ML20148J181

Model Study of Sequoyah RHR Sump. Concludes Initial Design of Sump Was Satisfactory for All Design Conditions,Incl worst-case 50% Screen Blockage,Excepting Air Entrapment During Initial Filling

ML20148J181
Person / Time
Site:	Sequoyah
Issue date:	10/31/1978
From:	Fain T TENNESSEE VALLEY AUTHORITY
To:
Shared Package
ML20148J126	List: ML20148J121 ML20148J175 ML20148J181 ML20148J188
References
WM28-1-45-102, NUDOCS 7811150196
Download: ML20148J181 (29)

Text

(l

. Tennessee Valley Authority Division of Water Management

'. Water Systems Development' Branch 1

1.

i I 1- l l

4 1

' MODEL' STUDY OF THE SEQUOYAH RHR' SUMP 1

. Report No. WM28-1-45-102 l

1 i

i Prepared by l Theodoric' G. Pain 1 l

- Norris, Tennessee

. October 1978 1 1

l Y

- .c.

[

l

-l

-1

4 l

1 CONTENTS Page INTRODUCTION 1 PROTOTYPE' DESCRIPTION - 1 T

MODEL DESCRIPTION 5 i MODEL SIMILITUDE 7 Model Limits 7 Kinematic and Dynamic Similarity 7 Model Screens 12 Sump Loss Coefficient 16 MODEL -TESTS 18 RESULTS- 18 Air Entrapment 18 Vortices 18

. Sump Loss Coefficient 23 CONCLUSIONS 23 REI'ERENCES 25 i

l

'l l

f l

~

i l' )

P O

LIST OF' FIGURES Figure No.-

Title. Page 1 Original Containment Sump Design ' 2 2 Model Limits (Sequoyah Unit 1) 3 3 Overall View of'Model 6 4 View of Sump with Access Door Removed 6 5 Reynolds Number Ratio vs. Scale Ratio 11 6 Reynolds Number Ratio vs. Velocity Ratio 13 7 Loss Coefficient for Screens 15 8 Illustration of Sump Loss Coefficient Calculation 17 9 Arbitrary Vortex Strength Scale 19 10 Final Design with Vortex Suppressors and Other Modifications 21

~

11 Plan View of Containment Grating 22

. 12 Sump Loss Coefficient Reynolds Number Dependency 24 4

i v, .; , ,. ' ' " '

LIST OF TABLES Table No. Title Page 1 Range of Prototype and Model Reynolds' Numbers 10 2 Comparison of Model and Prototype Screens 16 3 Schemes Tested with Final Design and Results 20 1

1 l

l l

l l

l l

l l

l l

l O

a

l I

l

. l MODEL STUDY OF SEQUOYAH RHR SUMP I

INTRODUCTION '

TVA has verified the effectiveness of the containment sump for Sequoyah . Nuclear Plant by constructing and testing an undistorted

' ' scale physical model at the Engineering Laboratory, Norris. This  !

j report describes the model and presents test results. Objectives of the sump design were (1) to ensure that no free air would be trapped in  :

the sump during initial filling, (2) to ensure that no air-drawing vortices would form in the sump or containment during withdrawal through the sump, and (3) to empirically determine the sump loss coefficient for use with net positive suction head calculations. Results of the model tests on the final design show that these objectives were l fulfilled.

PROTOTYPE DESCRIPTION The containment sump, shown as initially designed in Figure 1,

4 has twin discharge pipes which can be operated independently.

Water reaching the discharge pipes passes through a screen at the sump intake and through a second screen with -inch mesh inside the sump. The screen at the intake, originally -inch mesh, was changed to h-inch mesh during the test period because of design considerations unrelated to. the model' study. The annular shape of the containment floor area - (elevation 679.78), shown in Figure 2, allows flow to ap-proach the sump from two directions simultaneously. The numerous equipment, supports, and piping in the containment can ' alter the pat-tern of flow toward tho' sump. The water surface in the containment l

l

I l

2 1 1

1

= l l F'--- T T T--~l

_ _ _ _ _ _ _ . _ _ _ q _J_ l l

' I L._ _ _ _ _.__._ l_ _

i I , J A i lI .I A h -_j ! l l I h r -- -- -- - - - - ~

i1- i i 1 i

~ VL __ __ __ _ . ._~. .__l_~- j i L _______f L__.__Ld-_._! .

1 I

PLAN . I SoLlo COVER PLATE CONTAIN MENT FLOOP l/2" SCREEN ON SIDES (El 679.78) \ ly it

, , l , o'*, t t,. ,- ,4 ,,. 4.'. 4 '

....' s 6 ,'b.. .

b..

. s. =' .

. BliFFLE PLATE ~ 6

..., 6

,',',0 ..s o . . .

,>. l. s l <,

Pl/4,I SCREEN l

' 5 ,'.

I I';,' ,

d

El

- [N . 669.78

' FLOW -h- -

y

,. s.

. u r L El. 667.48 4..,.f,.4, . ' ,,s., , .4 ,

SECTION A-A

..8_

i .

Y-

~

Figure 'l: Original Containment Sump Design f

1 (Unhs I and 2 Identical)

N

. 3 270*

l l '. '

DRAIN HOLES % .'.

(2) X-fj -- ' . _

RECIRCULATION PUMP (4)

l l CONTAINMENT SUMP

,i .,., _

i_

c .

v_'

~

STEAM GEN (4)

. LOOP .' '

LOOP 3-

-4

-  : y f ,

.. . .' fV

/

Y' AIR CLEANUP UNITS (2)

I 8 O' , -}. - .- -

O' CRDM COOLING

.. - l / UNITS (4)

. LOOP LOOP 2 I PRESSURIZER <

".. 1 IEF TANK ( '

.. DI MODEL LIMITS W

~ .3 ,

n r. ACTOR , . , . ..t SHIELD WALL CRANE WALL 90 N ~

2!

Figure 2: Model Limits.(Sequoyah Unit I) i o

4 area bounded by the crane wall and reactor shield wall is at elevation 693.0 during withdrawal through the sump. The pump runout flow rate through the sump is 9875 gpm per pipe. The maximum water tempera-ture during withdrawal is 160 P.

e e

a 9

e

5 MODEL DESCRIPTION The' model had the orientation of Unit -1 (Units 1 and 2 are opposite hand but"otherwise identical) and included the containment sump,._ the sump ' intake structure, the two discharge pipes leading from the sump, and a portion of the containment area between the crane wall and the reactor shield ' wall. The portion -of containment floor included (elevation 679.78) extended from the centerline of the refueling canal (azimuth- 270.0 ) clockwise to recirculating pump No.1 (aziumth 52.36),

as shown in Figure 2. .Within this space bounded by the floor, the crane wall, and reactor shield wall, all structures significantly affecting the flow up to elevation 693.0 were modeled. Figure 3 shows an overall view of the model. Figure 4 shows the sump and a portion of each discharge pipe. The crane wall, sump walls, and discharge pipes were

~

made of clear acrylic plastic to permit visual observation of the flow patterns.

Water .was supplied to the model through perforated plates at both ends of the model (azimuths 52.3 and 270.0 ), through the two drain holes located in the floor of the refueling canal, and through a movable pipe located in the containment near the sump. Flow rates for each of these sources were independently controlled by valves and measured with ' calibrated orifice meters. All water supplied to the mod el was withdrawn ' through the sump by a pump and recirculated '

'directly back to the model. The water level in the containment was

. controlled by regulating ~ the volume of water in the model and -piping system.

For testing-'at water temperatures higher than ambient, five

- 2.5.' kW resistance heating elements were installed behind each of the perforate'd plates at. the ends of the model. The maximum w'ater t iemperature possible was_ approximately 130 P.

c - -__

. 6

. ~.*

l M{f ', %- c ' n ,, ,Q[

,33-  ;- - a en ~ ~ m_- .~

?j.3 ,.a m

~'1

~~s_^

^

-_.. - _. L: m r : . . - - -

fl Q

~ z ~

4.ce .

); ,

ff , , N., . . ;

^

FIGURE 3: OVERALL VIEW OF MODEL

]

1. '$

3 i'$ . '

"? ,

peb W'1 I

,i [ f g

. , ;; ;f I~

/;Y[ jj  : ifIaj. ll

. M

}-

4

, gi "

i ;, l}-j:f;M I, }

4 ;.

.tJ, ii 4-

-h.

~

,p [!

gJ-l Sm

. FIGl1RE 14 : VIEW OF SUMP WITH ACCESS DOOR REMOVED ,

l

7 MODEL SIMILITUDE

- Model Limits

- Preliminary tests were performed to determine if a sufficient length of the annular-shaped containment area had been included in the model . This length would be sufficient if equipment in the flow outside the model limits would have no effect on flow patterns at the sump. To show this, water was supplied alternately to one end of the model only, to the . opposite end, and to both ends simultaneously. In each case, the flow pattern in the vicinity of the sump was observed with the aid of injected dye. No noticeable change in the flow pattern near the sump occurred. IIence, it was concluded that flow patterns at the sump being controlled by conditions relatively near the sump, and that the model limits were conservative.

Kinematic and Dynamic Similarity Because of the predominance of gravitational and inertial l forces in the flow processes involved, kinematic and dynamic similitude were achieved primarily by equating the Froude numbers1 of the model and prototype. The Proude number, representing the ratio of gravita- J tional to inertial forces, was defined as IF = (1) lii 1

--where V =-discharge pipe velocity

' g =.' gravitational acceleration

- s = submergence of the discharge pipes below the free surface

. and was made equal.in the model and the prototype:

m Y Y m. -?C-' +'71 m - -

___<m__ --___

9 1

8.

^

1

- F{=[F= 1- P (2) ,

where subscripts m, p, and 'r represent model, prototype, 'and ratio i

between model and prototype, respectively. Velocity, flow rate, and '

time, V, > Q, and t, respectively, were expressed in terms of the chosen

, geometric scale:

L m 1 Lr"L P *T e

t where L refers to length. By use of Equations 1 and 2 with g', = 1, m- = Lr,} ..PV = 0. 5p V V

4 Qn* =L b Qp-= 0.03125 Q r p t

m =L tp = 0.S t i

r p n

The flow field depended to a lesser extent on viscous and a

possibly. surface tension effects. The~ relative magnitudes of these forces to fluid inertia were reflected in the Reynolds and Weber num- J bers, defined respectively as R = Vd -i and w = PV*S

, where

, V = discharge pipe velocity d = discharge pipe diameter 1

v = kinematic viscosity 6

d

~

, , , , , , -.--- <, , m..

-s .

9 p = density o = surface tension s = submergence of discharge pipes below the free surface.

Because the gravitational force was the single largest force governing flow through the sump, the model flow rate had to be determined on the basis of equal model and prototype Froude numbers. Consequently, the Reynolds and Weber numbers could not have the same values they would have in the prototype. Any deviation in similitude of the flows attribut-able to viscous and surface tension forces was called scale effect. l Surface tension effects were insignificant because strong vortices were i not present in the model and the free surface was essentially flat2, ,

Vortex formation was therefore predominantly a function of Froude number, with possibly a minor scale effect because of the reduced model Reynolds number. The model Reynolds numbers were high enough that 1

l the flow in the model was fully turbulent, i.e., in the same regime as that in the prototype. Therefore, scale effects due to viscous forces were negligible. This condition was assured by employing the fobowing modeling techniques:

(1) The large model Reynolds numbers were achieved by choosing a large geometric scale ratio for the model (1:4).

(2) The model was operated at pipe velocities higher than Froude scale to increase the model Reynolds number, but without exces-sively violating Proude scaling criterion (07 maximum = 2.6).

(3) The model was operated at water temperatures higher than ambient to further increase the model Reynolds number (9 max = 130 F).

(4) The model was operated at water levels lower than the design 6

s

.. - - . - - - _ - - - = . . ._.. - - .. ._ __

. 10- .

minimum' to exaggerate the propensity for vortex formation.

The Reynolds number ranges of the model and prototype are given 11.

. Table' 1.

TABLE 1 RANGE OF PROTOTYPE AND MODEL 1R (Qp= 9875 gpm per pipe)

~

Prototype R Model R F r=1 (Ey2~ Fr=2.6 T=60 1.59x10 6 '1.99x10 5

3.97x10 5

5.17x10 5 T=130 3.47x10 6 4.33x10 5

8.66x10 5 1.12x10 6 T=160 4.38x106- __ __ __

Figure 5 shows graphically the effect of scale ratio on I

. Reynolds number ratio for -a range of model temperatures, with equal  !

Froude numbers in the model and prototype. The resultant curves can be used for comparing the Reynolds number reduction resulting from the chosen model scale (1:4) to' the reduction associated with other possible scale choices. For example, a 1:1 scale with ambient . water temperature (about 60 F) would have a Reynolds number about 0.3 l times the prototype value. ECCS sumps at other nuclear plants3 ,4 have.

been studied with 1:3 scale models at near prototype temperatures.

Figure 5 ~shows that increasing the scale of the Sequoyah model from

~

1:4 to 1:3 at a model temperature of 100 P would only increase the 1' Reynolds number ratio from about 0.08 to about 0-12, which would have insignificant effect on the already turbulent flow field. The model cost

.. would be increased considerably. Also, the literature contains examples

. _ - . ~ . - - _ , . _ .

. 11 1:1 l l l I l MODEL TEMPERATURES, 6m:16 0 0.9 - - - - - - - - - - -- -- - -- - -

0.8 b=1 IF p k 6p = l60* F E 0.7 . / -

13 0 *

.g -(

o' F 0.6 y SCALE RATIO l .*2 m

- 10 0*

y 0.5 ,

. E

/,/ ,

\

/ /

_ 70 o,3 33

• I 'O 1:4 /

0.2 /

'7// / A*

0.1 /g y/ /

40

?

O O 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 i SCALE ratio, Lm/Lp

.8 .

7 ty Figure 5: Reynolds Number Ratio vs Scale Ratio '

lT

'I

o

. 12

. of vortices prc'dicted in intake models as small as -1:60 which were later

- verlfled in the respective prototypes5 ,6 ,

7 4 In ~accordance with common practice , the model was operated with measured velocitics in the discharge pipes higher than Froude scale and observed for vortexing tendencies. The highest velocity.

obtained in the model was 17.4 ft/sec; the velocity in the prototype at maximum pipe discharge will be about -13.4 f t/sec. Under these condi-tions, V !V w s. approximately 1.3. Also, the water temperature in <

m p the model, usually in the range 40 to 90 F, was raised to about 130 F Tor some tests. These two steps increased the Reynolds number ratio .

to about 0.26, as shown in Figurc 6.

It .was necessary to create vortices in the model in order to verify the effectiveness' of the vortex suppressors.

Therefore, tests were performed with the water level in the containment lowered 5 feet below the 13 foot design level where, with the maximum possible flow 2 rate in the' model' and certain combinations of screen blockage, it was possible to generate relatively strong ~ vortices at the free surface, although of short and random duration. Inflow through the movable pipe was directed toward ' the sump from 'various locations to help

. determine worst-case conditions. Tests .in this series were run with and without vortex suppressors in place.

Model Screens

Screen material for the model was selected which had the same pressure loss ' coefficient as the prototype screen, with the wire diam-eter and opening. width as near 1:4. scale as possible. Thus the effects of 'the two model screens on pressure loss and on velocity profile modifi-

. - ~ . - + r - . ,- -- ,,

. 13 6p = 160*F MODEL SCALE = I:4 l l l MODEL TEMPERATURES,6m =l60 IFm /lF p= 3 4

a 0.4 MODEL LIMIT (Vm/Vp = l.3( -

- 13 0 g l 6 IFm /lFp = 24 F- 0.3 I 10 0 IFm /lFp = la s ,/ i 5 0 .2 -

. 70' l

O. I -

f A

[. C 40a o

0 0.5 1.0 1.5 2.0 VELOCITY RATIO, Vm/Vp E'

j' .

T .- Figure 6: Reynolds Number Ratio vs Velocity Ratio lT.

4 a

a- 14 cation were properly simulated.

~

The pressure loss coefficient for. the~

screens, defined as.

K=#

.v2/2g_

was computed from the expression8,9 l

r K= 1 103.4 (d a)2 .

6.24 A R A 3c2 e I

where AH = pressure drop through screen 1

j. v3= unobstructed upstream velocity l

d ' = wire diameter A =' width of opening in screen

, a = ratio of screen surface to screen volume Rg =- Reynolds number based cn the wire diameter and the I t.nobstructed upstream velocity c = volume ' void fraction (ratio' of empty volume to total i screen volume).

. Pressure loss coefficien+3. for both prototype and model screens are shown as functions of Reynolds number in l'igare 7. Table 2 compares the characteristics of prototype and model screens. Later in the model .

study, it was noted that tM sum of the computed head losses through the . two screens represented less than five percent of the total sump l head loss. Thus 'the differences between prototype and model screen pressure loss coefficients were insignifi: ant. l i

-__ _ _ _ _ _ _ - - _ _ - _ - _ _ __ = - _ _ _

wu 2 8 - -es - 102.7 -

= s

• = . .

I l Op =l60 F vd s

9 -

Om: 40 -130*F Re= y Op = 19750 gpin I-z .8 '

K=

~W vs2/2g o

.7 -( '

{ I W

O PROTOTYPE O .6 ( 3 SCREENS

$ \ l S

3 w 7 -

a-cn 4 -

[ j/ k- PROTOTYPE y [ ,

MO%L SCREENS INTAKE SCREEN "

a: PROTOTYPE

' .3 ,

/ clNTERNAL z SCREEN w

w '

.2 ,

• A-BOTH MODEL SCREENS

• i RAr*SE OF MODEL

/ OPERATION FROM TEMPERATURE VARIATION

.O U 200 400 600 800 1000 1200 1400 SCREEN REYNOLDS NUMBER, R e (BASED ON WlRE DIAMETER)

G Figure T : Loss Coefficient fOr Screens

, 16

![.

TABLE 2 l COMPARISON OF PROTOTYPE AND MODEL SCREENS l l

I

% Open d E (in) (in)

Prototy; e 70.9 .047 .25' Model 74.6 .017 .108 l

Sump Loss Coefficient, The sump loss coefficient was determined 10 by extrapolating the measured discharge pipe pressure gradient to the pipe inlet and computing the head loss, hg , as

' 2 hg = Ah - V /2g where ah is the static head change between the free water surfece and the pipe inlet and V 2/2g is the velocity head in the pipe. The sump loss coefficient was defined as h

I-C g = V2/2 9 l l'igure A shows a typical evaluation of Cgfrom pressure gradient data.

l .

6;

17

~~

" o l TEST #I4 g' WATER SURFACE Ah = 32.669 cm he IN MODEL y 2/29 = 21.031 cm ht : Ah -V2/29 :II.638 50 -

CL= = 0.553 e V2/2g N ,

N t '

g0 ' TOTAL ENERGY '%

N l

l

- ah GRADE LINE IN PIPE ]

l ti o l V2/2g h "

. E e

~

20 EXTRAPOLATED g r PRESSURE GRADIENT fM E ASURED %

PIEZ0 METRIC HEAD IN DISCHARGE PIPE 10 0

0 50 10 0 150 200 250 DIST/,W E FROM PIPE INLET, cm f

g. .

T'

• o- .

J- Figure 3: Illustration of Sump Loss Coefficient Calculation T

3 ,

18 MODEL TESTS Systematic tests were performed under a variety of depth, flow, temperature, and geometric ce ditions, with and without vortex suppressors. For each test the flow patterns were visualized with dye and surface floats. Vortices, if observed, were rated according to the arbitrary scale shown in Figure 9. Table 3 shows a summary of the ,

various test conditio: using the final dnsign.

RESULTS

. Air Entrapment

)

i The final design, recommended to ensure escape of free air I from the sump during initial filling and to prevent the formation of vortices, is shown in Figures 10 and 11. Inside the sump, the plate over the intake pipes was vented near the sump wall to allow air to l l

escape. The rear wall opposite the discharge pipes was sloped for- '

ward, beginning at a point 5 feet above the sump floor, to eliminate the overhead horizontal surface which could trap air. The solid cover on the intake structure was sloped and vented to prevent air entrapment. l Vortices When low water level and/or screen blockage conditions were established to intentionally create vortices in the model, they tended to form. at the free surfaces and at the solid boundaries, in agreement with Chang2 . These tendencies were eliminated by placing grating (2 "

x 3/16" bars on 1-3,'16" centers) under the free surface and on the walls inside the sump, as shown in Figures 10 and 11. Although no surface rotational tendencies were observed with the water surface elevation at the design elevation (elevation 693.0), the grating shown at elevation 688.5 was recommended for conservatism.

l 1

NUMBER DESCRIPTION ' I9 l

1 i

? -

0 NO ACTIVITY ,

l l SURFACE SWIRL I

[^ l l

l v 2 SURFACE DIMPLE l

y. 3 DYE CORE l

4 VORTEX PULLING TRASH 9UT NOT AIR j

,' TRASH 5 VORTEX PULLING AIR BUBBLES I

\.

• .,AIR BUBBLES 6 VORTEX PULLING AIR CONTINUOUSLY

!~ .

f* ' Figure 9 : Arbitrary Vorjex Strength Scale a

W-

20 TABLE 3 SCIIEMES TESTED WITil I'INAL DESIGN, AND RESULTS Observed Vortex Strength

• l' low Direction **

• Left Side Only 0 Right Side Only 0 Both Sides 0 Plow Rate Maximtun Possfole Model Flow 0 Scleen Blockage Worst Case, 50 percent 0 Water Depth Design Depth (13 feet) 0 Reduced Depth (8 feet) 1

~ 1 Single Pipe Discharge Left Pipe Only 0  ;

l l

l l

I 1

• See figure 9 for definition l

• • Looking Downstream l

i Note: Design water depth and pump runout flow rate were used, except where indicated. l l

l .

f 1

l

1, 21 EL 693.0 -

g V

EL 688.5 r GRATING p

iiiiiiiiisiit i11 iil l l 1 il liiLU j CRANE-WALL

's

, GRATING UNDER COVER PLATE 9

AIR ESCAPE AT SLOPI,NG COVER END OF' GRATING 8

,,, PL ATE- >

.y 7 INLET tiTRUCTURE WITH

.4 CONTAINMENT i r;Pf p /41 MFSH SCREEN FLOOR , :S EL 679.7'8s / l- .I*

,.~ ~.:...'

,1 C'....

.9 -

. , y '. .y'1 y*

BA Fl.E PLATE b -

, WITH VEN" HOLES, - ) ,

AT TOP - -

BAFFLE PLATE 1/4" MESH , ,' .

I

,h-SCREEN- , ' ,,

, s [ ,, u i  : .

,y' .

... 'D -

l

. 5,

_ FLOW Y' l 1_:

~

.1 .. .

'9 . ,

,. ,: I v '

, i r - . u

'.i:. ' ; ', : . i : . 'y ' . . ' 'c GRATING ON SIDES AND REAR WALL OF SUMP (NOT TO SCALE) e A . o.

g J,

Figure 10 : Final Design With Vortex Suppressors -

Y and Other Modificotl0ns 7 ! .-

m' n

v , - -

'o' 22 CRANE WALL i 6

1 *

• l \

o, .

.'  % ._. I

,4' REACTOR SHIELD WALL

/

.. .' \,

, ,' STEAM GENERATOR"4 ,,

~' ' ,,

/ PPROXIM ATE s ! /. -.b '#~ *

~T '

LIMITS OF GRATING ,1,

/ / /

a <

{ '. , ,> l

. ', A i

% /

{L /

s

. 'f . $ / 4 C ' s, ~

r,,, 1 I .? 's 1

. .. / ss\s\) ss / ,

l

:!g: N ./ /%

lc s.3,' .

\

, t . . .:,'. -

i

..v - -

. \

'y,' ,

8..  ::: . ,4 ,,

\,

')', ,i r$$ $ $N[Nhb:~': '

CRDM COOLER }..L)

• k ", *

• pCROSSOVER ,

c PlPE ' '9 '. s a

P, ,,*o

.A'_ .' .

e

, . =,,l 4

, . N.

,.T

. '. s. 4 . .. - . .

,. , , ; ,A e ,' 4 -

, s .. .

.. . 4

. e. s'.

a i .~

/J \

4 f:) l

<^ s )i T.

rr

.., l'

\ f ' N.

0

., l -

) l , ~a '.

• u \

N/

.n

,s ,

. s ,,- ,

RC PUMP # 4 , ,. .. ' ' ^ , , .

... ,s .

,a, ,

'- u.

.. ..u, ,' .

/...Id'..

u ,'.. . .,

. o.

t, . ,. ' Figure 11 : Plan View Of Containment Grating l.

~O tu Id 0

. 23

. The grating under the cover of the sump intake structure prevented vortices from forming en the underside of the cover and extending down into the sump. The grating on the vertical walls inside the sump prevented vortices from forming at the walls and extending into the discharge pipes.

Sump Loss Coefficient The sump loss coefficient, C,g is shown as a function of discharge pipe Reynolds number, in Figure 12. The value of Ct in the prototype at pump runout flow rate (19,750 gpm) and design water  ;

l temperature (160 F) will be about 0.45.  ;

CONCLUSIONS It was determined that the initial design of the sump was l satisfactory for all design conditions, including worst-case 50 percent screen blockage, with the exception of air entrapment during initial  ;

filling . The modified design incorporating sloping of previously hori-l zontal surfaces and venting of crevices was developed. Vortex suppres- '

sion devices were added for conservatism. This modified design gave satisfactory hydraulic performance under all possible conditions.

I O

.w M 28 4 5 - 102.12 ,

a- . . ,

FINAL DESIGN WITH UNBLOCKED SCREENS, EXCEPT WHERE NOTED.

.7 .

.6 j

(WORST CASE)

H (50% SCREEN BLOCKAGE b ~5 c

O h o o o ve w 7 q

{ .4 w

7 t/ O l o

u ,3 m

8 c'. 2 2

o w .I oI I05 106 REYNOLDS NUMBER, R St Figure 12 : Sump Loss Coefficient Reynolds Number Dependency

25 O

REFERENCES

1. Pao, Richard H. F., Fluid Mechanics, John Wiley & Sons, 1961.

2. Chang, E., " Review of Literature on Drain Vortices in Cylindrical Tanka," the British Hydromechanical Research Association, Cran-field Bedford, England, TN1342, March 1976.

3. Durgin, W. W. , et. al. , " Hydrodynamics of Vortex Suppression in the Reactor Building Sump Decay Heat Removal System - Three Mile Island Nuclear Station , Unit 2," Alden Research Laboratories, llolden, Massachusetts, February 1977.

4. M. Padmanabhan, M., " Hydraulic Model Studies of the Reactor Building Sump, North Anna Nuclear Power Station - Unit 1," Alden Research Laboratories, Holden, Massachusetts, July 1977.

5. Andre Gagnon, et. al., "OUTARDES 4, Etude sur modele des prises d' cav," Lasalle Hydraulic Laboratory, Ltd., Quebec, LHL :-412, April 1966.

G. Quach, T. T. , " Intakes of Manicouagan 5 and Outardes 4 Develop-ments, Inspection Report," Hydro-Quebec, March 1973.

7. Chang, E., " Review of Literature of the Formation and Modeling of

, Vortices in Rectangular Pump Sumps," the British Hydromechanical Research Association, Cranfield Bedford, England, TN1414, June 1977.

8. J. C. Armour, et. al. , " Fluid Flow Through Woven Screens," J.

AIChE, Vol.14, No. 3, pp. 415-420, May 1968.

9 Tennessee Valley Authority, " Flow Through Screens," Report No. i 87-8, Norris, Tennessee, May 1976. I

10. Daily, J. W., and D.R.P. Harleman, Fluid Dynamics, Addison-Wesley , 1965.

l 1

l l

1 O

e D