CD CD

p-4 CO CO O V4CO CD

44 ,n• H B

CO 3 2co dCU

J-4

CO3 • •

o •HPL,

HOS3Or

<D CO

IDHZ

34

QJ

OC

• co

a) ,c• d

•h qj

co rC

QJ

cn siCO uu o

,£>

O J-i

M-l O

Oco •

PL, O

P4QJII

4->

O CO

oo

QJ

CO

4-1 •

iH •->

3 4-1

CO C

e< 4-i^

a.-o .

o

CM

QJ

a

OcO

I-l

CO

CO

o

o4-1

II

CL)

CJ

cC

oCO

•HucO

6O T3 H-l

CO 3QJ CO

O QJ

cO cO -H8 J! CO

M CO iHo a.*U OJ

oj co aPM co cO

cr.

QJ

U

•H

OV-i

<4-l

CNCM

U0)

e

&QJ

acO

M-l

M

WHOis

©I COOLIO.

35

oII

•HCO

^"" rH> PU rH

rH• TJ CO

cu C• CO CO

•H CNv_^

CU T3,£>

a CO

cu CO

CO J3 ft

CO a CNCJ> CU CN

r-l rC #*

o 4-1 CTi

IW o rH

/—

>

ncon 00

P> o rHH m55 -

ft

0) rH •

£o Oin •> CN

a • mv_/ o rH 13

oc0)

CO

4-1

rH

ii

PO

CO CO

CO

M rH

DCCO

cu

Pi • W cu

r-l r-l

a 4-1 pq cu

o cCO CO H g•H 4-1 3M CO s aCO a oeu o r-l CO

e u IH CU

o ou II TJ CO

cu mcu Pm 5-4 V-i

o cu 3a ** o ,£> CO

c0 4J CO

B C cu 3 cr-i CO CJ C cu

O 4-J CO CU

<4-4 CO M-l to ts

r-4 C r-l CU 4-»

CU o 3 o cu

Pm o CO CO rOl»-l

u cu• 3 -Ho C/3 rH

rH

CU • •

COo

U

M•H

Ho

36

•HII ctfH£> P.H cu

^ x) •HC rH

• n3

CU rH. tj rH

•H d) uJ^ Oc P~)

nJ CNo

a •>

CU CM0) CNto ,ccd 4-» •v

C_> O cr\

rQ rHrJ

o u *S

m o 00M-l rH

0) «

> = vD*->. o rHwn

> ° •>w o mrH o

CO || O4J CO CNrH U aJ

3 Q T3CO rH ti

CU M rtJ

Pi •

/^"X W r»c 4J rJ rHo cco cfl sa T3•H U H OJ

u to Hco a 6 CU

CX O o ,2& o u E=

o 14-1 3O II

idG

<u rn cu CO

o u 0)

G - cu uCO -U o ,a nJ

6 £ cu S IHrJ Ctf o 3 nown (3 3m co cw (0

V-i P! U CO

1) O 3 cu CPh a co o CU

CO CUm &• ^ 4J

rH d a)

rH CO rQ

CU

M • •

3 w00 H•H OP^ »

37

o. a.

38

diameter of 0.50 inches. Comparison of each individual surface to a

smooth surface having the same nominal diameter as the enhanced

surface is the best type of comparison, if the comparison is under-

taken for surfaces having a common nominal diameter. When the plate

spacing varies, this comparison method is invalid Comparison of

many surfaces to a single common smooth plate nominal diameter permits

a relative comparison between surfaces having different nominal diameters,

Through use of this method, all plate-fin surfaces from K-L (4) will

be considered. Since a nominal diameter of 0.50 inches has been fully

investigated, only those D = .50 inch surfaces that bound the

maximum or minima for each case will be shown in Figures 13 through 17.

Figures 13 and 14 show the case for the same shape and V =

constant or case a. Figure 13 shows the ratio of ordinates (NTU /

NTU ) vs. Re and Figure 14 shows the ratio of abscissas (P /P )s ns b e s

vs . Re ons

Figure 15 shows the case P = constant, V = constant or case b

(i e NTU /NTU vs„ Re ).e s n

Figure 16 shows the case NTU = constant, P = constant or case

c (i.e Q V /V vs B Re ).s e n

Figure 17 shows the case NTU = constant, V = constant or case

d (i.e. P /P vs. Re ).s e n

Higher ratios are preferred for Figures 13, 15, 16 and 17 while

a lower ratio is preferred for Figure 14

39

Calculations were a made f jhe five pin-fin surfaces of

reference 4. Only the best of these surfaces is shown in Figure 7.

They are not shown in any of Figures 8 through 17. Since the best

four have the same performance despite having different nominal diameters

Surfaces 28, 29, 31 all follow surface 30 of Figure 7 while surface

32 lies atop surface 1.

40

Q) Oco mc0 •

u oM II

ohh dO

co

H 4->

"~^ *CD

e sS c0

}H

co 34-> COrH3 43CO 4-J

QJ OPi O„ eC cooCO o•H 4-1

1-1

co /'--.

CX QJ

e sO 3U rH

QJ >aa -aco ae co

uO QJ

<4-i a,U CO

QJ 43CX, CO

rH

43•u•H£

CO

QJ

Oc0

UHJ-l

CO

a•HMH •

I/~-

QJ >^4-1 rHCO QJ

rH >a, -h

•u>H OO QJ

mh a,CO

f3 QJ

3 r<

O

CO

qj aU to

in

E H13

cd

CO

i•HXCO

e =oqj m43 •

4.) O

eo Q

a) ••

^ W3 HbO O•rH £3

41

i i i i i i—

i

r 1 1 i i i i—

r

oo

ID

O

O 0)

11 I I » I I I

COo

QJ £1iJ

co •Hco £

, co

CD QJ• CJ

•H CO

• h~ 3cO O CO

m• a

QJ O •HCO CMcd II 1

•

U QJ ,—

s

d +J >.Q CO rH

u rH QJ

o Cu >m ,d •H+-> U 4-J

^-^ «H O CJ

co£ IW QJ

Pm PI-

-->. QJ

BCO

QJO QJ

P-< co o W**s H-l rd

r-l co <rco 3•P CO QJ T3rH H d3 X CO cO

co j-i

QJ O rH v£>

Pi o CO rHB 6

d w •H CO

o d uCO o •H QJ

•H +J 6 £>^ £cd /~v T3 3Cu QJ d de e CO

O 3 QJ

U rH rH >O cO H

QJ > E 3CJ H od Td X »—

'

as d cO

e co e ~

u OO 0) QJ UO

u-i a. 42 •

u co +J oQJ rCfa CO

rHII

d d• o Q

<rrH

##

QJ wH H3 OW) 2•HPn

0). CO0L |Q.

42

x:• • 4-J

cu r •H• o &

•H mN*-^ • CO

o CD

OX> II cO •

a r-l >.QJ Q 2 rHCO CO 0J

co ,d >O 4-1 d -h•H •H 4-1

£ M-l Or-l 1 CU

o cu cu a,iw o 4J CO

cO co a)

/—s iw rH J-l

U)U a,t3 3 CNH co J-l rH5z; O"">- X! <4-l •>

0J4-J -<r

to o cH O 3 T3is 6 O d^-y CO

CO

CO o r^~

c •P 4-J

rH0) rH

0) 3 /-v cO *>

CO +J 00

cc a) dPi 0)

cO

B co

4-1 •H >-i

d w d cj

o a H ,0CO o 6 B•H CJ dM ti d03 II da, co cu

B > >o CO r-l

u •> e d4-1 •H cj

0) d x ^O cO cO

C 4-> s =CO CO oS d ai mC o X, '

o o 4-> om^ II >^ il

oj rHPn CU d d

O Qo

m • •

rH

(1) OM »360HtH

43

om ^• 4-»

o •H

II

• CO

0) c CU

• Q CJ

•H cO •

s-^ CW /—N.£ r-l ^4J d rH

CJ -H CO a)

£ >a •H

a) cu •H +Jco o <-H CJ

CO cO 1 0)

O M-t CD o<M U CO

3 CO a)

M CO rH >-i

O P*M-l ,£ CN

P M H/-% o O

cuo lM n

> b -cr

CO Cj

CO 3 13> o CJw o rC CO

4-> CO

CO r--

•U /-^ CD rHrH U S-i

a a cO r>

co c0 00CI) -u CO

Pi to B CO

C •H >-l

a o c CU

o o •H ^>CO b B•rl || 3M T3 ti

CO PM aa. cO CU

B - >O -U cO uu d B 3

c0 •rl oCU JJ X v-'

CJ CO cO

a c B zcO O ob o CD m>-l rC •

O II U oIH>-l ^ >^ II

Q) H rHPM !3 a ao o•

\o • •

rH

a) OH S3360•HPm

CO.

> >

kk

10

ii 1 1 i i—

r

UN I I I—I

Hill II I

O

I I I 1 I I L mill i l n i ii i i i

CO

COo

Cs!

O O

o 4JIT) •H

• &oW

• II QJ

QJ O• c cO •

•H O IW <-~N^—y *-< ,

rG 3 iH4-> CO QJ

X) -H >£ a H

•H 4J

a) qj LW Ow o 1 QJ

cO co QJ a.U 4-1 4-1 0)

!-i CO QJ

3 iH u>-l CO a,o CNM-) ,d u H

•U o/-N O u-i •N

a)o <rPm e d"-^ co 3 TJw O d

(X, X cdw O w4-1 r-»

in QJ rH4-> /^N M<H 4J CO #>

3 a COW cO CO

qj +-> B wPh w •H V4

a C QJ

a o •H rOo o E eCfi 3•H II T3 d>-< dCO > cO QJ

a, >e - cO uO 4-> 6 3u d •rl o

cd X v_^

QJ 4-1 CO

O W 6 za d ocrj O QJ mB O .C •

^4 4-1 oO II

4-4 >. II

U JZ) iH0) H c dPH 55 o p

# , B

r^- wiH HOQJ J23

u3toO

•HPn

45

IV. COMPARISON WITH SAND GRAINED ROUGHENED SURFACES

Many enhanced surfaces have been designed to increase the per-

formance of heat exchangers „ Dipprey and Saberski (3) determined

the heat transfer coefficient and friction factor experimentally for

rough tubes containing a granular type surface with rougbness-height-

to-diameter ratios (e/D) ranging from 0.0024 to 0.049. The best

of these surfaces from the standpoint of higher heat transfer coefficients

for lower friction factors (i.e. e/D = 0.049) was converted to the

3proposed performance parameters and plotted as j Re vs. f Re

in reference 11. Figure 18 contains a plot of the performance parameters

2 3 4j Re /D vs. f Re /D for the Dipprey and Saberski surfaceJ n n n n n n rr j

using D = o 50 as well as the best two K-L plate-finned surfacesn

having D =0.82 inches and 0.50 inches respectively. Also shown

is the smooth surface performance curve for D =0.50 inches.n

46

Tff»o

t3Gca

co

CU

OCOt>JM» mo 3

CO

Tm

Plate-fin

rface.

CM 34-J CO

r=aaiCO

O CU CT>

PQ <f

T"1

v;-

OO

t"! C •4-4 li

1

"—i

CO QCU ^> CU

UCO 3

V" s-~\

*b cu enQJ'—'4-1

^ i DCC Q CU -H

cO co

S-i l-i

»»— 1

nee

Pa

&

Sabe

o cU

U <U'6 O 5-i

U-J EL,

0)gure

18,

PerDip

r-°

47

Vo RESULTS OF COMPARISON

. 2The higher a curve lies on the performance curve of i Re /Dr n n n

3 4vs. f Re /D the better the surface performance. Inspection of

n n n

the curves for all surfaces suggests that the following surfaces, or

groups of surfaces are best in order of decreasing performance.

Ranking Number Surface Number Plate Spacing

1 27 .41

2 25 .41

3-11 8, 15-20, 23, 26 .25(8), .41(1)

12-14 21, 22, 24 .25(2), .49(1)

15-20 13, 14, 28-31 .24(1), .25(2),

.40(1), .50(1)

.75(1)

21 6 .33

22-27 4, 7, 9-12 .25(5), .42(1)

28 5 .48

29-30 1, 32 .47(1), .51(1)

31-32 2, 3 .41(1), .82(1)

48

VI. CONCLUSIONS

1. The comparison method of La Haye et al. (5) was modified

to compare heat exchanger performance on four different bases:

a. Same shape and volume of heat exchanger.

b. Same exchanger volume and pumping power

c. Same pumping power and NTU

do Same volume and NTU

2. All of Kays-London plate finned surfaces, unfinned surfaces

and a sand roughened surface were compared on the above bases.

3 The "best" surface of those compared is found to be the

wavy-fin plate-fin 17.8 - 3/8 W with plate spacing of 0.41 inches.

49

REFERENCES

1. Bergles, A. E.; Bunn, R. L. ; Junkhan, G. H. , "Extended PerformanceEvaluation Criteria for Enhanced Heat Transfer Surfaces." Lettersin Heat and Mass Transfer (1974), Vol I, No. 2.

2. Bergles, A.E.; Blumenkrantz, A. R. ; Tabolek, J., AIChE reprint 9

for 13th National Heat Transfer Conference (1972); Paper FC6.3for Fifth International Heat Transfer Conference(1974)

.

3. Dipprey, D. F.; and Sabersky, R. H. , "Heat and Momentum Transferin Smooth and Rough Tubes at Various Prandtl Numbers." InternationalJournal of Heat and Mass Transfer (1963): Vol. 6.

4. Kays, W. M. ; London, A L. , Compact Heat Exchangers . 2nd Edition,McGraw Hill Book Co c , New York (1964).

5. La Haye, P. G.; Neugebauer, F J ; Sakhuja, R K. "A GeneralizedPrediction of Heat Transfer Surfaces " Journal of Heat Transfer

,

ASME (November 1974): Vol 96:511-517.

6. Mack, W. M. , "Evaluation of the Heat Transfer Performance of ThreeEnhanced Surfaces Including an Enhanced Surface Heat ExchangerPerformance Comparison Methodo" Ph D dissertation, MassachusettsInstitute of Technology, 1974.

7. Moody, L. F. , Transactions , ASME (1944): 66 -671„

8. Rohsenow, W. M. ; Choi, H., Heat, Mass and Momentum Transfer.Englewood Cliffs, Prentice-Hall, Inc c 1961.

Thesis IE 378S663S54 Soland

Performance rankingof ptate-fi nneH uea*

exchanger surfaces.

thesS663954

Performance ranking of plate-finned heat

3 2768 002 01665 1

DUDLEY KNOX LIBRARY