ML20062B186

Mark II Design Assessment Rept

ML20062B186
Person / Time
Site:	LaSalle
Issue date:	10/16/1978
From:	COMMONWEALTH EDISON CO.
To:
Shared Package
ML20062B143	List: ML20062B139 ML20062B186
References
NUDOCS 7810230258
Download: ML20062B186 (31)

Text

4

~

~

LSCS-MARK II DAR Rev. 3 10/78 go - 313 LA SALLE COUNTY STATION 37y ' -

MARK II DESIGN ASSESSMENT REPORT INSTRUCTIONS FOR UPDATING YOUR LSCS-MARK II DAR To update your copy of the LSCS-MARK II DAR, insert pages and figures as indicated.

INSERT Appendix B After page B.2-61 Pages B.3-1 through B.3-31

(

1

(

1 7 &/o A3 0 AF8 A

LSCS-MARK II DAR Rev. 3 10/78 B.3 QUESTIONS OF JUNE 30, 1978 The index below provides the location of the detailed re-sponses to each of the questions. Following this index is a list of all the questions along with the specific DAR response, or reference to the DFFR.

NUMBER

• QUESTION

• KEYWORD INDEX TO QUESTIONS DFFR DAR 020.64 Downcomer Chugging Lateral Load X X 020.65 Flanged Downcomer Chugging Lateral Load X X 020.66 Static Equivalent Load and Down-f comer Natural Frequency X X 020.67 Multiple Downcomer Lateral Loads X X 020.68 Maximum Pool Swell Elevation X X 020.69 Drywell Floor Negative Differen-tial Pressure X 020.70 Submerged Structures Drag Load

( with Pool Swell Blockage X X 020.71 Drywell Pressure History as Input for Pool Swell Model X X-020.72 Pool Swell Impact Loads on Small Structures X X 020.73 Margin for Calculated Pool Swell Velocity X X

020.74 4T Chugging Load and FSI "R ingou t " X X 020.75 Steam Condensation Submerged Structures Load X X s.

• Series 020: Containment Systems Branch.

B.3-1

~

LSCS-MARK II DAR Rev. 3 10/78 QUESTION 020.64 "The data base from which chugging loads on downcomers was developed indicates that lateral loads were also observed at vent clearing. These loads were as high as 3.5 kips (See Table 3-3 of NEDE 21078-P) . Therefore, it is our posi-tion that a design load not less than 3.5 kips be specified

'for downcomers during vent clearing. This static equivalent load should be used for each plant with a vent natural fre-quency less than 7 Hz. For a vent, natural frequency greater than 7 Hz a higher vent clearing static equivalent load should be specified and justified."

RES PONS E In the 4T test series, no significant lateral loads were observed between the start of the the test and the onset

( of steam condensation loads such as condensation oscillation and chugging. However, in the referenced tests (Table 3.3 of NEDE 21078) , static equivalent measurements of lateral load- up to 3.5 kips were observed. These loads are thought to be unique to the test setup (Figures 31, 31A, 32, and 33 of NEDE 21078) and not applicable to either the 4T fa-1 cility of the Mark II containment.

In the test facility where the 3.5-kip static equivalent loads were measured, there is effectively no drywell volume except the air occupying the vent line. In contrast to the 4T facility or a Mark II containment, in the referenced tests only a very small quantity of noncondensible gas is vented to the pool, following which steam condensation occurs immediately. In these tests without a drywell, the vent pressure typically rose to approximately 1 to 1.5 atm while the small air volume cleared. It then dropped approx-imately 2 atm from this point as condensation of the steam commenced. This reduction in vent pressure is evidence B.3-2

LSCS-MARK II DAR Rev. 3 10/78 of the collapse of the bubble at the vent exit and an attendant reentry of water into the vent.

The bubble collapse (similar to a chugging event) causes the lateral load (during vent clearing), which would not have occurred if a drywell were present, as in the 4T facil-ity and the Mark II containment. In this case, continuous air flow to the bubble at the vent exit would be gradually.

diluted with a larger-flow of steam (which in itself is capable of maintaining a positive bubble at the vent exit) .

In the absence of a collapsing vent exit bubble, significant lateral loads would not be expected to occur during the 4T or Mark II vent clearing transients, and this was con-firmed in the 4T tests.

Nevertheless, since the 8.8-kip static equivalent load used

, to evaluate the downcomer and drywell floor is greater than

( the 3.5-kip load postulated to occur at the instant of vent clearing, it is clearly demonstrated that the governing design case for the La Salle containment has been considered.

The results of this evaluation, which shows the structural adequacy of the downcomer and drywell floor to accommodate this 8.8-kip load, are presented in the DAR, Section 4.3.

i t

4 l B.3-3

~

~

, LSCS-MARK II DAR Rev. 3 10/78 QUESTION 020.65 "The data base (NEDE-21078-P) from which the chugging load specification for downcomers was developed was obtained with a vent configuration unencumbered by flanges or other protuberances located in the vicinity of the vent exit.

It is our position that these load specifications are not applicable to any Mark II plants with vents which are flanged at the vent exit. Either the vent exit flanges should be removed or additional steam tests should be con-ducted with a vent exit flange."

RESPONSE

Flanges or other protuberances located in the vicinity of the vent exit have been or will be removed.

d B.3-4

LSCS-MARK II DAR Rev. 3 10/78 QUESTION 020.66 "The static equivalent load for a downcomer depends on the natural frequency of the downcomer. The current load spe-cification of 8.8 kips was obtained in a test facility with a downcomer natural frequency of about 7 Hz. This load has not been demonstrated to be conservative for downcomers with a higher natural frequency. For a vent natural fre-quency greater than 7 Hz a higher lateral load should be specified and justified. Additional information is needed to establish a static equivalent load for downcomers with a natural frequency greater than 7 Hz. In addition, we require that each Mark II plant provide an evaluation of the downcomers utilizing the dynamic forcing function in Task A.13 in the Mark II supporting program as confirmation of the static equivalent load evaluation. The static equi-valent and the dynamic loads for the downcomers described 5

above are based on tests with downcomer diameters of 24" or less. Additional information will have to be provided to establish lateral loads loads for downcomer with a larger diameter."

RES PONS E As stated in the LSCS-DACR, the natural frequency of the downcomer is less than 7 Hz. If the hydrodynamic mass is considered, the frequency would be even lower, thus the 8.8-kips bounding lateral load is applicable for La Salle.

l Since the downcomer diameter for La Salle is 24 inches, the dynamic forcing function defined in NEDE 23806-P is applicable.

The evaluation of the downcomers utilizing the dynamic forcing function in Task A.13 (Reported in NEDE 23806-P) in the Mark II supporting program has already been provided B.3-5

- - - ---- - l

l

'LSCS-MARK II DAR Rev. 3 10/78

-in the closure report (DACR) and is reproduced'below:

I Load Max. Vent Moment Max. Vent Shear 8800 lbf 301 ft-kip < 1. 0 k s i (static) 10,000-30,000 lbf 65 ft-kip < 1. 0 k s i 6 msec-3 msec

.(dynamic) i 1

r i l

1 .

f 8,

)

1 4

i B.3-6

, LSCS-MARK II DAR Rev. 3 10/78 QUESTION 020.67

" Referring to Section 4.3.2 of DFFR (NEDO-21061-P, Rev. 2) .

(1) It is noted that the force magnitude distribution em-ployed for the probabilistic analysis of multiple down-comer loading is taken from Table 3-6 of NEDE-21078-P.

These data were obtained during steam blowdown with significant air admixture (test 5 and 7). Thus, they do not correspond to the " worst" loading case (0% air admixture) which yields the 8.8 kip maximum lateral load specification. We require that the multiple down-comer loading be modified so as to be consistent with this worst case distribution.

(2) Since the direction of the combined loads from multiple downcomers is arbitrary, assumption 2 of the analysis

, is unjustified. We require that the magnitude of the resultant of all forces be employed to~ define multiple downcomer loads. The analysis and results (Figures 4-10 and 4-10a) should be modified accordingly.

(3) The results shown in Figure 10-4'-a implies that a single .

downcomer will experience an infinite loading. This is obviously incorrect and suggests that the referenced figure is in error. Provide a corrected version of this figure."

RESPONSE

a. In order to be consistent with the worst-case distri-bution for the multiple downcomer loading, the load values obtained from Table 3-6 have been increased by a factor of 8.8 = 1.26. The La Salle plant is currently being assessed for the impact of these loads.

B.3-7

LSCS-MARK II DAR Rev. 3 10/78

b. Because of the random nature of the chugging phenomenon, the lateral load magnitude anf. direction and also the number of downcomers which chug at the same instant vary at random.

For a conservative assessment of the loads from multiple downcomers, the load on each downcomer was assumed to act in any direction so as to induce the maximum force at a given design section of the drywell floor as well as the downcomer bracing system.

c. Figure 10-4-a of DFFR Revision 2 (GE Company Proprietary) has been corrected and is provided as Figure 4-53* in the Proprietary Supplement to DFFR Revision 3 (NEDE 210 61-P) submitted to the NRC on June 30, 1978. This correction was effected by exchanging titles on the ordinate and abscissa while retaining the orientation

'[ of the curve and extending it to the limits of the graph.

(

B.3-8

LSCS-MARK II DAR Rev. 3 10/78 QUESTION 020.68

" Based on our review of the 4T test reports ( NEDE-13 4 4 2 P-01, NEDE-13468P) and the Phase I, II, III Applications Memoran-dum dated January 1977, it is our position that the speci-fication for maximum pool swell elevation account properly for observed trends with submergence and state of the blow-down fluid. We require that the maximum pool swell elevation specification consist of the maximum of either 1.5 times a submergence or that predicted by the p001 swell analytical model using a polytropic exponent of 1.2 for wetwell air compression."

i

RESPONSE

Subsection 3.3.2.2 of the DAR specifies on page 3.3-4 that the pool swell loads be applied up to an elevation 1.5 times

( the submergence depth, above the initial pool surface eleva-tion. This elevation of 1.5 times the submergence depth is greater than the elevation predicted by the pool swell analytical model using a polytropic exponent of 1.2 for wetwell air compression. Elevation above pool surface is 19.25 feet using the criterion of 1.5 times submergence

( depth (1.5 x 12.833 feet) , as compared to 17.2 feet predic-ted by the analytical mode) using polytropic exponent of 1.2. Therefore, the design basis for La Salle bounds the case with a polytropic exponent of 1.2.

B.349

LSCS-MARK II DAR Rev. 3 10/78 f

2PESTION 020.69 Our review and analysis of the data base (4T tests and EPRI results) for upward P suggests that the 2.5 psi specifica-tion is inadequate for certain Mark II plants. We require that the current specification be replaced by:

A PUP = 8. 2 - 4 4F (psi) 0<F< 0.13 A PUP = 2. 5 (psi) F> 0.13 where F is a plant unique parameter defined by AB*AP*VS F=

VD * ( AV) 2 l

with AB = break area AP = net pool area

{

AV = total vent area VS = wetwell air space volume VD = drywell volume

RESPONSE

t Although we do not fully understand ncr agree with the source of the suggested NRC load specification, an evaluation was made for the La Salle County Station.

For the La Salle plant, 2

AB = Break Area = 3.139 ft ,

AP = Net pool area = 4685 ft 2, 2

AV = Total vent area = 295 ft ,

VS = Wetwell Air Space Volume = 166,400 ft , and VD = Drywell Volume = 221,513 ft 3 ,

B.3-10

LSCS-MARK II DAR Rev. 3 10/78 Using the equation, p, (AB) * (AP) * (VS) *

(VD) * (AV)

= (3.139) x (4685) x (166,400)

(221,513) x (295) 2 t

= 0.127 Since the calculated value is 0 < F < 0.13, the upward pres -

sure differential on the floor is; a pup = 8.2-(44) (0.127) = 2.62 psid The-La Salle drywell floor has been designed for an' uplift pressure of 5 psid, LSCS-DAR Table 1.1-1, and is presently being analyzed for a 9 psid uplift pressure.

i 1 1

1 4

'l

, B.3-11

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

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

~

LSCS-MARK II DAR Rev. 3 10/78 QUESTION 020.70 "The DFFR (NEDO-210 61) methodology for estimating steady state drag loads on submerged structures is unacceptable for those cases where the structures represent significant blockage to the pool water slug motion. We require that the drag coefficients used to compute the loads be modified according to traditional literature references which take account of the effect of blockage."

RES PONSE During pool swell, the pool water slug motion is vertica17.y upward. Significant blockage of the pool motion is affected by obstructions which are transverse or perpendicular to the flow direction and not by objects oriented parallel to the flow direction. The large majority of structures

( in the suppression pool and pool swell air space are oriented vertically, i.e., parallel to pool water slug motion. These structures include ECCS piping, SRV discharge lines,.down-comer vents and drywell floor support columns. As was re-ported in Section 4.8.1 of the LSCS-DAR, the downcomer bra-l cing which lies in the horizontal plane is well above the

( peak pool swell height and is therefore of no effect. The pool water slug motion has been observed in tests (Compar-ison of 1/13-Scale Mark II Containment Multi-Vent Pool Swell Data with Analytical Methods, NEDO-21667, August 1977) which include such structures, and no flow blockage has been ob-served. A recent review of the La Salle drawings confirmed that there are only a few dispersed pipes which are trans-verse to the pool water slug motion. Suction strainers submerged in the pool are below the region of bulk pool swell motion. The short horizontal runs of pipe connected to them are widely dispersed to avoid possible blockage i and to satisfy separation criteria. Some structural com-ponents within the pool swell zone are horizontally oriented, B.3-12

a .

LSCS-MARK II DAR Rev. 3 10/78 but it is largely dispersed. The downcomer bracing system, which comprises the bulk of the horizontally oriented struc-tural members, is located above.the maximum pool. swell eleva-

$ tion and hence does not contribute to the. blockage of'the pool water slug; motion.

The total area of the dispersed horizontal piping is less than 5% of the pool water slug area. No flow blockages are encountered for the pool water slug motion-in the La Salle wetwell. Hence, the values of the drag coefficients need not be and have not been modified on account of blockage

! effects.

j 4

f 1-l

(

l e

i i

i i

i B.3-13 i

, , . . . . ,_,-4 , - - - . , . . . , _ - . .

. - , , , .--r, . _ -

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

LSCS-MARK II DAR Rev. 3 10/78 QUESTION 020.71 "In the response to Question M020.58 (4) (questions dated January 18, 197 ~/) it is stated that " Calculations of pool swell for Mark II containments using the analytical model utilize the appropriate calculated drywell pressure response (NEDM-10320) as an input." It is our position that the specification of pressure history is an essential element of the DFFR methodology and that the particular choice cited above has not been demonstrated to be appropriate.

To justify such use, we require that pool swell-calculations be made for selected 4T tests using drywell pressure response computed according to NEDM-10320 in lieu of the measured drywell pressure histories. The selected 4T tests are the two saturated liquid blowdowns made during the Phase II tests series (Runs 36 and 37). The response (pool swell l ( elevation, velocity, bubble pressure) calculated in this manner should be compared with measured values and with similar calculations made using the measured drywell pressure histories."

RESPONSE

For Mark II plant design assessment, the calculated drywell pressure response (NEDM-10320) is conservatively chosen to input the Mark II pool swell model. To illustrate this conservatism, the pool swell model calculations are performed for 4T tests runs 36 and 37 using the computed drywell pres-sure response (according to .NEDM-10320) as input. These calculations are compared to both the measured 4T pool swell response and the pool swell calculations using the measured 4T drywell pressure histories. The corresponding pool swell elevation, ve'locity, and bubble pressure are plotted in the attached figures.

B.3-14

t LSCS-MARK II DAR Rev. 3 10/78 The pcol surface elevation curves.(Figures 0020.71-1 and 0020.71-2) show that results using NEDM-10320 drywell pres-sures yield pool swell elevations which exceed the measured 4T test data and also exceed the results obtained using the 4T measured drywell pressure.

The pool surface velocity (Figures 0020.71-3 and 0020.71-4) curves illustrate that the maximum velocities calculated using NEDM-10320 pressures are greater than (as in run 37) or comparable to (as in run 36) those measur'ed in the 4T tests. Furthermore, pool surface velocities using NEDM-10320 input consistently exceed the velocities calculated using the 4T measured drywell pressures.

The bubble pressure curves (Figures Q020.71-5 and 0020.71-6) shcw that bubble pressures calculated using NEDM-10320 dry-well pressures provide the upper bound on both the measured

( 4T data and on the calculated values using the measured 4T drywell pressure.

Based upon the above results, it is concluded that using the calculated drywell response (using NEDM-10320) for pool swell calculations results in conservatively predicted pool

, swell parameters for design assessment.

B.3-15

- Rev. 3 10/78 18 16 14 12 7 W ',

t, /

=

O P lO 2 /

l d /

ge f ecr /

( a /

_, 6 k /

/

4

/

/

/

2 /

/

/

/

O .

O .2 .4 .6 .8 1.0 1.2

• • TIME , SEC.

• REFERENCED TO POOL HT AT VENT CLRG. 23.63

• • TIME AFTER VENT CLRG.

RESPONSE WITH 4T MEASURED LA SALLE COUNTY STATION DRYWELL PRESSURE INPUT MARK 11 DESIGN ASSESSM ENT REPORT s R SPONSE WITH NEDM-10320 FIGURE Q20.71-1

--- 4 T TEST DATA POOL SURFACE ELEVATION - RUN 36 B.3-16

Rev. 3 10/78 18 16 14 ,

F /

U l2 /

z /

9 h

g lO w

u /

es /

! /

/

E6 /

/

4 /

/

/

2 /

/

O '

O .2 .4 .6 .s 1.0 1.2

• • TI M E , S EC.

• REFERENCED TO POOL HT. AT VENT CLRG. 23.63

• • TIME AFTER VENT CLRG.

LA SALLE COUNTY STATION RESPONSE WITH 4T MEASURED u ARx is otsics assessment REPORT

+ DRYWELL PRESSURE INPUT RESPONSE WITH NEDM-10320 FIGURE Q20.71-2 INPUT

^ ^ "- ""

--- 4 T TEST DATA 8.3-17

Rev. 3 10/78 26 24 7

/

/ \

20 / \

/ \

j 18 3

! \

R 16 t \

g' 14 o

\

3 12 \

s 10

/

( 5 /

\

m 8 j \

6 \

\

4 \

2 O

O .I .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.2

• TIME,SEC.

+ RESPONSE WITH MEASURED

• TIME AFTER VENT CLEARING DRYWELL PRESSURE INPUT RESPONSE WITH NEDM-10320 INPUT LA S ALLE COUNTY ST ATION

- - 4 T TEST DATA MARK 11 DESIGN ASSESSMENT REPORT

( FIGURE Q20.71-3 P00L SURFACE VELOCITY - RUN 36 B.3-18

Rev. 3 10/78 30 28

'26

[

/

24

/ T 22 l \

l \

3 20 / \

I- 18 -

/ \

D 16 8_J / \

s14 l \

j a

@ 12 i

/

5

• 10

\

a 8 /

\

a 8

/

\

6

\

4 / .

\

2

/

O O .I 2 .3 .4 .5 .6 .7 .8 .9 1.0 1.2

• TIM E , SEC.

• TIME AFTER VENT CLEARING LA SALLE COUNTY STATION MARK 11 DESIGN ASSESSMENT REPORT

_ RESPONSE WITH MEASURED DRYWELL PRESSURE INPUT FIGURE Q20.71-4

~ RESPONSE WITH NEDM-10320 INPUT P00L SURFACE VELOCITY - RUN 37 4 T TEST DATA

. Rev. 3 10/78 9

~

s

\

N N m.

\ -

N N

\

' \

/

/

<J N N'

/ o ow

/ ~

cn

( l

/ W

/N  ?

s 9F 1  %

N.

N W.

\

\

\ T.

N 5 N

\ -

\

\

\ o e m , N o m o e N O *

• 1 m m m n A N N N N N VISd-38nSS38d LA SALLE COUNTY STATION

+ " ' " ^ "

R LP UE fPUT RESPONSE WITH NEDM-10320 FIGURE Q20.71-5 INPUT 4 T TEST DATA BUBBLE PRESSURE - RUN 36 B.3-20 -

Rev. 3 10/78

_9

\

N

• x N

N

\

N t T

)

/

,, / N f _

\

,J J

r 00

/ -w

/ cn V W

/  ?

( 9F

! c.T%

> ^w

-4'

% 5 N

N l '

\ * .

\

)

\

g N.

\

l

\

\

1 O e O

N O 5

• o ll';

o E0 N

• e N m

~

W T

l ylsd-3snss38d LA SALLE COUNTY STATION DR LP SUE NP T RESPONSE WITH NEDM-10320 INPUT FIGURE Q20.71-6 4 T TEST C':m BUBBLE PRESSURE - RUM 37 B.3-21

] LSCS-MARK II DAR Rev. 3 10/78 i

5 OUESTION 020.72 "The DFFR impact load specification for small structures is inadequate. The current load specification consists of a peak pressure-velocity correlation developed from the Mark'II PSTF tests. The peak pressure is used is conjunction with an " aver.cge" 7 msec duration to completely define a pressure pulse. The use of the same 7 msec duration for

! ,all situations has not been justified, thus the gyrrent specifications is incomplete. We require that th'e load specification be modified so as to establish a conservative ,

pulse for all Mark II anticipated situations of target geo-metry, target size, pool flatness and pool approach velocity "

RES PONSE l The impact pressures presented in Figures 4-34, 4-35, and

( 4-36 of DFFR are actual test data. These pressures do indeed' I

depend on the width of the target and the flatness of the approaching pool surface. In the Mark II supression pool the pool surface is relatively flat; therefore all of the PSTF impact test data using only circumferential targets

, are prototypical of the Mark II conditions. Furthermore,

( the type and the sizes of the targets tested were also proto-typical of the Mark II plants. The load specification in DFFR is therefore a direct application of the test data without any extrapolation.

i With regards t the duration of the load, the DFFR specified 7 msec. The 7 msec duration was chosen because it is the i most representative time duration based on tests using cir-cumferential targets. The duration for these targets ranged from 2.0 msec to 7.0 msec for 17 out of a total of 18 test cases. The one other case had a duration of 15.7 msec (NEDE-13 4 26-P) .

B.3-22

LSCS-MARK II DAR Rev. 3 10/78 The effect of the pool approach velocity on the duration of the pressure pulse is also accounted for in the DFFR

- specification of 7 msec. All of the 18 tests with circum-lI ferential targets were conducted in the velocity range of 22 ft/see to 38 ft/sec. The La Salle pool surface velocity

' is within this range and therefore no extrapolation of data

! was needed.

I The use of an " average" 7 msec pulse and associated peak pressure is justifiable for all structures in the pool impact region.

Structures which have a fundamental frequency below 70 Hz

! (this includes all piping) have a dynamic load factor (DLF) which is proportional to the duration of any 7 msec or shorter pulse. Therefore the static equivalent of a 7 msec

- or shorter pulse is constant. A pulse of greater duration

( than 7 msec results in a DLF which increases at a slower rate than that associated with a 7 msec pulse. Thus the static equivalent decreases as the pulse duration increases.

, The fundamental frequencies of the structural steel in the impact region are around 110 Hz. For these flat targets i ( with a 6-inch width exposed to the impact, the DFFR pulse is conservative for any system frequency less than 130 Hz.

i n

4 A,

B.3-23 .

LSCS-MARK II DAR Rev. 3 10/78 QUESTION 020.73

" Based on ou,r review of 14 test reports, application memo-randums and the pool swell analytical model report (NEDE-2154 4-P) , it is our position that the specification of pool swell velocity according to the analytical model prediction does not provide sufficient margin to cover uncertainties in the measurements. We estimate this uncertainty to be on the order of f 10%. Accordingly, we require the addition of a 10% margin to the values predicted by the analyses for pool swell velocity."

4 RES PONSE In report NEDE-21544-P, the pool swell phenomenon was modeled according to two sets of assumptions.

( The first set of assumptions, called the "best-estimate assumptions", is presented in Section 6.4. The objective of this set of assumptions is to model the pool swell pheno-menon as closely as possible to match the 4T test results.

In the context of this set of assumptions, the uncertainty in the computation on the maximum pool swell velocity is on the order of f 10% when compared with the test results.

The second set of assumptions, called the " applications assumptions", is presented in Section 6.7. This set of assumptions is recommnded for design calculations. According to this assumption, the predicted maximum pool swell velocity is higher than the observed maximum velocity from the 4T tests.

The major difference in the two sets of assumptions is in the modeling of the flow from the drywell to the wetwell i through the downcomer vents. Under the best-estimate assump-tions, an air-steam mixture is allowed to flow from the

-B.3-24' -

LSCS-MARK II DAR Rev. 3 10/78 drywell to the wetwell after the expulsion of the air in the vents. In the applications assumptions, only air is allowed to flow through the vents. The latter results in an increased bubble pressure and higher acceleration of the pool surface, thus producing a higher pool swell velocity in comparison with the results from the 4T tests.

The analytical model for predicting the La Salle pool swell phenomenon used the second set of assumptions where only air was allowed to flow from the drywell to the wetwell.

This is a very conservative assumption that results in a computed pool swell velocity higher than what might be ex-pected in the actual case. The La Salle calculations thus provide sufficient margin to cover the uncertainties in the computed pool swell velocity.

Notwithstanding the conservativeness of the pool swell model,

( the pool swell impact loads calculated for the La Salle plant include a 50% design margin as recommended by the DFFR Revision 3. A description of this load is given in I

Section 3.3.2.2 of the DAR.

The 50% design margin adequately covers the concern raised 1 by this question. For example, if instead of imposing the

design margin of 50% on the pressure, a margin of + 10%

were applied to the velocity and the model in the DFFR were then applied, the loads obtained by imposing the 10% margin on.the velocity would be lower by about 27% than loads obtained by applying a 50% margin on the pressure.

\.

B.3-25

LSCS-MARK II DAR Rev. 3 10/78 l QUESTION 020.74

)

"The current chugging load specification consists of an oscillatory pressure load derived from a conservative chug

in the 4T facility. This load includes the FSI related i " ring out" of the test walls. The actual load is an impul-i sive load resulting from collapse of steam bubbles at the exit of the vents. To confirm that the direct application of the pressure signal to containment walls ~is conservative

additional information is needed. Wall pressure measurements during a conservative chug at the plane of the vent exit should be used to construct an impulse load at the vent exit. The impulse load specification should be used in the coupled fluid-structure analytical model of the 4T facil-ity described in NEDE 23710-P to confirm the conservative nature of the current chugging wall load specification."

I i RES PONSE f

A' study confirming the conservative nature of the current chugging wall loads specification compared to an " impulsive" specification was demonstrated in a May 17, 1978, Mark II containment meeting with the NRC. A proprietary letter report " Lead Plant Containment Response to Improved Chug-

< ging Load Definition" was submitted on July 7, 1978.

l t

I

! N +

B.3-26

l I

LSCS-MARK II DAR Rev. 3 10/78 QUESTION 020.75 "The supporting program report NEDO 21297 includes an LTP effort to define main vent condensation submerged structure loads. The current DFFR and lead plant program do not in-clude a definition for this load. Either provide a load for steam condensation - submerged structure drag for the STP or justify deferring this item to the LTP."

RESPONSE

The submerged structures in the La Salle suppression pool are currently being assessed for loads due to main vent steam condensation loads. Loads due to both phases of main vent steam condensation, i.e., condensation oscillation and chugging, are being evaluated. The Mark II Owners Group and General Electric Company are currently pursuing a meth-( odology for main vent steam condensation load evaluation.

An interim procedure is being employed for the LSCS to pro-vide a conservative design basis consistent with the evolving formal methodology currently under development.

The same basic approach and fundamentals that are applied s to air-bubble-induced loads are applied to steam bubbles (see NEDE-21471 and NEDE-21730). The steam bubbles are treated as stationary, finite-sized, multiple sources.

The resulting potential gradient distribution within the bounded pool is determined by utilizing the method of images.

The total drag loads due to both standard and acceleration drag are determined for each submerged structure as follows:

Standard drag, F s

= D^x0 ll c

Acceleration drag, F =0 A 2)

C B.3-27

LSCS-MARK II DAR Rev. 3 10/78 where:

p = pool water density, A

x

= structure's area normal to flow direction, C = standard drag coefficient, D

Vg = acceleration drag volume, U = fluid velocity, and U = fluid acceleration.

The source strength for steam condensation oscill'ation loads is derived from the 4T data (NEDE-13 4 68 P) . The maximum pressure oscillations for this phenomenon were observed on the bottom of the tank and are bounded by a 1 5-psi value.

The load history is considered to be sinusoidal, with an amplitude of f 5 psi and a frequency range of 2-7 Hz. An equivalent source strength at the downcomer exit is derived from the maximum observed load on the bottom of the tank.

I The derivation considers the finite size of the steam bubble and the 4T tank and downcomer vent configuration.

The derived source strenth is then used directly in the La Salle-specific submerged structure load determinations.

Condensation oscillation is treated as occurring simulta-

, neously and in phase at all the downcomer vents.

In an analogous fashion, chugging source strengths are de-rived from the 4T data. Again, the maximum observed loads on the tank bottom are used to establish bounding source strengths. The bounding positive load used is +20 psi, and the bounding negative load used is -14 psi. The equivalent source strengths at the downcomer exit for each of these loads are derived as in the case of the condensation oscil-lation loads. The frequency range used for chugging loads is 20 through 30 Hz. Main vent chugging is a stochastic s phenomenon in both its occurrence (timing) and load mag-nitude. Thus, the number of possible permutations for B.3-28  ;

k LSCS-MARK II DAR Rev. 3 10/78

/

multiple main vent chugging combinations considering relative timing and load magnitude is immense. A process of this type is frequently addressed by a Monte Carlo simulation.

The relative locations and orientations of the source and the submerged structure of interest are a major consideration in defiring the forcing functions on that structure. These considerations are complicated when multiple sources are pretent, since their relative phasing is important. This rec 2),ts directly f rom the fact that the load on a submerged strudture is produced by the differential pressure across the structure. Hence, when multiple sources are considered, the situation where sources on opposite sides of the sub-merged structure are out of phase will produce a larger load than the situation where the sources are in phase.

The following conservative approach is being utilized as an interim method for the La Salle plant. The randomness in multiple source timing and phasing is accommodated by k considering the worst case. The worst case is defined for each specific application such that the pressure gradient across the given structure is maximized. It is recognized that as the number of main vents considered to be partici-pating in the chugging event increases, the probability of the assumed bounding source configuration decreases rapid-

, ly and soon becomes incredibly small. Thus, as the number of participating main vents increases, the source strength for each is diminished by a factor as shown in Figure Q20.75-1. For example, if five main vents are being con-sidered, the derived 4T source strength for each vent is reduced by multiplying by the factor 0.57. The curve in Figure Q20.75-1 is based on Figure 4-10b of NEDE-21061-P, which has been extended to range from one to one hundred downcomers and then normalized. The source strengths are thus specified as a function of the number of main vents considered to be contributing to the loading on the given

< i structure. This procedure is applied by initially consider-ing only the main vent of closest approach to the given B.3-29

LSCS-MARK II DAR Rev. 3 10/78 structure to be chugging. The ?. cad is' determined for that configuration. Next, the counding combination of the two closest main vents is used to define the load. This pro-cedure is repeated, and one additional downcomer is con-sidered each time in the load determination. Since each successive downcomer is farther from the given submerged structure, after the third or fourth downcomer, the incre-ment of load increase is small. At the same time, the like-lihood of each specific configuration decreases as the number of vents increases so that the source strengths also decrease.

The load on the submerged structure thus goes through a maximum as the number of main vents is increased. The com-bination which produces the largest load on the given struc-ture is used for the design assessment.

Since the procedure described above is excessively conserva-

, tive, it is considered to be only an interim procedure.

k Nevertheless, in an effort to comply with La Salle's licensing schedule, this interim procedure has been used to evaluate steam condensation loads on submerged structures in the La Salle suppression pool. However, analyses currently I in progress or soon to be initiated will utilize a Monte Carlo simulation technique similar to the one used for bound-t ary loads and described in NEDC-21669-P. In this case the

, source strength and timing are determined by the Monte Carlo simulation. The remainder of the load calculation follows the procedures set fot th in NEDE-21471 and NEDE-21730.

J B.3-30

~

3 2$- " Re 0

g 0

1 g 0 9

g 0 8

g 0S 7R E

M O

C N

g 0W 6 O D

G N

I G

g 0G 5 U H

C F

O g 0 4 ER e B M

U N

l 0 3

l 0

2 l' 0 1

q - - - O 0 8 6 4 2 O

1 0 0 0 O eoFo<ta O < 3 o Z 5 a 5 Q W ! a_< scoz t

r> m>rr ooc 2 <- Y O2 r>a* _ o Q 6z ,* I z^ :A5

-; ngEm asLra

- E l

gG ?m s3z c <Ea nEEsa - tSo 9E2 l

Pub i

(