__

,

___

__

fi

4

)

h

fl!!:s

Tile EFFECT OF SWIRLING FLOWtvi ON PIPE FRICTION LOSSESy

':

I$

.

hd

by} Carl R. Jani!.to

Mahadevan Padmanabhan

3!;;

il4eF

Research Sponsored by

Yankee Atomic Electric Companypj Public Service Company of New flampshireyfil

'

Y ALDEN RESEARCH LABORATORY: -

7 WORCESTER POLYTECHNIC INSTITUTE

vLCg 26-81/M296KF

I7ebruary 19hl)

~ |] | C I N :p u : ; ; .,_ ^vocu uwen;.,

PDR

_-

. ,,

TIIE EFFECT OF SWIRLING FIDW

ON PIPE FRICTION IDSSES

by

Carl R. JanikMahadevan Padmanabhan

Research Sponsored by

Yankee Atomic Electric CompanyPublic Service Company of New flampshire

George E. Elecker, Director

ALDEN RESEARCH LABORATORY

WORCESTER POLYTECHNIC INSTITUTE

IIOLDEN, MASSACllUSETTS

February 1980

. . . . . . . . . . . .

- - - -.

.

. ,

., ..

ABSTRACT

Results obtained from tests conducted on the Scabrook Containment Sump model

at the Alden Research Laboratory (ARL) revealed small amounts of swirlingflow within the sump suction pipes. The effect that swirling flow has en

pipe friction losses, and in turn on the Net Positive Suction Head require-monts|of the Emergency Core Cooling Recirculating pumps, was not understood;

therefore, it was decided beneficial to conduct a separate study in order todetermine the effects of swirling flow on pipe friction losses. The study

was conducted at ARL and sponsored by Yankee Atomic Electric Company.

An experimental facility was constructed consisting of: 34 ft of 10 inch

diameter pipe containing a swirl generator; a series of pressure taps; a

flow meter; and capabilities for placing a vortimeter either upstream ordownstream of the pressure taps. The system had the ability of producing

swirling flow from set swirl generator values of from O' to 40' and for

Reynolds numbers in the range of 1.0 to 3.5 x 10 .

Prnliminary tests of the system revealed that the discharge coefficient ofthe orifice type flowmeter was dependent on the set swirl angle. There wasan approximate 5% increase in the discharge coefficient for 40' swirlingflow compared to non-swirling flow. Results indicated that swirling flow

had practical effects on the frictional losses; for example, frictionlosses in the pipe due to swirling flow with 20* swirl vane settings wereapproximately 40% higher than those for non-swirling flow, at the corres-ponding Reynolds numbers. Swirl decay was found to vary with swirl angleand also to depend on Reynolds numbers and the results obtained were con-sistent with the findings of other researchers.

)

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ - _ _ _ .

-,. .

11

TABLE OF CONTENTS

Pago No.

-ABSTRACT i

- TA13LE OF cot 1TE!1TS 11

IrlTRODUCTION 1

SWIRL AllGLE 2

TEST FACILITY 3

TEST PROGrWt 5

RESULTS 6

CONCLUSIONS 8

ACKNOWLEDGEMENTS 9

REFERENCES 10

PilOTOGRAPilS

FIGURES

|

1

i'

. . . . .

-m

'

-

INTRODUCTION

5.

~

'llydraulic model tests on.a 1:4' scale model of the Scabrook Nuclear Power

JStation ~ containment sump, conducted at the Alden Research' Laboratory (ARL)of Worcester Polytechnic -Institute (WPI) (1)' showed the existence of small

amounts of swirl in the flow within the suction pipes when operating with

> partially blocked vertical sump screens even though no strong free surfacevortices of concern were observed. The swirl angles, as indicated by avortimeter,-were mostly less than 5 degrees, which may' be considered toosmall-to cause' pump impeller vibrations. However, realizing the necessityof accurately evaluating the available Net Positive Suction Head (NPSH) for'the recirculating pumps, the effect of this swirl on the entrance loss and

-pipe friction loss was of concern. Determination of total intake. losses,

including entrance losses, was a part of the sump model study (1) . However,

it was decided to conduct a separate investigation on the effect of swirlf .on pipe friction-losses so that differing degrees of swirling flow could be

established to provide more general information on the subject. This wasparticularly desired since swirling flow in suction pipes could also becaused by combined bends in piping systems as well as from any rotationalflow in the sump (2) .

Although~several studies have been completed on swirling flow in pipes,most of.the research has been conducted at small pipe diameters and highswirl angles (3, 4, 5, 6) . These studies are not representative of theconditions that exist in a reactor sump where pipe diameters are large,flowrates are high, and swirl angles are relatively small, less thaa 15*.Therefore, specific results concerning the effect of low intensity swirl-ing flow on the frictional losses in sump suction pipes are not readilyavailable.for a designer, and the need for an additional study is evident.

' A separate experimental set.up was constructed at the ARL, and considering.the usefulness of the study to the design of the Seabrook ECCS system, the-

project was sponsoted by the Yankee Atomic Electric Company (YAEC) . In

J

___ _ , _ _ _ _ _ _ - - - - - - - . - - - >

_

2

order to accurately determine the effect of various degrees of swirling flowon the pipe friction factor, a systematic test program was set up, wherebyboth the degree of swirl and the flowrate could be varied within a 10 inchpipe. The degree of swirl was varied using a four vaned swirl generator,enabling an accurate and easy method of setting the swirl angle.

SWIRL ANGLE

Swirling flow in a pipe can be considered as a combination of vortex motionand axial motion along the pipe axis, which means the existence of axial com-ponents and tangential components of velocities at any point in the flowfield. For the problem of concern in the present investigation, the swirl-ing flow is turbulent, incompressible, and steady. Along the length of the

pipe the swirl level decays and the velocity and static pressure distribu-tions change with axial position along the pipe.

For the present study, the rotational speed, e, of a vortimeter which con-sists of a vaned rotor containing four blades, was used for swirl measure-

ments. The swirl level was calculated using the measured w values and ex-

pressed mainly as an indicated swirl angle e defined as (7) :

0 = tan- " ! (1)u

where d is the internal diameter of the pipe and u is the average axial

(cross-sectional) velocity of the flow. Equation (1) can be rewritten in

terms of vortimeter rotations, n per second:

-l u n d0 = tan (2)

u

The use of the term w or n to indicate swirl level was considered conven-ient and easily understandable. A more commonly used fluid mechanics term,

the angular momentum of flux, K, could be derived using the rotationalspeed e of the vortimeter (6), as given by:

&

IuI III M I I HI I i4 i dM l I i i I

. . _ . . ___

|

| 3|

2 2 2K=mpurw (r +r g )/2 (3)

where o is tne water density, r is the pipe radius, and r is the rotor hubg

radius of the vortimeter. Equation (3) assumes negligible losses and rigidbody motion. With the axial momentum expressed as M = w p u d /4, theswirl icvel could then be defined by a parameter K/Md.

Combining Equations (1) and (3) , a relationship between 0 and K could beobtained as,

~l K0 = tan [ ] (4)2 2 2wpu r (r +r y )/2

The swirl decay between two locations in the pipe at an axial distance ofAx apart has been shown by many investigators (4, 6, 8) to follow an expo-

.nential function which could be defined as,

K -0 Ax/d""K (5)o

where K and K are the upstream and downstream angular momentum fluxes, re-9

spectively, a distance Ax apart, and 6 is a decay parameter which depends onthe Reynolds number, Re = u d as well as the upstream swirl' level indicated,y

bv K /ttd.' o

TEST FACILITY

The test facility used in this study is shown in Figure 1 and Photograph 1.Water was drawn from a sump through a 12 inch line by a 10 lip pump. The flowthen passed through a 12" x 10" reducer and a series of flow straighteners,initially consisting of two perforated plates separated by one pipe diameter.In order to ensure a flat velocity profile in the pipe, a pitometer traversewas obtained downstream of the straighteners. The initial velocity traverse

showed a higher velocity region away from the center which was remedied by

I

. _ _ _ - _ .

4

placing a ring, constructed of perforated plate, between the plates of theflow straightener, the,reby directing the flow towards the center of the pipe.The velocity traversing points and the 'inal velocity profile obtained are :

shown in Figure 2.

The swirl generator was located 8 diameters downstream of the flow straight-ener and, as previously stated, consisted of 4 adjustable vanes. The 4 vanes

GO' apart in the circumferential direction, one diameter long and occu-were

pied 90% of the diameter of the pipe. Vano angles were adjustable from 0*

to 40* with respect to the axis of the pipe, and had an attainable settingaccuracy of 11/2*. The swirl generator is shown in Photograph 2.

A series of 10 static pressure taps were located starting 11 diameters down-

stream of the swirl generator. The pressure taps were set one diameter apart

and connected to air-water manometer tubes (Photograph 3) . In order to re-

cord the piezometric head from each tap at a particular instant of time, pho-

tographic documentation of all the taps was made and measurement of the pie-zometric head was obtained from the photograph. -

Both upstream and downstream of the pressure taps, spool pieces were providedfor the placement of a cross-vaned vortimeter. Photograph 4 shows the vorti-

meter with its 4 perpendicular vanes, each 4 inches long and 8 inches wide, at-

tached to a rotor hub of radius 0.9 inches. The revolutions per any desired

period of time were counted using an optical sensor and an electronic counter.

To keep the pipe length of the system constant, which was necessary in order to

obtain accurate swirl decay data, a blank spool section equal in length to the

vartimeter was placed in the space not occupied by the vortimeter. To ensure

that the vortincter did not affect the pressure gradient measurements, read-

ings of the gradeline were taken while the vortimeter was in the downstream

position.

An orifice meter, coupled to a differential manometer, and located downstream

of the pressure taps, provided a means of monitoring flow throJgh the system.

Since the orifice meter was located downstream of the swirl generator, it was

necessary to establish the effect of swirling flow on the flow meter discharge

coefficient for various swirl angles. In order to estabifsh this, the sectionof the facility from the swirl generator to the meter v.as removed and calibrat-

ed in ARL facilities. An accuracy of 10.25% is expected for flow measurements.

; /I

-

5

In order to be representative of large scale reactor sump models, a rela-

tively large flowrate was ner ded, although the naxinum flowrate attained. in

the model was limited to 2.5 cfs. A control valve was located downstream

of the orifice plate and was used to vary the flow through the system. Flow

was routed back to the sump, thus completing the cycle.

TEST PROGRN4

In order to accurately determine the effect of swirl on the pipe friction

losses, a test program was set up using six swirl angles, 0, 6, 10, 20, 30,

and 40 degrees, each tested at 12 flowrates vf between 1 and 2.5 cfs. At

each swirl angle and flowrate, pressure gradients for 9 diameters of pipe

length were obtained. Vortimeter readingo were obtained both upstream anddownstream of the pressure taps for the above swirl angles and also for 15,

25, and 35 degrees.

The, friction factor for the specified pipe lengt1 was obtained from the

Darey-Weisbach Zormula:i

[ f = Ah ( ) ( ) (6).,

,' / u,

/ wh rg,.

a,

'

f = calculated friction factor, . ,a f /

'

Ah = masured pressure differential over 9d!s

i

( d a diamet ar of pipe, 10.0 inches4

e, -L = n'ina,dianeters length = 9d'

''i ,u = nwyrage velocity calculated from u = Q/A

23 / ,i+.5 Q measured by the orifice meter and A = nd /4w,

,

<<

,

t- 1

y; Anjiverage pressure gradient for the 9 diameter length was obtained from a*v \ <

ilt.st 'sriares fit through it.dividual readings taken at one diameter spacings.[ // i /\

' { /,' i'

f , Knofn'ef 'the'humber of rotations of the vortimeter, n, the indicated swirl3

.\- ''

angle,', A wad' c,btained from Equation (2) . By determining the indicated swirlL

,

/ i y ,"

tagle at>the two vortimeter locations, a percent swirl decay over 17 pipe I3

/ '. 's,

IUlameterleihithscouldbedeterminedfrom:/

3 j, ,<

' \

s '( ' .< ,

t i

/- [E . ' , ,<o <, ,

1 f p' \ ! 'I{ $*'

,,

.s /'s . f ). j ' } *'

. _}f . [- 9y~ ' ' ~ q. ._ - ', _- .. ' , , -: i ' } [ '' ~.

' * '

_

._ >

6| e

,

:b. i. 0 -O

upPercent Swirl Decay = " x 100 (7)

6.up4.]E

where 0 and 0 are the measured swirl angles upstream and downstream of

the pressu;e taps, respectively. Decay rates were determined for all flwsand swirl angles. The values of the angular momentum fluxes were also calcu -

.

lated using Equation (3) and the decay parameter S was evaluated using Equa-

tion (5). As S is a function of R and K,/MD, curves were obtained for Sversus R, with K,/MD approximately constant and S versus K /MD with R, ap-proximately constant. An average K value (K,y) was obtained from Equation

1 L={f K dx, and K was used to denote the average swirl level(5) as Kover the pipe length of L (L = 9d for frictior, measurements) . O was de-

rived from K . using Equation (4) . This enabled the plotting of percentage

increase in friction loss due to swirl versus 0 curves, for constant Rey-

nolds numbers; These curves may be used by designers to predict approximstelyj

| the frictional losses'for pipes with swirling flows, if the approximate swirl

angles and the frictional losses for non-swirling flow at the Reynolds number~

!s of interest are known.;

g-.

6

RESULTS

I 'As shown in Figure 2, a fully d)veloped velocity profile was obtained up-

| stream of the swirl generator by , roper flow straightening devices. No re-

sidual swirl was indicated by the vortimeter for a zero setting on the>

'' swirl generator.

, , ;

IO The results of the orifice meter calibration are shown in Figure 3, in which'

,

3 -the angles set at the swirl geaerator (0 ) are indicated to denote the.

''

swirl level. At the largest swirl angle tested, 0 = 40', the presence of

swirl increased the discharge coefficient by approximately 5% for all flow-

- rates tested. The effect of smaller. swirl angles.was progressively less, as-

shown in Figure 3.J

,j.[ -In order to verify the operation of the swirl generator, vortimeter readings

were taken at~the upstream location and' transformed into indicated swirl an-gles using Equation (2) s The = indicated 0 values (0 ) were then compared to

7

, t :7". .

-. . . - - -..

7

'to the set readings (0,, ) on the generator. As shown in Figure 4, for setswirl angles less than approximately 15*, the indicated swirl angle is less

than the' set swirl and for larger set swirl angles, the indicated swirl was

seen to be slightly. greater than the set swirl. This fact could be caused

by adverse flow patterns past the swirl generat'r involving flow separationo

at the vanes.

Because.of the swirl decay, the pressure gradient in a pipe with swirling flow.

} might not be linear. However, in this investigation ove the nine pipe dia-meter length, the gradient was assumed to be linear since the non-linearitycould not be accurately defined due to experimental limitations (piezomstricheads.could be measured only to an accuracy of 0.005 ft). Equation (6) was

used to cciculate an equivalent friction factor f.4

Results of the effect of swirl on the friction factca are shown in Figure 5.As shown in the figure, the friction factor ir. creases substantially with setvane angic. The impact that these results would have on calculating pipe fric-tion losses can be seen more clearly in Figure 6. In this figure, the per-

contage increase in the friction losses.for a particular average swirl angle,0, , compared to zero degree swirl, is plotted against the average swirl an-gle for the region of no R dependence (i.e. , R > 2.0 x 10 ). At the swirleangles of interest for the Seabrook sump (0 up to 5 degrees), an 1. 1

of up to 10% in the friction losses, over that for non-swirling flow cat .

expected..

Figure 7 presents data for swirl decay over 17 pipe diameters. As seen in4

the figure, there is a higher decay rate for.small swirl angles and a lesserdecay, 20%, for larger swirl angles. Thus, in the range of interest (up to

20 degrees), more than 60% of the swirl could be expected to decay in 17 pipediameters. Dependence of decay rates on swirl intensity and Reynolds number

~

was also observed by other researchers (5, 6).

.

~.

, . . . - . . - -

,-

,

|i

'

8-i

Figure 8 shows the variation of the decay parameter, 6, w'ith the pipe Roy-

i: nolds number for various set' swirl generator angles. 8 is more or less-independent of R 'for R greater than 2-x 10 and swirl angles greater than

- 25*. 6 increases with decreasing R, for a particular set swirl generator; vane angle (6 ). Also given in Figure 8 is a comparison of the range of

|- (K,/Md)_for a given set swirl generator angle. This Reynolds number depen-'

donce of 8 is consistent with the results in references (5, 6).

CONCLUSIONS

The following conclusions can be drawn from this study:|

1. Swirling flow hac an effect on the perfomance of an orifice meter.

At the highest swirl angle of 40 degrees (swirl generator vano set-

ting) used in this study, the orifice meter discharge coefficient

( increased by about 5 percent compared to the same flowrate with no

! swirl. For small indicated swirl angles (less than 10 degrees), the!

| effect of swirl on orifice meter perfomance was found to be negli-

| gible (less than 1 percent).

2. The effect of swirling flow on the friction loss is dependent on

swirl intensity. For an average indicated swirl angle of 5 degrees

(nurmally encountered in suction pipes due to inlet rotational flow),3

the iricrease in the frictional loss s. ,ul 1 be approximately 15% com-. pared to that for non-swirling flow at the same Reynolds nu der. Pc

| a_ higher swirl angle, the increase in friction factor would'oe greate.|

3. Svirl decay with distance is a function of Reynolds number as well~| as initial swi.*1 angle. Ilowever, for R greater than-2 x 10 , swirl'

decay is more or less_only dependent on the initial swirl level' for

the larger swirl levels tested. The decay parameter, 8, decreased

i' 'from 0.056 to 0.014 for increasing swirl generator vano angles of 20to'40 degrees (or for K/Md values from 0.10 to 'O.37) for R gr ator

e5

than 2 x 10 ..,

J

m

, . .. . --.

9'-

;,

ACKNOWLEDGEMENTS

~

,1

The. help provided by all-members 'of the staf f at the Alden Research. Labora-

tory in' conducting the study is,' acknowledged. Particular thanks are due to~

:

Professor George E.: Hecker for his helpful suggestions during this .investi-.gation.-

.

. Special mention must be made of the work done by Messrs. J. Noreika hndM. Majcher'who con'.ributed greatly to the. design and construction of the

'

experimantal setup and carried out the experiments witl. considerable care.-

; Financial support for this research was provided by Yankee Atomic Electric !

Company, 2nd their interest and support are greatly acknowledged,

r

b

'

.

.

\''

____. __ ._ ______ _ . _ _ _ _ _ _ _ _ _ . - . _ _ _ _ . ._ -_ _ _ _ _ _ _ _ _ _ _ _ _

10

REFERENCES

1. Padmanabhan, M. , " Investigation of Vortexing and Swirl Within a Con-tainment Recirculation Sump Using a Hydraulic Model - SeabrookNuclear Power Station," ARL Report, May 1979.

( 2. March, P. A. , and Noreika, J.F. , " Investigation of Swirl and Axial

Velec.ty Distribution in Suction Piping," ARL Report No.127-77/M220AF, August 1977.

3. Kreith, F. , and Sonj u, O.K. , "The Decay of a Turbulent Swirl in aPipe," Journal of Fluid Mechanics, 22, 1965.

4. Youssef, T.E.A., "Some Investigations on the Rotating Flow With a

Recirculating Core in Straight Pipes," ASME Paper 66-WA/FE-36,

1966.

| 5. Rochino, A.P., and Lavin, Z., "An.51ytical Investigation of Incom-

pressible Turbulent Swirling Flow in Pipes," NASA CR-ll69, 1968.

I' 6. Daker, D.W., and Sayre, C.L., " Decay of Swirling Turbulent Flow of

Incompressible Fluids in Long Pipes," Flow - Its Measurement and

Control in Science and Industry, 1974.,

t

7. Hattersley, R.T., " Hydraulic Design of Pump Intakes," Journal of

Hydraulics Division, ASCE, November 1974.

I!

8. Davis, R.W. , et al. , " Numerical Solutions for Turbulent Swirling

Flow Through Target Meters," ASME Winter 7.nnual Meeting, San

Francisco, California, August 1979.(

t

n . .e -w a, xwa .-,a.. - xs _.4.,a a,s. ..- .a a. . ,w~-a w a--a ,,-+.s. . .s.u,a a _ .-s. a . n-.a = a .- e a_.. ,-u. s.-, - -n-, -

'i

e.e-

'

.

_

r

!

!'d

J

,

'

i,

1:(4

I

J

2

i

f

f ^f.

r

-_ h

(tI ,i

p - >

T Io

I

'!a

1-'

b

! PHOTOGRAPIIS,

t

!.

i .i4

8

.I: e

k f

I

!

! *

id

t.

. .t'

1

.

.

b

b

^

.

: i

'. --- , . , -., . , , ..r...E..--. , , . , . . ,,',.#..-.-_ . . - .E_.--,, - . . _ . , . - .~.e,. m., . . . . . . , . . . _ _ . .- .-. - . . . . ~ , .-- - - - - e,.-. . - - ..E -

_ _ _ _ - - _ _ _ _ _ . _ _ _ _ _ . ._ . __ , _ __ _. . - __ __ _ -

1.

E 11(

.:

%'__ ' % , ,. 1

.

s - . . . .

s ' T,A; |- ~~ , ;4...

-' , y. i sg, .

., 76 ..

.

s - '4. - -;

N i st? - q )'

j

/$bb.s a

a) Upstream of Pressure Taps,

l

7. , w., ,.

''I )' ' * -4 .. .

I'

*,

| !! i

b .

[:; .Ig;.|i

---

b !_ |c.- e

7 1 _.i . -

X- YV.

,,.

,-, g

I

b) Downstream of Pressure Taps

ihotograph 1 Experimental Set-Up

.-. _ _ . ._ . . _ _ _ - _

_ _ _ .__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ . . _ . _ . _ _

-

|u . . .yc,. . . .. : + c. . . -

-- ,- ... . s ,x ,. : ..

g . ,s . . e ,. s.-;- .v r .,7 - p.x;

. '

., .

s4 '. .?. .g.:. t

' -. ' 55( ^- .sO.

..

J'q

_g!_ -: 3.f-. . g, . f r. ' " :. 4,,,k, y- gg;

g p..

3. . . . c'f. ., ., s.i.y - ,, , .

'

7 4 Y .!]f .. ;% 2

le yy . , .s,

g . .- _ .;_ ~: _ ., .. ,- .

.. - . ,n s. g - A z.n. . .r

,

,.:... . ,c ... ; 1 .(, s. . .-

.

r, 3 1.;. ,

-. ~ . ,p--

.s; N :+, ..

g p. : ,.,y .- . .y, ,

c N 9 -

* T ,f.: (,y? ' % _*

./ a. . .

-..

;4,. . . )', G. ' ; ' _.

V .'1 -". ~: :, . :. .. :~. i * ., e ; ' . + ' . + , . e g-. t.= :-

o- .~ w . .

3- ;,.., ; . .::, ; ^ ~ - '.;j;'_ { q2 . [..

. ,.,;. y-t ..

..

,. . pr , . :., . .,--.;.F . - - . M ,._5

g, . - .. | :; 1.; *.._ _

| p- .. ; _-| '. w,

a s: ,

'' ' A ,,z g,. .,y % . "f)f . - ' Q;. .

.

;-n,.

. p < .

.,

, . '-

. . : ". ~' .$ ' _C. j...;

...

._ f . , . Q .+ . . . 3. .4g?. .q. . . , . , . . .,.

g. p t. . . v -.

. <. .u. s . :,"1.2.'U' 3 .'

. , ,.

.-

.

_|~~ in' 'Q.), ;, NygJ, g. g,

.

s ,as. . , ': ~ ' "Yi. ]:,;. - || z.(-f Q~, . .

< , ~.;: ,L.,, )W. W.7 -| g

V^.- 3_

?.?? : ~;: . .._

? -m m- . . - ' 2 ; . v . . .: . . ..,

l'hotograph 2 Swirl Generator)

|'

f

w.m.,- 7-m-:q77

t-

i i5 i .

a*

~

.- - :n.4- , u i, , 1mo- ,

a. 3 - ,

.j, .

j. \p so .f) w.x .. . . .

. ..

- s; .. i-,,{t< k ( 's i

. .. [-fst| ['; s,- 5,

...

,.m - cc .

! L a q ::.9 c,umM i 19.yl h uj 0 . , 5 ; 5 . ' <{.

1:

' 4 T M: ; :. ' >.

w imb t :-;-

.,.,ci XL .

j-

7 ,9 1 * % 3 h a: .a g ti b ;; "I!5 S h Ljk ( ' a. Sj i

g >t{. '1 n >, . , d j f. . 'jt.i H'-:y-1!n ; ;. or i..,...:g: n -|:. ..

:a ,4, . c sp u-n<

! g.c ,

1 . ;..-M1 - p

. ."'$:E

i .

I~.

i Jt i

|1

i,

!

.I'liolograph 3 Typical Pressure Gra<lient

5Oseg = 10 : Itc = 2.15 x 10

|,

d

__ __ __ _ _ _ _ _ _ _--.-_-.---se,-----w, w-----w.w-_,.<--w,-- - - - - , - - . - - - . - - , . - - * - - - - _ _ - - - - = _ _ - - - - - - - _ _ _ _ _ _ _ _-

!

I1

a' M

''1mrsy.'-

c.,[ 3 - t, - . ..

,1 t. ' t

. Q Q- , .

| f'/ ~.) ' kb s' '

V - yA .

/ ., , ,.

. - t,~

.-

- :. .,

.

# .

|+,-

, ,+

% i ' ''

|^ ,\ ' :

9.,. .

'.. . . . - . ,/, s ')] .

,

-

' i s" ~-

^ '*'

Q .

4_ y , .. _ - f., , ..

4. 1 y , , -~.., ~

,- .,/

m,n p.,g' . i.e'~,1%.' .-..'

f ,..N '... i .' ,.

.

_ / . :. ,. ./ ig , y 9%L' w ;Jit.'

o c e sc h.4.ty.i, c.-s.:

.- . . . . - A. N . .. - - - wr4 ''.-

.

: :-,e

c f*~f.'j' .'", ..

..b j c;fd j mf.q4. ..

,, . . . ' , '.. p gv ry,gI, - -

..

' -n erM % .

w r %$ (f at .- . . . w -@p g yf e ,41.

.g. .y.

=

,

_ , 4 ; -y.

'

.[.-

. s f ~-. ',

.&:.. .,

, eg;wt.y. ; . ..ofgs c, s. yy .;,. , . . _ -a

,

. , - .k <. y.q.. ,.'y,, .;.- :

| - .. :, .,rrf-ks a .,W;n ,r W: *p "r Q 'g v? g... . ' '.t.?.-. ,.4, .r .- ;.,,.

'?#...-

,4 ,e- W .~ A~,--

.-

.,

-~ me.: . .a::.rc . . .,.y + . > . ,~fre s .ev/ ,k y .u, $ h j. . :; .

w .

-....

. s . g.M..-w+ 1%{p f sys.1)<

:, , .., s.

_

% ,3 n s -j s } ' :, st y,; '

| - . ;- ~4. ,h ; }| u ; ' 'f ? Y'

... -t _ % m w .n ..: m .. m - . ;. .- m .

1

|

l.

Photograph 4 Vortimeter Used for Swirl Measurementst

,

i

1_ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ . __ , - . -

- - -

- - -- -. - - w _,_ _ ~ ..-.- -.- - ~ -___ . ~. . - _ _ _ . . - -- . . ., .n-.- ... .. _.n..,~. ~.. . ~ . - - -- -.

6

I

!..Ir

!

l

!

!t

!ie

I

!!

,

|.

FIGURES,

,

!. !r . p

,

- !,

I >

4-

ta

h

.

I

,

.(

' i

a

vtIt6

I*

.

a

,

L,

. W

ie

' 't* ,

|6

. l

.

I

t

5 ,.. x.--,. ,m.-..-. ,-,,,e w-=,..w,-.e.*---omw.=%..-,.--.,+4. ==%-,--. -- 1--.s...,-. -..-4.---- - - - - - . . - - . . ~ , -

- -

. . .-

. I

i

|i

l

10 HPPUMP

_

F LOWc SUMP

10" DIA.PIPE

---- 2.7 ' ----+

If: 8D : 1D - 6D- = - 16D ID +-3D : SD :; :

I|| -

" PR ESSURE TAPS|1L 'J]

, ..

lL__ #a I--~1 FLg |_

J t--- w n'---~ - &, J . . . . . . . ... 2.,.

., .

i

D + 5D-+ 1 D -* *

" VORTIMETER VORTIMETERPOSITION POSITIONUPSTR EAM DOWNSTR EAMSWIR L

GENERATORORIFICE

PITOMETER CONTROLVALVE

FLOWSTRAIGHTENER D = 10"

PLAN

FIGURE 1 EXPERIMENTAL FACILITY

M

-- -- _ _ -- . __

L' / / / / / / / / / / / / / / / ////////E

\i .

-

2 - !

l 1=--

- - 4 - -

b -

6 -

7 -

(|<

,-

0 5

u (FT/SEC)

a) VERTICAL TRAVERSE

| '/////////////////////A

1 - <

2 - >

3 - '

- - 4 - L -

f -5 -

6 -

# -

| \tE///////////////////////\

0 5

u (FT/SEC)

b) HORIZONTAL TRAVERSE

1 ,

1j._

2--

3--j. '

-=f FLOW = ---j -- / ! !4' ! !- -

fT T 1 2 3,_56 7

6--7- -

1 SECTION 1-1

i

;

FIGURE 2 PITOMETER TRAVERSING POINTS AND VELOCITY PROi lLES

M

_ . _ _ _- --

_ _ _ _ _ __-_.._--_-____- ___ _ _ __ _._. . . _ _ _ _ _ .-_ _ . . . .

- .

l

!

|

|

|

- SWIR LGENERATORVANE ANGLE (SET)

*

g g g O - 4 04** *( i ., _ g

u C & 4 & 25*

1.03 -

259 1.02 - O O O 20*o sIL *

O

g 1.01 -

us

a C O O 10'1.00 -

A E OE A A A 0*g - - - -

OO NON-SWIR LING

o 0.90 - FLOWOF PIPE DIAMETER = 10"I 0.80 - ORii:!CE METER $ RATIO = 0.6

I ' '0.701.00 1.50 2.00

REYNOLDS NUMBER, Re x 10-5

FIGURE 3 ORIFICE METER CALIBRATION WITH SWIRLING FLOW,

_. _ -. _

1

l

.

~ O <' e

OO

8^40 - Q

,

3 @ A <

$k b EQUAL INDICATED

J AND SET SWIRL ANGLESE 30 -

5 00 @5 REYNOLDS NUMBERSE .

53 8 1.00 x 1020 -

5TJ d O 1.25 x 10

5O 1.75 x 105I o 2.00 x 10

O o 2.25 x i0s

10 - - g

O

h I i i i2

'0" 10 20 30 40 50

SWIRL GENERATOR VANE ANGLE, OSET

,

FIGURE 4 VORTIMETER COMPARISON OF INDICATED SWlRL ANGLEUPSTREAM LOCATION VERSL*S SWIRL GENERATOR VANE ANGLE

e

i

-- ,

|

0.10 -

|'

_\ -

$ 0.08 -

t&

5P9m 0.06 -

5

$ 0 e=

T \s 0.04 -

.o O g g 40

e

s A,_ e$p A 303

_ p- 20*$ 0.02 -

1 10* aaw = = =

* '

O

SW1RL GENERATORVANE ANGLES (SET)

| | | | I I I

O 0.5 1.0 1.5 2.0 2.5 3.0

REYNOLDS NUMBER Re x 10-8

FIGURE 5 FRICTION FACTOR VERSUS REYNOLDS NUMBERFOR VARIOUS SWIRL ANGLES

M

- - - - - - - - - _ _ - . - - - - - - ----- -----

160 -

I

" -

439u.

$ 120 -

=5$RwS 100 -

0>

$S'

do 80 -

P9E

Ew

h 60 -

c o><

EOw

E; a

w 40 -

O REYNOLDS NUMBERSE O 2 x 10 5

5O 2.25 x 10

20O

i i i i it0 10 20 30 40 50

O \AVG

u

FIGURE 6 PERCENT INCREASE IN THE FRICTIONAL LOSSESFOR AVERAGE INDICATED SWlRL ANGLES Q

,

. _ _ _ - - - _ _ _ _ _ - _ _ _ _ _ _ - _ - _ _ _ _ _ _ _ _ _ _

)

SWIRL GENERATORVANE ANGLE4

1001

8-10x

s 75 -

o0 g 153t %

$20* jo

||

50 -

3

u -

8

$ 0m -

p 25 - - 25

6 , " O A A ^ '* 30o - ^A. ____ 35*$ "- ,40* w w

n.

O I I I !-

0 0.5 1.0 1.5 2.0 2.5

0REYNOLDS NUMBER, R, x 10

I! 1

'

r

I

t

I

| FIGURE 7 SWIRL DECAY OVER 17 PIPE DIAMETERS VERSUS REYNOLDS NUMBERl

- _ - _ _ _ _ _ --- _ _ _ ----- _ -

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

SWIR LK

GENERATOR o

VANE ANGLE RANGE SYMBOL

0.14 - 10 0.051-0.062 8 g15 0.073-0.101 520 0.101-0.153 @25 0.099-0.211 $30 0.203-0.262 b

,

0.12 -35 0.206-0.317 @40 0.256-0.370 @

<

0.10 -

.

ndY .

$ 0.08 -

12 -

.

>-

.

8 -

J 0.06 -*

*5g ..

,

0.04 -.

.

0.02 -

,,

! I I I0.00.5 1.0 1.5 2.0 2.5

REYNOLDS NUMBER, Re x 10-5

FIGURE 8 # VERSUS REYNOLDS NUMBER FOR SET SWIRL GENERATORVANE ANGLES AND I;o/Md AANGES

I

. _ . _ _ _ . _ _ . _ _ . - . - . _ . _ _ _ . _ _ _

_ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _

!

;E

E

,

- --- WORCESTER-

POLYTECHNIC- INSTITUTE

- -_ _

ALDEM RESEARCH LABORATORYHOLDEN, MASSACHUSETTS 01520

l

i_

O+

_ _ _ . _