ML20091J695

Analysis of Clogging of ECCS Sump Trash Racks by Debris & Paint Peels Following Accident

ML20091J695
Person / Time
Site:	Comanche Peak
Issue date:	03/31/1984
From:	EBASCO SERVICES, INC.
To:
Shared Package
ML20091J682	List: ML20091J695 ML20091J703 TXX-4189, Forwards Amend 50 to FSAR & Supporting Studies Which Substantiate That Containment Coatings Not Significant to Plant Safety
References
NUDOCS 8406060104
Download: ML20091J695 (41)

Text

_

[

TEXAS UTILITIES GENERATING CO.

COMANCHE PEAK S.E.S.

ANALYSIS OFTHE CLOGGING OF ECCS SUMP TRASH RACKS BY DEBRIS AND PAINTPEELS FOLLOWING AN ACCIDENT FOR TEXAS UTILITIES GENERATING CO.

BY EB ASCO SERVICES INC.

2 WORLD TRADE CCNTER NEW YORK, N.Y.10048 f$0A$0ck4S$88?j3 c

PDH

ERRATA FOR EDASCO REPORTS June 1, 1984

p, = desalty of the peel A = area of the,paal thicknees c't the peal t =

( = dras coefficient for paint Peel motien in vertical direction

,C dras coefficient for paint peal motion in horizontal

=

g direction

(" = coefficient to account for the lift forces on paint peal motion l

p,= density of water v,, pool water velocity paint peal vertical velocity u =

paint peel horizontal velocity y =

These equations (1 ) and(2-) are numerically solved for paint peal horizontal and vertical terminal velocities, The resulta have been plotted on Figure 7.

An angle of aero would result in the longest vertical t. ravel time. The furtbust horizontal travel occurs at the angle for which the value of the multiplication of the horizontal velocity and the vertical travel time is the largest. As seen from Figwe 7, the longest travel time occurs at a small angle. Bowever, the equations used to model the paint peel fall are inappropriate at angles near zero cnd ninety degrees since boundary layer offects on drag have been ignored in the equatione. Consequently, results at angles greater then 85* and less than 5' are not considered to be representative of actual conditions.' Despite these cenissption problems, the average horizontal travel distance for say paint peel can be calculated to f all between 22 and 24 feet using the information presented in Figure 7.

17

3/84

\\.

e

(

)

e 1

I l

l

/

I I

l 0

v s

e c

U t

ag i k I

i l

I Figure 6 Paint Peel Force Model i

1 l

=46-5/84 I

h

,.m_.._

l he teaction at the metal surface equals 1031.3x0.3, or 315.4 lbf.

l Assuming a coefficient of friction of 0.1, the frictional force equals 2

31.54 lbf. The peal of area 1 ft and thickness 0.03" weighs 0.2 lbf.

l Bance, the peal would indeed cling to the surface of the actuen verti-i r

cally.

l Doing Fig. 4, Consideration will be given to a vertical peel striking the, water surface at a horizontal distance of x fe; from the trash rack. It will be conservatively assumed that the incident velocity equals zero.

Since the drag on a vertically descending paint peal is =mm11 and negligible, the time t taken by the peal to fall through a distence of 16.875' is given 2

by is st = 16.875 and is found to be 1.02 secs. While the peel descends ver-ticalf, it would be transported to the screen by the velocity of tha water.

The water velocity is 0.161 ft/sec at the screen and it reduces with distance.

It is conservatively assumed that the water velocity is a constant equel to O.161 f t/sec and that the peel is transported without slip. The distance z l

1s then equal to 0.161 t which is about 2" with e equal to 1.02 secs. Thus, the paint peels incident within 2 inches of the trash rack could be expected j

to clog the acreen vertically.

Estimation of thc Averame_ Horizontal distance traveled by the ' Paine p 1= inci.

l dent on the surface of the unter at different anales:

Figura 6 illustrates a paint paal at an angle 8 to the horizontal moving i

under water. The weight og acts vertically downwards. The peel would fall f

down under the influence of ag opposed by drag and aerofoil type of lift forces. The ' equations of motion in the vertical and horizoncal directions respectively are:

d" 2

a

• as -

0,Acos0 U (1) d p,ASin9(v-v,) h-vj (2)

A Cos0 U m

=

Where, m = mass of the psal = Atp, and sfa4 b

l:

l TEXAS UTILITIES GENERATING CO.

COMANCHE PEAR S.E.S 1

ANALYSIS OF THE CLOGGING OF ECCS SUMP TRASH RACKS BY DEBRIS AND l

PAINT PEELS FOLL0k'ING AN ACCIDENT

-l

''' ^ ""' '"'-

("~

2 WORLD TRADE CENTER NEW YORK, N.Y.

10048 l

0

+

[-

[

[

March, 1984

[.j:J

{.

E_

Executive Summary Clogging of the trash racks covering the ECCS sumps, by paint peels and debris following an accident has been investigated, in order to assess the resultant effect on the performance of the ECCS pumps.

Debris other than paint was assumed to fill half the depth around the trash racks as per the requirements in Regulatory Guide 1.82,

Revision 0, even though application of R.C. 1.82 Rev. 1 shows that the plant possesses little or no possibility for debris transport.

s Factors like clogging of trash racks by paint peels sticking verti-cally to it, transport of paint peels.to the vicinity of trash racks and the packing ratio of paint peels were investigated under accident

[

conditions.

L The ECCS pump performance was adjudged not affected, if the pressure l

L drop across the trash racks were much less than the suction head of the pump.

It is concluded that the performance of the ECCS pumps would not

[

be adversely affected, if the trash racks were blocked 83% or less.

Such a blocking limit could be reached if (1) debris exceeding about I' ~

4000 ft fill the space around the trash racks, in addition to the total L

quantity of paint filling the same space with a packing ratio of 0.75 or, (2) if the entire quantity of paint fills the space around one trash rack with some of the paint peels clogging the screen by vertically stick-ing to it in addition to debris covering half its depth. Since the accum-L ulation of all the paint inside the containment over and above the debris blocking half the depth of the trash rack is insufficient to exceed the h

83% blockage, ECCS performance is adjudged not capable of being affected s

by trash rack clogging.

r W

r u

I L,f r

TABLE OF CONTENTS Page 1.

Introduction 1

2.

ECCS Suction System Description and Inputs 1

3.1 Initial Blocking Analysis 4

3.2 Supplementary Analyses 11

-4.

Conclusion 19 5.

References 20 Partial Report of Sump Model Tests Appendix A Paint Packing Ratic Estimation Appendix B

.d 4

[

E

[

g,

(

L...

1.

Introduction The main purpose of the ECCS pumps is to supply the containment sprays and reactor cooling system with water from the sunps located in the containment following a LOCA.

These containment sutas are covered by trash racks 6'3" in height located at an elevation of 808.0 (Figure 1).

Following a LOCA, debris and paint peels could accumulate around the trash racks causing an impediment tc the flow of water to the ECCS pumps. This condition has been analyzed in order to determine the amount of blockage of the trash rack by paint peels and the effect on the performance of the pumps.

f 2.

ECCS Suction System Description and Inputs There are two trash racks, one covering each sump at 808.0' ele-vation (Ref. 1,2).

Each trash rack is in the form of a partial sector in plan, subtending an angle of about 30 at the center of the containment, 8' in width and 6'3" in height. The perimeter of each trash rack is found to be about 80' and the area in plan is 2

about 235.0 ft.

The trash racks possess a free flow area of 70%

f (Ref. 3 and Appendix A).

The nondimensional irreversible pressure loss coefficient for the trash rack calculated from experimental maximum pressure drop data with 95% confidence (Ref. 3 ana Appendix A) was found equal to 28.0 with reference to the flow through a half

{

clogged trash rack with 12500 gpm. As indicated in Table 1 (Ref.4),

the volume of paint on concrete equals 831.25 ft (285000 ft with an average thickness of 0.035").

The volume of paint on steel equals

{

204.14 ft with an average thickness of 0.01'.

Further, there are about 17000 ft of unqualified paint (Ref. 5) with an assumed thick-ness of 0.01"; the volume of this paint equals 14.2 f t The total volume of all paints equale 1049.6 ft The following information

{

were not available from TUCC and were assumed: (1) The dry density of paint equals 80 lbm/ft. ;

this is the lower limit paint density from EBASCO constructed plants (2). The height of flood water above

(-

808.0' equals 20.0'.

[ r

'; ^ '-

y -;.-~ q j

p

[glk!fl)fff N

l l'.. i,. l l

$[

. '.'i k

• I

'.4

. li

,a f

ijd, f

i 1

M i I, q.

. n~.,..

4

[lpi.!ii!!!igi,b. W!! Hilj!!'jll1 Ft k, l' "t'.' :-

4 a d.i. 3,. ul2 7 O

ili

!i !

11; r

i l

l' !!

s 1: 1. I j

+ T, ~

h!!d k;'r,M f

l ia

. v; > {::.. _l,y+ :.

,,VMfg

1

"_t 3.j! k,.'t p-W

,M.m h- - t

_3 -

i I-J.'

g g.

i b [.l 'h.h.,.c*n".~*~U[Mkn r.-erm.

.s.*. w.~g,i J b. ufl f

fi t w. -

4-i c.

i wr. <

m l

,4,..

j,a lyl,. [ m fbl,; f.U[:f.' '

[

b I.

i'

. N, 21.II'.T'i * - [ y;Mr.ij l

j

,i

.f.dmM; spi %,Jt!wu,"i5t$$.D y !th

+e.Ii Ib O!

(,l i

k r..-

g

. bp

.e q,rtip!p%

iJ e 9

+ > ' - d.

pimw vy.),y:n a

j 9,

x Ar%

w s

3:,rg[u...wd4 mwar al.w.uc~. m

, - 4;b -

g p, e.; y I

N$,$hWhk.'b !!

whyPh,, e.y,b by

(

' ]d Jb i

W pp u.#

mq A=N. ]

F k k f hg3 h

f.

S

%'N hb p y n ) k h,r.,

$e ], k,h _I k $a$

e a h

m l

f.

k.h j Jp c

.gg. j g p g' f*i 1

u

[

. # ff

• --Il

• g W @ 1

.:

• MWOM"li me;if 'E if f

C

-e

~

, i=.

-A g:

4 J.

u__

IUlidp$g$-$ 1-

-i m-i) t

-[c.

• r J

jg(

i i

i gi[i!kt

},

l k

[

el f< s Q'p J. a e.

a ;4 s-w e

t 1 p_1 ir i

a c

q h

'.e hi.

.d*hmU_ -

d A-

[

(]

p /

h fl j.

• 1 3.

M:

g 4g.

g

.e 3

$.e 1;

%.J p

^.

r 3%

4l

-=

m. a 5

-[

t r-

.s g

. ~.

[

[

2 1

f

TABLE I PAINT DETAILS DESCRIPTION Painte,d Area Average Paint Volume, ft' Thickness, ft3 u

inches

1. Concrete 285,000 0.035 831.25

2. Steel Liner 145000 0.01 120.83 Pipe Supports 4520 x 11 = 49720 0.01 41.43 Cable Tray 755 X 11 = 8305 0.01 6.92 Supports Conduit 4812 X 8 = 38496 0.01 32.08 Supports Miscellaneous 2500 0.01 2.0S Miscellaneous 87 X 11 = 957 0.01 0.80

3. Unqualified paint 17000 0.01 14.20 Total 1049.59 s

(

(.

[

-F

3.1 Initial Blocking Analysis As shown on Figure 2, and based on geometry considerations it was esti-mated (Ref. (4)) that an area of about 500 ft was available around each trash rack for paint and debris accumulation.

It was conservatively assum-ed that one trash rack was operational, and that a minimum of half of the depth around both trash racks would be covered with debris other than paint.

The assumption of 50% blockage is consistent with reference 3 and with the requirements of Regulatory Guide 1.82 Revision O(10)

The requirements of

} states that calculations show that Regulatory Guide 1.82 Revision 1 the accumulation of debris will not result in a loss of the available NPSH exceeding "50% of the NPSH requirements." By specifically applying the criteria of R.G. 1.82 Rev. 1, Comanche Peak is classed in criterion 3 since the flow velocity at the trash rack screens is less than.15 ft/sec for the unobstruct ed flow case, and the water level is above the trash rack sump screens. Thus, Comanche Peak, under criterion 3 possesses little or no potential for debris transport for the three major types of insulation, fibrous, reflective metal or closed-cell (encapsulated). This conclusion is further supported by NITREG-0897 ) which states that plants having large screen areas can tolerate large quantities of transported debris and that a 0.2 ft/sec flow velocity was required to initiate the motion of indivi-ual shreds of insulation.

Comanche Peak's large trash rack area of 500 ft and surface velocity of about 0.08 f t/sec verifies this conclusion since f

pool velocities far from the trash racks, although not specifically determined, will be smaller. Thus this study, with a conservative 50% debris blockage, will concentrate on paint effects alone. All paint was also assumed to accumu-late as peels.

Since the zine primer paint would most likely disintegrate as a powder, this assumption is conservative with regard to trash rack clog-(

ging since a powder could flow through the trash racks and screens. The top of the trash rack is blocked and is unavailable for flow. Various

(

quantities of paint peels were assumed to pile on top of the debris with packing ratios of 1.0 to 0.75 with no flow area available through the paint peels. Justification for the lower packing ratio of 0.75 is discussed later in this section.

The uncovered depth of the trash rack was computed in each case and the pressure drop due to a flow of 12500 gpm through this uncovered depth was

[

( I

I i

H I

u,.. '. ' L

~

e,

, =

i g,g,/

C91;N

.y) -,.iy s

y21 c.

l.,,

i -

.. x h

. [ g.,..,....

..uw e

gf<.

d e

,s.--

e' :-t.:t.d,iZ i (di

.O.

ip..c is ~ i;i: 19.. t..'hl.

i

.;,e h(' M b f>:. - (r L, '

i. j,Y

'h'

.,e a,

3 e

[i 1 T/

.[] ('[J.

, s. n. n...

w-s

(

MtttIllL; g !'.., '

. f

..lb; ~.;.,,,.1.' f,.

.: m

, I

g, W

.yy.

.....w w

y -

1 g ".a R"\\

i 'g- ;fF id ' ',.$!/c,'.: 77

1",,.f,

.s 1

xvws Q j i

1

}t[$ *[

}[ - 0 I

-[1 g

s

..',$.%e;..;;+.A-

.m V@.

y}dy_i..

d~ I

..' \\'

i{

f f:

\\ g. Q'ii.l.

3,

.p F 3 g

t

.M

'* g

.?

i b :.

,:. '.A..i h:ll t

=

.e s,

se g

-~ ?.s s 28 0

V % j. Y,f !\\9"4[9.'s' f*j c.3Q;t'j'.CA,L,)

gf"m

$3 7.1 N

i i 5s

.9.

.g.c s

au o

-)

W

{ mCL 2,

'c,~.,,~??'g:K

'Jy \\

~

-mC

=-

'i.

a.n i ;..'.'-

yin, P!9aj. 4(k.

$ (f.,,. :.'TR f.

\\

~

i

( - -- r ' -

,e if f

,I n

!. rs-a i+w w-H w..

t

\\\\

,p g

ou g

.?

a

)

1,+

71

.1

! 1.

s.

.,, / / f 8 -

U

//

t,

/ <j.

t

.h;5t../,

(

\\'

~. : ? -..?.),w '-llll, u.a 4

1 l

• s'-

f t

h' ' '

,?

f.

7 c '.'. ' si ll $

'l.

e 1

.,p -

JJ

)1 i

' i,.. ' :?

,E 1 ?.

2,,

~

't I '

N-

• l g

g A,.

y.

3 v

..io y

.;'L51.- JJ

.._13

.V ' 'I"/

.e

/.1 s

?

.m Table II No Vol. of paint Vol. of debris Total Percentage Percentage Ap, across blocking the blocking the blocked screen screen trash rack, screen, ft3 3

screen, ft depth, ft.

blocked by blocked by ft of water debris debris and paint 1

1 1049.6 3125 4.175 50 66.8 0.025 2

1049.6 4541 5.59 72.6 89.5 0.25 Y

3 1049.6 4967.7 6.02 79.5 96.3 2.0 4

1049.6 5053.4 6.10 80.8 97.6 5.0 Note: Percentage screen blocked = (depth of screen blockol x 100/6.25)

Packing ratio for the paint = 1.0

calculated. The pump performat:ce would be adjudged not affected if this pressure drop were much less than the suction head of the pump (8'8").

The results for a packing ratio of 1.0 could be found in Table II.

It is seen that the pump performance would be affected when the screen is blocked 90% and over.

It was assumed in Case 1 of Table II, that half the depth of the screen is blocked by debris other than paint, and the total quantity of paint, 1049.6 ft, was assumed to stack on top of it.

In cases 2 to 4, the volume of debris other than paint, was increased beyond the value in Case 1.

It is seen with these assumptions that the pump performance would be affected when debris in excess of about 4000 ft collected and the total quantity of the paint peels fill the space around the trash racks.

Realistically water will be trapped between the paint peels with the result-ing packing fraction less than 1.0.

Two least probable configurations for packing of the paint peels exist. As illustrated on Figure 3 they are:

(1) All peels packed tightly together with no water space between them.

(2) All peels packed in an alternate fashion to allow maximum amount of water between peels.

The first case represents a packing ratio of one where packing ratio is defined as the ratio of the volume of paint to the total water and paint volume. The second extreme represents a packing ratio of 0.5.

The distri-bution of paint peels in the pool of water is a random process; therefore, it is appropriate to define the average packing of the peels as equal to the mean value of the two extreme cases. Consequently, it is appropriate to use 0.75 as the average mean packing ratio for the accumulated paint in the sump.

Assuming water viscous effects would not appreciably affect the paint packing ratio, a packing ratio of 0.75 has been considered to obtain the pressure drop results in Table III.

For conservatism no flow was assumed possible through the spaces in the occupied region.

Even with this lower value of paint packing ratio, the conclusion regarding pump performance remains unchanged. That is the pump performance would be affected only if the total quantity of paint and debris in excess of 4000 ft accunulate around the trash racks.

The spacing between paint peels is inversely proportional to the pack-ing ratio.

The average distance between stacked paint peels.

l 1

lll 7

kt k

eMV//

f

/A 2

/

/

GN

/mV//AM I

K

/'

C

/

A P

A/mMM TN 7/

IA P

/4eMV/A 0

/wFW/

0 n

/

A 1

o i

ta

/

ru

/

g i

f 3

no E

C RU g

G n

m M I

i F

kca P

AW M l

G e

N e

P I

K a

t V/A C

n A

i P

a P

TN

/A VA IA

/

P V

0 m V//A 5

'A VA 1'

I1l l\\

ll lllll l'

llt llIll[Il!

ill lll

h Tai le 111 No.

Volume of paint Volume of debris Total Percentage Percentage Ap, across blocking the blocking the blocked screen screen t rash rack, 3

screen, ft screen, ft3 depth, ft.

blocked by blocked by ft of water debris debris and paint 1

1049.6 3125 4.52 50 72.3 0.037 h2 1049.6 4192 5.59 67.1 89.5 0.25 3

1049.6 4618 6.02 73.9 96.3 2.0 4

1049.6 4704 6.10 75.3 97.6 5.0 Note: Percentage screen blocked = (depth of screen blocked x 100/6.25)

Packing ratto for the paint = 0.75

k 1

of thickness 0.035 inches with a packing ratio of 0.75 is 0.012 inches. Smaller packing ratio would result in larger spacing distances between paint peels. The larger spacing would permit water flow to reach the trash rack surface, thus, in effect, increasing the flow area to the trash rack screens. At a packing l

ratio of 0.5, the average spacing between the peels would actually equal the paint peel thickness. Thus the actual pressure drop through the trash racks would be expected to increase with decreas-i l

ing packing ratio up to a maximum value at a critical packing ratio I

which would offer the largest resistance to flow. Then a decrease in value wc ild occur for smaller packing ratios as more flow area be-h came available between peels. Thus the pressure drop calculations in Table III for a 75% packing ratio and no flow area between peels can be considered conservative since in actuality the blocked paint height would contain about 25% flow area for water. An approximate method to estimate the actual packing ratio due to kinetic theory

[

is contained in Appendix B.

t L

L e

~

L r (:

j

(

l t

3.2 Supplementary Analyses It was conservatively assumed in Section 3.1 that all the con-tainment paint accumulate in the area immediately around the trash racks and fill the space above the debris. Questions about the motion of the paint peels within the water and the ability of the peels to stick to the vertical trash rack surface are addressed here.

It is assumed here that only one sump is operative and half the depth around the trash rack is covered with debris other than paint.

The screen presents a free flow area fraction of 0.7, that is 70% of the face area is available for flow and 30% is covered with metal.

Paint peels incident normally near the screen surface would be pressed against the screen by the hydrostatic pressure. The weight of the peel tending to pull it down would be resisted by the friction between the paint peel and the metal surface. Calculations indicate that the friction is much higher than the weight of the peel and thus the peel would indeed cling to the surface of the screen vertically.

It was estimated from the dynamics of vertically descending peels, that the peels incident on an area within about 2" of the screen surface could be pressed against the screens. Details of this estimation are des-cribed at the end of this section. For conservatism, 25% of the paint incident on top of the trash rack was added to this category. The re-sulting area blocked by vertically sticking paint peels was calculated f

to be approximately 72 ft ; this area corresponds to a depth of about 0.9'.

Since half the depth is already assumed blocked by debris, the total depth blocked is now 4.025'.

The suction to the ECCS pump will now pass through the remaining avail-able depth of 2.225'. The flow stream lines would be normal to the screen surface near the trash rack.

Paint peels falling elsewhere in the pool f

at different angles of incidence, would be transported under the water velocity, gravity and drag to the screen subject to buoyancy'and other effects.

(- f

r.

They would ultimately tend to align themselves with the stream lines.

It was estimated that the peels et an average distance of 24 f t from the screen could be transported to the vicinity of the surface of the screen.

Details of the estimation can be found at the end of this section. This distance corresponds to an area about 922 ft surrounding one operating trash rack.

If all the containment paint peels block the trash rack in addition to the already blocked depth of 4.025' with a racking ratio of one and no flow assumed through the paint, a further height of 1.14' would,be blocked, leaving a height of 1.08' for flow.

The pressure loss would be equal to 0.09', corresponding to a blockage of about 83%.

Thus the pump performance would not be affected. The blockage of the trash rack occurs due to two processes: (1) peels sticking vertically which will be called adhesion and (3) peels accumulating in the surround-ing volume which will be called accumulation.

If the entire paint blocked height due to paint peel adhesions and accumulation were made available to, pair.t accmulation,for the 4.025' blocked height, a packing ratio of 0.56 would result which is close to the lower limit packing ratio. Thus the conservatism of the above assumptions is justified. Additionally, calculations indicate that the flow velocity on top of the paint is insufficient to lift it due to.a venturi effect.

Estimation of the distance 2" from the surf ace of the trash rack, for vertically clogging peels Referring to Fig. 4, the trash rack has a height of 6.25'.

Half this depth, equal to 3.125', is assumed filled with debris.

The flood water level is assumed to be 20' above the base of the trash rack or 16.875' above the surface of the debris. The hydrostatic pressure due to this head of water equals 1051.3 lbf/ft Referring to Fig. 5, showing a peel pressed against a typical grid, a unit grid area of 1 ft can be cens id ered.

Since the blockage factor is 0.7, the metal area is 0.3 f t and the balance,0.7 ft,is flow area. The hydrostatic pressure will tend to hold the peel against the screen. The metal area would of fer a reaction.

f The weight of the peel would tend to pull it down against the frictional force offered by the bearing surface.

If the frictional force were higher than the weight, the peel would cling to the surface of the screen vertically.

l

.u.

t A

-FLOOD WATER LEVEL o

B ai

-TRASH RACK

,~,~~

~~~,%

~

1r

~~

o.

%, x A. _, -

d i o

N

.t PAINT PAH4T

.to Y

A a

e DEBRIS DEBRIS i r I

P t

i k

1 i

t I

i I

4 Figure 4 i,

J 6

Trash Rdek Elevation View I

6 1

1 1

4 i

i 1

d n FRICTION pR

[

METAL AREA ~

REACTION R 0.3 FT2

/

HYDROSTATIC FLOW PRESSURE AREA ~

0.7 FT2

~ PAINT PEEL A

r "WElGHT mg Figure 5 Paint Peel Adhesion to the

[

Trash Rack Screens t

The reaction at the metal surface equals 1051.3x0.3, or 315.4 lbf.

Assuming a coefficient of friction of 0.1, the frictional force equals 31.54 lbf. The peel of area 1 ft and thickness 0.03" weighs 0.2 Ibf.

Hence, the peel would indeed cling to the surface of the screen vertically.

Using Fig. 4, Consideration will be given to a vertical peel striking the water surface at a horizontal distance of x from the trash rack.

It will be con-servatively assumed that the incident velocity equals zero.

Since the drag on a vertically descending paint peel is small and negligible, the time t taken by the peel to fall through a distance of 16.875' is given by gt

=

16.875 and is found to be 1.02 sec. While the peel descends vertically, it would be transported to the screen by the velocity of the water. The water velocity is 0.161 ft/see at the screen and it reduces with distance.

It is conservatively assumed that the water velocity is a constant equal to 0.161 ft/see and that the peel is trc.asported without slip. The distance x is then equal to 0.161 t which is about 2" with t equal to 1.02 secs.

Thus, the paint peels incident within 2 inches of the trash rack could be expected to clog the screen vertically.

Estimation of the distance of 24' from the trash rack regarding peels incident on the surface of the water at different angles:

Figure 6 illustrates a pair.t peel at an angle 6 to the vertical moving under water. The weight mg acts vertically downwards. The reaction R acts normal to the surface. The peel would fall down under the influence of mg opposed by R Sin 6; at the same time, the component R cos 6 would tend to push the peels towards the trash rack subject to drag in the horizontal direction. The equation of motion in the vertical and horizontal directions are:

m

= mg - R Sin 6 2

= mg - C /2 u A cos 6 D

D w

= R Cos 6 - C /2 A Sin OC (v-v )

m y

o Where, m = mass of the peel = p Aw p

b (

I I

I I

l l6 R Cos6 g

n R

R Sin 6

" mg Figure 6 Paint Peel Force Model (.

i p

= density of the peel A

= Area of the peel

= Thickness of the peel w

C D

E#

• p

= density of water v

= pool water velocity 9

The vertical terminal velocity for any angle 0 is obtained by setting m

equal t zer. An angle zer d

w uld result in the longest verti-cal travel time. This furthest horizontal tragel occurs at the angle for which the value of the multiplication of the horizontal velocity and the vertical travel time is largest. As seen from Figure 7, the longest travel time occurs at a small angle. However, the equations used to model the paint peel fall are inaccurate at angles near zero and ninety 'egrees since boundary J

d layer effec-ts on drag have been ignored in the equations. Consequently, re-aults at angles greater than 85 and less than 5 are not considered to be representative of actual conditions. Despite these assumption problems, the average horizontal travel distance for any paint peel can be calculated to fall betwee a 22 and 24 feet using the information presented in Figure 7.

l c

, l

3 <-

i I

/

tI i

n 2

Vertical gf

'u

/

6 j

j w

(

\\

/

a

~

/

8 p orizontal H

i

/

~o

/

1 i

n, 3

s'

{

E=

l 0

l 0

90 Angle (6) (degrees) 60.

30 m

U

\\

e w

u v

o y

\\

z Time

,~

\\

/

\\

r

=

.20 "'

1. 40

,e o

N

\\

,c N

E

/

s e

o Distance 10 "

m 3

E e

w 3

'N E-w w

l N

S

\\

5 0

l 0

90

\\

Angle (6) (degrees)

Figure 7 i

( f

4.

Conclusion It is estimated that the perfor=ance of the ECCS pump: would not be adversely affected, if* the trash racks were blocked 63* or less.

The pressure loss across the trash rack due to a blockage of 83', is much less than the normal suction head of the pump.

Such a blocking limit would be reached if (1) debris exceeding about 4000 f t' fill the space 2

around the trash racks, in addition to the total quantity of paint fill-ing the same space with a packing ratio of 0.73 or, (2) if the entire quantity of paint fills the space around one trash rack with sete of the paint peels clogging the screen vertically, in addition to debris covering half the trash rack depth.

Since the accu =ulation of all the paint inside containment plus the assumption of having half the trash racks blocked with debris is insufficient te exceed 837. blockaEe, ECCS performance is not considered capable of being affected by trash rack clogging.

L'

_19_

1

)

)

5.

References 1.

Texas Utilities Services Inc., C.P.S.E.S.

Drawing CPD-1402-1, 3/5/81 2.

Texas Utilities Services Inc., Comanche Peak S.E.S.

Drawing 2323-SI-0564, 6/12/79 3.

"Model Testing of Recirculation Sump Containment", November, 1981 Western Canada Hydraulic Laboratories Ltd.

4.

Memorandum from R.M. Kissinger to Lisa Bielfeldt, TUGC0 QA, Dal.las, 12/29/83 5.

Telex from C. Dupre of TUGC to V. Thiagarajan, Ebasco Services 2/6/84 6.

Sumps for Emergency Core Cooling and Containment Spray Systems, Proposed Revision 1 to Regulatory Guide 1.82, May 1983.

7.

Streeter, Wylie, Fluid Mechanics, McGraw Hill,1975 8.

USI A-43 Resolution Positions, NUREG-0869 (for Comment), April, 1983 9.

Containment Et arg ency Sump Performance, NUREG-0897, April,1983 10.

Sumps for Emergency Cooling and Containment Spray Systems, Regulatory Guide 1.82, 1974.

11.

A.W. Adamson, Physical Chemistry of Surfaces, John Wiley, 1976

+

)

1 Appendix A Partial Report of Western Canada Hydraulic Laboratories LTD.

Regarding

[

Model Testing of the Recirculation Containment Sump

~.

-rr I

I i

COMANCE PEAK STEAM ELECTRIC STATION t

l UNITS l AND 2 MODEL TESTING OF TE RECIRCULATION CONTAINMENT SUMO I

l FOR TEXAS UTILITIES SERVIGS INC.

i l

l I

i t

BY WESTERN CANADA HYORAULIC LABORATORIES LTD.

I PORT COQUITLAM, B.C.

73044 NOVEMBER,1981 Al

(_.

':, i t E = ', : :. :. :. ~ ~ ~ = :., : Le t ::~ ~= ic

~~

I.0 PURPOSE OF STUDY The purpose of the hydtculic model studies wcs to test one modify, if necesserv.

the Recirculction Contcinment Sumps one intokes of the Comenche Deck Steem Electric Station, Units I one 2, and ae nonstrate that the acceoted cesign wovid not be subject to vortices or other hydroulic phenomenc thct would degrace their perf ormcNe or p*oduce odverse effects on the ECCS pumps. The tests else mecsvred en infeke heediess for ecem intoke te demonstrate thct intoke beod losses were within the design value for the plo,t.

and measured head losses ocross trcsh rocks end screens.

I s

A2 Is

s

~.

4

-t ? =. : v.: : - : :: :..:. q 7 e ~:= tc. :

i i

i

2.0 INTRODUCTION

The U.S. Nuclect Regulatory Commission in Reguictory Guice 1.7* states thct "A comprehensive preoperational test program on the Emergency Core Cooling System and I

i its components should be performed to provide cssurcnce thet ECCS will occomplis5 its i

intended function when required". Furthermore, "the (preoperctienc0 testing should include taking suction from the sump to verify vortex control end ecceptable pressure j

drops ceress tresh rock with screens and in volved svetion lines", and "the testing should j

reavirec ct j

j. v1rify thet the cvcilcble net positive suction hecd is grecter then thct i

occident temperature."

A sctisfoctory in-plant test of the Comanche Peck Stecm Electric Stefion Units i j

I l

cnd 2 wcs not fecsible due to logistical problems of flooding the containment and the loc <

j cf cecess to the sump f or observation to ensure proper vortex control.

The citernative, es presented in this report, was to construct cnd test models of the sumps md intokes to verify vortex control and to determine the heedlesses cssocicted I

I I

with the trcsh rock, screens ed pipe inlets.

I i

Tests for the Comanche Peck Units I cnd 2 recirculcting intokes were corried out l

on c 1:1 secle model of a pair of 16 inch diometer intckes end their contcinment sume end l

trcsh rock structure et flows rates greater then mcximum postulcted values and et weter The conteinment geometry one cli depths equivclent to the minimum postulated levels.

Model tests j

significent items in the vicinity of the trosh rock were clso modelled.

underteken for the Dcvis Besse, J. M. Forley, ANO-2, San Onofre, Midlend one Pelo j

plants have demonstrated the effectiveness of a svitcbly-placed grcting l

Verde nuclect preventing the development of adverse flow conditions which could feed to coge it.

The effecti.eness of a similcr grating coge on degrading effects en pump performance.

the Comanche Peck Units I and 2 intokes was demonstrated during these tests.

The rctionole for the test program is presented in this report together witn c o discussion of effects which could degrade pump perfer-

{ description of the intekes, l

mcnce, o description of the test fccility, the testing progrem, test results end j conclusions.

I t

A3

. :::: :: :3.--

v.;: E=. :: :::

4..

i i

3.0

SUMMARY

AND CONCLUSIONS The recirculation intakes for Comanche Peck Units I onc 2 were testec using c 1:1 3,1 The model was teste::

scale model of a pair of intokes cnd their sump and trosh rock.

und:r the following conditions:

Postulated For Tested l

Plant LOCA Minimum wcter depth 8

ebove sump floor 8' 8" 8'I"-9'l" l

Mcx flow-one intoke ope-eting 7,200 (CS) 12,348 5,300 (RHR) 8,65!

1 Mox flow-both intakes operating 12,500

! ?,582 i

44-1T1 Water temperature CF 247 Intake pipe Reynolds No.

5.82 x 10 (CS)

From 1.05 x,106 6

to 5.25 x 10e 6

I 4.25 x 10 (R-iR)

(Seth CS cnd RHoJ t

I l

Blockage of screen crec, percent 50 90 8

1 The results of the tests on the single sump cre cpplicable to cli four sumps (two 3.2 each in Units I and 2) because of their similar geometry, flow rctes and depths of submergence.

l f

The differences between train A cnd trcin 5 sumps in ecch un:t cre es follows:

i I

f I)

The trosh rock frame on train A sump in ecch unit is of heevier cons +r.ictior then that on train B sump. There ere inter,cl broces on the train A sunc ticsh rocks which reduce effective screen crec, and hence increcse coorooc'. flow velocities. In addition these broces are o potentic! source of flow disturbcnce.

A4

.n v.I: Er-c:.::: s::.. :.: E e O: E!. :

2)

The trosh rocks on trein A sumps are also mcrgincl'y smaller onc e-e l

osymmetrical. Trcsh rocks on train B sumps are symmetricci ebout a radio! cxis through the center pair of columns of the trash rock.

l The model configuration is that of the train A sump and the surrounding j containment crea in Unit 2. This is o mirror image of, and hence hycroulicelly identice:

to, the train A sump in Unit 1. This mode, configuration was chosen because:

l l'

i.

the optimum arrangement within the test focility for testing ed oese vo.

tion could be attained; f

li.

the trosh rock is smaller with heavier and more numeroJs support members.

and thus more likely to generate poor hydraulic conditions within the sump.

Following the completion of tests on trein A sump, the containment c ec odjace.?

l to train B sump wcs modelled, retaining the more conservctive train A sump trosh rock.

A series of sensitivity tests was corried out to establish any effects due to these changet Further sensitivity tests were corried out by placing a superfluous 6 in. H-becm in the vicinity of the sump to delineote the effect, if any, of odditional structural members.

i 3.3 Severe cavitetion occurred of the lip of each intoke when tested with the origi,ci i design, je with the intoke pipe cut off et 560 to the pipe exis. This cavitction wcs I! eliminated by odding a 200 cone to each intake. (See Section *.2.2 and Figure 10).

I

'. 3.4 The tests showed that without the grcting coge in place, cir-entroining free l

l surface vortices cs well cs vortices originating from the sump walls occurred even withoJi 9

l trosh rock blockage.

Large free-surface vertices could also be produced by various j combinations of screen blockages. No vortices developed under any tested conditions

{ when the intoke was protected by a suitcble grating coge.

3.5 The trcsh rock and screen loss wcs found to be less then 0.021 ft at c pretotype

total discharge of 12,500 gpm.

I I

- 3.6 The maximum intoke head toss coefficient measured with the grating coge in piece l

i i

t and 50 percent blockage corresponded to a loss of 0.52 f t for o prototype disc 5crge of I

7,200 gpm.

A5

. = : :- :: g t 7

..;r c e-

..m: w v -: :.,...

3.7 A separate series of tests under a variety of 50% blockage conditions, desig ied to estchlish the mean intake loss coefficient and the 95 percent confidence intervol, con be i

summcrized os:

Prototype flow rete 95% Confiden:e Limits 50% blockooe opm Intoke i 5,300 0.153 f t < head loss < 0.175 f t intake 2 7,200 0.102 f t < heed less < 0.126 ft L Trcsh rock crid 12,500 0.011 f t < heed loss < 0.01 i f t screens 3:8 The sensitivity tests showed that no significant effect on heed losses or vortex action occurred with a substanticily citered contcinment configuration or the inclusion of c superfluous structural member, j 3.9 The effectiveness of the grating cage in providing vortex control wcs l

d2monstrated repectedly.

I l

i l

A6

v.u

=.:v,. : - m c...: -:ere: r= n. :

410 DESCRIPTION OF CONTAINMENT SUMPS AND INTAKES Eoch unit of the Comanche oeck plent contcins two recirculction su nps locctec c-the west side of the building.

The sump configurction of Unit 2 is o North-Sosts mirror-imoge of thct of Unit 1. (See Figure 1).

Each sump is 5'5" in width ed 6'0" deep, and subtends an engle of !!.c 50' to tne conteinment centerline. The side walls of cil four sumps are cres wit" rodii of St.' II" cnd 60' 4", and the end walls are radial to these cres.

l Two intokes are located on the outer well of each sum:, with their ce-ter'iees l 4o45' each side of the radial cxis of symmetry. One intoke in ecch su-.c seacties the C5 pumps ed the other supplies the RHR pumps. These intakes drew 7200 gam end 5300 gom i

rGspectively. However the intokes themselves cre identical.

The trosh rocks are asymmetriccily loccted with respect to the sumps, cnd their i construction differs between the two sumps in each unit. The tresh rocks on train A sumps ccrry pipe supports ed seismic restreints, and hcve odditionc! interr'cl bracing.

1 The spacing of the columns is also osymmetricci on these tresh rocks. Tresh reeks on the train B sumps are lighter in construction and have co!vmns regulcrly spaced et intervc!s of g 7030'.

i I

The tresh rcck bc s ed cocrse screens are mounted in fremes made of steel c,gle, I and these in turn are bolted to the outer faces of the tresh reck columns. Tne fine screens are mounted in separcte fromes of steel ongle and fict bcrs which cre bolted to e

fienges extending from the centrol web of ecch column.

l The tresh rock bcrs are 5/16" diameter, et 4-5/16" O.C. Due to manufceturing end supply problems, o number of changes were mode to screen specificctions during construction of both model and prototype. These are reflectec in the following tebie:

o l

A7

v:t = ~ E = =. : *:~: "v::: ;. *.: : ::~ = ii.~:

I COARSE SCREEN FINE SCREEN Specified

.500 in. Sq. Openings 7 Mesh By Client

.105 in. Dic. Wire

.020 in. Dic. Wire 1

68.3% Open Arec

.123 in. Wide Ope.ings 7i..I % Open Arec Trash rock Supplier 2 Mesh 7 Mesh 14 CA. Wire

.026 in. Dic. Wire

.420 in. Wide Openings

.I 15 in. Wice Ope-:n;s 70% Open Area 64.6% Open Are:

Tsst Model

.500 in. Sq. Openings 7 Mesh

.105 in. Dic. Wire

.02S in. Dic. Wire 68.3% Open Arec

.115 in. Wice Open!. ;s 64.6% Open Are:

Since most of the screen head Icsses occur ceross the fine screen, the slight difference in cocrse screen size is not expected to be significant. In any cose, this j difference on the model will tend towcrds more conservative estimates of screen losses.

i l

WCHL constructed its tresh reeks with the horizontet bars laid over the top of the

, vertical bcrs. On the prototype trosh rocks, the supplier interwove these bers. This diff erence is not significont in terms of the vclidity of test results.

i I

I i

I l

A8

l Appendix B Paint Packing Ratio Estimation

u.

Entimation of the Packing Ratio Under Static Conditions By considering the accumulation of an ideal uniform stack of paint peels j

with a packing ratio of one without separation and an average thickness t=0.03", 400 peels would exist in a one foot depth.

In reality however, molecular forces will result in producing a separation between the peels, varying with the depth of water and weight of peels above the point of con-cideration.

If the number of peels from the surface of the water to a particular depth is defined as n, the downward force (F) on the water at the considered depth could be computed as the difference between the weight of the peels and the buoyancy on the peels.

The area of a peel of width b and depth d is A = bd.

Therefore, the resultant force is:

j F = n Ato

- nA t p i

y

=nAt(o-p) p The pressure due to this force is computed by F/A = nt (o -p ).

Since the p y fluid is assumed static, the pressure is the same in all directions. The resultant force in the horizontal direction is equal to nt (p - p )r d for p

y a small separation (r).

This force is balanced by the molecular surface forces between the water and the peel which decrease with increasing separation r.

These forces act on the two bounding paint surfaces. These surface fcrees are expressed in units of force per unit length c.

In FPS units, c will have the dimensions of lbf/ft.

El

l[

t l.

This force is balanced by the total hydrostatic force.

Therefore, I

/

nt (o

-p) rd=2bo p

y b_

2 d o

r= nt (o - p) y In the above equation b,d and t are in dimensions of length, o has units of force per length, n is the number of peels, c and o are the weight densities y

of the paint peel and water with the dimension of Force /(Length).

khen square peels are considered, 20 r= nt (o

-p) p y

The surface-fluid interface constant c for water and paint peels is not known precisely. However, o for air interface is 0.005 lbf/ft. (Ref. 7).

Adamson (Ref.11) reports values of 72 dynes /cm ((0.0049 lbf/f t.) for water air inter-face; he also reports values as low as 20 dynes /cm (0.00137 lbf/ft) for interfaces between water and certain surfaces. The packing ratio may be defined as the ratio of the height of vertically stacked paint peels without separation and the height with separation. The packing ratios were found to be 0.31 and 0.39 with o's equal to 0.0049 and 0.00137 lbf/ft. respectively.

The separations calculated for the first 10 peels were found to be high, thus the assumptions made in the derivation of the separation distance calculated by this method may not be valid.

Hence, the packing ratios were also calculated omitting the first 10 peels. The packing ratios were found to be 0.39 and 0.76 with c=0.0049 and 0.00137 lbf/ft. respectively. Thus the pack-ing ratios could be expected to range up to 0.76.

82

_o

3.0 Figure B.1 Separation Vs. No. of Peels

~

c = 0.0049 lbf/ft.

2.5 2.0 Eii 3

1.5 E3 a

?

v1 1.0 0.5

=

1 1

n a

i 0

100 200 300 400 No. of peels 1

i 1

1 i

1 22.5 26.0 31.0 35.1 38.9 Depth, inches B-3