ML20034D516

Requests Addl Info on WCAP-10924-P,rev 2,vol 2,addendum 3 to, Westinghouse Large Break LOCA Best Estimate Methodology: Upper Plenum Model Improvement

ML20034D516
Person / Time
Issue date:	12/10/1992
From:	Rosalyn Jones Office of Nuclear Reactor Regulation
To:	Tritch S WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
References
TAC-M82099, NUDOCS 9301050185
Download: ML20034D516 (8)

Text

'

7 December 10, 1992 Mr. S. R. Tritch, Manager Nuclear Safety Department Westinghouse Electric Corporation Box 355 Pittsburgh, Pennsylvania 15230-0355

SUBJECT:

REQUEST FOR ADDITIONAL INFORMATION ON WCAP-10924-P, REV. 2, VOL. 2, l

ADDENDUM 3 (TAC N0. M82099)

In our review of WCAP-10924-P, Revision 2, Volume 2, Addendum 3," Westinghouse large Break LOCA Best Estimate Methodology: Upper Plenum Model. Improvement,"

and clarifications and corrections, there are several items which must be clarified and resolved for us to complete our review. These items are identified in the enclosed questions for your response.

Please contact Frank Orr (301-504-1815) of my staff if you have any questions regarding this request for additional information.

/s/

Robert C. Jones, Chief Reactor Systems Branch i

Division of Systems Safety & Analysis Office of Nuclear Reactor Regulation

Enclosure:

Request for Additional Information cc:

L. Hochrieter DISTRIBUTION R. Ankney Central Files S. Bajorek SRXB R/F RJones MCaruso F0rr F0rr R/F

Contact:

Frank Orr, SRXB, Ext. 504-1815 SRXB:DSSA*

SRXB:DSSA*

SRXB:DSSAgm i

F0rr: Bah MCaruso RJones

/

12/01/92 12/09/92 12//3/92 Doc.Name:

G:\\W10924.PAI

• See previous concurrence W

,v OSd W 170038

%0

/

vh

\\

i Review of the Westinghouse UPI Model Improvements l

UPPER PLENUM INJECTION TEST I

a it is stated that no attempt Regarding the scaling rationale, 1.was made to model the slots that might be in the guide tubes or irregularities could promote support columns.

These slots or additional breakup of impinging large drops while also causing disturbances of the liquid falling films thereby promoting an increase in re-entrainment. Both of these effects would act to remove water contributing to core cooling by reducing the water content of the fi1=s formed on upper plenum structures and breaking up the large drops into smaller ones. Please provide justification for the lack of these modeling effects on the UPI phenomenological behaviors.

e.The test is intended to simulate the upper plenum internal geometry for the Point Beach and Prairie Island Plants. While several structures in the tests were positioned to represent the few of the orientation of the injection upon contacting a select tubes and/or columns along the direction of the jet, the other i

In the actual plant, the l

neighboring structures were not modeled.

upper plenum contains a multitude of guide tubes and support columns which was not represented in the tests. Because of the the drop sizes and liquid omission of the neigthoring structures, sheets emanating from the initial jet impingements would incur further collisions and impingement further reducing drop size. In view of these omissions, please justify the maximum drop sizes from the tests in view of the fact that these larger drops and sheets will undergo additional collisions with the upper plenum t

structures.

l I

3. The jet velocity was approximately 50 f t/sec. What are the range of jet velocities expected for the Prairie Island and Point Beach j

plants during injection? Are these velocities characteristic of one or two pumps operating? What injection ttmperature was used in the and what temperatures are assumed for the Prairie Island and Point Beach plant analyses?

Could the proposed change in UPI test l

modeling affect the worst single failure analysis or assumed j

i for the plant analyses and if not, please injection temperatures explain?

J 4.

Still photographs were taken of the tests.

Were these in view of this l

photographs used to measure the drop sizes? If so, i

drop size measurement technique please address the following 3

a. Given the large depth of field required to capture the drop concerns:

nagnification error due to different object lengths can be How was the problem of the large range in positions i

sizes, significant.

and what were the measurement and hence magnifications overco=e j

I

b. It is common practice to identify particles in focus by errors?

making a subjective judgement regarding the sharpness of the images i

involved. What was the basis of the choice of the drops to be 2

measured and what are the errors associated with this approach when out of focus drops are also included?

c. Visual sizing and counting methods are very tedious and suffer from bias imposed by the operator who usually has to use subjective judgement in placing the individual particles into a group size. This bias error could also be large (as high as 20%).

etc. If so, what Are the drop sizes identified the mean, largest, is the above bias error. What fraction of the total population do the quoted drop sizes of 1/4 inch and 1/2 inch represent? Since the full range of drops are produced, mist to 1/2 inch, what is the distribution of sizes for this test? Does the omission of the other upper plenum structures, as per Question 2 above, bias the drop distribution toward the larger sizes?

d. What is the drop size measurement error?

Considerable liquid was observed to fall outside of the water 5.collection boxes. While this may suggest that additional water may be available for fallback in the actual plant, the lack of a detailed representation of the upper plenum structures (only 6 vertical structures were modeled in the test compared to 100 plus vertical structures in the plant) supports the possibility that less water may be available for fallback. These deficiencies will produce larger and nany more drops leaving the test rig. The additional structures are expected to produce more collisions and smaller drops which can be entrained by the steam exiting the core.

Please justify the drop size distribution in view of these omissions. What is the spacial variation in drop sizes across the upper plenum?

6. The top support plate contributes to fallback in the test.

However because the wall above the UPI nozzle was missing in the test rig, the potential for return flow from the upward diverted jet from impingement on the outer rows of the guide tubes and other support columns, could not be assessed. The backflow of water back on to the UPI jet could breakup a portion of the jet emerging from the nozzle creating smaller drops or additional sheets which are easily broken up upon hitting the interhals. Please justify the omission of this phenomena.

7. Please also discuss the effect of neglecting the following phenomena in the tests:

No steam environment nor steam upflow was simulated..The a.

lack of steam-water interaction will result in larger drop sizes.

Please justify the drop sizes in view of steam upflow velocities expected in the UPI plants?

What are representative steam velocities and steam temperatures at the start of and during UPI injection?

b. The upper plenum surfaces will be hot initially. What effect would the hot. guide tubes and other structures have on jet breakup, drop breakup and accumulation of UPI water on these structures?

c. The use of plywood to simulate these structures will accentuate the accumulation of waterfilms and drops since plywood absorbs water? What effect does the plywood have on the jet breakup

compared to the metal in actual plants (also address surface roughness and temperature effects)?

Plywood would tend to in more water accumulate more water for film buildup and result accumulated in the collection boxes.

d. A second UPI injection nozzle in the actual plant produces a jet that directs flow into the opposing jet. What is the interaction of these two jets and how does it affect the drop distribution in the upper plenum?

8. Regarding Figure 2-21, in the continuous liquid jet, how is drop What causes the increase in drop size at size determined by TRAC?In Figures 2-29 to 2-34 when does UPI injection about 75 seconds?

begin and show the clad temperature resconse for the PCT location for each of the cases presented in the figures. What is the effect of only the drop size change on PCT for the Prairie Island and Point Beach Plants? How is liquid apportioned in TRAC to collect en vertical surfaces from the UPI jet versus entrained liquid from the Since some large drops are core and upper plenum that form films?

formed from the UPI, how does TRAC entrain and carry out the smaller drops and distinguish them from the larger drops? Are drops that splash water from the films accounted for and if not why not?

Figure 2-21 shows the range of drop sizes of 0.02 to 0.04 ft.

whereas the discussion identifies the range of drops to " vary in size from small and mist like to 1/2 inch".

Please provide the drop distribution density and spatial variation in the upper plenum and identify the drop size range that could be entrained by steam and the fractional population that contributes to down flow cooling. What fraction results in generating films on surfaces and of this what fraction is re-entrained and carried out the hot legs?

How is drop breakup in the steam flowing through the upper plenum accounted for? Are drops that are sufficiently small that they follow the stream lines and exit the upper plenum into the hot legs accounted for? What does the entrained fraction of liquid in Figure 2-20 represent, does this include entrained drops from liquid films on vertical surf aces, entrained water from core and all drops from Is the UPI which are small enough to be carried out by the steam?

condensation of steam on the jet and resulting pressure result in partial jet fluctuations accounted for and could this breakup prior to impingement on the upper plenum structures?

How are the surfaces modeled in the upper plenum in regard to liquid film development and entrainment? Are the films that develop in the interior treated the same as the surfaces near the periphery? Would a p' ant with more vertical surfaces produce moretha films and rentrainmer.

structures?

What data or methodology is used to determine how the liquid jet is partitioned between that which contributes to films and that which remains as drops? How was the size of the TRAC outer upper plenum region, receiving the jet, justified? See Question 17 below also.

Dentrainment Phenomena on Vertical Tubes in Droplet NUREG/CR-141, that the de-entrainment Cross Flow, provides data demonstrating

efficiency for dreplet flow is 0.19 for vertical tube arrays. Is the TRAC modeling cf deentrainment consistent with this data? If not please justify.

other supporting data or references in the 9.

Please provideinvestigated drop size and spatial distributions literature that and film formation resulting from jet impingement on a vertical array of rods?

CODE CORRRRECTIONS f

Did the runs presented in Figures 3-1 through 3-6 include the j

10.

change in UPI modeling as a continuous liquid field or are they based on the earlier modeling approach?

I HOT LEG-VESSEL CONNECTION MODEL

11. As stated in Section 4-3, the upper plenum flow patterns neglected the effects of upper plenum internal structures. What I

f effects were neglected? If the structures were not accounted for, would accounting for such structures reduce the flow area and I

result in higher velocities farther away from the nozzles producing a larger gap region for entraining drops? Please explain.

since this approach is to be used during UPI injection

12. Also, would the presence of large amounts of liquid in the steam region of the upper plenum result in increased steam velocities above that free of liquid? Also with computed assuming a steam environment most of the liquid near the periphery of the upper plenum, wouldn't the reduced f'.ow area in the upper plenum produce a higher velocity profile near and extending farther from the inlet to the hot leg than the idealized case presented herein? Please justify that the gap area chosen based on flow patterns and vectors calculated for this idealized case bound that for the conditions that would exist during UPI for the Prairie Island and Point Beach Plants,particularly when liquid is injected and water begins to pool up near the periphery of the upper plenum. For example, when the effective water pools near the periphery of the upper plenum, gap size is expected to expand into the interior of the upper for steam flow exiting the core will plenum since the area If films are more readily formed near the could this reduction in flow l

effectively decrease.

hot leg exits which may be thicker, area increase the steam velocities in this region and is this also accounted for?

13. What is the effect of this single change on PCT for the Prairie Island and Point beach Plants?

14. In Section 5, the continuous liquid field and interfacial drag Did the error corrections were included in the CCTF comparisons.

l CCTF comparisons also include the changes to the gap model for hot leg drop entrainment?

_-- _-_- _ __---------_----_____------- __--- -__--_---__-_____--------_ __-_j

ADDITIONAL QUESTIONS the CCTF tests modeled the UPI jets in the interior of

15. Also, the upper plenum and not as a jet that enters from the periphery.

Because of these differences, the drop size distribution and spatial variation of the drop size population would be expected to That is, be different between the UPI plants and CCTF facilities.

for UPI plants the larger drop sizes would be expected to occur only near the periphery of the upper plenum with the smaller sizes in the interior. In the CCTF, larger drop sizes would be expected near the interior where the jet is located with the smaller drop sizes near the periphery since the jet is directed outward. Thus, in CCTF, the potential for downflow of drops is greater since they are larger in the interior while the potential for downflow of drops near the periphery is also greater since the drops are smaller but the steam flow is lower. The higher steam flow in the interior further suggests liquid film downflow is more restricted in this region. The film downflow on the periphery may be greater due to the lower steam flow however less liquid is expected here for CCTF since the jet emanates from the interior producing smaller drops and a lesser ameunt of liquid once it reaches the periphery.

As a consequence, the potential for droplet downflow appears to be greater in CCTF than the UPI plants so that modeling the downflow of liquid into the core via films may predict clad temperature data but may be the incorrect mechanism. Furthermore, in view [these potential differences and the lack of data on liquid film formation how is the modeling of and drop size distributions in the tests, the UPI jet as a continuous liquid, which places all of the liquid on the surfaces, justified since information on the drop and film distributions in the tests is not available? How is the use of the continuous liquid jet modeling approach justified since no information is available on the partitioning of the jet into a drop field versus that which collects as films? The potential different expected for UPI plants drop / film behavior between CCTF and that could mean that modeling the UPI jet as a continuous liquid results in predicting the CCTF data for the wrong reasons.

16. Should Eq. 4-1 on page 4-2 include sin (theta,) ?

Please provide a comparison of the UPI flow per wetted surface i

17.

area between the plant calculations, CCTF, and UPTF tests presented in the report.

18. Please provide sample output of flows, velocities, and void fraction in the upper plenum at the time of PCT.

19. CCTF Run 76 showed that the region of strong down flow occurred the UPI injection location only, on the periphery of the core atsteam upflow exited on the opposite side of the uppe plenum. As such, f or the CCTF tests please indicate the location of while the intermittent and strong downflows for the WCOBRA/ TRAC simulations.

Illustrations of the CCTF tests, similar to that for the UPTF Test and through 2-4 should be provided.

results given in Figures 2-2 indicate on the figures the location of the UPI injection nozzle, broken loop, and the location of the strongest steam Please upflow. Also, what is the sensitivity of the location of the Please is modeled as a single annular zone.

outer global channelexplain why additional nodal detail (i.e divide the outer cha is not required since the CCTF tests into two or more quadrants) showed an asymmetric behavior with the down flow occurring on one side of the upper plenum with steam upflow on the opposite.

in CCTF Run 76 was The net water flow rate at the core inlet there was a net 20.

negative. Although water accumulated in the core, of liquid to the downcomer. Please discuss the CCTF Also, core cooling downflow WCOBRA/ TRAC results for this part of the test.at the time bottom reflood started.

Did was observed in the test WCOBRA/ TRAC predict this behavior? Please explain. What is the time of bottom reflood in the WCOBRA/ TRAC analysis of this test?

21. Please also provide the location of the intermittent and strong liquid down flows in the plant calculations discussed in Section 7.

Figures similar to that provided in Figures 2-2 through 2-4 are desired.

22.

Does the CCTF UPI impinge directly upon the upper plenum located just down stream of the injection nozzle Please discuss the orientation of the injection structures simulation?

nozzles in the CCTF tests relative to the impingement on the upper and justify the applicability of the CCTF nozzle j

structures orientation and jet behavior to PWR analyses.

states that On page 5-11 the last paragraph in Section 5-2-7 the results of the simulations showed that the amount of liquid 23.

retained in the upper plenum is under predicted and the loop flow l

rates over predicted. Please clarify that this is due to the 4

excessive hot leg entrainment and underprediction of condensation in the upper plenum and not excessive downflow of liquid into the 4

core.

24.

WCOBRA/ TRAC does not limit the liquid downflow based on countercurrent flow limit (CCFL) correlations.

Since the flow of steam and liquid is inherent to the drag modeling, please explain what is done in the code to prevent countercurrent downflows from violating CCFL. Are warning messages printed for cases when such conditions occur? How can the WCOBRA/ TRAC assured that CCFL limits have not been violated? Please explain.

3 a

25. The UPTF tests modeled upper plenum injection by injecting the flow via the hot leg. Because UPI in Westinghouse plants is placed directly into the upper plenum, please justify the applicability of the UPTF test results to that expected for a two-loop UPI plant.

maximum clad oxidation is

26. In the plant analyses of Section 7, uncertainty regarding oxidation is presented.

given however no Should peak clad temperatures occur approaching 2200

'F, oxidation becomes more appreciable. Please explain why uncertainty regarding oxidation has not been evaluated. Note that the time of quench is 5-15, 5-16, underpredicted by WCOBRA/ TRAC (For example, see Figures and as a result clad oxidation percentages will be 5-42, and 5-43) underestimated, particularly if peak clad temperatures are pushed toward the 2200 'F limit.

P

{

t

)

l

.