ML20154R377
| ML20154R377 | |
| Person / Time | |
|---|---|
| Issue date: | 05/27/1988 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1670, NUDOCS 8806070289 | |
| Download: ML20154R377 (99) | |
Text
_ _ - - __ - - _ _ .._ _. _ _ _ . . . . . . .. ._
- C/SF /6 70 r'
. CRI3\A-W UNITED STATES l
NUCLEAR REGULATORY COMMISSION I
l
, s am m as am m em = =a m es = = = = es se s= = = = = am m en e ss am m am e = = = m m mm m e s= = m us e m e = = = ,
I l
l
' ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
)
i In the Matter of: )
! )
SUBCOMMITTEE ON THERMAL )
HVDRAULIC PHENCMENA )
(OPE SE l
O l /l O rm@f)a y~.+7Q fgqQ)t prt,np q /x 9.;c 3
u ~ . . a uq x l
0 dof Rem 0Ve from AChSp.. \ A DATE: May 27, 1988 PAGES: 1 through 56, 252 through 268 i
LOCATION: Washington, D.C.
= = = se m am en e m ma ss an e mm en m en sm es um sm as am m am aa i
HERITAGE REPORTING CORPORATION O owam 1229 L Seset, N.W., Sake 600 WanMagton, D.C. 20005 (102) 0 8806070289 800527 PDR ACRS T-1670 PDR
PUBLIC NOTICE BY THE G
1 2 UNITED STATES NUCLEAR REGULATORY COMMISSION'S 3 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 4
5 6
7 The contents of this stenographic transcript of the 8 proceedings of the United States Nuclear Regulatory 9 Commission's Adyisory Committee on Reactor Safeguards (ACRS),
10 as reported herein, is an uncorrected record of the discussions 11 recorded at the meeting held on the above date.
12 No member of the ACRS Staff and no participant at 13 this meeting accepts any responsibility for errors or 14 inaccuracies of statement or data contained in this transcript.
(~ )
15 16 17
, 18 l
19 20 j 21 l
l 22 23 l
24 i
l 25 O Heritage Reporting Corporation (202) 628-4888 i
1 W 1 UNITED STATES NUCLEAR REGULATORY COMMISSION 2 ADVISORY COMMMITTEE ON REACTOR SAFEGUARDS 3 )
In the Matter of: )
4 )
)
5 SUBCOMMITTEE ON THERMAL ) OPEN SESSION HYDRAULIC PHENOMENA )
6 )
7 Friday, May 27, 1988 8
Room 1046 9 1717 H Street, N.W.
Washington, D.C. 20555 10 The above-entitled matter came on for hearing, 11 pursuant to notice, at 8:00 a.m.
12 BEFORE: MR. DAVID A. WARD
( ') 13 Research Manager on Special Assignment
' ' ' E. I. du Pont de Nemours & Company 14 Savannah River Laboratory Aiken, South Carolina 15 ACRS MEMBERS PRESENT:
16 DR. WILLIAM KERR 17 Professor of Nuclear Engineering Director, Office of Energy Research 18 University of Michigan Ann Arbor, Michigan 19 MR. CHARLES J. WYLIE 20 Retired Chief Engineer Electrical Division 21 Duke Power Company Charlotte, North Carolina 22 CONSULTANTS:
23 I. Catton 24 M. Plesset V. Schrock 25 p
l 1
I HERITAGE REPORTING CORPORATION -- (202)628-4888 l
l 2
1 ACRS COGriIZANT STAFF MEMBER:
2 Paul Boehnert 3 fiRC STAFF PRESEtiTERS:
4 G. Hsii 5
6 7
8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 O
HERITAGE REPORTING CORPORATIO!; -- (202)628-4888
7---
3 .
1 PRQCEEDI!JG S.
( }
2 CHAIRMAN WARD: The meeting will now come to order.
3 This is the meeting of the Advisory Committee on Reactor 4 Safeguards, Subcommittee on Thermal Hydraulic Phenomena.
5 I am David Ward, the Subcommittee chairman. Other 6 ACRS members here are Mr. Wylie. We expect Mr. Kerr a little 7 later. And we have consultants Mr. Catton, Mr. Plesset, Mr.
8 Schrock, and I believe Mr. Sullivan will be here a little 9 later.
10 The purpose of the meeting is to review the revised 11 ECCS evaluation model developed by Westinghouse for two loop 12 upper plenum injection or UPI plants. Most of the meeting (s
'\J 13 will be closed to discuss the information being proprietary by 14 Westinghouse. Paul Boehnert is the cognizant ACRS staff 15 member for the meeting. .
I 16 Rules for participation in today's meeting have been 17 announced as part of the notice of the meeting previously 18 published in the Federal Register on May 24th of this year. A 19 transcript of the meeting is being kept, and we request that 20 each speaker first identify herself or himself, and speak with 21 sufficient clarity and volume so that she or he can be readily 22 heard.
23 We started the meeting a little early, eight 24 o' clock, because there is a lot to squeeze in, and especially O 25 because it being Friday, and most of us are out of towners, e
HERITAGE REPORTING CORPORATION -- (202)628-4888
4 1 We would like to leave the meeting, actually leave the room no lllh 2 later than 4:45 today, so I comment especially to the 3 consultants I think and members, we would like to try to keep 4 to the meeting agenda as closely as we can so we can cover 5 everything.
6 Let me make a couple of comments, address these to 7 the, to the Subcommittee. The reason we are here today is to 8 play a role in mending what has actually been a gap in the 9 fabric of regulation that has existed and been allowed to 10 exist for a number of years.
11 Six particular plants have an ECCS system 12 configuration which provide for low pressure injection in the
~'; 13 nozzles directly feeding the upper plenums of the reactor LJ 14 vessel, and this is in contrast to the most PWRs design where 15 liquid UPI feeds in the cold leg piping. This isn't 16 necessarily a problem, but we really don't know for sure. At 17 least we don't know officially, because the codes developed 18 for "CCS performance have, evaluation have not adequately 19 modeled the UPI configuration, and thus it has taken 20 considerable engineering juogment, perhaps a, somewhat of a 21 taith leap to conclude that these six plants are in compliance 22 with the ECCS rule, and even this has been an interim 23 conclusion.
/x 24 Why this situation has been allowed to exist for ten
( 1 LJ 25 or twelve years or whatever the length of time is, without HERITAGE REPORTING CORPORATION -- (202)628-4888
5 1 pressure from the NRC staff to correct it, you will have to lg 2 ask the staff. Why the situation has been allowed to exist 3 for the same period of time without the licensees for these 4 six plants taking an initiative to correct it, you will have 5 to ask them.
6 One reason might be that while there has certainly 7 been a hole in the fabric of regulation, why there is a hole 8 in the fabric of reactor safety is less clear. Over recent 9 years, we have come slowly to the perception that large break 10 LOCA has been overrated as a threat to reantor core integrity, 11 and perhaps more importantly, that facility has been overrated 12 even as a surrogate for more real threats to reactor core. At c'x 13 any rate, our job today is to review what the staff and
( )
w; 14 licensees and their agents have done to repair the gap and 15 come to some conclusions about whether we think that repairs 16 are appropriate.
17 And I want to call your attention to the fact that 18 while the licensees have chosen to apply the new freedom using 19 realistic or so-called best estimate method of analysis, this 20 isn't the central issue today. The central issue is whether 21 the UPI configuration, even if it were through an EH approach, 22 is now being properly modeled.
23 Would any of the members or consultants like to 7-24 comment at this time? Ivan?
N~] By making 25 MR. CATTON: I have a number of comments.
HERITAGE REPORTING CORPORATION -- (202)628-4888
6 1 them early on, they can address them during their 2 presentation.
3 CHAIRMAN WARD: Are these primarily for the staff or 4 for--
5 MR. CATTON: I guess it is both, because the staff 6 sort of likes it, and the staff is sort of caught up in all of 7 these things.
8 First, the business of UPI and having sat there for 9 12 years, whatever, I think both CCTF and UPTF point to the 10 original Westinghouse assumption as being a correct one. The 11 water just sort of runs right down into the bottom. So in my 12 view, ' hat issue was addressed when they did experiments.
. I
(~~S 13 have a number of questions or criticisms of the package that i
14 we have.
15 First, we didn't receive a QA document or any 16 document describing what is inside the code. I can find very 17 few equations at all in almost I guess three inches of paper.
l 18 Specifically, when he took a look at core 19 nodalization, with the LOFT they use four nodes. In the core 20 CCTF, they used six. In FLECHT they used 12. You probably 21 argue gee, it doesn't make any difference anyway. You 22 separate the blow-down peak f ron, the reflood peak, take a look 23 at it, you find that on the LOFT they underpredict the reflood 24 peak by about a 150 degrees, which that was four nodes, the
(
7-)3 25 CCTF, which had seven nodes, without the UPI they had a slight HERITAGE REPORTING CORPORATION -- (202)628-4888
1 underprediction, and for FLECHT ACRS THTF, it had 1.
lll 2 they over-predict. NRU, which had 12 nodes, they als - 'cere 3 it was sort of marginal, looked reasonably good.
4 It seems to me they have either got bad physics or 5 cad nodalization, and they either ought to do everything in 4, 6 everything at 6, or everything at 12 to meet the requirements 7 that were laid down by the TPG for establishing uncertainty of 8 a code.
9 Second, the COBRA / TRAC predicts top reflood. THTF 10 top reflood doesn't exist. That's kind of an interesting 11 process. One has to wonder where they get the water for the 12 top reflood.
13 When they looked at CCTF, the pressure distributions
(~~..
14 were terrible, to put it mildly. Further the results look 15 unstable. In some cases you see the oscillations growing as 16 they do their computations. To me that means that there 17 should be somewhere in all of this documentation a time step 18 study to tell me whether or not that's a case because then if 19 your answers look reasonable, you have unstable code, that's 20 not acceptable in my view.
l l 21 I could give some reasons why I think they i
l 22 underpredict in CCTF. We will wait and see what they have to l
t 23 do.
l 7- 24 Fourth, they have found that they have excess upper t ;
25 plenum de-entrainment. Now anybody who has looked the TRAC, HERITAGE REPORTING CORPORATION -- (202)628-4888
8 1 PF1 or at any of the Brookhaven code assessment reports, which llh 2 by the way, are not referenced at all in any of the 3 documentation, you know that that's due to the constant Weber 4 number, and if you de-entrain everything in the upper plenum, 5 you don't put anything into the steam generator.
6 In another part in the report they talk about too 7 much water in the steam generator. I don't know how to get 8 around that conflict. At least I can't reading it. They have 9 G tests in which were UHI. It seems to me that the physics of 10 the process says that those G-2 tests are irrelevant. You 11 don't have countercurrent flow with UPI. Just runs down in 12 one region of the core, comes up somewhere else. G-2 test r 13 puts it in the top. They took it out the top.
^}
14 I think they have to do an uncertainty evaluation 15 for the blow-down and the reflood peak separately. They put 16 all their numbers on to one chart, and this mixes things up, 17 so when you do reasonably well in blow-down, you take credit 18 for doing well in reflood, and I don't think that they should 19 be allowed to do that.
20 The bottom line is the staff bought off too easily.
21 CHAIRMAN WARD: Thank you.
22 MR. SCHRCCK: I share what Ivan bas said with regard 23 to concern abcut the documentation, and perhaps in a little 24 different way the process here.
(
() r 1 y 25 It seems to me that the staff has relied heavily on HERITAGE REPORTING CORPORATION -- (202)628-4888
9 1 the engineering judgment of the Sandia people that did the lllg 2 analysis for them. One of those people I believe was a major 3 developer of, of the COBRA code, and it is not at all 4 surprising that then to hear him say that the COBRA code is 5 capable of doing all these things.
6 I remember when we had the advanced code review 7 group still working, the reviews of the COBRA codes 8 performance for upper head injection analysis were still 9 really just in their early phases as the activity of the 10 advanced code review group wcs discontinued, and I don't think 11 that that code has ever had the same kind of scrutiny that we 12 have seen for versions of TRAC and RELAP 5.
(~'y 13 The language that is used in the Sandia report I J
14 find to be just questionable from the standpoint of good 15 engineering practice. There are many places in it where 16 phrases that are used are what I would consider to be 17 inappropriate. Conclusions are stated kind of throughout the 18 thing with regard to the fact that it appears that the spirit 19 of the SECY 472 is being met by this analysis, and statements 20 of that kind, so my reading of it is like Ivan's I guess. I 21 am just not comfortable with what I have read here, a good 22 quality documentation supporting this analysis of the behavior 23 of upper head injection.
7,.,
24 My other point has to do with the process of N_)
25 qualifying a computational procedure under SECY 472--in my HERITAGE REPORTING CORPORATION -- (202)628-4888
10 1 judgment, quite a deviation from what was originally intended llll 2 there, certainly as we have heard it in previous ACRS 3 subcommittee meetings. There is not an assessment of the 4 uncertainty in the best estimate code used at all. Instead 5 what you have is an engineering judgment superbound so-called 6 which is said to, on the basis of really simple calculations, 7 be more conservative than one, the 95 percent probability 8 uncertainty application to, to the best estimate calculation.
9 So while you say the key issue here, Dave, is to 10 look at whether or not the analysis of upper head injection is 11 or isn't being done adequately, at this point in time, it does 12 seem to me that there is another issue here that needs to be
,-'; 13 addressed. Maybe today isn't the time to do it, but that is
( !
t-14 how well is the new regulatory process being put to work?
15 As I see this particular presentation, there is some 16 rather serious questions in that area that ought to be 17 addressed, and of course I have always got to come back to my 18 pet peeve, and that is people don't read the ANS standards and 19 they don't do calculations in accordance with the ANS standard 20 in a straightforward way.
21 The documentation here that we have for how they 22 calculate the decay power which is certainly always the first 23 order influence on the results of any loss of coolant accident
,-3 24 is in fact not consistent with the ANS 1979 standard for the i 4 25 best estimate calculation. We can go into more detail on that HERITAGE REPORTING CORPORATION -- (202)628-4888
11 1 later if you like. You may consider it not appropriate for lllh 2 today's discussion either way.
3 CHAIRMAN WARD: Bill?
4 MR. PLESSET: I don't disagree uith my distinguished 5 colleagues here. You certainly are experts, but I don't know 6 why they are so sur'arised that COBRA / TRAC is not much good 7 because I expectr,d that, approach the discussion of it with 8 that already in mind. I think we have to look beyond that, 9 and in so doing, I'm fairly well impressed with what the staff 10 and the applicant have gotten. It does give a little ray of 11 sunshine.
12 CHAIRMAN WARD: Okay. Ivan?
'^') 13 MR. CATTON: Just to balance that, I think Bill is
,J L
14 probably right. The SER had a really strong caveat in it.
15 This is accepted only with the UPI plants, and if we only 16 accept the supermounting calculation, period, it would be 17 okay. Then you could--the rest of this is just you are upset
- 18 because they don't do it the way you want. That is not what 19 the SER says. The SER sort of accepts this as a large break 20 LOCA code, period. And I think that's where the concern comes 21 in.
12 CHAIRMAN WARD: Okay. Well, your points--I agree 23 these other plants are important, and we do want to try to 24 consider them today. I just didn't want them to completely b('s 25 overshadow the other issue. That's the reason I said that.
HERITAGE REPOI. TING CORPORATION -- (202)628-4888
12 1 Okay. Anything else? Let's go now to the rest of the agenda.
llll 2 And I believe Bob Jones is going to lead off for the staff.
3 MR. JONES: I just want to make a few opening 4 remarks from the table here.
5 First off, I hear your comments, and hopefully we 6 will addrese them, We'tinghouse, s during~the presentation. If 7 not, have out the experts and to make sure your questions are 8 answered.
9 Just in relationship to Mr. Catton's final comment 10 about whether this is just for JPI, the review of the model 11 was restricted only as a LOCA code to be applied to UPI 12 plants, because obviously a lot of thir, the uncertainty r~~ 3 13 evaluations, et cetera, would have to be repeated and
)
14 re-examined for application to just a normal LOCA analysis 15 although much of the work would probably be used for that 16 purpose, but it would have to be done, at least the 17 uncertainty part for sure, in application to another plant.
18 The reason we are here is simply this is essentially 19 the first application of a best estimate code to a PWR type 20 plant. This is a long-standing problem, as Mr. Ward has 21 already stated, which we are finally resolving for at least 22 four of the six plants. This review only covers the Kewaunee 23 Prairie Island and the Point Beach plant. Kewaunee and 24 Prairie Island are pursuing different model with Combustion 73 25 Engineering, whi!h the work is just starting on. We have some HERI". AGE REPORTING CORPORATION -- (202)628-4888
. _ . _ _ _ = _ _ . . _ -
c 13 1 serious questions in that review which we are trying to l 2 resolve right now.
3 We are not here to seek a letter ~from the committee.
4 It is simply-a briefing to bring you up to speed as to how we 5 have finally resolved this .ang-sta.1 ding issue. If you feel a 6 letter is appropriate, of course that is your prerogative, but 7 we are not here asking for one.
8 With that, I will turn it over to Jiom Hsii for an 9 introduction.
10 CHAIRMAN WARD: Bob, your first comment responds to 11 Dr. .Catton. Do you think that the SER clearly states that 12 this is a limited application of the--I mean you are limited 13 in the application of COBRA code to this particular--
14 MR. JONES: I have to go back and look at the words 15 again. I thought it did, because it does say it is the model 16 of two loop UPI plants. That was-the scope of the review and 17 that was the title of the report, so we fully intended it to 18 say that, and if it doesn't, we will go back and modify the 19 right or we will go back and look to make sure it says that.
20 CHAIRMAN WARD: All right. Thank you. Mr. Hsii?
21 MR. HSII: My name is Jim Hsii. I am the staff 22 member of the Operations Branch in NRR.
23 I just want to give brief introductory remarks on 24 the history of the concern with respect to two loop UPI ECCS 25 model, and give an overview of NRR safety evaluation on the <
HERITAGE REPORTING CORPORATION -- (202)628-4888
14 1 new Westinghouse ECCS model w.ith respect to UPI plant.
2 (Slide) 3 MR. HSII: I will give a very brief description as 4 to how the UPI configuration is.
5 The hot high-head injection pump is injected into 6 the cold leg and accumulator into the cold leg. The upper 7 plenum, the lowhead injection was injected into the upper 8 plenum at about 90 degree from the hot log. They are two 9 parts, and they are about the same elevation of the hot leg 10 and cold leg nozzle.
11 And the reason for this two split upper plenum 12 injection configuration is because the purpose of, in case f ; 13 there is a postulated large break LOCA in the cold leg, the L, '
14 single value will not defeat the purpose of the ECCS. That's 15 what, that's the reason they put that in the upper plenum.
16 (Slide) 17 MR. HSII: I will talk about the history of concern 18 with two loop UPI ECCS EM. The original model, there are two 19 models. We have rejected so far one 1975 ECCS model and the 20 other one is modified UPI ECCS model submitted in 1979.
21 (Slide) 22 MR. HSII: The original model was the 1975 approved 23 model. That model was approved for the cold leg injection.
- 24 In that application of UPI they treat the upper plenum 25 injection the same as the cold leg injecticn. Essentially HERITAGE REPORTING CORPORATION -- (202)628-4888
> 4 3 a nz - au- .a~ a.s sa-.~--w.i s .-sa a ..~-~~---..ss s m. can.u +- .- - u--
15 i 1 what they are doing is they are assuming the upper plenum 2 injected water was highly concentrated in.the-localized area 3 and fall'directly'into'the lower plenum resulting in 4E 4 interaction with the core or steam.
5 MR. CATTON: That is not a bad assumption, is'it, 6 based on what we know today?
7 MR. HSII: Some of them, that's-true, and but on the
- 8 other hand, there are some phenomena that is overlooked; for 9 instance, the entrainment phenomena.
- 10. MR. CATTON: Entrainment is other than where the 11 water is falling.
12 MR. HSII: Original model did not consider 13 entrainment. The evaporation--in the licensee's contention 14 that was a conservative model because they neglect the 15 interaction up to the core, was with the core, and the heat 16 removal and the injection water falling down, the staff start 17 getting concerned in 1976, '
77 area because of the data. More
- 18 research data was evailable at that time, so the staff start 19 looking at the effect of interaction and find out the model 20 was unrealistic, incorrect or non-conservative, because the 21 UPI water evaluation was due to the heat transfer of the core, 22 all the entrainment by the stes:m leaving the core, and 23 horizontal entrainment to the l'ot leg, all those would
< 24 increase the total way to the h.it leg, and with the heat l
l 0 25 transfer from the steam generator, would generate more steam, l-HERITAGE REPORTING CORPORATION -- (202)628-4888
16 1 increase the steam binding, and that will retard the bottom
) 2 reflood,-and that in turn will increase PCT, so the greatest 3 effect was not conservative, but the staff rejected this model 4 but the plant was allowed to continue operation based on the 5 justification provided by the licensees. They used the 6 interim model to take into account of, the adverse effect of 7 this UT! water. It is a boiler fix, so the staff allowed them 8 to continue operation but asked them to figure out a new.
9 model.
10 (Slide) 11 MR. HSII: In December of 1979, Westinghouse come in 11 2 with a modified ECCS model specifically to address the UPI 13 question. What it did is take the original EM, modify it, and
.14 they incorporate a UPI model in the W code. That is the code 15 they used to calculate the heat transfer phenomena, so 16 essentially what it do is try to incresce the hot leg flow by 17 the UPI water evaporation entrainment, take into account of 18 this effect, and calculate the increased steam binding, the 19 effect on the bottom reflood and on the PCT.
20 The staff reviewed this modified model and in 1985, 21 staff find out this model again was not acceptable, and in l
22 1975, the staff send later to the, describe the deficiencies 23 of the code; in addition to the reason of inadequate i
! 24 documentation, lack of sensitivity study, the other point of 25 the deficiencies, with respect to the code in that region that HERITAGE REPORTING CORPORATION -- (202)628-4888
17 1 UPI phenomena. For instance, the model does not have a water l 2 pooling in upper plenum. All the UPI water that is not lost, 3 it was not lost by the vapor or entrainment, assumed to fall 4 down to the lower plenum and on the way down they condense the 5 rising steam from the bottom lever.
6 The model also had another assumption. They assume 7 a fixed percentage of core was covered, covered by the upper 8 plenum injection water, and essentially the core was divided 9 into two regions. One is covered region. The other is 10 uncovered region. There are very simple heat transfer 11 mechanisms within the two regions--not within the two--in each 12 region. This transfer model is too simple, only address the
) 13 decay heat removal or the removal at different regions.
14 Depends on whether it is the quench region or, quench region 15 or the steam region.
16 MR. CATTON: How can you tell whether or not this is 17 simple or whether this is more complex? You can't see the 18 coding or the--or have you looked at the internals of this?
19 MR. HSII: I am talking about the previous model 20 that we rejected.
21 MR. CATTON: I understand, but it is not wrong just 22 because it is simple.
23 MR. HSII: Well, the reason we say it is simple,
,/^3 24 because the phenomena is not based on the mechanistic model
's J 25 that you could calculate heat transfer based on that, and it HERITAGE REPORTING CORPORATION -- (202)628-4888
18 1 just say if it is a, in the quench region, the heat removal (q,)
2 will be the, how many degree of, how many degree reduction in 3 the temperature. There are also assumptions of the top down 4 quench time to the mid-plenum. Assume it is linear function 5 of the UPI per--and also the KO fraction developed for the 6 cold leg injection model is used for the calculation of the 7 UPI plenum, and also the horizontal entrainment is assumed 8 fixed percentage of total UPI injection.
9 MR. JONES: If I can add something to Dr. Catton, 10 while these are very specific items we are talking about here, 11 when we did the review, and we had the CCTF data at that time, 12 they just appeared the model would not have been capable of 13 predicting the CCTF data and that was one of our concerns. It 14 still did not appear to have the appropriate physics. It 15 seemed to be grossly simplified. It wasn't really much of an 16 improvement over what we were using as the interim models, and 17 we just felt that the model needed to be upgraded to reflect 18 the appropriate phenomena. And that, while these are specific 19 line items and examples of what the problems were, these are 20 essentially the same problems we had with the 1977 model.
21 MR. CATTON: My only--the comment was directed at 22 the statement that somehow simple meant bad, and it doesn't.
23 What you have got here is a much more complicated code, and in 24 some respects might even be a little bit better, but there are 25 certain things that this code can't do any better than the old HERITAGE REPORTING CORPORATION -- (202)628-4888
19 1 one could, and somewhere you have got to put assumptions in,
,. en
( ,) 2 and if you talk to the Germans who were dealing with UPTF, you 3 will find this business, the area where the down flow is, that 4 can't be calculated very well by any of these codes, so what 5 do we have now? We have complicated COBRA / TRAC. It can't do 6 these kind of calculations any better, so where are you? You 7 have substituted complexity for simplicity and now it is not 8 even scrutable. So what the hell do you do with it? You just 9 accept it.
10 MR. HSII: The reason we say simple, maybe it is not 11 a good term. The simplified model means is not a, does not 12 have phenomena.
/~ 13 MR. CATTON: I understand. I understand where you L)'
14 are coming from. I just took that as an opportunity to 15 preach, but it got lost in the translation.
16 MR. HSII: In the 1985 data, the license, the staff 17 also request that the licensees provide the staff with plan to 18 resolve this issue, and to develop a new model, and also the 19 staff also give a guideline as to what are important phenomena 20 with respect to UPI plenum that are being looked for.
21 So in 1985, Northern State Power and Wisconsin 22 Electric Power Companies, those two are the licensees for 23 operation of the Prairie Island and Point Beach, they 24 contracted Westinghouse to develop a new UPI model, so the 25 Westinghouse approach is, to staff, use the best estimate HERITAGE REPORTING CORPORATION -- (202)628-4888
20 1 model. Use'WCOBRA/ TRAC as code, and also they decide to use lllh 2 the SECY 83-472 approach to deal with the problem, the EM 3 calculation. Those models are documented in WCAP 10924, 112.
4 i (Slide) 5 MR. HSII: I will give you overview of the NRR o ,.
evaluation on Westinghouse new EM. I will give a background 7 and talk about the scope and bound of review and the method 8 and result of review.
9 (Slide) 10 MR. HSII: I just want to put it in perspective what 11 SECY 83-472 approach is, and talk a little bit about the 12 COBRA / TRAC, and I understand that Westinghouse will provide 13 more detailed presentation on this, and also talk about the
}
14 basic review questions and talk about the documentation that 15 has been submitted by Westinghouse for staff review.
16 (Slide) 17 MR. HSII: SECY 83-472 approach is an approach 18 developed by the staff in 1983 to allow for the systematic 19 reduction of some conservatism in the existing, in the EM 20 model. Those conservatisms come from the Appendix K
- 21 requirement, come from the staff required conservatism for, to r 22 plan every model in the code, and also some conservatisms that S
23 were reserved by the venders to, in order to simplify their 1 24 model.
( %
'25 MR. CATTON: Will Westinghouse tell us exactly which i
HERITAGE REPORTING CORPORATION -- (202)628-4888
21 1 conservatisms are reduced and how?
() 2 MR. HSII: Well--
.3 MR. CATTON: I assume so?
4 MR.JHSII: They will give us,the overall approach to 5 the--the conservatism in every model is not incorporated in-6 there. It is the overall conservatism of the model.
7 MR. SCHROCK: Can you state for us how the staff 8 came to the conclusion that the use of an engineering judgment 9 superbound can logically replace the requirement of SECY'472 10 for doing an error assessment?-
11 MR. HSII: The SECY 83-472 approach to calculation 12 of determination of conservative PCT, that is at least 95 q .
13 percent, at least 95 percentile PCT, and we believe the 14 approach by Westinghouse using this so-called superbounded 15 calculation is, does provide at least 95 percent PCT because--
16 MR. SCHROCK: How do you form that judgment?
17 Failing to do the error assessment?
18 MR. JONES: Let me try to take a shot at that. We 19 did have as part of the overall review in Westinghouse model 20 specific sensitivity studies on many of the parameters which 21 were assumed to be at their bounds in the 95 percent 22 calculation. Some of these were new sensitivity studies.
23 Some of them were based on some old histvry from an 24 engineering judgment standpoint we felt would still apply or 25 would be appropriate.
HERITAGE REPORTING CORPORATION -- (202)628-4888
22 1- SECY 83-472 does require you to come up with at If 2 least a 95th percentile PCT. It is silent on.how you get 3 there. You can do it by doing a rigorous statistical 4 treatment of all the errors or we have been integrated in the 5 past in like the GE model review for their SAFR code where 6 some items were treated statistically, others were not.
7 Others were set at bounds.
8 And in this case, Westinghouse took an approach 9 where they said easentially--I won't say all, but the vast 10 majority of the parameters are set at the upper bounding and 11 we felt that that was sufficient to demonstrate the 95th 12 percentage.
13 MR. SCHROCK: How do you select an upper bound that
}
14 is equivalent or better than a 95 percent probability level 15 without an analysis to support that choice?
, 16 MR. JONES: In essence, what we were dealing with 17 primarily were input parameters to the code such as say stored 18 energy or even the decay heat, although that was treated 19 somewhat statistically in the 95th percentile. For che PCT it 20 was still set at a very high value.
21 By setting those values at what would be considered 22 to be bounding values in, essentially for the majority of the 23 input, we felt that that was in our opinion, guaranteed so to 24 speak in upper bound.
25 I would like to make one other comment on that, HERITAGE REPORTING CORPORATION -- (202)628-4888
~. - -
23 1 though. Westinghouse aid take their nominal calculations that A
( p) 2 they had done, and.they took some of the uncertainty or 3 sensitivity evaluations they had done to try to come up with .
4 what I would call a very crude estimate of the 95th percentile 5 PCT, and this was in an attempt to show that indeed the 6 superbounded -case was indeed at least a 95th percentile upper 7 bound and that is discussed in the SER where there are three 8 temperature range quota base on how they looked at coming up 9 with an estimate of 95th percent. We did not review that to 10 assure it was statistically perfect, just to assure it gave 11 reasonable answers as to wnat a likely 95th percentile PCT 12 would be, and as I remember the numbers, there is several 13 hundred degrees additional margin between the superbound and 14 the, these rough estimates of the 95th percentile, so that 15 provided us additional confidence that this is indeed an upper 16 bcund problem, and that is discussed in the document.
17 MR. SCHROCK: I hear what you are saying, but if I 18 could restate it in a simpler way, there would a very few very 19 simple calculations done that indicate that the upper bounds 20 that have been selected are indeed above the 95 percent level.
21 That's hardly evidence to establish that you know that.
22 That's the point that I am making.
23- MR. JONES: I do also want to make another comment 24 that although we have talked the 95th percentile number a lot, 25 that is not the absolute regulatory standard. It is, in HEP.ITAGE REPORTING CORPORATION -- (202)628-4888
24 1 essence it is a, you can argue it is a high probability or
() L2 whatever which'is undefined. If you look at the new ECCS 3 rule, you will see that's in there. We believe that this is 4 indeed a very high number, and as'such, we are comfortable 5 with it from the standpoint of providing reasonable assurance.
6 MR. SCHROCK: What bothers ne about this is that the 7 regulatory process has been improved to permit a good 8 scientific engineering basis, and now what we are doing is to 9 turn around and say we are incapable of doing this problem on 10 a good sound scientific engineering basis and therefore, our 11 fallback position, it is to do what we have always done, and 12 that is to say that we know the answers and we have various 13 ways of knowing the answer, but we can't explain this in a 14 scientific engineering way. That's where we are I'm afraid.
15 MR. HSII: We believe the approach Westinghouse 16 taken is conservative because the way they apply the 17 uncertainties using deterministic method, they put all the 18 uncertainty at their adverse limit and input it in at the same 19 time so all the effects are considering the same time rather 20 than the statistical approach that you would use.
21 MR. SCHROCK: That's the mentality that produced 22 Appendix K.
23 MR. CATTON: That's right.
24 MR. HSII: I think our objective is to see the 25 overall model whether it is conservative or not.
HERITAGE REPORTING CORPORATION -- (202)628-4888
.i 25 1 MR. CATTON: You had an overall model before. It lh 2 was called EM model. In many respects I don't view this as 3 being much.different. When you look at the--best estimate 4 means you take your best shot. When you look at the UPI flow 5 or the way they thought it.would be, it doesn't look like what 6 we have observed in CCTF or UPTF, so how can you put a bound 7 of any kind on that?
8 I am not disagreeing with the fact that that 9 superbound is indeed a superbound, but that's, again that's my 10 judgment. I can't put my finger on anything that would prove 11 that I'm right, nor can you, so you are not doing anything any 12 different. You are dealing with just a different EM model.
13 MR. HSII: We are to use EM conservative model oc
.}
14 best estimate model, the best estimate model, and use the 15 statistical approach. It is up to the licensee to decide 16 because they may consider the cost of doing an analysis, they 17 would rather go to the more conservative approach to--
18 MR. CATTON: I hear you, but when I read your SER, 19 almost every other sentence says best estimate. This is a 20 great code, it calculates all important phenomena, and that's 21 nonsense. It doesn't do that. You are saying it based on the 22 superbound, and you should say so.
L3 MR. HSII: When it is best estimate, we mean the 24 code is realistic code that can calculate important phenomena 25 in the code rather than putting some priority assumptions.
HERITAGE REPORTING CORPORATION -- (202)628-4888
~ . . - -. - - . . . . . - . . . . .. - .
26' 1 MR. JONES: Let me try_to state that slightly--
-,A
( ,I 2- MR. CATTON: Let me just answer that one sentence.
.3 .MR._ JONES: I was going.to try_to answer your 4 statement'.
5 -MR. CATTON: You have put assumptions into it. You 6 have not_let the code do the job. I mean not you, but 7- Westinghouse, has not let the code do the job. Westinghouse 8 has put a whole lot of assumptions in there in order to get 9 the job done, so that violates what you are saying.
10 MR. JONES: If I could just--how did we accept the 11- model? I think your premise is we have accepted it on the 12 superbound calculation only.
13 We are stating that we believe that the COBRA / TRAC
}
14 model does provide a reasonable best estimate calculation of 15 the important phenomena. Where we end up then is we think the
-16 physics are basically modeled. Then the next problem is how 17 do you select the specific plant input? And it is in that 18 area, in reflecting the plant, that the vast majority of the 19 questions are, the conservatisms are showing up in the 20 superbound calculation in the input reflecting the plant.
21 It could have been treated in two ways. It could 22 have been done on pure statistical approach, or in this case, 23 Westinghouse decided to skew it in the superbounded 24 calculation, and we believe that does bound what you would 25 have gotten statistically, but we do believe the code with the HERITAGE REPORTING CORPORATION -- (202)628-4888
27 1 biases we have obtained from the experimental benchmarking of
( ). 2 the code, does demonstrate or provide a reasonable best 3 estimate calculation of the important phenomena.
1 4 MR. CATTON: You could not have done it on a purely 5- statistical basis. I don't believe that. And I don't see a 6- single reference where in any of the documentation either by 7 the staff or by Sandia or by Westinghouse that refers to the 8 work that was done at Brookhaven on code assessment for TRAC 9 PD2. There is nothing. And I recollect lots of difficulties 10- with TRAC PD2.
11 MR. HSII: The Westinghouse does reference the 12 assessment work done by the, either the development assessment 13 or the independent assessment by other people on COBRA / TRAC O 14 and that is, the COBRA / TRAC is the code that, based 15 on--COBRA / TRAC is the combination of--
16 MR. CATTON: I know what COBRA / TRAC is, when you 17 look at the pressure distributions that were calculated for 18 CCTF, frankly they are terrible. All right? They are just, 19 they are just no good at all. Things run off in the wrong 20 directions, and not by small amounts, by large enough values 21 that could drive a lot of, flood a lot of places. This is the 22 code that COBRA fits in. I mean your COBRA is the core. The 23 rest of the systems doesn't seem to me it is being handled all i 24 that we well with a IRAC PD2 yet this is not addressed at all, O 25 and certainly your contractor, Sandia, could have looked to HERITAGE REPORTING CORPORATION -- (202)628-4888
28 2 1 those-reports and'said something about them. Either they were
./A.
( ,) 2- -wrong, they made a mistake, and it is really a lot better than:
3 they said or som.ething. You have got a long history for TRAC 4 PD2. Also got a long history for COBRA, some of which 5 Professor Schrock mentioned, and none of these things are 6 brought in at all. It-is~almost as if we are starting from.
7 ground zero with the work that Westinghouse did, and that is
. 8 not true.
9 MR. HSII: I think Sandia review was looking at the, 10 also the COBRA / TRAC. I'm not sure if--John, can you answer 11 that question? Have you looked at TRAC PD2?
12 MR. KELLY: Dr. Catton is right in a couple of 13 respects. We did look at WCOBRA/ TRAC independently of all the 14 previous history because.there were a number of modifications 15 made to the code and they did not rely explicitly on the "
16 previous assessment validation work, so we, they did refer to l
i 17 that, though, in the document. I believe that's Section 5 of l 18 Volume 1 or Section 4 of Volume 1.
19 MR. HSII: Section 4.
20 MR. KELLY: As far as our assessment, we are looking j 21 at the specific code validation exercise that they carried out 22 as part of the submittal.
l- 23 MR. HSII: Westinghouse have identified in Section l
24 4, Volume 1, the assessment work done by others, and that 25 covers a lot of phenomena, i
i I
l HERITAGE REPORTING CORPORATION -- (202)628-4888
_ _ _ . ~ _ . _ .-. . -. __ .-_
29 1 MR. CATTON: The only reference there is in Chapter l 2 4 is a report, COBRA / TRAC thermal, hydraulics code for-3 transient-analysis, NUREG CR 3046, Volumes 1 through 5, so you 4 -are mistaken, or your contractor is mistaken in his: reference.
5 MR. YOUNG: There are references to work done by 6 Brookhaven in other sections of the report and we.can identify 7 those for you later if you want. There are appendices where 8 we looked at specific areas such as break flow uncertainty and 9 so on.
10 CHAIRMAN WARD: Was that the assessment of TRAC PD2 11 that Dr. Catton was talking about?
12 MR. YOUNG: Yes.
13 MR. HSII: I think when Westinghouse make 14 presentation, a lot of questions may be answered at that time.
15 We are talking about the SECY 83-472 approach. The 16 key element in SECY 83-472 approach is you have got to have a 17 best estimate code capable of calculating the important 18 phenomena, and also capability of, to assess the uncertainties 19 associated with PCT calculation, so the end result is that the 20 PCT calculated by the code plus the uncertainty will provide 21 at least 95 percent, at least 95 percent PCT.
22 The accident criteria for SECY 83-472 approach is i
23 that the way to do it is the licensee would be required to l
24 calculate a conservative PCT, that is at least 95th percentile l
25 PCT, and they also have to calculate the Appendix K PCT. That l
HERITAGE REPORTING CORPORATION -- (202)628-4888
30 1 is PCT calculated with the best estimate code put in nominal value without considering uncertainties, plus the code will be lllh 2 3 augmented by the Appendix K requirement. The resulting 4 Appendix K PCT is required to be greater than 95th percentile 5 PCT and less than 220 degree.
6 (Slide) 7 MR. HSII: I will talk about WCOBRA/ TRAC. I think 8 Westinghouse is going to be, they will talk about it in more 9 detail.
10 WCOBRA/ TRAC is Westinghouse version of COBRA / TRAC 11 with Westinghouse modifications, those modifications to 12 improve some of the, some part of the code and also to, with r'; 13 modification to specifically handle the problem with respect C/
14 to UPI problem, and also some of the nodifications to comply 15 with Appendix K requirement.
16 The code does not have a priority assumption like 17 the previous model have, assuming some percentage of core is 18 covered by the UPI water. The code is mechanistic 19 phenomenological model that is capable of calculating 20 important UPI phenomena.
21 This phenomena include water pooling in upper 22 plenum, upper core spray, top down quench, heat transfer to 23 UPI, water in upper plenum, diffusion, steam binding on core, 24 heat transfer and steam generation, entrainment, the 7~
- I
,j 25 entrainment, non-equilibrium heat and transfer.
HERITAGE REPORTING CORPORATION -- (202)628-4888
31 1 (Slide)
~
a ] ,g/- 2 MR. HSII: Our basic review questions are geared f-
! 3 toward SECY 83-472 approach. We went on to make sure the code 4 WCOBRA/ TRAC is capable of calculating important phenomena-5 dealing in large break LOCA in the UPI ECCS plenum, and we are 6 to ensure whether--
7 MR. PLESSET: Could you raise that up a little bit?
8 MR. HSII: We are going--the quees..on is whether the 9 uncer';ainties associated with PCT calculation can be 10 adequaraly accounted for. In other words, when the PCT is 11 called by the COBRA / TRAC code plus uncertainties, does it 12 provide a 95, 95th percentile, at least 95th PCT? The 13 Westinghouse is so-called superbounded PCT, whether this PCT 14 is at least a 95th percentile PCT. The other question is 15 whether the Appendix K are complied with.
16 (Slide) 17 MR. HSII: The documentation provided by 18 Westinghouse in the staff review--I don't want to get into the 19 detail of this list, but basically we have got two volumes, 20 and the response to the questions, and there is a Volume 2, 21 Division 1 that is. That supersedes the original volume 2.
22 MR. CATTON: Your bottom line on those three 23 questions that you had on the previous slide is that indeed 24 they did do those things, right? ,
25 MR. HSII: Yes.
HERITAGE REPORTING CORPORATION -- (202)628-4888
= . - . . . - , - ..- - -. - . - - . . - . - . - . . . - _ . . . . - - _ - . . -
i 32 1 MR. CATTON: I'm still confused as to how in the
. n.
( ,) 2 face of what is in Volume 1 you were able to conclude that.
3 MR. HSII: Volume 1 or Volume 2?
4 MR. CATTON: I am talking about Volume 1 where they 5 show the pressure traces as compared to the experimental 6 pressure traces, and they are so widely divergent, how could 7 you conclude anything? You obviously have something that you 8 are basing your judgment on. Could you share that with us?
9 MR. HSII: Well, we are looking at the overall 10 product. The overall product showed the code indeed provide 11 conservative prediction of 95th percentile peak.
12 MR. CATTON: That sounds like a Michigan car with a
(T 13 good paint job. It could be all rust underneath.
G 14 MR. JONES: We will have our consultant address that 15 one when he makes that presentation.
16 MR. CATTON: Good. I looked through the list. When 17 is he going to talk?
18 HR. JONES: I believe it is early afterncon where 19 he--late afternoon when he does the audit calculation, but we 20 will expand that in a few more details.
21 MR. CATTON: I see it--3:20 p.m..
, 22 MR. JONES: Also challenge Westinghouse, too, when i
23 they get up.
24 MR. HSII: Okay. I will talk about the scope and g w-
\-
25 bound of our review.
l l
HERITAGE REPORTING CORPORATION -- (202)628-4888
33 1 MR. CATTON: Could I just ask one more question?
t,g ); 2 Then I will stop.
v.
3 I asked the question early on about separating the 4 peaks, blow-down from the reflood. You must have given this 5 some consideration.
6 What was your base for lumping all of them together?
7 MR. HSII: We thought about that, too, and--
8 MR. CATTON: I find no mention of this difference 9 anywhere in any of the reports.
10 MR. HSII: We thought about that. The conclusion 11 we had was we think a lot of phenomena, they have same 12 physics.
/') 13 John, do you want to answer the question?
(/_
14 MR. CATTON: I can't hear you. Could you speak up a 15 little?
16 MR. KELLY: John Kelly from Sandia--we did question 17 that, why weren't reflood and blow-down peaks considered 18 separately? Westinghouse response to that was that the basic 19 or the local heat transfer mechanisms, that the fuel rod are 20 local flow conditions and--local rod conditions, and the same 21 correlations are used independently of whether the code is 22 reflood or blow-down phase. Therefore, we are not looking 23 specifically at what we call a blow-down peak or reflood peak.
r3i 24 We are looking at the heat transfer mechanisms whether they 25 were appropriate to the particular flow regime.
HERITAGE REPORTING CORPORATION -- (202)628-4888
1 34 1 MR. CATTON: One of the things that came out of this
(,-)
2 attempt to try to establish the uncertainty on the TRAC _ code 3 was that that view was a little too elementary, that blow-down 4 you could argue very well will scale independent, and I think 5 that they did tha. very successfully and that the blow-down 6 peak is fairly easy to come by and you could get, you can get 7 at it pretty quickly, but the reflood peak was a different 8 story entirely. There were a lot of different mechanisms that 9 played a role, and so you could have one value for the 10 blow-down peak and an entirely different value for the reflood 11 peak.
12 And another reason is time. Even if I get the heat 13 transfer coefficient reasonably well, I have got a lot more
}
14 time for it to run me off on a different track. So that 15 argument is not adequate to rationalize the peaks are the 16 same. That's a conclusion, by the way, of the effort to find 17 uncertainties of TRAC PF1.
18 DR. KERR: Excuse me. I want to understand. You 19 seem to be saying that the peaks are the same. I didn't--
20 HR. CATTON: No. I hope not. There is a, you can 21 get--
22 DR. KERR: You said that's the reason the peaks are 23 not the same. Are you saying that report indicates the peaks l
( 24 are the same, or just the same method is used to calculate
)
25 both of them?
HERITAGE REPORTING CORPORATION -- (202)628-4888
35 1 MR. CATTON: They treat them as if they are the
() 2 same. They take, they take the LOFT experiment which peak 3 during blow-down, and they take another case like FLECH or 4 CCTF which is just reflood. They look at the delta Ts 5 associated with and plot them on the same curve. You can't do 6 that. You should plot two curves, and depending on whether 7 you peak at blow-down or reflood, you use the appropriate one.
8 You cannot combine them. There is too many more mechanisms 9 involved in one than the other.
10 MR. KELLY: Again, the code looks at the situation 11 on a local basis, and my contention will be that the local 12 mechanisms is what is governing the heat transfer and that is 13 is what is being assessed. It is not looking at the overall
[v')
14 thing and saying this is blow-down peak, that is reflood peak 15 and we are determining the heat transfer mechanisms and that's 16 whet leads to the peak. The integral type of parameter we are 17 looking at is assessing the local parameters.
18 MR. CATTON: But you are plotting all of these delta 19 Ts on the same ground. One occurs earlier and one occurs 20 late. And the one that is late is going to be more impacted 21 than the one that occurs early. You have got to plot them on 22 separate graphs. You can't treat CCTF reflood, FLECH reflood, 23 NRU reflood the same way that you treat LOFT blow-down. You 24 can't do that.
25 HR. SCHROCK: In the non-UHI plant, at the peak HERITAGE REPORTING CORPORATION -- (202)628-4888
y e 36 ,
1 train blow-down.
( ) 2 CHAIRMAN WARD: You mean non-UHI?
3 MR. SCHROCK: Non-UHI, the peak during the blow-down 4 is in many cases the highest peak, and in fact, it has nothing 5 to do with ECC systems. It has to do with the inherent 6 thermal hydraulic characteristics of the cold system. Now it 7 seems to me that an important question to be examined here is 8 is that still true for the UHI plant? Is the peak--
9 CHAIRMAN WARD: You mean UPI.
10 MR. SCHROCK: UPI; what did I say? I'm sorry. Is 11 the peak during blow-down in these upper head injection plants 12 caused by the injection of ECC water, or is it simply the
/' 13 consequence of the inherent thermal hydraulic characteristics
(_)T 14 of the plant? Do you know the answer to that?
15 MR. KELLY: I do. You would look at a realistic 16 case, which was the 50 percentile number, in fact blow-down 17 peak is ADT when you apply the many conservative assumptions L 18 on behavior of ECCS system. Then in the bounded cases you, we l
19 get reflood, but if you look at a realistic case which is, I 20 guess is, experiments are considered realistic cases, you will 21 get a blow-down peak, and that's what the code does.
22 MR. SCHROCK: It is an interesting answer, but it is 23 to a different question.
24 The question that I asked is does the peak during 25 blow-down occur as a consequence of injection of ECC water, or HERITAGE REPORTING CORPORATION -- (202)628-4888
. - . - - _ - _ . = . .
37 1 does'it occur as a-consequence of the inherent thermal
() 2 hydraulic characteristics of the system?
you didn't have any ECC system?
Would it occur if I think the_ answer is yes.
3 4 MR. KELLY: Yes.
5 MR. SCHROCK: So what has all the argument about the 6 behavior of the UPI system got to do with the determination of ,
7 that blow-down peak? Nothing.
8 MR. CATTON: That's right.
9 MR. SCHROCK: My point is you have got to separate 10 these things. What we are talking about here is the, ic this 11 UPI system being adequately analyzed? Why confuse the issue 12 by muddying up comparisons of peak clad temperatures in time 13 domains where in fact the UPI system isn't even active with 14 peak clad temper 6tures that occur after the UPI water 15 injection begins? Aren't you just automatically confusing 16 yourself when you do that?
17 HR. HSII: The blow-down phase affects that later 18 calculation in the reflood phase, too.
19 MR. SCHROCK: I'm sorry?
20 MR. HSII: The result of blow-down phase will affect 21 the result of the reflood phase.
22 MR. CATTON: You shouldn't be, I think what 23 Professor Schrock is saying, you should not be weighting in i
24 the deviations associated with the blow-down peak into the 25 same batch of deviations for the reflood peak, yet you do.
HERITAGE REPORTING CORPORATION -- (202)628-4888
, , ,_ 4 ,. _ _ _ . _. . -.
l
-38 1 YouLare confusing.the two parts of~the problem.
ll h 2 MR. HSII: Well, we- -
- 3. MR. JONES: Let me make a comment on that. I think 4 we' understand your concern. We are obviously not going to be 5 able to totally address it here, at least even try to separate 6 'them today. Let.us go back and take a look at'it a little bit 7 and see whether or not we think the approach used--I think we 8 stand behind the approach we used and we understand your 9 concerns and we will go back and we will take a look. We will 10 see if we can separate what the analyses say for blow-down and 11 the reflood peak, see what kind of uncertainties that might 12 give us, and see whether the approach we used was reasonable 13 or at least there is sufficient margin in the superbounded to
}
14 cover that difference if we are indeed in error with the way 15 we approach the problem, but I understand your concern. We 1
. 16 will go back and take a look.
17 MR. CATTON: Just take a look where they show the 18 reflood peak for LOFT. Forget the initial blow-down peak and E19 you will see it is underpredicted by a hundred to 150 degrees. ;
, 20 That's for LOFT. What am I going to judge for reactor? I L want an adder of 150 degrees. Either that or you have to do 21 I
i- 22 the physics right, and the physics aren't done right.
23 MR. JONES: We understand your concern. We will go ;
l' 24 back and look at it and see what, if any, modifications need i
! 25 to be made. We will address it somehow in the final SER.
I HERITAGE REPORTING CORPORATION -- (202)628-4888 -
39 1 CHAIRMAN WARD: I guess that's good. I am a'little 4 2 -surprised. This seemed like kind of a fundamental question.
,,/
3 I'm not sure why the present SER doesn't deal with it.
4- MR. JONES: I think that the conclusion we came to 5 is the approach we used was sufficient approach. We 6 can--right now, to some extent, it is difficult to address 7 your comments without getting and showing some of the specific 8 details that form the foundation for our conclusion.
9 DR. KERR: Why not give him a chance to have another 10 look? I made a mistake once. I bet you did.
11 CHAIRMAN WARD: I will allow for that possibility!
12 I guess I am a little bothered by I guess Mr. Kelly's response 13 here. He seems to be still talking evaluation model language.
}
14 If I understand what you said, it is that if, if you make a 15 best estimate model analysis, clearly the blow-down peak is 16 larger than the reflood peak, but if you then want to look at 17 the uncertainties involved, you come to the conclusion that 18 the uncertainties on the reflood peak are so large that when i
19 you add an adder, the reflood peak may become larger than the l
20 blow-down peak. Is that the--
i 21 MR. KELLY: I think that's right. It is due to 22 facts like assume that is degraded, sa fety injection that is 23 degraded, for instance.
24 A key one, of course, you are also much higher decay 25 heat level. You are about 50 percent.
HERITAGE REPOR11NG CORPORATION -- (202)628-4888
40 1 MR. CATTON: I don't think you are answering the lllh 2 question.
3 MR. KELLY: I am trying to explain why, why the 4 boundary calculations give you a higher--
5 MR. CATTON: Different patterns for blow-down and 6 reflood; you are not using different patterns.
7 CHAIRMAN WARD: I mean as I understood what--it was 8 stated that in a best estimate calculation, the blow-down peak 9 is consistently greater than the reflood peak.
10 MR. KELLY: In a realistic case when you do bounded 11 case, you degrade the system performance, the behavior turning 12 areund. It is due to the physics. It is noc due to any kind 7
, 13 of add, fission adder. You degraded the system, and so the
( _/
14 performance of the cooling system isn't an good. You are in 15 much higher decay heat level, and these two things tend to 15 drive the reflood higher.
17 MR. SCHROCK: It also drives the plenum heat flood.
18 MR. KELLY: That is true, but in a relative sense.
19 The reflood peak will become higher. It is not due to--it is i
20 due to the ad hoc input assumption. It is net due to any 21 modifications to the code or any artificial adders that have 22 been put in.
23 MR. CATTCN: You are adjuuting the plant conditions.
s 24 You are sort of changing the envelope that the thermal LJ 25 hydraulics is working within, and that results in a higher l
HERITAGE REPORTING CORPORATION -- (202)628-4888
41 1 reflood prak. Is that what you are saying?
(. ,, 2 tiR . HOCha6tfER: You are changing--this is Harry 3 Hochreiter of Westinghouse. We will be going through in our 4 presentation u calculation called bounded calculations and 5 superbounded calculations, and what you will see is how we 6 have changed some of the parameters that go into the code.
7 The code uses initial or boundary conditions, and in many 8 cases, those parameters will affect a reflood condition more 9 than they will affect the blow-dowc.. When we go through our 10 presentation, I think you will see that.
11 MR. CATTON: So then the question is--
12 CHAIRMAN WARD: But you are taking this approach
~
/ 13 instead of an approach where you more explicitly calculate an C)/
14 adder. You are using the, I mean the sur iaound approach is 15 different from--
16 MR. HOCHREITER: We are taking the uncertainties 17 together and put putting them into one calculation, and we 18 have used sensitivity studies as a basis to determine what the 19 effects are of a given parameter and we have taken a more 20 conservative approach by putting them all together into one 21 calculation on certain boundaries, but that is just an 22 approach.
23 I mean we could have used another approach if we 24 wanted to where we did a statistical design of experiments
('s]/
u 25 type study uhere you did a number of calculations looking at HERITAGE REPORTING CORPORATION -- (202)628-4888
_= ._ . . ._
42 1 things one at a time, two at a time, or in combination, and ,
(m ) 2 then developed a response to calculate the 95th percentile 3 PCT. We felt that for this particular application, and it 4 being the first application, that perhaps a superbounded 5 calculation would be an easier thing to go in with and get 6 licensed.
7 CHAIRMAN WARD: Now if you had taken the other 8 approach, and you come up with a statistically based or 9 estimated adder, if your present approach is revealing 10 something approximating the truth, then this other approach 11 would have given you an adder for the reflood peak which is 12 substantially greater than the adder for the blow-down peak.
g' 13 HR. HOCHREITER: I think that is true because what (s
14 you see with the reflood peak is that, you kn<w, as I think 15 Dr. Catton pointed out, you are looking at a time effect as 16 well, so when you include uncertainty like the the decay heat, 17 uncertainty, it is going to have small effect on the blow-down 18 peak because you are looking primarily at stored energy. When 19 you look at reflood, depending upon the length of transient, 20 you have a larger effect. When you degrade the 87 system, it 21 will not affect the blow-down peak, but it will affect the 12 reflood peak, so the uncertainties that you include will 23 affect one peak more than the other, and in general, time is 24 penalizing you when you look at the reflood peak. When you (V
25 combine these together, that tends to drive the reflood peak HERITAGE REPORTING CORPORATION -- (202)628-4888
43 1 up relative to the blow-down peak, but we will be going llll 2 through the logic that we used, and how we developed these 3 conditions for these calculations.
4 CHAIRMAN WARD: Okay. Thank you. Where were you?
5 MR. HSII: We were talking about the scope and bound 6 of our review. I just want to go through these items that we 7 reviewed, we contracted. We reviewed COBRA / TRAC and 8 modification--
9 MR. CATTON: When you looked at the modifications, 10 those are modifications from COBRA / TRAC. Now when we looked 11 at TRAC PF1, we find that there is huge differences between 12 the early manuals and what was in the code at any given time:
(~) 13 so when you say modifications, those are modifications from
~ . -J 14 what snd how did you know what it was?
15 MR. HSII: Westinghouse kept track of what version 16 they adopt from COBRA / TRAC. I think it is cycle 15 or 17 something.
18 MR. CATTON: There was no QA document of that cycle 19 13, so how did you know what your initial conditions were?
20 MR. HSII: We look at the report in NUREG CR 3046.
21 That is a COBRA / TRAC documentation.
22 MR. CATTON: Is that documentation adequately, is it 23 anything at all like the TRAC correlations and models document 7, 24 or the RELAP 5 correlations and models document?
0] 25 MR. HSII: For that document five, for our five HERITAGE REPORTING CORPORATION -- (202)628-4888
44 1 volume, five volume, the--
lllh 2 MR. CATTON: I have not seen those documents. I am 3 just asking a question.
4 MR. HSII: Yes.
5 HR. KSLLY: They are not exactly the same, but they 6 are comparable. They state the basic conservation equations 7 that are solved. They show the methods for the relationships 8 and they describe the methods of solution, so they are 9 comparable.
10 MR. ZUBER: They are not comparable. Just for the 11 record, they are comparable to the extent that manuals of 12 TRAC, the TRAC PD2, they are, have no relations to, similarity
('
L_;'
) 13 with the QA document.
14 MR. CATTON: If I were to read those documents, 15 would I--
16 MR. ZUBER: You will drau the same conclusion as you 17 would draw by PD2, I mean it may be--relates to what? The 18 code it raay not , depending on at the point in cine or the 19 development.
20 MR. CATTON: I conclude you did not know the initial 21 conditions and just having a list of modifications made by 22 Westinghouse is insufficient? You need a QA document or you 23 can't make the reasonableness test of the code.
, . . 24 MR. HSII: Well, Westinghouse realized in Appendix A 25 to--
HERITAGE REPORTING CORPORATION -- (202)628-4888
45 1 MR. YOUNG: We will cover this in a little bit more k 2 detail later, but the fact is that we reviewed the code in 3 essence line by line to assure ourselves that the 4 documentation reported in the NUREG is accurate. The 5 equations there do in fact appear in the version of the code 6 that we picked up.
7 MR. CATTON: You were very fortunate. That hasn't 8 been the case for the recent coae until the QA document.
9 MR. SCHROCK: When you say that, you mean all 10 equations are as documented?
11 MR. YOUNG: As documented; there were one or two 12 instances where a coefficient was slightly different, and I 13 think we so stated in our documentation which appears in 14 Volume 1 of the appendix.
15 CHAIRMAN WARD: Okay. Let's see. I think we better 16 let Gene finish, and get on with the Westinghouse presentation 17 because I think some of these same questions will come up.
18 MR. HSII: We also reviewed a modification to the 19 code. Those modifications are individual in Appendix A to the 20 Volume 1. Of course, the Appendix A to Volume 1 is very brief 21 description as to what the modifications are, and there are 22 some important models. We think we should know more. We are 23 asking the question, we get response from them to help give us 24 more detailed description of what the model is.
25 MR. CATTON: In your uncertainty qualification and HERITAGE REPORTING CORPORATION -- (202)628-4888
46 1 methodology, there are counterpart experiments between CCTF, llll 2 and FLECH, and scaling is one of the big questions in this 3 whole business.
4 Why didn't you require that they do some counterpart 5 calculations to demonstrate the efficacy of their code?
6 MR. HSII: We did talk about scaling effect, 7 uncertainty associated.
8 MR. CATTON: You talked, but you could have done 9 more than talk. You could have required them to do certain 10 calculations that would demonstrate it, and I want to know 11 what basis did you have for not doing that?
12 MR. HSII: We thought--okay. We did, we just c'~ T 13 thought about this as scaling uncertainties, and we are 7
)
14 talking to Westinghouse about this issue, too, but the 15 Westinghouse has done the scaling since--scaling started by 16 putting the measure to predict the difference in PCT power 17 scale, and they found out the effect of scaling is not a big 18 effect.
19 MR. CATTON: Well, it certainly isn't for blow-down 20 but it could well be for reflood, couldn't it?
21 MR. HSII: Overall, for different scale, that I 22 don't recall whether it is important or, it is overall effect, 23 whether it is important or--
73 24 MR. CATTON: Yes, because you mixed peaks in that
(,)
25 plot or Westinghouse mixed the peaks in that plot and they HERITAGE REPORTING CORPORATION -- (202)628-4888
47 1 have LOFT and it is a blow-down peak and you have FLECH and it n
I, ,) 2 is a reflood peak and you put it on the same graph.
3 MR. HSII: Again, we just thought about, we didn't 4 think of whether it is appropriate to mix the blow-down peak 5 and other peak.
6 MR. CATTON: We just heard the contractor say that 7 it wasn't appropriate to mix them because there was a time 8 factor involved, yet you accepted--
9 MR. HOCHREITER: Excuse me. I didn't say that, not 10 for the uncertainty.
11 MR. CATTON: We will argue about it when you get up.
12 MR. HOCHREITER: I'm sure we will.
13 MR. SCHROCK: Before the staff finishes their
()
v 14 presentation, I would like to ask them about the upper or down 15 calculation with regard to the decay power. According to the 16 documentation, the 1979 ANS standard was used as a part of 17 this upper bound cal:ulation, is that correct?
18 MR. HSII: Yes.
19 MR. SCHROCK: Now my question is this--do you 20 understand how they used the 1979 standard to get the decay 21 power for the upper bound calculation?
22 MR. HSII: I didn't get into that detail. Maybe our 23 contractor--
24 DR. KERR: I think the answer to your question is 25 no.
HERITAGE REPORTING CORPORATION -- (202)628-4888
~,.
48 h .
N -!-
MR. SCHROCK: Let me point out to you that the staff 1
went through a great deal of evaluation on the GSAR system'as lllh 2 3 to how it would be acceptable to use the 1979 ANS standard to 4 do the statistical analysis, part of the SECY 472. That's not 5 a trivial problem. In fact, in this application, from the 6 documentation, it appears that there are twozvery important 7 missing elements in that assessment, and the first or.e is that 8 in the power-plant, fissioning is a consequence of presence of 9 several different fissionable nuclides, and they each have 10 different decay power functions sufficiently different at some 11 time as they influence the large break LOCA.
s 12 The documentation of the COBRA / TRAC code shows that 13 in fact the only information in COBRA / TRAC is the ANS standard
)
~
14 decay power functions for uranium 235, not for the other 15 fissioning nuclides, so there is no account taken of what the i
16 composition of the reactor core is at the time of shutdown, at i 17 the tima of the accident.
18 Secondly, there is the question of how does this, 19 the developing spatial variation in the fuel composition, j 20 influence the decay power locally within the reactor core?
l 21 The ANS standard gives you a relationship between the history 22 of fissioning locally at a point in the core, and the decay I
l 23 power thct occurs following shutdown at the same point in the 24 core.
25 There is an additional uncertainty that has to be ,
HERITAGE REPORTING CORPORATION -- (202)628-4888
49 p ~1 allowed if you'are. going to use a generic curve. One single gh 2 decay power represents the core-wide situation during the 3 accident. That takes a considerable amount of evaluation. It 4 is not an overwhelming amount of cost to do that analysis, but 5 it has to be done if one is going to claim that you have a 6 power function which is input to this calculation which is 7 realistic in the context of the requirements of SECY 472.
8 Do you understand what I have said? .
9 MR. CADEM: I think someone back here does.
10 MR. WARD: Go ahead.
11 MR. STUCKER: Dave Stucker from Westinghouse--indeed 12 the document and the reactor heat source models in COBRA / TRAC 13 address the different fissile isotopes as per the 1979 L 14 standard of the uranium 235 fissile isotope 23 and the uranium l
15 238 as directed by the standard.
16 Further as directed by the standard, the heating is 17 directed, it is calculated as per the standard. These fissile 18 isotopes are related to the local composition of the fuel as 19 modeled by WCOBRA/ TRAC.
20 The--could you repeat your second question?
21 MR. SCHROCK: Let me point out to you that the 22 documentation had a different bit of information than you have i
23 just indicated. If you look in the appendix of your document l
24 which deals with the decay power, you find that the only l
25 information listed there is the expoential fit for the decay HERITAGE REPORTING CORPORATION -- (202)628-4888
o 50 1 power function for.U235. No mention is made of the existence
() 2 of any,other contributors to the decay power.
3- CHAIRMAN WARD: Could you get him the page number?-
4 MR. Stucker: Appendix B, page 4.
5 MR. SCHROCK: It you have it in front of you, you 6 can see what I am saying is in fact' correct. In addition to 7 that, the spatial problem is nowhere addressed that I can 8 find, place in the documentation that indicates that you have 9 addressed that question of the spatial situation. There is 10 further nothing said in the documentation about the 11 uncertainty in the decay power function.
12 MR. STUCKER: I would like to respond to that
- - question, however. In some consultation time with your
~13
}
14 colleagues I will respond to that.
15 CHAIRMAN WARD: Why don't you come back sometime at 16 the appropriate time during the Westinghouse presentation?
17 MR. STUCKER: Thank you. !
18 CHAIRMAN WARD: Thank you. Okay. Gene, how much 19 more time do you think you need if we leave you alone?
20 MR. HSII: Depends how much questions I get.
21 CHAIRMAN WARD: If we leave you alone, how much 22 time? ,
, 23 MR. JONES: Why don't you do the results of the i
24 review and the conditions for acceptance because that is l 25 really the bottom line? I think the rest is pretty much l
l HERITAGE REPORTING CORPORATION -- (202)628-4888
.. . __ _ _-=_ ._ _ _-.-_ _ _ . _ _ _ . _ . _ _ . _ - _ _ . . _ _ . . _ . _ - - _ , - -
51 1 self-explanatory now.
h 2 MR. HSII: The review method, we have answered basic 3 questions we have, and we have technical assistance from our 4 expert consultant from Sandia. During the review process, we 5 have constantly meet with Westinghouse to be informed of their 6 model development, their analysis, the analysis to convey the 7 staff concerns, and we also send them, the formal set of 8 questions to them to respond, and we also perform an audit 9 calculation using TRAC PF1, Mod 2, to compare the results.
10 MR. CATTON: You used TRAC PF1. Did you give any 11 consideration to the work that is going on in RES at the 12 moment in establishing the uncertainties of TRAC PFl?
8 13 14 MR. KELLY: We used TRAC PF1, Mod 1, Version 14, and it was strictly an audit calculation. We tried to as best we 15 could simulate one of the bounding calculations of the 16 Westinghouse with their code, do assessment of relative 17 similarities and differences between the two codes.
18 MR. CATTON: Have you been keeping up with TRAC PFl?
l 19 I think they are having a great deal of difficulty in doing 20 some of the calculations at least for SCTF. How could you use l
21 it to audit when you don't know what it is doing itself?
22 MR. KELLY: The version we used was the 14. The 1
23 calculations were done about a year ago, and I am not aware of 24 current calculations that are being done to TRAC today with a 25 different version, HERITAGE REPORTI!E CORPORATION -- (202)628-4888
52 l 1- MR. CATTON: it seems to me if the staff is going to 2 have somebody do audit. calculations-for them, it is incumbent
(
3 for the staff to see that they are kept up to speed with ,
4 respect to what is being done with that code. I get the 5 feeling that wasn't done.
6 HR. HSII: The purpose of the,-all the calculations 7 is to see, to compare how two different codes behave, and to 8 explain, if there is any difference, can we explain the j 9 phenomena of.what cause the difference?
10 MR. CATTON: You use the results of the audit 11 calculation to come to some conclusions. How can you come to 12 the conclusions that you do without knowing something about 13 the capability of-the code or at least the people who are
}
14 doing the computation for it, have some feeling for these i
15 things?
! 16 CHAIRMAN WARD: I don't think you can answer that. .
l 17 MR. CATTON: That is just a statement.
18 CHAIRMAN WARD: Just go ahead.
1 19 (Slide) ,
20 HR. HSII: The results of our review--we believe r
l 21 COBEA/ TRAC with proper plant modeling is capable of 22 calculating important phenomena during the large break LOCA in 23 the UPI PWR plant. We think the uncertainties associated with 24 the PCT calculation have been properly accounted for. The
.O 25 four channel superbounded calculation is a conservative HERITAGE REPORTING CORPORATION -- (202)628-4888 .
. _, _ .. n -. _ . _ _ , - . . . _ - _ . _ . _ . _ ~
53 1 estimate of the 95th percentile PCT, and we believe that the
,e
( ) 2 Appendix K requirements, except for the three items, have been 3 complied with. Those three items are one is the Westinghouse 4 did not have a, did not brought the nuclear during the 5 blow-down phase. Two items are related to this. Of the two 6 items, one is 1D3. The Appendix K rejected required the 7 carry-over fraction defined and used to calculate the exit 8 fluid flow at the core. We believe that the comment was 9 intended for the cold leg injection plant.
10 The WCOBRA/ TRAC does not define the carry-over 11 fraction.
12 Another item is the reflood heat transfer
/'T O
13 requirement, calculation requirement by Appendix K, and again 14 we think the Appendix K requirement is geared toward cold leg 15 injection plant. It is not applicable to the UPI plant.
16 And also there is some other conditions. We will 17 talk about that later, but we think if those conditions are 18 properly incorporated and implemented, the Westinghouse new EM 19 model will satisfy the SECY 83-472 approach, and it is 20 acceptable for licensing application.
21 (Slide) 22 HR. HSII: Talk about the conditions for 23 acceptance--maybe some of these conditions should be addressed
- 24 later on up to Westinghouse presentation, make it more clear.
v 25 HR. CATTON: Gene, there is some calculations in HERITAGE REPORTING CORPORATION -- (202)628-4888
54 1 Volume 2 using a four channel model. Did all these
[%s, ) 2 oscillations go away? You don't know?
3 MR. HSII: I didn't get into detail.
4 CHAIRMAN WARD: Were you indicating someone could 5 give you an answer, or will that come out of the Westinghouse 6 presentation?
7 MR. KELLY: I don't believe any four channel tests 8 are documented. This is fouir core channels.
9 MR. CATTON: It is talked about in here in that 10 document.
11 MR. KELLY: My understanding is there is no--
12 MR. YOUNG: Are we talking about--
()
s_/
13 MR. CATTON: Go ahead. I will take a look.
14 MR. HSII: I will just read these conditions. Maybe 15 up to Westinghouse to make the presentation. These conditions 16 will become more clear.
17 Worst condition we have, we require Westinghouse to 18 reanalyze at least UPI test using four channel model, and if 19 necessary, do quantify uncertainties.
20 Two, we require them to perform a plant-specific 21 sensitivity study to determine--we require licensee, we 22 require licensee to perform plant-specific sensitivity study 23 to determine the hot leg, hot assembly location.
24 Three, we require Westinghouse to install a flag in bgx 25 WCOBRA/ TRAC to block return to the nucleate boiling during HERITAGE REPORTING CORPORATION -- (202)628-4888
~
,/
55 1 blow-down for Appendix K calculations. This is a requirement g h 2 by Appendix K.
3 Four, before the licensing application, we require 4 the licensee to request for exemption to Appendix K items 1D3 5 and 1D5.
6 CHAIRMAN WARD: All right. Thank you very much, 7 Gene. Let's go right now to the Westinghouse presentation, 8 and this is going to be Mr. Parker?
9 MR. PARKER: Correct.
10 CHAIRMAN WARD: And I believe at this point we need 11 to close the meeting.
12 MR. CATTON: I found the chapter. It is Section 3-4 13 of Chapter 2, plant vessel model, four channel model, and the 14 question had to do with the oscillations.
15 CHAIRMAN WARD: Will that be answered as part of 16 your presentation today?
17 MR. YOUNG: Not mine, but we certainly can address 18 it when it becomes appropriate because we will talk about the 19 four channel.
20 CHAIRMAN WARD: Why don't we address it at that 21 time, unless there is a real quick answer like yes or no?
l 22 MR. CATTON: There is no real results from the four 23 channel model in the report. It is discussed in and 24 justified.
- 9. 25 MR. YOUNG: We will discuss it in more detail later.
HERITAGE REPORTING CORPORATION -- (202)628-4888
56 1 CHAIRMAN WARD: All right. Fine. Mr. Parker?
2 (Please refer to page 59 of the closed transcript 3 for continuation of the proceedings.)
4 5
6 7
8 9
10 11 12
( 13 O] 14 15 16 17 18 19 20 21 22 23 24 25 HERITAGE REPORTING CORPORATION -- (202)628-4888
252 1 P R O C g-E D I il G S 2- CHAISV.AN WARD: We are going to have to stop at
)
3 4:30. I want to have a few minutes to talk about what we are 4 going to do next.
5 Would you just give us a ten-minute overview?
6 MR. KELLY: I can give you bottom lines on the audit 7 calculation, and try to indicate why we think the differences 8 between TRAC and COBRA / TRAC that you will see handouts.
9 As far as the objectives of the study, it was to 10 verify using audit calculation that the predicted large break 11 LOCA thermal hydraulic response predicted by COBRA / TRAC was 12 consistent with current understanding of phenomena, and to do 13 that we ised TRAC PF1, Mod 1 code. This was Version 14, and 14 this is the version that was released about a year ago.
15 HR. CATTON: Isn't there a Version 14.37 16 MR. KELLY: There is a subsequent version after we 17 made these calculations.
4 18 MR. CATTON: 14.3 is the frozen version, and it is 19 the one that is undergoing this uncertainty analysis by 20 Research?
21 MR. KELLY: 1 2 are right. The funding for audit 22 calculation came at a time which was a little over a year ago.
23 Some calculations were done at that time. Subsequent funding 24 for further repeating those basic calculations with the new 25 version of TRAC, I am just explaining why we didn't redo the HERITAGE REPORTING CORPORATION -- (202)628-4888
j' 253 1 calculation.
2 MR. CATTON: I am not suggesting you go find 14.3 3 and do the calculations. What I am suggesting you do is that 4 you review the results of that before you come to any 5 conclusions.
6 MR. KELLY: Yes. The main conclusions I am coming 7 to in this talk are just why the code is different. Okay.
8 MR. CATTON: I don't think you can do that on the s 4
9 calculation.
10 MR. KELLY: Just to, first to model, identify what 11 are the differences in the behavior, and then look at the 12 modeling differences that lead to that and then do some kind 13 of assessment of the modeling differences that lead to predict 14 behavior, whether it is significant for the overall peak clad 15 temperature prediction.
16 (Slide) 17 MR. KELLY: The next couple handouts go to the 18 nodalization. I guess it is important to point out that the 19 two codes are different and you can't do things exactly the 20 same in the two codes.
l 21 In TRAC, we used in vessel three radial rings, 1
22 azimuthal sectors. In the core we used 6 axial nodes, with 2 l 23 radial rings and 2 azimuthal sectors, so nodalization scheme 24 is different between the two codes.
i 25 We had to make a choice about the difference between HERITAGE REPORTING CORPORATION -- (202)628-4888 l _ _ _ . . _ . _ _ . . _ . _ - . . _ - . _ _ _ _ _ _ . _ . _ _ _ _ . . _ . _ . . - _ . _ - _ _ _ .- _ ___ _ ___ _ _
254 1 the assembly types, that is between the guide tubes support
/~x 2 column open holes. We almost used all the guide tube 3 assemblies in our central ring, and the outer ring in the 4 core, we put the support column open hole assemblies. TRAC, 5 unlike COBRA / TRAC, had more difficulty modeling hot 6 assemblies, and hot rods. Within each core cell we model one 7 of each type, but we are severly limited in the sense that the 8 gap size has to be the same tor all rod types, and as you will 9 see in a couple slides down the road, this has impact.
10 HR. CATTON: It would take a simple change to get 11 around that.
12 MR. KELLY: We are not in the--that may be a simple 13 change. I am not sure how simple it is, but it was not in our
14 charter.
15 MR. CATTON: Gene knows all about that. It is being 16 done in order to do the uncertainty analysis. They changed, 17 they used different gap, different properties of the fuel, 18 different--
19 HR. KELLY: These fuel rod times are different in i 20 each assembly, but there are limitations in the release 21 conversion we are using that you just couldn't do it without 22 making code modifications I suspect.
23 With this model we are able to look at three 24 dimensional azimuthal effects to see if they are important, t
25 but we did use less detail in hot assembly model. To do it HERITAGE REPORTING CORPORATION -- (202)G28-4888
255' 1 correctly with TRAC, you would have to have one very small 2 channel in the center of the core to represent equivalent hot
{s'- )
3 assembly, and we couldn't justify the cost of doing that.
4 This gives you an idea what the nodalization looks 5 like.
6 MR. CATTON: Does the diameter mean it was spli.t in 7 half?
8 MR. KELLY: Which one are you looking at? This is 9 the vessel. This indicates here the, where the breaks and 10 injection systems. This is the broken hot leg, broken cold 11 leg. UPI injection was on the face of this ring in the upper 12 plenum. This is the in tact hot leg and in tact cold leg.
13 These are indicating where the connections are. I also have a 14 looped diagram I don't think I included.
15 MR. CATTON: Two azimuthal sectors?
16 MR. KELLY: Yes; 2 radial rings in the upper plenum, 17 three radial rings total.
18 (Slide) 19 MR. CATTON: It seems to me there is disconnects 20 between the use of the NRC tool and the work NRC does on its 21 tools. That's just a statement. You don't have to respond.
22 MR. KELLY: It is difficult to assess.
23 HR. HOCHREITER: This work was done before the other 24 stuff started.
i O 25 HR. KELLY: There is timing questions. These i
HERITAGE REPORTING CORPORATION -- (202)628-4888 i-
256 1 calculations were done sometime ago, and we had actually done 2 previous UPI calculations before. I don't want to say that 3 the latest version of COBRA / TRAC would not give different 4 results. It may have more capabilities. What I am trying to 5 do is compare what we had at the time with what was at that 6 time the best available tool, and you know, things progress.
7 TRAC is different, too. We didn't have the opportunity to go 8 back and redo it all, though.
9 MR. CATTON: Did you compare any of the results with 10 the NRC calculations using the TRAC, TRAC with the fine 11 nodalization?
12 MR. KELLY: We looked at NRC calculations as a part
,, 13 of assessment because that was a reference in the report.
> a 14 Just as--but there again, you are looking at COBRA / TRAC, 15 comparing it to the WCOBRA/ TRAC similar type problems. The 16 trends that are reported in the documentation are, I believe 17 are correct.
18 Now what we tried to do as far as matching, we go to 19 the rod temperature distribution to match the average rod 20 response. We did have a hot rod, and I will show those 21 initial comparisons in a second. What we wanted to do is 22 match total power in the total core and match the average rod 23 temperature distribution, and we did this by adjusting the gap 24 and keeping the power and flow constant. Got reasonably good
! f
-' 25 agreement on predicting average temperature.
HERITAGE REPORTING CORPORATION -- (202)628-4888
257 1 (Slide) 2 MR. KELLY: What we wanted to do was to make sure
- 3 that our peak linear heat rate in the hot rod was the same 4 between the two codes. We felt this was very important from 5 the reflood peak at least. And so the peak linear heat rate 6 occurs in the mid-plane and we see we did a reasonably good 7 job of matching that, but by doing that and having the 8 previous, and I believe Dr. Catton will say it is an old 9 constraint, in the gap modeling and TRAC, this led to higher 10 initial temperature for the hot rod in TRAC which has the 11 implication that the TRAC calculation started out with a 12 higher stored energy.
,_ s 13 (Slide)
( \
14 MR. KELLY: So when we look at blow-down, I think 15 the best thing to compare in this case is the average rod with 16 the same stored energy, and then we are looking at differences 17 in thermal hydraulic heat transfer behavior to do an 18 assessment of the response of the average rods, and what we 19 see is that the average response is in fact quite good. TRAC 20 does show non-asymmetric behavior, and we do see difference, 21 depending on whether we are on the in tack or oroken side of 22 the core. This is azimuthal effect arises. We see in both 23 calculations the guide tube assemblies that are assemblies 24 under guide tubes have better cooling during blow-down, and P 1
25 this is consistent between both.
HERITAGE REPORTING CORPORATION -- (202)628-4888
258 1 What we find is that WCOBRA/ TRAC, using different 2 model for the minimum stable film boiling temperature, the
/^}
3 point at which you will go into quenching, actually using 4 higher value, that is based on better correlation of the data, 5 and higher T min leads to earlier quench. Again the fact the 6 hot rod started with higher stored energy leads to higher 7 blow-down peak for the hot rod.
8 The next few slides demonstrate that. For each 9 curve, we have the one COBRA / TRAC calculation, which 10 indicates, which shows at least the guide tube. In this case, 11 it is an open hole support column, and this is compared to the 12 two track. Two track has two azimuthal sectors for the same 13 rod type, and the slide wants to go down. This is too light 14 and it wants to float. We see that the comparison here is 15 excellent in this, for this particular channel. This is the 16 WCOBRA/ TRAC and the blue here is the TRAC on the break even 17 side. The in tact side gives better blow-down cooling and 18 reaches its peak earlier and turns arcund.
19 CHAIRMAN WARD: Better skip over those.
20 MR. KELLY: Let me point out you see similar 21 behavior in other channels. We see in both cases that 22 WCOBRA/ TRAC quenches due to the higher film boiling time, 23 minimum stable boiling temperature, and if you compare the two 24 plots, you will see the blow-down peak is lower with the guide 25 tubea, and again for the hot rods, we see that TRAC actually HERITAGE REPORTING CORPORATION -- (202)628-4888
259 1 gets a higher blow-down peak. Due to higher stored energy, 2 you see in fact in TRAC that it actually starts out in steady
)
3 state film boiling when you monkey with the gap like that.
4 As far as in the reflood,'we are going to be 5 entering reflood with different temperatures in the average 6 rods, and this is going to affect the behavior. Qualitatively 7 you can see the same type of response in the two calculations.
8 The main differences arise due to the fact that we 9 have better blow-down cooling, and/or mainly due to difference 10 in blow-down cooling, that is, that COBRA / TRAC has the higher 11 T min, start reflood at lower thermal state. Then later in 12 reflood we see two effects coming on. One is that TRAC PF1 13 has larger effects of the accumulator nitrogen. That is, when 14 the nitrogen comes on, it tends to pressurize the system, more 15 and more widespread.
16 Then COBRA / TRAC, this is I believe due to the fact 17 that it is an explicit model for the gas fuel, and transport 18 that condensible field in the more realistic sense, and the 19 other important item is that the liquid entrainment modeling 20 WCOBRA/ TRAC tends to sweep more water out of the core and 21 affect the water inventory in the core region.
22 I think the--
23 CHAIRMAN WARD: How about going to your summary 24 slide?
O 25 HR. KELLY: Okay. Summary slide just repeats it. I HERITAGE REPORTING CORPORATION -- (202)628-4888
260 1 think really the key--let me show you one that I think is the
. 2 key. !
%/ 3 The key is what is the water inventory in the core !
t 4 during reflood? Okay. I think you are going to understand j 5 the initial condition being different because of the t 6 difference in blow-down cooling, but what we see is that, at 7 this time period is when the nitrogen hits the system, it 8 -causes more pressurization in TRAC, which forces more water 9 into the core. This shows liquid level. It is an indication 10 of the total core liquid mass.
11 Forces it in. It causes more steam generation rate, t
12 The water goes out, and the TRAC comes back in and goes out 13 again, and later in time we see a more rapid build-up of water
-- 14 in the TRAC PF1 calculation whereas it is more or less 15 quasi-steady out here in the WCOBRA/ TRAC.
l 16 If we go to a couple more slides down the road, we l 17 see the difference is due to difference in break flow during 18 this period, where on a blown-up scale, this is the pump side 19 of the break, they reconsider COBRA / TRAC over this time, or 20 reflood it, consistently predicting higher blow out the break 21 than does TRAC, the reason being entrainad in the core was 22 swept up in the steam generators, Icading to higher pressures 23 in the core. This is the pressure on the top of the core. .
24 See higher pressure at the top of the core.
I 25 What is really important is the delta P between the ;
r
[
I HERITAGE REPORTING CORPORATION -- (202)628-4888
~ ,- , , , - . . , - - , - - - , ~ ~ - - - - - - , , - , - . . - . , - . - - - - , , - , _ - , - - ---m,- - , - , - . . - , , , . - - - - - _ . - , _ - - . . - - . - . . . . - - , - _ . -
261 ,
1 top of the core which are these two calculations, and this is l 2 the containment pressure. The delta P is higher and you are 3 getting more mass lost in WCOBRA/ TRAC directly as the result 4 of the modeling, explicit modeling of entrainment, which is 5 not in TRAC PF1.
6 It is clear when you look at the thermal response 7 that they don't agree exactly. However, I don't believe that 8
8 we saw anything inconsistent in the COBRA / TRAC response.
9 There were differences in how things are modeled that couldn't 10 'be traced directly to why there were differences in the clad 11 temperature, and I think that we can understand these, and in 12 terms of whether this is a realistic bounding case, it 13 appeared to be. Since it did have two additional conservative 14 assumptions, I guess that tended to make the reflood peak 15 higher than TRAC and did seem to bound, certainly bounded TRAC 16 in the reflood case.
17 So I think we can conclude that it is realistic
, 18 bounding which is the case that we did try audit.
19 HR. CATTON: You can't conclude that at all, not i
20 from comparing it with TRAC. TRAC is having some terrible 21 difficulties in trying to model the Japanese test results.
22 They don't have void fraction right, or the heat transfer is 23 wrong. Entrainment isn't correct. Even have gone in and 1
24 nultiplied some of the interspatial drag coefficients by i
O 25 factors of 20, and they still can't get it right.
l HERI'1 AGE REPORTING CORPORATION -- (202)628-4888 l
262 1 So you can't, you can't really do that. If you are j 2 going to try to make some arguments about realistic or 3 non-realistic, it seems to me you have to compare, you have to 4 look at CCTF calculations that were done by Westinghouse.
5 Then you can make some conclusions. You can't conclude based 6 on TRAC nuclear, unless you tell me TRAC does better beyond t
7 CCFT than they do, and I don't think you know that.
8 MR. KELLY: I don't know the answer to the question 9 on CCTF. I know we have asked Westinghouse to go back and 10 redo the CCTF calculation because we are--
11 MR. CATTON: You are coming to conclusions based on 12 the comparison with TRAC.
13 MR. KELLY: No. I think you are wrong about that. <
C) 14 MR. CATTON: You are saying realistic.
15 MR. KELLY: Based on this audit calculation where we
- F I 16 compared the same case, the same bound, realistic bounding j 17 case, got excellent agreement in blow-doun qualitatively. We 18 could see the differences and understand the difference 19 quantitatively in reflood, and there was nothing in that phase
) 20 that indicated to me that there was extremely bad models or 21 something was cotally inconsistent with our understanding of 22 the chenomena.
23 They are looking at the picture differently, though, 1 !'
1 24 and this is, leads to the differences in behavior. Our 25 question or our purpose of that was to make sure that the i
HERITAGE REPORTINO CORPORATION -- (202)628-4888 i
[
263 1 calculation that Westinghouse was telling us was a bounded 2 case was in fact consistent with a bounded case, and I think 3 we can make that conclusion.
4 HR. CATTON: Don't you first have to say TRAC is 5 good or bad? If you say TRAC is good, then you make your 6 comparison.
7 HR. 13LLY: I did not want to make that statement.
8 It is an independent assessment tool.
9 MR. CATTON: If you don't want to make a statement 10 about TRAC, then the comparison is just another exercise ar.d 11 doesn't allow you to come to any conclusions.
12 MR. KELLY: Our calculation was part of our overall 13 conclusion. Let me state that the, it is a different tool.
7_
( i 14 It looks at the same problem in a different way, and there is 15 some validity I believe in making those kind of calculational 16 comparisons becausa you get at the root of why codes do what 17 they do and why there are differences between the codes.
18 I don't want to say that either code is right or 19 wrong. I want to say that in a, in a sense of predicting 20 thermal hydraulic behavior, I do not see anything in these 1
21 calculatior.s that i>uld lead me to believe that some major 22 phenomena had been overlooked in COBRA / TRAC. I don't know if 23 I could say the same thing about TRAC in view of the, the 24 entrainment behavior th?.t WCOBRA/ TRAC predicted. That was out
('m.
( )
iJ 25 of my jurisdiction to go and decide whether TRAC was the best HERITAGE REPORTING CORPORATION -- (202)628-4888
264 1 or WCOBRA/ TRAC. It is an independent method to look at the same calculation, and that's all it is.
) 2 3 MR. CATTON: In the particular review, could you 4 comment on the bad pressure predictions of CTFF,.why I should 5 not be concerned about that?
6 MR. KELLY: In the original--let's see. I think you 7 have got revised Volume 1. The original documentation, that 8 we received the data, one of the misconceptions you have is 9 that this one document is all we have ever received. We 10 received draft documentation quite early on. We sat down with 11 W0stinghouse, with consultants, and we came up with a list of 12 38 questions in Volume 1, plus--those are formal questions.
es s
13 Plus there were various other exchanges trying to get them to 14 improve the quality of the document, the technical content, so 15 that we could do our review properly.
16 It is really unfortunate you didn't' receive the 17 questions and responses to Volume 1, and there is also a whole 18 set for Volume 2 because it turns out that the questions you 19 were asking today have already been asked during the review 20 process because if you look, I agree if you look at Volume 1.
21 those questions need to be asked.
22 As far as the CCTF specifically, I think the code 23 shows the oscillations in there. The data today did, too, and 24 we couldn't read it. The frequency of the data was so bad we l
[J 25 couldn't tell what the pressure predictions--we asked l
HERITAGE REPORTING CORPORATION -- (202)629-4888
265 ,
1 Westinghouse to clarify it. Their solution was to smooth it
-m 2 out so-you can't see it. It would take a fairly detailed
)
3 analysis to go back and look at amplitude and frequency of 4 those oscillations, but in a qualitative sense, the code was, 5 shows that kind of behavior and what the problem is. In the 6 data in the report it is not shown because it was smoothed out 7 in just a simple grapics technique.
8 I would like to stress, though, that the questions 9 you raised, if you go and look at our set of questions, have 10 been asked, and Westinghouse came up with initially responses.
11 We in general did not find their initial responses acceptable 12 in any case, and required them to go back and do them over.
s 13 There was an iterative process where we eventually got this j j 14 complete set of documentation which is now formally submitted.
15 and it is part of the docket, so it should be made available.
16 I think that the other point is that as far as the 17 definition of the superbounding case, during our review we 18 identified many of these system parameters with Westinghouse, 19 but we tried to highlight things we thought were important and 20 had them go back and look at them with the sensitivity 21 studies. It was through this review process, iterative review 22 process, that they came up with this long list of sensitivity 23 studies that they finally did, and this has been over a couple 24 of years, and ultimately made the decision on how to do their 25 superbounding case.
HERITAGE REPORTING CORPORATION -- (202)628-4888
266 1 MR. WARD: Okay. Thank you very much, John. I just 2 have a couple minutes. I think the Subcommittee is going to 3 need to hear more. There were a number of questions that we 4 have. Some are perhaps due to the fact that documentation 5 isn't complete and we didn't see everything. We didn't dig 6 through it thoroughly enough, but I think there are a number 7 of other questions which I'm not sure that there are--well, 8 until we see--I suspect the answers aren't in the 9 documentation that's available. Maybe they will be.
10 MR. JONES: Some of the answers are not.
11 CHAIRMAN WARD: So what I would like to do is 12 arrange for another Subcommictee meeting. Right now it is .ot 13 real clear to me exactly what that agenda ,ill be. What I
(' 1
~'
14 would like to suggest is that Mr. Boehnert will collect 15 suggestions. You have heard our questions and you folks have 16 heard our questions, and there may be places where we just, 17 you haven't been able to make us understand what is there, but 18 if you could, if you have suggestions to Mr. Boehnert for how 19 we could approach the next meeting agenda, he will collect 20 those, including from you folks.
21 MR. CATTON: That means you want a report?
22 CHAIRMAN WARD: Yes, I would like to have one, but 23 in particular, I would like to get something, your suggestion
! 24 for the agenda for the next meeting to Paul early enough so
, s 25 that he and I then get together and shape up the actual l
HERITAGE REPORTING CORPORATION -- (202)628-4888
267 1 agenda.
2 We can use a date which we had scheduled for this 3 Subcommittee for another purpose. That's June 21st, ::.nd 4 pastpone, I think we can postpone the subject we had scheduled 5 for that.
6 MR. BOEHNERT: Yes.
7 MR. WARD: Let's tentatively say June 21st we will 8 meet again on this subject.
9 MR. JONES: We will also provide you the additional 10 documentation that you didn't get, the questions and answers, 11 and we will get that to you ASSAP.
12 CHAIRMAN WARD: The answer to both volumes?
7 13 MR. JONES: Both volumes; we will make sure that
( )
14 gets out to you ASAP.
15 CHAIRMAN WARD: Let's get tha".
16 MR. CATTON: Does th,it mean a revision to these 17 documents or what?
18 MR. JONES: Those are questions and answers to that i
19 document. I do not believe they have revised the Volume is 20 and 2s.
21 MR. HOCHREITER: There were sections that were 22 rewritten for Volume 1.
23 MR. JONES: As part of responses; the Volume 1 24 answers are going to be several huadred pages.
\/ 25 MR. CATTON: This is several hundred pages, and I HERITAGE REPORTING CORPORATION -- (202)628-4888
268
', 1 shouldn't have bothered to read it.
i.
, /i 2 MR. HOCHREITER: These are the responses.
3' MR. CATTO!i: Okay.
4 CHAIRMAll WARD: Okay. Well, thank you very much.
5 Have a nice weekend.
6 (Whereupon, at 4:35 p.m., the meeting was 7 adjourned.)
8 9
10 11 12 13 14 15 16 17 18 19 20 1
21 22 23 24 25 HERITAGE REPORTIllG CORPOP ATIOli -- (202)628-4888
1 CERTIFICATE
('f u
h 2 3 This is to certify that the attached proceedings before the 4 United States Nuclear Regulatory Commission in the matter of:
5- Name: ACRS--Subcommittee on Thermal Hydraulic Phenomena, Open Session 6
7 Docket Number:
8 Place Washington, D.C.
9 Date: May 27, 1988 10 were held as herein appears, and that this is the original 11 transcript thereof for the file of the United States Nuclear 12 Regulatory Commission taken stenographically by me and, 13 thereafter reduced to typewriting by me or under the direction 14 of the court reporting company, and that the transcript is a O true and accurate record of the foregoing proceed ngs.
(/ 15 16 /S/ ' O. m 17 (Signature typed): Catherine S, Boyd 18 Official Reporter 19 Heritage Reporting Corporation 20 21 22 23 24 f
25 O Heritage Reporting Corporation (202) 628-4888
e .
SNL AUDIT CALCULATIONS JOHN E. KELLY RUPERT K. BYERS i
( > SANDIA NATIONAL LABORATORIES ALBUQUERQUE, NM 87185 l
l ACRS MEETING MAY 27, 1988 I
( >
Sandia National Laboratories- ,
l
I
< i OUTLINE OBJECTIVES OF AUDIT CALCULATIONS TRAC-PF1 MODEL
, BLOWDOWN RESULTS 4
REFLOOD RESULTS
SUMMARY
t >
Sandia National Laboratories-
1 OBJECTIVES OF AUDIT CALCULATIONS l
- (
VERIFY THAT THE PREDICTED LBLOCA THERMAL-HYDRAULIC BEHAVIOR OF WCOBRA/ TRAC IS CONSISTENT WITH CURRENT UNDERSTANDING 0F PHENOMENA l
- USED TRAC-PF1/ MOD 1(V14) AS AUDIT TOOL t
q >
l IDENTIFY MODELING DIFFERENCES THAT LEAD TO DIFFERENCES IN PREDICTED BEHAVIOR ASSESS IMPORTANCE OF MODELING DIFFERENCES dI Sandia National Laboratories-
TRAC-PF1 MODEL i '
N0DALIZATION l
VESSEL N0DALIZATION USED 17 AXIAL N0 DES, 3 RADIAL RINGS AND 2 AZIMUTHAL SECTORS CORE REGION USED 6 AXIAL N0 DES, 2 RADIAL RINGS, AND 2 AZIMUTHAL SECTORS GUIDE TUBE ASSEMBLIES MODELED IN CENTRAL RING 1
SUPPORTCOLUMN/0PENH0LEASSEMBLIESMODELEDIk2NDRIN q >
EACH CORE CELL HAS AN AVERAGE ROD, A HOT ASSEMBLY ROD, AND A HOT R0D AZIMUTHAL EFFECTS CAN BE ASSESSED LESS DETAILED TREATMENT OF HOT ASSEMBLY t >
Sandia National Laboratories-
O
, , , , , 11.2 4 TRAC VESSEL NODALIZATION I I i i i t 1 i t t L__L
_ a _ _ L _ _ hI _ _ a _i _ Til _ _i i
i i
l i
i e i i i l I I I i i s j l l
'__[__. __l___ ___[__- 9.63 15 ! ! !
l ! l 9.03 1 1 I 14 i ' '
I i l i e i --
__r--- 7--- r--
i i i l l I i 1 i
! BHL [CL
__7__,, _ _ ,! _ _ , 7_ _ ! _ .
i i i
_. __e_____q._____e_. .
i i I 1
_. __y_____m______y.. ..
I t q I i-p, 6.27 l
_. ._ _____ ______L__. -
E 4' 2 1 13 5 i
I I I
e i i i
._5,_____a______5__
i i l I I i !
1 I i
+
_. ._e_____y______e__._
- l l t I I l ,
I
._e_____ _,I______ir_ _..-
BCL [HL l l 1 I I I i i r---- 1----- r--
- i 1 4
y.I I. I. U
. . . . 2.61 l t I l I 1 3 ' ' '
i 1 I l 1 1
____p_____a______p___ 1,3s i I I I I
( l 1 1 i '
l l l l l
_ q _ _ p _ _ _ _ _ q _ _ _ _ _ _ p _ _l_
I I I I I l l 1 I i O.00 0
TRAC-PF1 MODEL i >
INITIAL & B0UNDARY CONDITIONS i
MATCHED TOTAL POWER, FLOW, AND AVERAGE R0D TEMPERATURES AS WELL AS POSSIBLE l
HOT ROD PEAK POWER DENSITY MATCHED AS WELL AS POSSIBLE I
LIMITATION IN TRAC-PF1 LED TO HIGHER STORED ENERGY IN l
HOT ROD l
USING WESTINGHOUSE SUPPLIED FLOW TABLES AND BREAK PRESSURE, THE CALCULATED INJECTION FLOWS AND BREAK FLOWS WERE WELL REPRESENTED DURING THE TRANSIENT l
l 4 >
Sandia National Laboratories-
t 1
l
/
r; I
AVERAGE ROD FUEL TEMPERATURES AT STEADY STATE 1.6 , , ,_ , , , , , , , , , ,
1.5 - ,
,- 's ,
1.4 - -
/ \
\ -
1.3 -
\
/
\
/
g i .a - /
\
\
o /
O / \
ga 1.1 - g a:
/ \
, V t .o - -r s -
\
/
M i ca \ -
6-- o.9 - /
\
\
- ~ s
~ \
' s I s '
' \ -
o.7 - /
' \ \
g j \
\
N' -
o.6 -'-
o.5 , , , , , , , , , , , , , ,
o.oo o.25 o.so o.75 1.00 1.25 t So 1.75 2.co 2.25 2.5o 2.75 3.00 3.25 3.50 3.75 CORE ELEVATION (m)
--- WESTINGHOUSE CENTERLINE WESTINGHOUSE SURFACE TRAC CENTERLINE TRAC SURFACE O
. . _ , . _ . . . - . . _ . . - , , , . . . . . . . _ , . . _ . . , . . . , . , _ . _ . . . _ . _ _ , _ . . _ _ _ . . . . _ . _ _ _ , . . .,_,m.-- _ _ _ . . _ .
aJS . _ _ - - 3.#eA 4: 4 6 4 a.a. 4 4 6h ---a14 4 - A.4aJ.. 5 A- .L ,,.4--1 o
. i 'l l
c HOT ROD LINEAR POWER DENSITY AT STEADY STATE 4
60.0 - - -
a
- == g SCO- s
' s
/ \
/ \
/ \
/ \
/ \
40.0- , y
/ \
/ \
v' / \ [-
/ \
30 0- ,
/ \
, /
, \
\
g '
/ \
{ / \ -
20 0-
/ \
/ \
/ \
/ \
/ \
10.0- / \
/ \
\
/
\
/
\
/
r i i 0.0 , , , .
3.75 0.00 0.75 1 50 2.25 3.00 CORE ELEVATION (m) ,
WESTINGHOUSE DATA TRAC (AVERAGE ROD PROFILE WITH PEAKING FACTOR = 1.8308)
O
E t
HOT ROD FUEL TEMPERATURES AT STEADY STATE.
2.8 , , , , , , , , , , , , , ,
2.4 ~ '
. ~'~, .
s
' s e N
/ \
2.2 - / \ -
/ N
/ \
- \ '
2.0 - / \
- \
k '#
C'"> 1.8 - \ -
- \
- \
Cal Ct: /
1.8 - f
'\ -
llE 1.4 - / g
- \
\
\ .
i 12 ,
' \
- \
1.0 - I g l' _ . __ _ __ --- - _ _ _ ___ \
' ~ \
( # '
\
l 0.8 - - -
- \
e i
s \
l s "
l 0e , , , , , , , , , , , , , ,
O'.00 0.25 0.50 0.75 1.00 1.25 1 50 1.75 2.00 2.25 2 50 2.75 3.00 3.25 3.50 3.75 CORE ELEVATION (m)
WESTINGHOUSE CENTERLINE
- - - WESTINGHOUSE SURFACE TRAC CENTERLINE l- TRAC SURFACE l.
O l-I:
l 1:
BLOWDOWN RESULTS l
> l 4
AGREEMENT OF AVERAGE ROD RESPONSE IS QUITE GOOD TRAC-PF1 SHOWS ASYMMETRIC BEHAVIOR l INTACT VERSUS BROKEN LOOP GUIDE TUBE ASSEMBLIES ARE COOLED BETTER IN BOTH CASES 4
WCOBRA/ TRAC USES A HIGHER TMSFB AND QUENCHES EARLIER HIGHER STORED ENERGY IN TRAC-PF1 HOT RODS LEADS TO HIGHER BLOWDOWN PEAK i
t Sandia National Laboratories-
TWO-LOOP UPI AUDIT COMPARISON 900 0 , , , , i i ,
P* -
850 0-
' s E
d
\'* -
C?
600 0- \
\
b 750 0 - l ,
s is
'd 700 0- . '9, / ' 5 ' "
)
D'.
m o 650 0- t 7.
600 0-"
~~ ~ ~ Y -
5 O
~ ~h. i O
h 550 0-W s O
5 W 500 0-s s
Q 'O -
N -
450 0-
% /
400 0 '15 0
-lb.0 -50 d.0 $0 lb.0 15 0 2b 0 25 0 TIME (s)
TRAC-PF1/ MOD 1 OUTER RING. INTACT SIDE
- - + - TRAC-PF1/ MOD 1 OUTER RING, BROKEN SIDE
-O- WCOBRA/ TRAC UNDER OH/SC o
TWO-LOOP UPI AUDIT COMPARISON 850 0 , , i p ., , ,
\
M. i s.
/C ii 500 O- *I -
I m '
\.
t2 7600-i\ig {
S ', '& ,A' , ,. .. _ .s 700 0- 'g j q x I i \.*'
s .: ,
i 650 0-e 600 0- i o ~ %' '
8 a
b 550 O -
W s
- C s ,O 500 0- C f
% i N I g
- C 450 0 ; , , , , , , r
-150 -10.0 -5.0 0.0 5.0 10 0 15 0 20.0 25 0 TIME (s)
TRAC-PF1/ MODI CENTER RING, INTACT SIDE
- - + - TRAC-PF1/ MOD 1 CENTER RING. BROKEN SIDE
-G- lyCOBRA/ TRAC UNDER GT (n) v,
l l
l TWO-LOOP UPI AUDIT COMPARISON 1.25 , ' - i e i 120- ,/ '\ _
s\
1.15- / { _
.s E '
- i.to -
, b _
t2 I
's 1.05- \ \ _
q '
'g \
n '
100- s -
w .
50 5- l & x....
o80- \, ,.'
- o s5- -4 e ,
E #
j
@ 0 e0 -
5 .
y o75-O i _
o x 8 0.70-
= ,
~
0 65- '
O-- --O---- '
O e0- d _
0 55 . , , , ,
2". 0
-15 0 -10.0 -50 0.0 5.0 10.0 15 0 2b.0 TIME (s)
! TRAC-PF1/ MOD 1 OUTER RING, INTACT SIDE
-+ - TRAC-PF1/ MOD 1 OUTER RING, BROKEN SIDE WCOBRA/ TRAC UNDER SC l
m l
()
m i
l
[
REFLOOD RESULTS 4
R0D RESPONSE IS QUALITATIVELY SIMILAR DIFFERENCES IN BLOWDOWN COOLING (AND INITIAL TEMPERATURES) ARE DUE TO HIGHER, THENCE,USED MSFB IN WCOBRA/ TRAC 8
l i DIFFERENCES IN CORE LIQUID INVENTORY ARE DUE TO LARGER i
EFFECT OF ACCUMULATOR NITROGEN IN TRAC-PF1 AND LIQUID ENTRAINMENT MODELING IN WCOBRA/ TRAC 4 >
Sandia National Laboratories-
1 . . . - - . . . .. ._ -
a.]-
- - / .
TWO-LOOP UPI AUDIT COMPARISON.
850 0 , , , , , , , , ,
fl . -
800 0- \
E I Cb , 750 0- g . s, s
u e
- a ,/
- n. * -
6 700.0- .
). / ,
w '\/
g a 4 .j 650 0- ( ,
C'O *
> l 8 Z 3
6000- ,
I a t
i g
/\
i l Q 550 0 - f ht !, -
I d '
I I
/
I g 500 0-g i
C N 9 '
i b
450.0- 'g
> 1 400 0- TOS-4_O- -
350 0 , , ,
200 0 180.0
-20 0 0.0 20.0 40 0 60.0 80.0 100.0 120 0 140.0 160.0 TIME (s)
TRAC-PF1/ MOD 1 CENTER RING INTACT SIDE
- + - TRAC-PF1/ MOD 1 CENTER RING, BROKEN SIDE O- FCOBRA/ TRAC UNDER GT ,
O y- ,w,,- y- --,-y-f m.~,-.,.-,,,.y,.,, , _ , , , , . , ,,-,,.,,,,7, 7 . , ,,,_.,,,,,-,,,,,..y ,._._-,_,.,..,,_-,-,__..,,,,-_,..,.,,..,-,,_,y.,....y- , . ,, _ , ,_,
TWO-LOOP UPI AUDIT COMPARISON 900 0 , , , , , , , , , ,
850 0-E .o00 A
g <>
m 750.0 - () -
E. . ,
w .
2o0.0 .\d W '/ .
g -
m \v..*i k 650 0-
-% O ' .'. .
? I
\ ,
l I '
.g 600 0- @ -
' I -
i g s
O C \. -
'O 5 550 0- 3 l ,- ' -
\ '
s I
! c lp '
o \ i m I / I -
p 500 0-O \ */ i a , o I o 450.0 -
g8 #
1 -
I I
/ Y,i
-v.$ - i, g
\,g 400.0- W y- N.D 7 ~~ .s.,..-.4
~- -
350.0 , , , , , , , , , ,
-20.0 00 20.0 40.0 60.0 60 0 100 0 120 0 140 0 160.0 180.0 2000 TIME (s)
! TRAC-PF1/ MODI OUTER RING, INTACT SIDE
- + - TRAC-PF1/ MODI OUTER RING, BROKEN SIDE t - e - WCOBRA/ TRAC UNDER OH/SC O .
~-_ , - . . - . . - _ - . . - - - . . . - _ _ . . - - - ._. -- - - . . - _ . - - . - . . _ _ .
I
)
- ,. I i
-I
.l TWO-LOOP UPI AUDIT COMPARISON i
1.3 , , , , , , , , , ,
- O%
g .g . ..
,,.D,s I,l
~
xa,~s 4 ..
( -
9er E s .
S / -
q 1.1 - i f,s . -
. /
9 1.0 - '\
\ ' .f\.m t
m g <> . $ *( f f, f 0.9- ') .
g<
4 O \ . _ . ^ . , k ",\
% i o
'h/ '\
~
j N- 'q Q '%
O 0 *9 - l *s, r
f.
O o
g s _e, t
N.N g U 06-
- O "
O l
!' CM .
0.5 - l
~
+
i 04- - . ::
E
+ 0.3 , , , , , , , , ,
200 0
, -20.0 0.0 20.0 40.0 60.0 80 0 100.0 120 0 140.0 160 0 180.0 TIME (s) [
TRAC-PF1/ MOD 1 OUTER RING, INTACT SIDE
-+- TRAC-PF1/ MODI OUTER RING, BROKEN SIDE
-G- ]LVCOBRA/ TRAC UNDER SC [
i i
!O t
e TWO-LOOP UPI AUDIT COMPARISON 3.75 , , , , , , , , , ,
3 50- _
3 25- _
3.00-m 2 75 -
b h d 2.50-D o J 1
2 25-sO 2 00- ,
t.75 ]
o -l 150-8 Q ',
w 125- ( D s' I ;
s1 If i til g, 'g .:
1.00- s ,
o '
p ,, ,i ; ,, ,'b ', ! b ',g' 'd'd
"- o
. ,y',iO o, lc 9 :; ,
0 50 - 8, .
8, s i ! J 0.25- g
\ \
- ~ '
O.00
-20,0 d.0 20.0 4b.0 6b.0 8b.0 1NO O 120 0 14'O 0 16'O O 18'O 0 200 0 TIME: (s)
- e - LycoBRA/ TRAC CHANNCL 11
.O m;
.y.
t,4
'TWO-LOOP.UPI AUDIT COMPARISON
'2.50 , , , , , , , , , ,
4 22S-
' 2.00 -
d .
(*
a 1.7S -
. e-0 1.50-n.
126-m y 1.00-() .
0.75-Ei ,
g- o .
D 0.S0 - 1 I
k
- I I % \
O.26-
' ' 'S ' ' ~ #
f g d.0 2b o 4b o 8b.0 8b.0 1d0.0 12'o 0 1/0.0 1$0.0 1$0.0 200 0
-200 TIME (s)
TR AC-PF1/ MODI
- e - FCOBRA/ TRAC O
a TWO-LOOP UPI AUDIT COMPARISON 10 0 , , , , , , , , , ,
i (D
90- 1 -
r 4
80- g -
7.0 - -
u s -
60-e 6 W 1 50- (9 i
i g 40- ,
( 30-i -
k 20- -
8 1.0 -
g 1
0 s o.o _ i o- ___ 3 _---e,_ __ o
- 1.0 - , , , , , , , , , .
-20.0 0.0 20 0 40 0 60 0 80.0 100.0 120 0 140 0 160.0 180.0 200 0 TIME: (s)
-G- ECOBRA/ TRAC o
l l
8 l
TWO-LOOP UPI AUDIT COMPARISON 50 0 , , , , ,
- \ /\ f' '
- b /
45 0- ' '
\
,\ 9'~ 6i s 9,
O b \ /
/
i 40 0-I
\ D /
)
- \
\'
/ ~ \g _ /
05 0- t l
\ g 000- A .
- t , s V D \ %
/
25 0-
= '
20 0-O e is o-T c.
% 10 0 - -
50-.q ,
o0- I _
I i ,
-10.0 , , , , ,
90 0 100.0 tto 0 ?20.0 130.0 140 0 t 50.0 TIME (s) i TR AC-PF1/ MOD 1
-G- WCOBRA/ TRAC O
i
, -lh U?
TWO-LOOP OPI AUDIT COMPARISON ,
. 300 0 , , , , , . , i i 289 0-278 0-i l
267.0 - @ "
,i i
I
-2560- 8' *
,b
% Q 'i n e ,s , i es C i ,s f g g g I
l\ \ 0 \- / Q ,0 g!~, I y
hg ' O bs' g
~
234.0- '*
f's s t j ,
\1 Q ' , ' Q l 223 0-
$' 212 0- - .4 '+- , "
2010- *-. ~
- --.,,,+ - . .
190.0 , .
i ,
165.0 175 0 s
?$.0 85.0 95.0 105 0 11S 0 125.0 135.0 145 0 t55 0 i,
TIME (s) l l TRAC-PF1/ MOD 1 AT TOP OF CORE '
-G- WCOBRA/ TRAC AT TOP OF CORE i' - + - BREAK PRESSURE b
1 I
I ..
f
! i r
t
.I l
i l [
t
+
l a.-.-....---.._-....-__-..-._,.- _ - _ _ ,
,,e
i
SUMMARY
WCOBRA/ TRAC AND TRAC-PF1 DO NOT AGREE EXACTLY l
l l HOWEVER, OVERALL RESPONSE IS NOT INCONSISTENT WITH KNOWN PHENOMEN0 LOGICAL BEHAVIOR t
t DIFFERENCES IN RESPONSE CAN BE UNDERST0OD IN TERMS OF MODELING DIFFERENCES THE WCOBRA/ TRAC CALCULATION DOES, IN FACT, REPRESENT A REALISTIC, BOUNDING CASE 4 ,I Sandia National Laboratories-
- - _ _ __ . . _ _ - _ - . _ _ _ _ _ _ _ . . _ _ - _ _ _ _-_