ML20205E293

Transcript of ACRS Subcommittee on Safety Philosophy, Technology & Criteria 851009 Meeting in Washington,Dc. Pp 1-78.Supporting Documentation Encl

ML20205E293
Person / Time
Issue date:	10/09/1985
From:	Advisory Committee on Reactor Safeguards
To:
References
ACRS-T-1457, NUDOCS 8510170256
Download: ML20205E293 (104)

Text

{{#Wiki_filter:_ OPJGINA' 00f87;/4S7 O UN11ED STATES NUCLEAR REGULATORY COMNIISSION IN THE MATTER OF: DOCKET NO: ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON SAFETY PHILOSOPHY, TECHNOLOGY AND CRITEEIA b O LOCATION: Washington, D. C. PAGES: 1 - 78 DATE: Wednesday, October 9, 1985 .m a y

:j Q %nny;;s.,,9y a

~, g y ~ m. It ?- 'g, L w d Jo ?:!ot Renmve fmm ACRS Office ace-FEDERAL REPORTERS, INC. O OfficialReporters 444 North CapitolStreet Washington, D.C. 20001 9";IO170256 851009 (202) 347-3M PDR ACRS T-1457 PDR NATIONWIDE COVERACE

CR24597.0 MPB/dnw 1 UNITED STATES OF AMERICA g I i i/ 2 NUCLEAR REGULATORY COMMISSION l 3 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON SAFETY PHILOSOPHY, TECHNOLOGY AND CRITERIA 5 Nuclear Regulatory Commission 6 Room 1167 1717 H Street, N.W. 7 Washington, D. C. Wednesday, October 9, 1985 8 The subcommittee met at 3:00 p.m., David Okrent, f 9 chairman, presiding. 10 l jj ACRS MEMBERS PRESENT: I DAVID OKRENT I2 l' FORREST J. REMICK JESSE C. EBERSOLE l ( 13 RICHARD SAVIO j '~ CHARLES J. WYLIE 14 WILLIAM KERR CARLYLE MICHELSON 15 l 16 1 17 l l 18 19 l 20 21 22 23 h Am-Federal Reporters, Inc. 25 1

PUBLIC NOTICE BY THE t UNITED STATES NUCLEAR REGULATORY COMMISSIONERS' ADVISORY COMMITTEE ON REACTOR SAFEGUARDS WEDNESDAY, OCTOBER 9, 1985 The contents of this stenographic transcript of the i proceedings of the United States Nuclear Regulatory Commission's Advisory Committee on Reactor Safeguards (ACRS), as reported herein, is an uncorrected record of { the discussions recorded at the meeting held on the above date. No member of the ACRS Staff and,no participant at this meeting accepts any responsibility for errors or inaccuracies of statement or data contained in this transcript. l t ) C) i i I

i ) 5970 01 01 2 MPBeb 1 PROCEEDINGS 2 DR. OKRENT: The meeting will now come to order. 3 This is a meeting of the Advisory Committee on 4 Reactor Safeguards, Subcommittee on Safety Philosophy, 5 Technology and Criteria; David Okrent, the Subcommittee 6 Chairman. Our ACRS members present at the moment are 7 Mr. Michelson, Mr. Kerr, Mr. Wylie, and Mr. Remick. 8 The purpose of this meeting is to continue the 9 Subcommittee's discussion of the two-year trial use of the 10 1983 safety goal policy, the future use of the safety goal 11 policy, and the NRC staff work on development of containment 12 performance guidelines. Maybe I should say the purpose was, S 13 since I think in fact it will only be the last item on that (J 14 list that the staff was prepared to talk about today. 15 Dr. Savio is the cognizant ACRS staff member for 16 the meeting. 17 The rules for participation in today's meeting 18 have been announced as part of the notice of this meeting 19 previously published in the Feo0ral Register on Wednesday, 20 September 25, 1985. 21 A transcript of the meeting is being kept and 22 will be made available as stated in the Federal Register 23 notice. It is requested that each speaker first identify 24 himself or herself, and speak with sufficient clarity and () 25 volume so that he or she can be readily heard. i l ACE-FEDERAL REPORTERS, INC. l 202-347-3700 Nationwide Coverage '600 336-6646 l

5970 01 02 3 MPBeb 1 We have received no written comments or requests 2 for time to make oral statements from members of the public. I 3l I should note tha't this Subcommittee meeting is 1 4 being held really on the suggestion of the staff who have 5 been starting to work on containment performance criteria 6 and wanted to come in to -discuss I guess what you would call 7 their plan for approaching this. 8 Do any of the Subcommittee members wish to make 9 any comments or raised any questions? i 10 (No response.) 11 DR. OKRENT: If not, I will call on Mr. Ernst. 12 MR. ERNST: As this Subcommittee is fully aware, 13 the subject of containment performance design objectives is 14 not a new one. The ACRS, in its documents some few years 15 ago on safety goals, suggested that such a design objective 16 be pursued. Generally the staff has been sympathetic to 17 that. Realizing that there are significant uncertainties 18 and the source term work going on and things of that nature 19 led to the conclusion that it might be premature as of a few 20 years ago. 21 As part of the trial evaluation, the group in NRC 22 looking at the safety goals and evaluating the safety goals 23 again considered the containment performance design 24 objective and did suggest again I guess that such a design () 25 objective would be useful to pursue, and should be pursued ACE FEDERAL REPORTERS, INC. 202 347 3700 Nationwide Coverage 800 336 6646

i 5970 01 03 4 [' MPBeb 1 at some time. ) 2 There have been subsequent discussion I think 3 between the EDO staff and the ACRS and others, and the staff 4 has been actively considering the design objective. The 5 requirement to develop a plan was given to Research by the 6 EDO a month or so ago. I've forgotten the exact time. And } 7 we have indeed prepared such a plan in full cooperation with 8 NRR and with the cognizance of the EDO staff. And that is 9 the plan we have today. 10 The presentation will be basically in two parts. 11 I will give what I will call the programmatic aspects, what 1 l 12 we are doing, the kind of time frame that we are thinking 13 about, the various considerations that we think are 14 important to the development of a containment performance 15 design objective, and things of that nature. f 16 Jack Rosenthal, I asked him to give the more 17 technical discussion on the appendix to the plan which 18 covers basically the important aspects of the accident 19 sequence analyses that might bear on containment 20 performance, the physical aspects of containment 21 performance and source term aspects of containment 22 performance, and if the containment performance were i 23 described in terms of public risk in some way, the kinds of 24 uncertainties or problems that might exist in that kind of a () 25 formulation. ACE-FEDERAL REPORTERS, INC. 1 202-347-3700 Nationwide Coverage 800-336-6646 . _ _.. ~

a l 5970 01 04-5 The staff does not have any recommended MPBeb' 1 2 formulation at this time. As we will see in the 3 presentation, there are several possibilities. That does 4 not exclude other possibilities. The staff would be very 5 interested in the ACRS' viewpoints regarding possible 6 formulations or methods of attack on this particular 7 program. 8 DR. OKRENT: By way of comment, I might note, 9 since we lost one agenda item for this agenda, I would like 10 to shoot for finishing by five o' clock if that seems 11 plausible. 12 MR. ERNST: That seems very plausible. 13 (Slide.) 14 First, the basic purpose of the plan, clearly and 15 fundamentally the principal purpose, is to develop a 16 proposed design objective and guidance, and I stress e 17 guidance. I think one thing that was learned in the 18 safety-goal work over the past few years is that unless 1 l 19 people have a fundamental idea of how such a design 20 objective might be implemented by the staff, it becomes very l 21 difficult to comment. 22 It is easy to comment on the concept. I 23 Unfortunately the concept is not what is going to trip up I i 24 the containment performance design objective, so we intend O 25 intend to develop hand-in-hand not only the design objective ACE-FEDERAL REPORTERS, INC. 202-347 3700 Nationwide Coverage 800 3366646

5970 01 05 6 (*g MPBeb l but.also the method of implementation. \\_) 2 DR. OKRENT: By " guidance" I assume you mean 3 something like what is given in the Standard Review Plan for 4 Evaluation of the Reliability of an Auxiliary Feedwater 5 System? 6 MR. ERNST: That type of guidance. I think there 7 is a substantial question about the degree -- the 8 description that would be required in any kind of-- There 9 are several ways one could implement. One could implement 10 in very general terms, without any specific ideas of what 11 one might call acceptable versus non-acceptable. There's a 12 lot of flexibility, in which case a lot of prescriptive 13 guidance might not be necessary. 14 There are various ways of implementing, and since 15 there are considerable uncertainties for all the 16 calculations, not only of the threat but also containment 17 performance, I think that some fundamental guidance is 18 necessary as to how the NRC staff would treat the various 19 uncertainties. I don't think it can be as prescriptive as a 20 shopping list for a weekly trip to a grocery store, but I do 21 think that some general idea of how the staff would 22 implement is warranted. 23 The plan also indicates various periods of public 24 comment af ter that, which would include in the detailed () 25 schedule an opportunity for a public workshop where we could ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800 336-6646

5970 01 06 7 N 1 get explicit feedback. (v]MPBeb 2 We then propose to the Commission a design 3 objective if it appears to be feasible as an addition to the 4 safety goals. As we will see on the schedule, we are 5 talking a time frame of perhaps a year for this process to 6 take place. 7 Then evaluate through a trial use similar to the 8 safety goal trial use, and then assess the experience and 9 make a final recommendation to the Commission. That's the 10 general approach that we are thinking about. 11 (Slide.) 12 There is a portion of the plan that has,I think g ten items that might be important considerations in 13 G 14 developing such a design consideration. I took a look at 15 that and it seemed to be a mixture of things. We were in a i 16 reasonable hurry to put this document together. In my own f 17 mind I thought that these would coalesce perhaps into 18 attributes that one might use to measure a design 19 objective, so I collapsed them into basically four kinds of 20 attributes. 21 Clearly the principal attribute would be the 22 degree of protection of public health and safety that would 23 be provided. In general there are a number of thoughts that 24 come to mind under this kind of an attribute. (]) 25 One thought might be does the goal formulation ACE-FEDERAL REPORTERS, INC. 202 347 3700 Nationwide Coverage 800 336-6646

5970 01 07 8 MPBeb 1 provide some kind of minimum level of safety or minimum 2 level of performance of the containment? Is the goal 3 structured in such a way to at least maintain performance, 4 performance that seems acceptable, or some kind of a process 5 for reasonable improvement of containment if it looks like 6 there's some marginal performance? 7 If indeed it might pass some minimum level of 8 safety test, it strikes me that the protection of public 9 health and safety also needs to consider cost-benefit of 10 marginal improvements past that point. i 11 In considering protection of public health and 12 safety, one does need to consider the uncertainties in 13 calculating the performance of the containment and there 14 needs to be some consideration of to what extent does one 15 have to go conservative in the face of uncertainties, or 16 what. 17 Therefore, one could say an answer to that might l 18 be some kind of prescriptive kinds of analyses. Another 19 possibility might be where there is a fair amount of 20 uncertainty, it is the-amount of weight that they give to 21 that particular parameter in considering the containment 22 pe rformance. 23 Another aspect of protection of public health and 24 safety is the consideration of external events as well as (]) 25 internal events. There is a significant question that l ACE-FEDERAL REPORTERS, INC. 202-347 3700 Nationwide Coverage 800 336-6646

5970 01 08 9 MPBeb 1 perhaps the consideration of external events might have to 2 be approached somewhat differently than internal events, but 3 that would remain to be seen. I 4 The bottom line basically of protection of health 5 and safety as far as I'm concerned is to maintain the needed 6 defense in depth. In the ideal case it would strike me that 7 if you essentially performed the multiplication of core melt 8 frequency times containment performance times consequences, 9 you should be at a level' of risk that is significantly less i 10 than the public risk design objective in order to maintain a 11 defense in depth kind of concept. 12 The way the Commission has taken a position for a 13 long time is that you want some offsetting things such as 14 because of the uncertainties involved, if your core melt 15 frequency is actually significantly greater than you 16 calculate, there should be a margin in the containment 17 performance, and vice versa. 18 I'm not saying to what extent that needs to be 19 taken care of. If there is no uncertainty in the 20 calculations, then there would be no problem about defense 21 in depth, so I think that is a fundamental consideration for 22 the protection of the public health and safety. 23 As the second general attribute of the 24 development of the design objective, it seems to me that you () 25 need to consider the integrated performance of the ACE FEDERAL REPORTERS, INC. 202-347 3700 Nationwide Coverage 800 336 6646

5970 01 09 10 MPBeb 1 containment. A design objective that just considered /Asd 2 structural integrity, it pretty much goes without saying 3 that that would not be a satisfactory performance design 4 objective. 5 One needs to consider the engineered safety 6 features that are designed to reduce the containment loads, 7 the safety features designed to reduce the fission products, 8 and there needs to be a certain amount of emphasis on the 9 timing of containment failure because of the attenuation of 10 fission products as a function of time. So it seems to me 11 like that's an important kind of an attribute. 12 Flexibility seemed also important. It is clear 13 to me anyway that in implementing safety goals there needs 14 to be some flexibility in considering uncertainties. I 15 don't think you can just take the design objectives and turn 16 the cost-benefit crank and come out with an answer that 17 clearly makes your decisions for you. 18 There needs to be built into the design 19 objective, as there is in the safety goals, a chance for 20 reasonable flexibility, reasonable kinds of engineering 21 decisions and public policy kinds of decisions. 22 This kind of flexibility might occur fn several 23 places. One would be to permit trade-offs between accident 24 prevention and consequence mitigation. As long as you don't f'T i 25 exceed the minimum level of performance as indicated above, s) ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336 6646

5970 01 10 11 g^) MPBeb 1 there could be some flexibility and greater emphasis on the v 2 more toxic fission products, small versus potentially large 3 releases. 4 What I mean there is that there might be an 5 option that says let's vent Mark 1 containments for example, 6 realizing that that would result in a limited public 7 exposure, the effect of which would be to keep the 8 containment intact, providing that'you-can then close the 9 vent when the time comes to close it. That is'a balance, 10 and more than just an insignificant kind of decision. 11 Thet, gain in flexibility I think uncertainties 12 plays a role. You have to have some flexibility as to how 13 to handle the various uncertainties in a given situation. 14 The final general attribute it seemed to me would 15 be one of efficacy, and that has several elements I think, 16 one being understandability, sufficiently understandable to 17 the public if they know what is going on, but sufficiently 18 meaningful and understandable to the engineer that they know 19 what to do with it. 20 Perhaps where you bridge the gap is the guidance 21 I'm talking about. If you have a design objective that is 22 understandable to the public, you can have guidance that is 23 understandable to the engineer. 24 It certainly has to be implementable. I think (]) 25 the test of that is whether or not you can write a i ACE-FEDERAL REPORTERS, INC. 202-347 3700 Nationwide Coverage 800-33 4 646

5970 01 11 12 /' MPBeb 1 guidance document that makes sense. %,]/ 2 And one has to consider the costs of 3 implementation. That is not the costs of backfit or 4 anything else, but you need to consider the costs of 5 actually doing the job, and how you are going to use that 6 number in real-life regulation if the use of the objective, 7 for example, would require a full-scale DRA on every plant 8 and the safety goals have already been structured in such a 9 way as to not require that. 10 So one needs to consider at least the necessary 11 resources to reasonably implement it. 12 DR. KERR: That seems to me to be a good, 13 representative group of things that need to be considered. 14 The impression I get is that to some extent, you want 15 perhaps a list of these in a mode of thought which looks at 1 16 the way in which existing containments perform. This makes 17 sense. 18 It also seems to me that one could ask how should 19 containments perform if we didn't have any containments. 20 That's a little too exaggerated because, as you say, the 21 thing has to be implementable. 22 Am I correct? Does this sort of come f rom a 23 position which is considering existing containments? 24 MR. ERNST: Yes and no. The Yes from the (]) 25 standpoint that in considering how you would approach the ACE-FEDERAL REPORTERS, INC. 202 347 3700 Nationwide Coverage 800-336-6646

5970 01 12 13 MPBeb 1 problem analytically and perhaps from a decision-making 2 standpoint, if indeed this objective is to have any utility 3 for operating plants, it seems to me like ene needs to at 4 least consider the various aspects that are important with 5 regard to presently designed plants. 6 DR. KERR: Yes. 7 MR. ERNST: That I think is the context that I 8 would consider present plants. 9 The flexibility comes about in trying to develop 10 something that makes sense and, if actually implementing in 11 an operating plant, could be implemented in some kind of a 12 way that makes good common sense. If you just have a 13 go/no-go on a number and say you have to meet it, and you go O 14 out to an operating as-built plant and find out you miss it 15 by a factor of two with no flexibility to say "Well, what is 16 the importance of that miss," then I think that would be 17 improper. 18 DR. KERR: This I think implies a rather thorough 19 understanding and perhaps a rather detailed analysis of 20 various existing containment systems. 21 MR. ERNST: Yes, I think we would draw heavily on 22 the severe accident work that is 'on-going now with the six 23 reference plants. This is one reason why we have a schedule 24 that is pretty much patterned after the results we expect to {} 25 get in on the six reference plants. ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336 6646

5970 01 13 14 /m MPBeb 1 DR. KERR: Mow suppose in the course of this O 2 investigation it should develop that either some existing 3 containment or some existing containment with slight .1 4 modifications was clearly superior to other kinds of 5 containments. Are you prepared to deal with that, or do 6 you think that is so unlikely that one need not worry about 7 it? 8 MR. ERNST: I would hate to judge the likelihood 9 of that until we get the results in on the reference 10 studies. 'The intent of this effort would not be to carry on 11 and say -- make judgments on a beauty contest between 12 containments or something like that. I think what we are 13 after here is some kind of an approach that makes sense in 14 which to measure containment performance. And you really 15 also need to measure it against what is needed. 16 For example, if all containments happen to meet 17 the necessary health and safety standards it becomes a 18 lesser important question if one or more happened to meet it 19 be t te r <. But it is a matter that would certainly come to the 20 fore. You don't have to do this to have it come to the 21 forefront, though, because the reference plant analyses 22 would show the same kind of thing. 23 DR. KERR: I'm not sure how we would derive the 24 conclusion but it does seem to me if one did reach a (]) 25 conclusion that clearly one type is superior, at least from ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverase 800 33(Hi646

5970 01 14 15 (V') MPBeb 1 the viewpoint of the staff, that it would be unfortunate not 2 to say that. 3 That to me doesn't mean that you force anybody to 4 build containments like that, or even that you are 5 necessarily correct. 6 MR. ERNST: I think that the work here and the 7 work on the six reference plants would indicate whether or 8 not it appeared that one containment type happened to be 9 superior to another containment type. I think the 10 analytical work would certainly indicate that. 11 MR. ROSENTHAL: Mel, may I interject? ,, 12 MR. ERNST: Sure. 13 MR. ROSENTHAL: My name is Jack Rosenthal, from 14 the Reactor Systems Branch, DSI, NRR. 15 We were mindful of the prior ACRS comments which 16 said that we should focus on what it ought to be as well as 17 how plants may perform today. We don't necessarily intend 18 to set a containment performance objective that we know all 19 the plants will meet. 1150 and the reference plant work 20 will give examples of how real plants measure up to proposed 21 CPO. 22 But that is not to say that we would expect that 23 we would adjust the CPO such that all plants meet it but, 24 rather, that there is this concept also involved of what it C 25 ought to be. ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800 336 6646

5970 01 15 16 MPBeb 1 By the same token we did recognize that if we 2 write a containment performance objective which we think is 3 honest and fair to the most robust of plants, then we may 4 well orphan existing older plants, and no decision has been 5 made as of this date not to do that. 6 7 8 9 i l 10 l 11 12 13 ';O 14 15 16 17 18 19 20 21 22 23 24 25 ACE-FEDERAL REPORTERS, INC. 202 347-3700 Nationwide Coverage 80 4 336 6646 l

5970 02 01 17 MPBagb 1 MR. ERNST: I think probably a good example of 2 that is the 10 to the minus-4 core melt goals is something 3 the Staff should know well but in DRA's there are 4 indications that some plants might not make that. It could 5 well be the case in this situation also. Not knowing what 6 the objective is and also not knowing what the reference 7 plant analyses may turn out to be, I have no way of knowing 8 what the answer might be. 9 DR. REMICK: I have a question: realizing that 10 the current safety goal is public-risk oriented and core 11 melt design objective is kind of a secondary type of 12 consideration, has any thought been going in that whatever 13 is developed here than can be developed should be somehow f3U 14 consistent with the overriding risk goal? 15 MR. ERNST: Clearly, clearly it would be part of 16 the safety goals. And I would certainly hope it would be 17 consistent with the safety goals. 18 I think where we are coming from on this is no 19 different than where we were coming from in a trial 20 evaluation and even prior with regard to implementation of 21 safety goals in general. There are substantial 22 uncertainties and there was a very significant question of 23 do you just take some best estimate and come up with a 24 number and if that meets some public risk design objective (v'i 25 is that sufficient? / ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336 6646

~ l 5970 02 02 18 /~ MPBagb 1 I think the answer has always been not ( 2 necessarily. Because of the uncertainties you have got to 3 understand the physics and the chemistry and everything else I 4 associated. And the natural flow of regulation would say 5 where there are uncertainties in knowledge, one tends to be 6 a little more conservative in picking a so-called best 7 estimate. I think it is that kind of a thing. And I think 8 in the trial evaluation and the safety goal evaluation again 9 there is a reflection of concern about uncertainties and 10 transforming that concern into some kind of band of numbers 11 that might be useful from the standpoint of implementation. 12 I would expect that this would follow along about the same 13 kind of a process. 14 DR. REMICK: So there would be intent to keep it 1 15 generally consistent? 16 MR. ERNST: Certainly. 17 DR. REMICK: Unless those numbers were changed. 18 MR. ERNST: Yes. 19 DR. REMICK: Unless the Commission changed them. 20 MR. ERNST: Yes. Clearly when you are talking 21 uncertainties, if I had a blackboard here I think it would 22 be fine if you took a mean estimate of core melt frequency, 23 a mean estimate of containment performance and then 24 multiplied those two together to come out with the public (]) 25 risk number. ACE-FEDERAL REPORTERS, INC. 202-347-3700 - Nationwide Coverage 800-336 6646

5970 02 03 19 MPGagb 1 Because of the uncertainties, it is not clear to 2 me that you wouldn't want some margin in the containment 3 performance as well as some margin in the core performance 4 in meeting that public risk standard, and that's what I 5 meant by there is some room for tradeoffs. But there ought 6 to be some minimum level of performance beyond which you 7 stop trading off. The tradeoff becomes pretty much the 8 margin that you maintain for the defense-in-depth concept. 9 DR. REMICK: Maybe I'm jumping ahead but in the 10 tradeoffs I think I can agree with that philosophy, but it 11 is conceivable that somebody might argue that they have, you 12 know, a very, very good containment. Can they back off in 13 that tradeoff with any kind of a minimum core melt O 14 frequency? 15 MR. ERNST: I think I have already identified 16 that as one of the things that needs to be considered. 17 DR. REMICK: Okay. 18 MR. ERNST: -- as an attribute. And I think what 19 we are trying to set up here is not go/no-go kinds of 20 processes but sort of figures of merit to measure various 21 proposed alternatives. There are some proposed alternatives 22 here which I think -- 23 DR. REMICK: Yes. 24 MR. ERNST: We can chat about that in a minute. 1 () 25 This is basically how it came out. And I think in the ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336 6646

5970 02 04 20 MPBagb 1 final evaluation plan we will arrange something more like 2 this where there are actually some attributes to look at a 3 figure of merit for various candidates. 4 (Slide.) 5 MR. ERNST: There are some what one might call 6 special issues associated with this. Whether that is the 7 best terminology or not, I don't know. There are 8 containment bypass events and I think there is a question in 9 implementing containment performance design objectives of 10 how one might treat leaving a door open or a valve open or 11 something like that. So I just listed it as an item that 12 should be considered. 13 The extent of reliance on predictions of accident 14 phenomenology is just anpther way of speaking about certain 15 kinds of uncertainties involved with the calculations and it 16 is really getting at the degree and type of prescription 17 that might be necessary in making the analyses. There is 1 18 the question that was generally touched upon a few minutes 19 ago, that is, whether there should be a different design 20 objective or perhaps a different scheme for implementation 21 for new versus old designs. 22 In this regard as I recall -- and correct me if i 23 I'm wrong -- I think the ACRS in their documentation 24 suggested the containment performance objective for new '() 25 designs or for CP -- new CP's, construction permits, and ACE-FEDERAL REPORTERS, INC. 202 347-3700 Nationwide Coverage 800-336 6646

5970 02 05 21 (V*] MPBagb 1 loft sort of open what one might do for operating plants. 2 So I think we had the same kind of basic consideration -- 3 and I call it flexibility, other people might call it other 4 things. 5 DR. REMICK: Now in answering that question, 6 wouldn't it be dependent on also what ultimately is done 7 with the safety goal, whether it is a policy statement or it 8 is a rule? 9 MR. ERNST: Clearly it would not be a rule, at 10 least in the current process for treating safety goals. And 11 this is intended to be a supplement to the addendum of the 12 safety goals at some time. 13 DR. REMICK: So your question is even if it is a O 14 policy statement whether there should be a different 15 objective for existing versus new plants? 16 MR. ERNST: Yes, you could either have a 17 different objective or a different approach, a different 18 degree of flexibility or whatever you want to call it. But 19 I think clearly since we fundamentally had just operating 20 plants or plants so far along in construction that you can't 21 do much from the basic design standpoint, for the near term 22 the real application would be for operating reactors. 23 The signal might be for future generations of reactors for 24 designers. (]} 25l But you know, it depends on what this is for. If-ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336-6646

5970 02 06 22 MPBagb 1 it is just a signal for the future, then that is one deal. 2 If it is going to be applied to operating reactors then 3 there are other kinds of considerations. 4 MR. MICHELSON: I need a couple of clarifications 5 on your slide. 6 The case wherein you have a loss of coolant 7 outside of containment and thereby have to isolate, is that 8 in your containment bypass events? 9 MR. ERNST: The so-called event V where your 10 failure is -- 11 MR. MICHELSON: Yes, is that considered 12 containment bypass? 13 MR. ERNST: No, not in the standard -- O 14 MR. MICHELSON: Because you're not needing 15 containment, you just need to isolate the break really. 16 MR. ERNST: Yes. 17 MR. MICHELSON: Is that in your containment -- 18 MR. ERNST: Event V's and containment bypass are 19 usually considered to be two different things: one is an 20 opening in the containment and the other is where you get 21 your failure of the isolation valve that goes and fails a 22 low pressure system and that causes not only the violation 23 of containment but also causes core melt. 1 24 MR. MICHELSON: The initial event could, of G(~T 25 course, be a failure of reactor water cleanup on a boiling ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336 4 646

5970 02 07 23 MPBagb 1 water reactor, for instance, outside of containment that 2 necessitates appropriate isolation. 3 Is that a part of your containment performance 4 objectives? 5 MR. ERNST: Well I think that would be an 6 interface between a reactor accident -- it should be 7 considered between reactor accident and containment 8 performance itself, namely, the appropriate operation of the 9 isolation valves. That really more affects the accident 10 probably. It is included. The simple answer is yes. 11 MR. MICHELSON: Because these are containment 12 isolation valves but the objective in this case is not the 13 normal concept of containment -- not really isolation of a 14 break but that would still be a part of your consideration 15 here? 16 MR. ERNST: Yes. 17 MR. MICHELSON: The other question is the 18 suppression cool bypass, is that a consideration here? 19 MR. ERNST: Yes. 20 MR. MICHELSON: And where is it covered and how 21 is it covered? 22 MR. ERNST: We will let you know in three or four 23 minutes. 24 MR. MICHELSON: Okay. But it will be in here 25 somewhere? ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800 3364646

5970 02 08 24 MP3agb 1 MR. ERNST: Yes. 2 MR. MICHELSON: Thank you. 3 MR. ERNST: Jack will have a few comments on 4 this. Clearly -- I don't want to steal any of his show but 5 clearly you are going to have to select -- well one approach 6 would be to select certain accident sequences as an 7 analytical tool to go in and measure performance. Another 8 approach might'be to consider every accident sequence, but I 9 think you have to be a little more prescriptive and you have 10 to decide, you know, what are the major sequencee that cover 11 the basic performances of the containment and use those as 12 basic tools for analyzing. 13 But I may be stealing some of your stuff. Cs) 14 (Slide.) 15 We did provide in the plan nothing really novel. 16 We had to fill in something so we did. We've got four 17 possibilities. There are probably dozens if you want to sit 18 down and consider permutations and things like that. 19 Clearly there is the ACRS proposal. There is a thought 20 about continuation with a certain degree of confidence for 21 isotopes that could do substantial long-term damage to the 22 environment. There is an option that talks about reduction 23 as a function of time, that might be a possibility. 24 The fourth option I think is sort of getting at ('j) 25 what Forrest was maybe talking about. You might have the ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800 336-6646

5970 02 09 25 MPBagb 1 probability, like for example I think what was quoted in 2 there was like a 10 to the minus-6 or something is desired 3 for a large scale release -- not condition on core melt 4 frequency or anything else but sort of a general thing. 5 And as a matter of interest if you had a large 6 scale core release, about a 10 to the minus-5 frequency 7 would be roughly comparable to the prompt design objective 8 for s3fe y goals. So if you said 10 to the minus-6, that 9 woud give some room for uncertainties and some room to play 10 around with tradeoffs between containment performance and 11 core melt as an example. 12 Then you might say but I also want the 13 containment -- which is similar, I think, to the ACRS ! () 14 proposal -- a minimum of good containment behavior for at 15 least 9 out of 10 core melts, then you could make tradeoffs 16 between those two numbers depending on how good the 17 estimate of core melt frequency was. If the core melt 18 frequency is less than 10 to the minus-4, you would want 19 pretty good assurance that the containments would fail 20 significantly less than one out of 100 times. If you had 21 good core melt, 10 to the minus-5 or less, then you might be 22 more willing to accept the one in 10 minimum kind of a 23 deal. So that's another possibility. 24 MR. MICHELSON: Is there going to be some (]) 25 definition of " failure?" l ACE-FEDERAL REPORTERS, INC. 202-347-3700 - Nationwide Coverage 800-33(H5646

~ 5970 02 10 26 MPBagb 1 MR. ERNST: I think the definition of " failure" 2 would be inherent in the various kinds of schemes. 3 For example, in the ACRS proposal they talk about 4 "large release" -- and I forget the exact words -- but that 5 was defined in terms of percent of inventory. I think each 6 of the-items there would have to define failure in some 7 way. 8 MR. MICHELSON: It would be some threshold above 9 which you now have failure. 10 MR. ERNST: Correct, whether it's in terms of 11 structural strength or cleanup, attenuation of fission i 12 products or a certain amount of leakage or what there would 13 have to be some basis. 14 MR. MICHELSON: But that will be a part of your 15 proposal? 16 MR. ERNST: Right. 17 DR. REMICK: Now the number four is one that 18 personally intuitively appeals to me. And I think you said 19 something that addressed my earlier question. Did you say .i 20 that if you have a large scale release of about 10 to the 21 minus-5 that's probably equivalent to the pump fatality 22 design objectives? 23 MR. ERNST: Roughly. 24 DR. REMICK: I see. Okay. (} 25 MR. ERNST: Roughly. l l ACE-FEDERAL REPORTERS, INC. l 202 347-3700 Nationwide Coverage 800-336 6646 L

5970 02 11 27 MPBagb 1 DR. REMICK: That does answer my earlier 2 question. 3 MR. ERNST: Yes. I think the number four is more 4 closely allied to the approach that is taken in safety goals 5 than perha the others. 6 DR.'REMICK: Yes. 7 MR. ERNST: Not to say that the others aren't 8 consistent. I'm just saying in our thinking process number 9 four is probably the closest. 10 Again I would like to stress that if there are 11 any thoughts on alternative approaches, we would much 12 appreciate hearing about it. As you see in the schedule, we 13 are anticipating a fair amount of interface with the 14 Committee and we would be vitally interested in any feedback 15 during the formulation process. 16 (Slide.) ~ 17 The overall plan as far as how we see this going, 18 the thoughts are that we would establish an interoffice 19 working group. We essentially have that now, with Jack 20 being basically the central focal point in NRR and 21 P.K. Niogi on my staff being the principal contact on the 22 effort. 23 We would also anticipate establishing some form 24 of what is called here a senior review group, whatever you {} 25 want to call it, with some kind of management overview at I ACE-FEDERAL REPORTERS, INC. l 202-347-3700 Nationwide Coverage 800-336-6646

5970 02 12 28 MPBagb 1 various phases of this to keep management informed of what 2 is going on. 3 The intent is not to require more research, the 4 intent is to develop a design objective based on extant 5 information which would draw heavily on the on-going 6 reference plant analyses. 7 MR. MICHELSON: Is there an intent to look at 8 pressurized water reactor containments separately from 9 boiling? 10 MR. ERNST: Whether it will be separately or 11 not -- 12 MR. MICHELSON: Would there be two sets of 13 performance objectives? (_/ s_ 14 MR. ERNST: I really can't address that. My gut i 15 feeling, personal feeling, would be no, that you would have 16 to have one set that would adequately encompass different 17 kinds of designs. When you get into it you might find that 18 for certain formulations you may not be able to go that 19 route, I don't know. But my gut answer to that would be no, 20 you would have -- 21 MR. MICHELSON: You are trying not to, in other 22 words. 23 MR. ERNST: That would be my gut feeling. 24 Jack, would you -- ( l () 25 MR. ROSENTHAL: Yes, sir. That is not to say ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-33H646 -__ =_ _- - -- - _ _ _

5970 02 13 29 <-s MPBagb 1 that we don't intend in a trial use period to measure the (_) 2 six reference plants against possible formulations of the 3 CPO and see how they stack up. We will surely consider 4 various plant types but as far as a simple communicable 5 statement to the public, the gut feel there would be -- 6 MR. MICHELSON: If you found that dry 7 containments were really the way to go for boiling water 8 reactors, would you say so? 9 MR. ERNST: Well I think we tried to answer that 10 question earlier. 11 MR. MICHELSON: Yes. I have a little more blunt 12 question, it is the same thing. 13 MR. ERNST: You won't get any more blunt answer. O 14 MR. MICHELSON: I see. 15 MR. ERNST: No, I really think the results we get 16 from the reference plant analyses which feed into NRR and 17 their judgments on the severe accident question for all 18 operating plants and this -- if there is such a thing as you 19 described, it would clearly come out of the analyses. What 20 you do about it, I am not addressing. 21 MR. MICHELSON: That will clearly come out if you 22 look for it. If you, however, are looking on boiling water 23 containments and you only look at the suppression pool 24 containments, you don't come up with it unless you purposely (} 25 look for the possibility of dry containments. And are you ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800 336-6646

5970 02 14 30 MPBagb 1 going to look for the possibility of dry containments on 2 boiling? 3 MR. ERNST: I don't think the intent would be to 4 look for new kinds of designs. 5 MR. ROSENTHAL: No. 6 MR. ERNST: The intent would be to see what makes 7 sense from the health and safety standpoint and then to see 8 what makes sense -- 9 MR. MICHELSON: Okay then I guess you just 10 answered my question, because simply if you are not going to 11 look at the dry containment in boiling water reactors then 12 you only can look at one answer and that is suppression 13 pools for boiling water reactors. That's all you can look 14 at. 15 MR. ROSENTHAL: Well this isn't a mitigation 16 study type exercise. 17 MR. ERNST: The severe accident work on the six 18 reference plants can get at this kind of question that you 19 are raising. What I am saying is between -- this work will 20 be compatible to safety goals with the reference plant 21 analyses. This effort will draw from the reference plant 22 analyses to the extent needed in order to come up with a 23 good implementation plan. I think what you are looking for 24 is the results of the reference plant analyses in NUREG 1150 l l () 25 and subsequent work. ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336 6646

5970 02 15 31

• ] MPBagb 1

MR. ROSENTHAL: Let me point out that if we just v 2 look at our GESSAR work and some of the reference plant work 3 that's coming in I think that part of the -- and forget 4 about the Mark I's and II's because whatever number of them 5 are built they are already in existence -- and what you will 6 find in the boilers is that as far as containment is that 7 you will expect to release the noble gases and for many 8 sequences little else. For the PWR's, the large dry 9 containments you will expect to retain -- for many sequences 10 there will be no failed sequences in which you will contain 11 the noble gases as well as the heavier isotopea. And that 12 will be a marked difference than between the boiler 13 suppression pool and the superlarge dry very strong. And 14 now you have to ask yourself what is your tolerance to 15 release of the noble gases. But that will be in my view a 16 marked difference between the two advanced designs. 17 DR. KERR: I guess I don't see how results from 18 the reference plants would give you much of an indication on 19 how good a large dry containment would be for a BNR since 20 none of them have that; nor will they give you any 21 information about how good suppression pools would be for 22 BWR's -- I mean for a PWR. I do not have a preconception 23 about whether there is a best containment and I was 24 interested, as Karl was, as to whether you are going to try (]) 25 to look for that. ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336 6646

5970 02 16 32 /'~gMPBagb 1 MR. ERNST: Okay. What I am getting at is the V 2 six reference plant analyses will get at the beauty contest, 3 so to speak, of one type of containment versus another type 4 of containmcnt, whether or not anything needs to be done 5 will be addressed also in the risk reduction kind of 6 options. I will agree though as far as I know there is no 7 risk reduction options for BWR's that says put them in a 8 large dry. So the specific answer to your question, I 9 guess, is no. 10 I don't think that means no in general though 11 because if you have substantial weakness that none of the 12 other options are going to take care of then I think one i 13 would then start looking around for other possibilities. As 14 Jack says, for future plants it doesn't seem like this is 15 going to be the case. 16 DR. KERR: Well to be, I guess, slightly 17 ridiculous, I have heard analysts say if I can maintain 18 containment integrity for the first 20 hours I don't have 19 anything to worry about. Now just suppose that one 20 particular containment type admitted ones could make that 21 sistement with a fairly high confidence, it seems to me that 22 would be a pretty good containment type and there are a lot 23 of caveats here. It is that kind of thing though that I am 24 -- it would seem to me that one would want to explore. (]) 25 MR. ERNST: That's right. And the chart is gone ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336 6646

5970 02 17 33 MPBagb 1 now but clearly at least one or more of the alternatives 2 discussed does take that into consideration. 3 DR. KERR: Yes, that was the impression I got. 4 MR. ERNST: In fact, I would say all of them take 5 that kind of thing into consideration. 6 DR. REMICK: Bill, there was something I failed 7 to see about your question. It seems to me that would be 8 very important information, but what do you do with it from 9 a design objective? Would you dictate everybody in the 10 future -- 11 DR. KERR: No, you would say our results indicate 12 that this containment has these very desireable 13 characteristics and it is there in the public domain. And 14 if people think that they don't need them or that they don't 15 care about them, maybe. 16 DR. REMICK: That wouldn't affect the design 17 objective. 18 DR. KERR: No -- I don't know whether it would 19 affect the design objective or not, it depends on what 20 design objective one finally formulates. It's possible that 21 one might reach a design objective that says new plants 22 would require it. At this point I doubt it, but it's 23 possible. 24 (]) 25 ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800 336-6646

5970 03 01 34 MPBeb 1 I was looking more as incidental. But, 2 nonetheless, it seems to be fairly valuable information if 3 it develops. 4 DR. REMICK: I agree with that. 5 MR. ERNST: Okay. 6 In developing the guidance regarding 7 implementation -- and Jack will be covering this in more 8 detail -- it seems to me like there are several pertinent 9 aspects to that. I do think you need to select the kinds of 10 sequences that would be analyzed for the containments to try 11 to narrow the analytical effort to some reasonable degree, 12 in any event. 13 There would be need, I think, at least in some O 14 minimal way, to indicate the kind of modeling assumptions 15 that are appropriate for the accident sequences, the 16 accident progression, the source term calculations, 17 containment behavior under certain loads, and consequence 18 calculations. 19 MR. MICHELSON: Will the accident sequences still 20 be based on doubled-ended ruptures of largest pipes, and 21 that sort of thing? 22 MR. ERNST: That hasn't been very risk-dominant 23 in PRAs. 24 MR. MICHELSON: No; it has been a fairly low /~' 25 probability. %s} i ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800 33H64

5970 03 02 35 MPBeb 1 MR. EENST: I think you would have to consider 2 existing containment requirements and things of that nature. 3 MR. MICHELSON: I was thinking in terms of the 4 leak-before-break concept. You won't get any kind of thing 5 but leaks: that's the biggest break that you look at, a 6 leak? 7 MR. ERNST: Your transients and your small break 8 LOCAs are big contributors to this. 9 MR. MICHELSON: A leak is not a small break LOCA 10 in normal analytical terms. 11 MR. ERNST: Yes. 12 MR. MICHELSON: LOCAs may almost disappear if we 13 ever get to the point where we believe, truly believe in the 3 u) 14 full break. 15 MR. ERNST: It is not clear to me that small 16 breaks would completely disappear. 17 MR. MICHELSON: Small breaks are hundreds of 18 gallons a minute. Thousands. 19 MR. ROSENTHAL: The advantage of doing a large 20 break LOCA in, let's say, the BMI context is that it answers 21 some of the questions about, "Well, wait a minute, how bad 22 do you think it will be?" 23 MR. MICHELSON: Yes. 24 MR. ROSENTHAL: With respect to containment () 25 performance for the transients and the small break LOCAs you l i ACE-FEDERAL REPORTERS, INC. i 202-347-3700 Nationwide Coverage 800 33 6

5970 03 03 36 MPBeb 1 expect some primary system retention as you melt the core, 2 the primary system retention of fission products. And it is 3 of some interest to test the design by saying " Wait a 4 minute, I don't want to account for that. Let me put all 5 the fission products in a containment and then say what the 6 containment would withstand." 7 Now, it hasn't been decided yet which way to go. 8 But the interest in the large break LOCA phenomenologically 9 is not because you expect it to occur but because of its 10 challenge. 11 MR. MICHELSON: I think steam line break is a 12 fairly large contributor; is that right? 13 MR. ROSENTHAL: No; steam line breaks, sir, are 14 comparable to-- 15 MR. MICHELSON: Primary system? 16 MR. ROSENTHAL: Yes; but the resultant 17 transients, as far as loss of decay heat removal which give 18 you -- they would be more benign with respect to fission 19 product retention than the large break LOCA. 20 MR. MICHELSON: Yes, sure. 21 Well, you're going to worry about it later, I 22 guess. 23 MR. ERNST: Yes. I think the intent here is to 24 cover the reasonable ranges of possibilities. (]} 25 I think an integral part of what we are doing, I ACE-FEDERAL REPORTERS, INC. - 202-347-3700 Nationwide Coverage 800-33MM6

5970 03 04 37 MPBeb 1 called it uncertainty analyses. And I expect that is the 2 wrong choice of words if there are any statisticians in the 3 room. Maybe I really meant sensitivity analyses rather than 4 uncertainty analyses. But clearly there needs to be 5 analysis of uncertainties by numerous statistical 6 uncertainty analysis procedures; which I don't think are 7 going to work very well because nobody knows how to do it. 8 Or the straight sensitivity analysis. 9 Then I think there may be a separate question of 10 should the approach here be any different for external 11 events than internal events. Then I think there should be 12 some indication of the decision analysis that might be used 13 in implementing the guideline. O 14 DR. REMICK: I'm not sure I fully understand 15 where we'll come out of looking at the six plants. Will you 16 come out with a feeling of some kind of containment failure 17 probability for the six plants? What will you get from the 18 six-plant study? 19 MR. ERNST: I think what you get from the 20 six-plant study is, you have gone essentially from soup to 21 nuts, you have looked at the accident sequences and plant i 22 damage states that might be important to containment 23 performance, you have looked at the physical walk-throughs 24 of the containment event trees to see how, for various plant ('} 25 damage states, the containment actually gets threatened and ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage - 800 336 6646

l 8 5970 03 05 38 MPBeb 1 performs. You have looked at the source term generation 2 from basically a chemistry standpoint to see what kind of 3 source terms get generated. And you get, through the 4 containment failure mode and the accident sequence, an idea 5 of the risk from those kinds of things. 6 So you do, as a function of each major 7 containment type, get an awful lot of insight as to how the 8 accident progresses and how the containment is threatened 9 and performs. 10 I think you need to draw on that rather large 11 wealth of information just for the technical insights that 12 you get. 13 MR. ROSENTHAL: I think that one could sit down 14 here and write a containment performance objective. You 15 wouldn't know how good it was. And what you do with the 16 reference analysis is that you test a proposed CPO, we will 17 have a proposed CPO prior to the reference analysis being 18 completed. But what we would like to do is test that 19 containment performance objective against the reference 20 analyses at some point before we go forward. + 21 MR. ERNST: Yes. I 22 MR. ROSENTHAL: A utility is-- 23 DR. REMICK: Can that be done during the trial 24 period? (} 25 MR. ERNST: I was just going to say it may or may ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336-6 686

5970 03 06 39 MPBeb 1 not: it depends on how complicated. I think one would get ^ s 2 some good insights about whether one were reasonably on the 3 right path in the kind of time frame we are talking about 4 for the development of the proposed goal. 5 So at least we would do that. 6 Whether we would actually, as a full trial or 7 not, I don't know. We might do the full deal during the 8 trial period. 9 DR. KERR: I assume that you have a good liaison 10 with the people who are going to do the results, going to 11 1150. Because at least from my understanding of the 12 presentations with 0956 it was not clear to me how the 13 containment modeling would be done in 1150. If one used a 14 BMI 2104 approach, for example, you would simply find the 15 failure pressure. If you don't use that, then perhaps the 16 people who do 1150 are going to develop contaimment 17 performance models. Is that what one looks to? 18 19 20 21 22 23 24 f'T 25 \\_) ACE-FEDERAL REPORTERS,.INC. 202-347-3700 Nationwide Coverage 800-336-6646

5970 04 01 40 MPBeb 1 MR. ERNST: Well, there was a containment 2 performance event tree for every one of the six reference 3 plants, for the physical processes. Then there were the BMI 4 runs on the source terms that should be reasonably 5 sufficient for the various physical characteristics of the 6 progression for each accident sequence. So those two things 7 are pretty well married. 8 DR. KERR: Is this event fault tree combination 9 in agreement with the containment performance working 10 group's results, or are they identical? 11 MR. ERNST: Yes. The containment performance 12 working group's results did net fill in all of the blanks on 13 the containment event tree, so in many aspects, judgments () 14 had to be made as to a number of the branch points, like 50 1 15 or 55 questions at the top of this event tree. And each one 16 has several branches. 17 I think if you work it all out there are 18 something like 10 to the 24th possibilities, which gets 19 pretty complicated. 20 9R. KERR: In effect, the containment modeling 21 now exists and what 1150 will accomplish is a response or a 22 model to the loads. Is that the additional information we i 23 will get out of the analyses of going to 11507 24 MR. ERNST: Well, there are certain assumptions (]) 25 on physical failure characteristics of containments that 4 l ACE-FEDERAL REPORTER', INC. S 202-347-3700 Nationwide Coverage 800 336 6646

r 5970 04 02 41 1 are built into containment event tree models, if I am ("] MPBeb V 2 understanding your question. 3 DR. KERR: My version was that I've heard -- and 4 that's the reason that I'm really asking -- that there now 5 exists a set of containment event trees. 6 MR. ERNST: That's right. 7 DR. KERR: So in that sense you already have 8 modeled containment performance for those representative 9 plants. What new information you will get as you do the 10 work that goes into 1150 is how this set of containments 11 responds to the accident sequences that will be studied. 12 Am I understanding-- 13 MR. ERNST: The event tree itself goes to 14 containment failure and it can take a number of paths. The t 15 containment threat and performance are all built into the 16 containment event trees. 17 Then, as a separate exercise, one takes typical 18 paths and takes the dominant accident sequences that starts 19 all of this and meshes those together and decides what BMI 20 runs do I need to run to see what the source term is for 21 that physical process that happens. 22 And then from that you go to risk, and 1150 will 23 take that, integrate it from the front end through the back 2* end, and not only look at the best estimate of what happens (') 25 but do uncertainty types of analyses to get at the variance x_s l ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-33MM6

5970.04 03 42 / MPBeb 1 of what happens also. ( 2 DR. KERR: So it is not unreasonable. I'll wait 3 to see. 4 MR. MICHELSON: In the past, containments have 5 been protected against the possibility of large break LOCAs 6 through missile shields and absorbing hardware of various 7 sorts so that the pipe was not breached during penetration 8 of the containment. These various devices in certain cases 9 are now being removed. 10 Now if you should experience a low probability 11 break such as a large break LOCA,'you may penetrate the 12 containment at the point -- at the time of the break, and 13 you may also inactive significant numbers of engineered O 14 safeguards which happen to be just outside of the 15 containment. 16 Has all of that been accounted for now in some 17 new thinking that still says large break LOCAs are not of 18 consequence? You can still perhaps draw that conclusion, 19 but I think until you have done your homework, gone back 20 through such analyses, now based on the absence of 21 restraints and so forth, we don't know what large break 22 LOCAs might do. You know, it is a very plant-specific 23 question. 24 MR. ERNST: Whatever work is being done is being (]) 25 done basically on a competing risk kind of process. ACE-FEDERAL REPORTERS, INC. 202 347-3700 Nationwide Coverage ~ 800-3364646

5970 04 04 43 / 1 MR. MICHELSON: Yes, but I'm wondering if the ( w MPBeb 2 people who are designing the risk are following the risk, 3 you know, doing the full examination of the possible 4 consequence of pipe break in terms of penetrstion of the 5 containment at the same time as the loss of engineered 6 safety features. 7 MR. ROSZTOCZY: Zoltan Rosztoczy, NRR. 8 I think you are bringing up a good point. There 9 are two parallel efforts going on. One of them is the 10 review of whether there is a need for the restraints for the 11 design basis accidents. 12 In parallel with that we are evaluating a severe 13 accident right now hopefully to prevent the bad core sample ^ 14 avenue. So I think it would be important that the specific 15 accident evaluation covers both cases, the case where you do 16 have the restraints there, as many of the plants do have 17 them, and also where there are no restraints there, to cover 18 future plants or to cover those that might remove them. 19 MR. MICHELSON: Apparently there is a great 20 interest in removing them-- 21 MR. ROSZTOCZY: Yes. 22 MR. MICHELSON: -- at every opportunity when they 23 come out. They would go back in if they had NRC approval. 24 So it is going to be a predominant thing in the future, not {} 25 just an. exception. I suspect most people are going to ACE-FEDERAL REPORTERS, INC.

D2 347 3700 Nationwide Coverage 800-336 6646

5970 04 05 44 MPBeb 1 remove those restraints whenever they have to remove them 2 for in-service inspection. 3 MR. ROSZTOCZY: The basic groundrule, at least 4 for our analysis of the reference-plant work, was to analyze 5 them as they are, and then as a possible preliminary study, 6 give credit for potential changes in the future, like 7 changes due to the ATWS work, and so on. So I think one 8 ought to look at in those changes both any new equipment 9 that might be put in as a result of some other requirement, 10 and also any equipment that they might be taking out as 11 changes in the regulation. 12 MR. MICHELSON: The thing I am trying to 13 emphasize, of course it isn't simply a loss of containment, O 14 it has a potential also for loss of certain engineered 15 safety features which are in the containment which now get 16 the containment atmosphere for which they are not qualified, 17 and for which they can fail very quickly. 18 So you have got to chase the steam all the way 19 through the building wherever it finally goes to vent 20 itself. And it gets pretty exciting if these are not in 21 place. And I am not sure I have ever seen a study that 22 postulated that there is no pipe break restraint and that 23 you do penetrate the containment and the old pipe just 24 shoots it out of the building. /~)'s 25 I haven't seen stch a study. I just wondered, 'w ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800 336 6646

5970 04 06 45 (O*) MPBeb 1 have you done that? That's a severe accident, and it ought 2 to be in the analysis before one writes off large break 3 LOCAs as non-contributors. 4 MR. ERNST: Of course you do have to overlay your 5 risk prospective. 6 MR. MICHELSON: Sure. That was really the 7 thing-- 8 MR. ERNST: You have a likelihood of a large 9 break LOCA, given no leak before break, and that's a small 10 risk to start with. And I guess the judgment is that 99 I 11 times out of 100 you're going to get a leak before a break, 12 or something like that. 13 If you take this very interesting accident, which 14 I agree would be very interesting, but it would still put 1 15 it-- 16 MR. MICHELSON: Yes, but from a risk perspective 17 you still have to make sure that although it is only 18 one-hundredth as likely,to happen, that you have the 19 consequence of one ten-millionth times what you were worried 20 about. 21 MR. ERNST: Let me try and wrap up here_because I 22 know Jack had an hour and a half. 23 (Laughter.) 24 "No," he says. O 25 (slide.) 1 j ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-33M646 -~ ~. - - ~ _. -.

i 5970 04 07 46 /~3 MPBeb 1 The general schedule. It is a little more (./ 2 detailed in the top part than the bottom. 3 This month we are now meeting with ACRS. 4 DR. OKRENT: You are right on schedule. 5 MR. ERNST: I always beat my schedules. 6 (Laughter.) 7 We also anticipate a final program plan.

Again, 8

if you have any comments on the plan I would like to have 9 them. Recognizing the schedule, we want to get this plan 10 out this month if at all poscible. 11 DR. KERR: I'm sorry, what is meant by " program 12 plan"? The whole package-- 13 MR. ERNST: Yes. 7-LJ 14 DR. KERR: -- or a schedule? 15 MR. ERNST: The program plan is the whole 16 package. 17 DR. KERR: Okay. 18l MR. ERNST: In November '85 we are anticipating 19 sending the final program plan to the Commission as an 20 information paper, no accion required. 21 This are sort of iffy, but we are figuring that I 22 probably along about February we ought to be well along in 23 formulation of this work, at which time it might be 24 appropriate to meet with the ACRS again. (]} 25 About a month or so from that time we would ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverase 800 336 6646

5970 04 08 47 MPBeb 1 anticipate issuing a draft design objective and guidance 2 documet? t for public comment, or at least basic elements of a 3 guidance document so somebody would have at least the 4 characterization of what might be involved in the analysis. 5 Then after that gets issued, if we feel a mutual 6 need we will meet with the ACRS again. If indeed there is 7 not much change from the February meeting, then that 8 probably wouldn't be necessary.- It would be basically a 9 mutual option. 10 In May, which would be a couple of months after 11 the publication, we would anticipate a workshop. 12 In June, meet with the group again to discuss any 13 insights that we got from the workshop, or from public J 14 comment. 15 MR. MICHELSON: Who do you envision in a 16 workshop? What kind of audience? 17 MR. ERNST: I would suspect sort of a mixture 18 like we had in Atlanta for NUREG-1050, the PRA reference 19 document, where you have a mixture of technical people for 20 technical insights on the one hand, and then more 21 decision-oriented, management-oriented people-- 22 MR. MICHELSON: But it would be a mixture of 23 the utility-- 24 MR. ERNST: Yes. f'JT 25 MR. MICHELSON: -- group, the industry group, et t ACE-FEDERAL REPORTERS, INC. 202 347 3700 Nationwide Coverage 800-336 6646

5970 04 09 48 MPBeb 1 cetera? 2 MR. ERNST: Yes. 3 MR. MICHELSON: The full spectrum you would 4 envision? 5 MR. ERNST: Yes. 6 Then after that, the redraft, and perhaps meet 7 again if it appeared to be appropriate. 8 A draft policy statement to the Commission in l 9 October, and hopefully publish them for public comment. 10 And then after that time we would anticipate 11 maybe another ten months or so from that point. I suspect 12 we would want to go for at least a 120-day public comment 13 period, so that takes four of the months, another four O t 14 months of working on things and talking to ACRS, and then 15 getting a piece of paper back to the Commission, which puts i 16 fiscal '88 and '89 for perhaps a trial evaluation period. 17 At least that's the thinking right now of this limited part 18 of the staff here. 19 MR. EBERSOLE: Mr. Ernst, let me ask you: t 20 We are talking about containment, aren't we, as 21 contrasted to the other safety systems of the plant? Would l 22 it be worthwhile thinking about the containment as a 23 dif ferent empire, so to speak, that deserved its own power i 24 package, its own discrete cooling network, independent of (]} 25 the network that is associated with core cooling and not i i ACE-FEDERAL REPORTERS, INC. l 202-347 3700 Nationwide Coverage 800-336-6646 -.. - -. ~., -..,

5970 04 10 49 [']MPBeb 1 thus intertied for a common gas gauge failure, and more or %) 2 less draw to one isolated room the concept of containment as 3 a discrete and independent function and not have it latched 4 up, as it always is, with everything else? 5 MR. ERNST: We would have a bunkered containment, 6 so to speak. 7 (Laughter.) 8 MR. EBERSOLE: Well, not exactly that, but one 9 that is clearly discrete, not one that is clearly I 10 intertied. You could pump it with small diesels or 11 whatever. You don't need a large one because it is just 12 going to be a heat remover, That's all it is going to be. 13 I have often thought why the hell have we got \\-)g 14 everything latched up together. 15 MR. ERNST: I think one could get perspectives 16 like that based on what we come out with in the six 17 reference-plant studies. 18 MR. EBERSOLE: Yes. 19 MR. ERNST: If it looks like, for example, the 20 loss of power to the containment systems is a very important 4 21 factor over and above them, it might be a considered option 22 or something. 23 MR. EBERSOLE: If I've lost the core cooling, I 24 don't think I have to lose the other cooling coincident with /'T 25 that, and I shouldn't, but yet that is what we do. (> l ACE-FEDERAL REPORTERS, INC. 4 202 347-3700 Nationwide Coverage 800-336 6646

l 5970 04 11 50 MPBeb 1 MR. ERNST: Yes. 2 Again this particular project would not be after 3 redesigning containment. What we are trying to do is get 4 some kind of standards and some kind of reasonable 5 implementation scheme that makes sense, that should address 6 things like this, or be able to address things like this, 7 and clearly it would. 8 That's all I have for the-- 9 DR. OKRENT: Let me throw a curve ball. 10 MR. ERNST: Oh, I was almost getting disappointed 11 today. 12 DR. OKRENT: Some years ago the California State 13 Energy Commission I think funded a study of underground O 14 containments and containments under berms and so forth, 15 presumably in an attempt to find the nearly perfect 16 containment. ~ 17 I don't recall any more whether those reports 18 gave enough information that you could put them together 19 with your six reference plants here, taking perhaps internal 20 events for the underground containment to be the same as one 21 of the reference plants, or so forth, so that you could look 22 at the performance of that containment to see whether it is 23 similar or markedly different from what you are getting from 24 what I will call your better current containments. (]) 25 Waac I do remember about history is something ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-336-6646

5970 04 12 51 {}MPBeb 1 that in my mind goes like this: 2 Someone, I don't remember who it was any more, 3 drew the conclusion you can do almost as well with a 4 filtered, vented containment as the underground containment, 5 so why continue work on the underground containment, thereby 6 getting rid of the California Energy Commission. 7 And then somebody else did a study that said 8 filtered, vented containments are not cost-effective and 9 they won't cover all possible accidents, thereby getting rid 10 of filtered, vented containments, leaving-- 11 DR. KERR: Indeed, you can almost get. rid of 12 underground containments without the vented containments 13 because there are a lot of disadvantages to having the thing 14 underground, not the least of which is you don't really 15 accomplish a lot in containment unless you also put the 16 power plant and everything else underground. 17 MR. EBERSOLE: Well, see what going underground 181 really does. l 19 DR. OKRENT: Well, again I am reciting what I 20 recall of history, and I think these qualitative 21 recollections are correct. I'm not trying to advocate 22 underground containments, but you might, while you are 23 looking at containment performance, go back and see what 24 seemed to be the qualities of whatever it was proposed then, l () 25 and see what it does for you. l ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-3366646

5970 04 13 52 ('/TMPBeb 1 I said I as throwing you a curve ball. s_ 2 MR. ERNST: Yes. I've got two or three parts to 3 a response I guess. 4 One is I do recall-- I don't recall the details 5 as you do, but I do recall the study, and also that there 6 were some clear questions about the usefulness of that type 7 of containment. 8 The reference-plant studies clearly would not get 9 into that kind of a thing because the very nature of the 10 reference-plant studies is to look at existing plants and 11 see if there are some risk reduction measures useful to 12 those particular plants. That in essence would do things 13 like preclude a detailed consideration of large drives for 14 BWRs, for example. 15 The information from the studies, though, if one 16 might postulate the large risk and no real cost-effective 17 fix for an as-built plant, could lead to some other kinds of 18 things like a large drive for a BWR or an underground 19 containment. So I think that would sort of be how I would 20 see these two ideas, and then maybe a follow-on or something 21 if you first determine that there is a substantial risk that 22 needs correcting. 23 But this study I don't think -- which is really 24 looking at what makes sense in trying to establish a public () 25 risk kind of design objective. Maybe we ought to write-- Acu-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 800-33M646

5970 04 14 53 MPBeb 1 This is not the purpose of the study. We have defined the 2 purpose. 3 Maybe we should say it is not the purpose of the 4 study to redesign the containment or come up with novel 5 ideas. 6 DR. OKRENT: The purpose is to come up with 7 something equivalent to 10 CFR 100, a guideline that will i 8 fit the existing sites. 9 MR. ERNST: That's an interesting thought. 10 DR. OKRENT: 'Well, you can find people saying 1 11 that on the record and off the record. l 12 MR. ERNST: Yes. 13 DR. OKRENT: Why don't we go on to Jack or he 14 will be cheated out of his punchline? 15 16 17 i 18 i 19 4 20 21 1 22 2 23 24 ' O '2s ACE FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coversee 2 336 6646 . ~ -... -..

5970 05 01 54 / MPBmpb 1 (Slide. ), 2 MR. ROSENTHAL: I'm going to speak for a few 3 minutes about defining what I consider the boundary of the 4 plant performance objective, define what I mean by large 5 scale release, talk a little bit about safety systems, a 6 little bit about containment, show you a sample containment I 7 matrix and the difficulty of taking a matrix and conver*ing 8 it to a simple statement. 9 And I have some information about seismic should 10 there be time. 11 (Slide.) 12 We really pushed to have an early meeting with 13 you people in order to get a -- throw out some strawman 14 concepts and get your feedback of whether we wers going in 15 the right direction or not. And so I really have only about i 16 twenty minutes, which allows time for lots of good 17 comments. 18 This should definitely be viewed as a strawman. 19 I would like to see the containment performance objective 4 20 address the performance of what I call the regulated 21 containment; that which we built as a containment structure 22 and reported in the FSAR. And you get into things like 23 Event V. 24 And various. groups would like to take credit for () 25 retention of fission products in auxiliary buildings. And ACE-FEDERAL REPORTERS, INC. 202 347 3700 Nationwide Coverage 800 33646d6

5970 05 02 55 1 those buildings are really there and there really is a 1-{}MPBmpb 2 potential and there really are fire systems out there that 3 can be turned on. But I personally don't think that the 4 CPOs should focus on that, but rather focus on that which we a 5 as regulators have some control over. And that is the -- 6 the containment is described as you saw. 7 Now I would leave room for the consideration of 8 other structures. And where I would look at that is in the 9 safety goals phase as distinct from the containment i 10 performance phase. 1 11 Going down the top, I think that we should give 12 credit where credit is due to primary system retention given 13 a core melt event because the primary thing is really there j 14 and we can at least pretend to know something about it. And 15 we should give credit for our understanding of engineered j i 16 safety features in containment, like pools and sprays, et 17 cetera. 4 t i 18 We would not be considering matters of siting or i 19 emergency planning. j 20 I for one feel that one might use a concept of 21 emergency planning in measuring your conformance with the 22 safety guide. But there was not enough space and we did not t } 23 look at that. I j 24 DR. KERR: Jack, I'm a little puzzled that you () 25 confine yourself to what you call regulated containment. ACE-FEDERAL REPORTERS, INC. Nationwide Coverage 800 - ---. -, _,_ --_, _.--_- _347-3700 - 202 . _ _ _ _,. ~,_ 3 36-6646. -. - -.. - -.. _ ~

5970 05 03 56 1 It seems to me if we are going to at least (}MPBmpb 2 numerically evaluate to some extent the performance of the 3 total system that you ought to look at the total system. 4 And wherever containment fits into that total system, 5 whether it is regulated-or not, you may feel that you are 6 more certain of the performance of reculated systems. And 7 if that's the case you take account of that. 8 But it doesn't seem to me that it makes sense to 9 ignore things just because they aren't regulated. I mean I 10 don't think one does that in the ERA, does one? 11 MR. ROSENTHAL: No, no, no. There one surely 12 looks at all structures, systems, and you go to ultimate 13 releases. But if we're trying to set any sort of ) 14 reliability of' isolation, or reliability of some leak rate, 15 we have some handle on it and that which is the containment 16 and gets reported to us. We have less information -- We are 17 receptive to what you said. 18 DR. KERR: Let me see. If I understood Bill's 19 comments -- and if I did, I agreed with him -- what he 20 seemed to be saying was that if one got the total 21 probability of release -- what you call a source term -- one 22 might do it simply by multiplying the likelihood of if stuff 23 gets into the containment or the likelihood that it gets 24 out; but that because of uncertainties you might want these () 25 terms that make up the product to be bigger than otherwise ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nadonwide Coverage 80033MM6

5970 05 04 57 MPBmpb 1 would be the case. 2 I have no quarrel with that. But it seems to me 3 to say that one is only going to look at that part of 4 containment that is regulated or that part of the plant that 5 is regulated. I had thought up to now, in talking with the 6 PRAs, one looks at what's there. 7 MR. ROSENTHAL: Yes. Well, basically. 8 DR. KERR: So I may be misunderstanding the 9 point. 10 MR. ROSENTHAL: No, no. We want that. Maybe I 11 am not clear enough. 12 We would like to set a containment performance 13 objective which is somewhat of the form that one should have

14 an expectation that the probability is something and that 15 the release is no greater than something else, or that the 16 leakage rate isn't any greater than something else.

In that 17 context I don't know what credit to give to the aux building 18 unless -- You could redefine your control volumes to be this 19 bigger area. 20 That's different from the assessment of risk and 21 a comparison of the safety goal where -- the proposed safety 22 goals you would go through it. 23 So I take it that you are proposing that we would 24 establish a containment -- when we say containment () 25 performance objectives -- ACE FEDERAL REPORTERS, INC. 202 347-3700 Nationwide Coverase 8003364646

4 5970 05 05 58 4 1 DR. KERR: I really misunderstood you. (}.MPBmpb 2 But, for example, in some of the analyses of 4 3 Event V that are reasonable for PWRs one does take -- some 4 people take credit for an aux building because it's there. 5 So that's the way you take credit for an aux building. 6 Now how you relate that to the containment is a 7 question that it seems to me one has to decide after looking 8 at the total system performance and deciding what fraction 4 9 of that you call containment systems. If you're saying I 10 don't want to call that fraction containment systems, I want 11 to call it something else, with that I have no quarrel I 12 because if you call it something else and take it into l 13 account, then that will account for probably what you are 14 trying to release. 15 So if that's what you mean by saying you want to j 16 exclude the contai-ient but you are still going to consider 17 it in your total risk evaluation, okay. And maybe that's 18 what you're saying. 19 MR. ROSENTHAL: Yes, sir. 20 DR. KERR: Okay. 21 MR. ROSENTHAL: I think if this is_still open, we 22 haven't resolved it. We want to think about it. But I did 23 want to bring up the issue. 24 I might also point out that if one meets the CPO, () 25 and if one meets the coro melt objective, then one would i ACE. FEDERAL REPORTERS, INC. 202 347 3700 Nationwide Coverage 800 33H646

5970.05 06 59 (]}MPBmpb-I have high cenfidence that we meet the proposed safety-regs. 2 And if you caa figure this properly, it may be possible to 3 examine a plant with respect to core melt, with respect to 4 containment performance and not carry out consequence l 5 analysis, offsite consequence analysis. And there is a 6 certain attractiveness in getting out of the body-count 7 business and pulling us back into worrying about the 8 containment. 9 DR. KERR: I don't see how you can avoid the 10 consequence-analysis unless you do it implicitly. 11 MR. ROSENTHAL: Right. 12 DR. KERR: The fact that you say that having met n 13 these two things I now meet the safety goals says something {} 14 about your assumptions -- 15 MR. ROSENTHAL: Yes, right. 16 DR. KERR: -- concerning consequences. 17 MR. ROS ENTHAL: Yes. Okay. 18 MR. ERNST: I think what Jack is saying -- and 19 I'm being sensitive to Forrest's point; and I agree with him 20 wholeheartedly -- is that the overriding consideration in 21 safety goals is public risk. I think what Jack is saying is 22 if you can come up with an engineering approach that looks 23 at containment and looks at coremelt, and say given that you 24 meet these two I have a high confidence that I have already () 25 met my public risk, then that would be a decent way to l l i ACE FEDERAL REPORTERS, INC. I 202 347 3700 Nationwide Coverage 800 33MM6

l 5970 05 07 60 1 consider and regulate it in a more engineering kind of a l (}MPBmpb 2 sense. ) 3 DR. KERR: But can you do that unless you do some 4 sort of generic consequence analysis? 5 MR. ERNST: No, you would have to do that, that's 6 correct. 7 MR. ROSENTHAL: Which we're almost at. 8 DR. KERR: If you can do it and bound things, 9 great. There is no point in having to do it again. 10 MR. ERNST: The problem is if you do it in a 11 bounding way then you get so conservative. And I would be 12 very reluctant from that standpoint. 13 DR. KERR: By bounding things I mean doing it in [} 14 a way that you are confident you are representing the risk 15 adequately. 16 MR. ERNST: Yes, exactly. 17 MR. ROSENTHAL: Some of the proposals would be in 18 terms of defining a large scale release and then trying to 19 have some sort of conditional probability. I was trying to 20 grope around and decide what I meant by a large scale 21 release. And the nomenclature in the RSS is PWR, one, two, 22 three, four, five, and BWR. 23 And here are the release fractions. Here is the 24 siting 9*.udy release fractions. () 25 (Slide.) ACE-FEDERAL REPORTERS, INC. Nationwide Coverage 800 336 4646 PNEN - m

5970 05 08 61 (' 1 V) MPBmpb You will note that siting study one is 2 horrendous, two is relatively small in terms of release 3 fractions. We have a lot of -- And so how you define -- I 4 wanted to throw these up because some of you work with them 5 a fair amount and it gives you some perspective. I would 6 say things greater than an SST2 or greather than a BWR 3 or 7 greater than somewhere between a PWR 3 and 4 I would 8 consider a large scale release. 9 (Slide.) 10 Now we've got a lot of crowded lines. So here's 11 the same table with another column added which is Jack's 12 Strawman Large Scale Release. And I expected we would have 13 a f air amount of discussion over this. 14 All of the nobles, ten percent of the iodine, 15 cesium, rubidium, tellurium, one percent barium, strontium, 16 ruthenium, and about a tenth of one percent lanthanum. 17 We've got crack calculations that come from 18 Brookhaven; we've got other crack calculations. I saw a 19 paper given by MUS. But one asserts that if the release 3 20 fractions are this large or smaller then independent of 21 evacuation, independent of whether you expect very few early 22 fatalities, you are no longer talking in terms of 23 conditional consequences or even, you know, a CCDF with a 24 tail that CCDF has added two or three thousand or some .() 25 horrendous number, but, rather, you are talking about some ACE-FEDERAL REPORTERS, INC. 202-347 3700 Nationwide Coverage 8043366646 _ ~ - _

5970 05 09 62 {}MPBmpb 1 very small, 08 or something. 2 DR. KERR: Less than ten. 3 MR. ROSENTHAL: Less than ten. 4 I have a number of eight for 50 weathers, no 2 5 evacuation, relocation at any point. 6 MR. ERNST: Your probabilities get very small for 7 those small numbers of fatalities because you have the tails 8 and your agility and your essential details of the 9 evacuation assumptions basically. 10 MR. ROSENTRAL: So you would have high confidence 11 that if your release was less than the strawman then you 12 would handle it to meet the proposed safety goal 13 objectives. { 14 MR. ERNST: And likely have no fatalities. 15l MR. ROSENTHAL: And these numbers yet, with i 16 exception, are large compared to some of the numbers that we 17 are seeing in BMI 2104. 18 Now this is based on a focus on the aversion of 19 early fatalities. 20 If you would like to aver concerns over land 21 interdiction then you have to add another zero to some of 22 the heavier elements. And notice that lanthanum is very 23 small. I 24 () 25 1 ACE-FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 233H646

I 5970 06 01 63 ^ MPBagb 1 I would like to point out that you don't need for 2 a boiler cooled to 10,000 in order to bring the iodine down 3 to .1, you need modest credit. And alternately for large 4 drys we have a pretty good concept that now it is one of the 5 more believable. The in-containment fission product 6 deposition code is one of the more believable and benchmarks 7 of the group and that shows that over time we were seeing 8 about a decade reduction over four hours. So it is credible 9 to believe .1, .01 type numbers. You don't have to have an 10 overreliance on some of the codes of technologies and 11 things. And my exception is lanthanum here. And I think 12 that there is a fair amount of RES work going on, trying to 13 close on lanthanum. Lanthanum would blow this whole thing 14 out of the water. 15 DR. OKRENT: A question for some information: On 16 the Xenon and Crypton iodine, when you say that with those 17 release directions you calculate very small early fatality 18 likelihood, is that assuming release -- well is there any 19 radioactive decay that has occurred in that calculation? 20 DR. KERR: Is the .1 .1 of inventory? 21 MR. ROSENTHAL: Of the original inventory. 22 DR. OKRENT: Of the original inventory. 23 MR. ROSENTHAL: Of the original inventory. I 24 have got some sensitivity studies, at least you can walk () 25 home with them. So in any case, this is my definition of ACE-FEDERAL REPORTERS, INC. 202-347 3700 Nationwide Coverage 800 336-6646

5970 06 02 64 (~S;MPBagb 1 large scale release and that hence -- if you knew that the w; 2 containment would 95 percent of the time given core melt, if 3 99 -- my number in my head is 95 percent, which I have a 4 numerical basis for it -- were sure that the release 5 fractions were less than this, then you would handily meet 6 the safety goal objectives on earlies, and earlies are more 7 of a problem than lates. 8 DR. OKRENT: The reason I am asking this, I 9 remember many years ago now TVA came in with some kind of I 10 suppose not too elegant calculation they had made of the 11 doses and the site boundary for one of their sites if they 12 just released all of the noble gases and the numbers were e'} 13 large and appeared to the LD-50 -- or at least it was larger k./ 14 than an LD-50 and they can't be right, it seems to me if you 15 are going to have a low fatality with more evacuation. 16 MR. ERNST: We can check that out. It is my 17 information to say it would be something different but we 18 can check that out. 19 MR. ROSENTHAL: Right. 20 DR. OKRENT: I think it was done for Sequoyah 21 back some time after Three Mile Island 2. 22 MR. ROSENTHAL: That the noble gas release alone 23 results at the site boundary in excess of 320 or 400? 24 DR. OKRENT: I'm quite sure the numbers were e~ (-)s 25 something like 7-or 800 if they were the filtered vents. l l ACE-FEDERAL REPORTERS, INC. l 202-347-3700 Nationwide Coverage 800 336-6646

5970 06 03 65 1 MR. ERNST: I wanted to check that out, that {}MPBagb 2 doesn't seem to jibe with our information. 3 DR. OKRENT: And it was one nail in the coffin of 4 supposedly how bad filtered vents were. 5 (Slide.) 6 MR. ROSENTHAL: I just have a few minutes on 7 sequences. We assumed you have to consider the spectrum of 8 sequences. It seems the Germans have felt this on station 9 blackout, the containment load working group was primarily 10 focused on station blackout and we think you ought to 11 consider blackout and event V and tube rupture and seismic 12 and not limit yourself to picking one design basis event. () Now there is a certain beauty in going the other 13 14 way, of saying okay I'm going to pick station blackout 15 because that is one where you have no active features and 16 you develop a test and then you test the plant against the 17 test. And the advantage of that is it is less costly, it is 18 more straightforward and you know where you are going to 19 land. The disadvantage, I think, is that it doesn't pick up 20 some features such as the risk dominance of event V or 21 consequence of tube rupture. 22 DR. OKRENT: I will support your position 100 23 percent. 24 (Slide.) () 25 MR. ROSENTHAL: We have to be careful in using ACE-FEDERAL REPORTERS, INC. 202 347 3700 Nationwide Coverase 800 336 6646

5970 06 04 66 1 the sequencers. I would like to give a hypothetical example }MPBagb 2 here. Let me say that I have in the first case a real plant 3 and I have two people look at the plant: TOUV is blackout 4 and TC is ATWS. And just for argument's sake take it that 5 TC gets bad release and blackout gives you more benign 6 releases. 7 The first analyst says that blackouts minus-4, 8 ATWS is minus-6. That analyst might be an analyst at 9 Limerick. He could conclude -- now let's say he takes a 10 conditional look at the containment performance. He comes 11 up with the perspective that 99 percent of the core melts 12 are going to lead to low releases and 1 percent of the core 13 melts are going to lead to bad releases and at that point ) 14 that plant is behaving at 99 percent retention. 15 The second analyst -- and maybe this is somebody 16 looking at -- a real group of analysts looking at a Mark I 17 -- says well blackouts is minus-5, ATWS is still minus-6 -- 18 or maybe it is two analysts looking at the same plant. And 19 he concludes that 90 percent of the time he has got a lesser 20 release, 10 percent of the time he has got a greater release 21 and then the strictly conditional containment performance is 22 .9 instead of.99 and he doesn't get as good a mark on his 23 report card. And yet this is clearly the better plant. And 24 the fact that he has managed to suppress his core melt () 25 frequency by an order of magnitude and he has been able to ACE. FEDERAL REPORTERS, INC. 202-347 3700 Nationwide Coverage 800 33HM6

5970 06 05 67 1 suppress it down the sequence. You wouldn't want to punish {}MPBagb 2 him somehow for having done that or create disincentives in 3 the system to suppress the plant. And that this sort of 4 argument leads one in part to a proposal four which said 5 let's combine the core melt and the containment reforms on 6 this to some kind of a 10 to the minus-6 conditional on both 7 with some absolute limits such that you don't allow full 8 risk-risk tradeups between prevention and mitigation. 9 The disadvantage of that is that of course it is 10 not a strict containment performance, a conditional 11 containment performance. 12 DR. KERR: It seems to me one might be different 13 if the containment performance took some f raction of the 14 time to get to a certain containment release. I still don't 15 see why it is not a containment performance. It might be 16 different numbers if you do it this way, but you still have 17l containment performance criteria, it seems to me. 18 MR. ROSENTHAL: If we were to come up with one 19 that said 99 percent of the time that you have a core melt 20 the release from containment should be small, let's say that 21 was the formulation. And let's say that this plant is 22 real. This guy meets it and this one fails it and yet this 23 plant has a lower core melt frequency. And that's the 24 disadvantage in my mind of setting a strictly -- of using () 25 the CPO strictly as a conditional containment performance. ACE-FEDERAL REPORTERS, INC. 202 347-3700 Nationwide Coverage 800-33m

5970 06 06 68 f'}MPBagb 1 MR. ERNST: I think this is the flexibility v 2 attribute that I was talking about. 3 DR. KERR: I'm not making myself clear. You 3 4 might have the containment performance criteria which would 5 be dependent on the core melt frequency, but it seems to me 6 it is still a containment performance criteria. It may be a 7 sliding scale. 8 Yes. 9 DR. KERR: That was the only point I was making. 10 And I made it because I thought I might be missing ~ 11 something. It sounded as if you were saying you would like 12 a sliding scale depending on how the plant behaves 13 othe rwise. That does not seem unreasonable to me. But it 14 seems to me you are still talking about a containment 15 performance criteria and even though it -- 16 MR. ROSENTHAL: Oh okay. 17 DR. KERR: Is that....? 18 MR. ROSENTHAL: Yes. And apparently there was no 19 dispute. There is a certain elegance of saying here is my 20! core melt frequency as a scaler, here is my i 21 containment-conditional consequence as a scaler, and here is 22 my risk, offsite risk, that number, from a communications 23 standpoint. 24 DR. KERR: It seems for me at least it is () 25 premature to choose one over the other. But at least at 4 ACE FEDERAL REPORTERS, INC. 202-347-3700 Nationwide Coverage 3364646

5970 06 07 69 1 this point I would not be unwilling to throw either one out. (}MPBagb 2 (Slide.) 3 MR. ROSENTHAL: This slide is extracted from the 4 back of the containment performance plan that you received 5 and we tried to summarize how we thought containments 6 behaved differently. And I think the ultimate point is that 7 we don't have the luxury of nations with one type of reactor 8 but we have all types and that the failure modes are 9 different for each type. We can spend a while on this or we 10 can skip it. 11 DR. KERR: I think the chart speaks for itself. 12 (Slide.) 13 MR. ROSENTHAL: I would like to make the point ) 14 that people were ready to write off large dry containments 15 because they looked pretty good and that we recently got in 16 some NRR work on Palisades and we find that the emergency 17 sump is right below the vasso. So if you fail the vasso -- 18 this little bit of concrete isn't very large -- you put the 19 core in the sump, you fail ECCS recirc or containment spray 20 recirc and you may have a path to the outside. So all 21 containment -- 22 DR. OKRENT: The French have a large dry which 23 has a vulnerability -- or did have a vulnerability below 24 the -- () 25 MR. ROSENTHAL: They had the diaphragm in the ACE-FEDERAL REPORTERS, INC.- 202 347-3700 Nationwide Coverage 800 33H646

t 5970 06 08 70 (}MPBagb 1 middle of the base mat. ) 2 DR. OKRENT: They had some penetrations that came 3 in in one of their containment designs. So certainly all 4 large drys are not equal. 5 MR. ROSENTHAL: Yes. ~ 6 (Slide.) 7 Here is the Davis-Besse -- 8 DR. KERR: Just out of curiousity, there was only 9 one sump or there are more than one? 10 (Slide.) 11 MR. ROSENTHAL: I think their big emergency sump 12 is right down here. Let me not speak more, I am just i 13 beginning the study. 14 DR. KERR: All right. 15 (Slide.) I 16 MR. ROSENTHAL: Davis-Besse has an interesting I l 17 design to do. This is not to scale. If you have a high 18 pressure blowdown -- or even a not so high pressure 1' 19 blowdown, you can lob the material right out of the in-core i 20 tunnel and you either drop it like a basketball into the 21 emergency sump or you can bank it against the wall of the i 22 containment. 23 Now that was the point. Now that doesn't mean 24 that we can't set a containment performance objective at ] () 25 all. What it means is -- I'm just pointing out the ACE-FEDERAL REPORTERS, INC. 202-347 3700 Nationwide Coverase ' 80433HM6

4 5970 06 09 71 MPBagb 1 difficulty in measuring performance. It would require sump j 2 to be defined by Zoltan's program in part. 1 3 DR. KERR: It would require that you understand 4 the containment performance. 5 DR. OKRENT: Well to some extent what you are 6 showing here suggests that your six reference plants are not l 7 fully representative any more than the original two. 8 MR. ERNST:- Maybe some more. 9 (Laughter.) 10 DR. OKRENT: Well okay. i 11 Go ahead. 12 MR. ERNST: Zoltan has the same problem, I think, t 13 in extrapolation of the plants. Whether you have a design . O 14 objective or not, these are the same kinds of questions that l 15 need to be addressed in the extrapolation. 16 MR. ROSENTHAL: And these are the kinds of i l 17 things, the kinds of considerations. How do we define the 18 containment boundary and how do we define how much 19 plant-specific looking do we have to do. It would end up in 20 an implementation document and that is the sort of thing 4 21 that is missing from the current safety guide. And I think 22 that's why we feel there is some incentive that it is so 23 important to have some sort of implementation document, a 24 document to go along with this, with more grandiose data. ( () 25 (Slide.) i ACE. FEDERAL REPORTERS, INC. 202 347-3700 Nadonwide Coversee 800 3364646

5970 06 10 72 MPBagb 1 This is the containment matrix -- 2 DR. KERR: Tell me what is meant by the phrase 3 " containment matrix." 4 MR. ROSENTHAL: Let me assume that I have a 5 vector here called core melt vector. 6 DR. KERR: It is not showing up on the screen. 7 You said this vector in the far left? 8 MR. ROSENTHAL: With or without numbers on it I 9 would call that a core melt vector. 10 DR. KERR: All right. 11 MR. ROSENTHAL: Okay. 12 And if I take that -- and then I would have a 13 matrix here which allows me to map of conditional 14 probabilities of containment failure by each of the elements 15 of the vector. And that allows me to map and to be in some 16 release fractions. 17 DR. KERR: Okay. Those numbers I see -- 18 MR. ROSENTHAL: -- the only one we would have is 19 a vector -- 20 DR. KERR: No, are those numbers the mapped 21 releases? 22 MR. ROSENTHAL: What I would like to do is say 23 that given TMLD prime, given station blackout, that there is 24 a.005 condition probability of containment failure due to (]) 25 early overpressure. And I will map the releases from that ACE FEDERAL REPORTERS, INC. 202 347 3700 Nationwide Coverese 2 0 336.#M 6

i i j 5970 06 11 73 MPBagb 1 into. bin one. I i 2 DR. KERR: Okay. 3 MR. ERNST: This is sort of what I was talking i i 4 about when I said you take the incoming event, go through 5 the containment event tree to see what happens and how it 4 l 6 threatens containment. From that then you run a VMI run, 7 which is the bin, to see what the source term is. l j 8 DR. KERR: Yes. That's enough for me. Thank i 9 you. 10 MR. ROSENTHAL: Okay. 11 My only point here is that the conditional 12 probabilities of failure by various modes and the mapping j 13 into release fractions has gotten progressively more and 14 more complex rather than simpler. 15 (Slide.) 16 Here is the corresponding bins that the core melt i j 17 vector was mapped into with the corresponding relief t l 18 fractions. 1 19 1 j 20 l 21 22 4 23 f 24 lO 2s l ACE FEDERAL REPORTERS, INC. i 202 347 3700 Nationwide Coversee sos.3364M6

5970 07 01 74 MPBagb 1 (Slile.) 2 Translating this matrix of conditional 3 probabilities into a bimple scaler was the challenge of 4 coming up with a performance objective. 5 MR. ERNST: These bins would be different for 6 Sequoyah than they would be for Surry and they would be 7 different for Zion. So instead of having a bin for PWR's 8 and a bin for BWR's like you had in WASH-1400, it gets 9 pretty complex. 10 I think the next challenge is how do you now 11 collapse this into some more useable information. 12 DR. KERR: I think at some point in your position 13 paper the probability was mentioned that you can begin 14 perhaps formulating standard problems and say something like 15 if you can read the results of this standard problem you are 16 okay. Neither of us is sure that one can or cannot do this 17 and it seems to me unlikely that you can. But I got the 18 impression that you were looking at that as a possible 19 approach if the analysis indicated that that might be 20 feasible. 21 MR. ERNST: I personally think that that would be 22 useful. Jack may have a different opinion. But the problem 23 with too simplistic a standard problem is that you may be 24 treating other kinds of containments too inequitably or O 2s someehing. ACE FEDERAL REPORTERS, INC. 202 347 3700 Nationwide Coversee 800.))MM6

l 5970 07 02 75 MPBagb 1 DR. KERR: Yes. 2 (Slide.) 3 MR. ROSENTHAL: Before we were talking about a 10 4 to,the minus-6, a large scale release. And I just want to 5 point out -- this is Millstone now. We have a seismic 6 hazard characterization program. i 7 DR. REMICK: Is that 1, 2 or 3? 8 MR. ROSENTHAL: 3, Millstone 3. And Dames and 9 Moore. And in terms of core melt frequency, using the shed 10 passing curve, I would expect to exceed the minus-6 11 number. 12 Let me say that, you know on this -- the plant 13 still makes the safety goals. I think that our early {) 14 fatalities here were like 17 percent of the safety. But 15 there is a problem with how do we consider seismic and how 16 do we match up seismic internally. I don't have the 17 answer. I just wanted to point out the problem. Dames and 18 Moore are at least an order of magnitude down. 19 Next the SE is a seismically-induced station 20 blackout pump seal LOCA in which you have a containment. 21 This work was done with RSS source terms, so I would expect 22 the risk from SE to be down. I would expect the risk from 23 AE. TE is one in which the auxiliary building fails and you 24 collapse the building structure on top of VSP. It is not () 25 unlike Dr. Michelson's LOCA which fails and that gives me L ACE FEDERAL REPORTERS, INC. r 202 347 3700 NedonwW Cowreer sk336aM6

5970 07 03 76 1 again a release, a seismically-induced core melt into an {}MPBagb 2 intact event. 3 On the other hand, V-3, is a failure of the 4 containment frame wall which takes out the containment, 5 takes out the primary system, gives you a large break LOCA 6 and for which, unlike AE, SE, TE the new source term is 7 going to suppress the risk, the new source term stuff is not 8 going to buy me anything for V-3 because I have early 9 failure of containment and I have, because of a large break 10 LOCA, little primary system retention. So the risk 11 significance of seismic will be there even when we apply the 12 new source te rms. 13 (Slide.) 14 I have on that considered this illustration 15 rather than a computer-genersted. The median fragility of 16 the containment frame wall was 1.82g and here is the seismic 17 hazard, the SHCP hazard function. What you are doing is you 18 are looking at the convolution of SHCP and fragilities and 19 events at half a g or higher. And so the uncertainty in 20 that convolution is quite high. 21 We know that we have to address seismic. We have 22 put it in our program plan that we would address seismic. 23 We haven't made decisions to date on how to work, how to 24 incorporate seismic in a performance objective. So we do () 25 feel that it is an issue that must be addressed. ACE FEDERAL REPORTERS, INC. 202 347 3700 Nationwide Coverage 8(n))MM6 u

5970 07 04 77 1 I am finished. (}MPBagb 2 (Slide.) 3 You will find the slide like this in the back of 4 your handout. And here is a base release fraction and 5 consequences. And this is moving with -- let me just take 6 this column here. With all releases equal to zero but 7 lanthanum at 1.1.001 what are the consequences. So it will 8 give you a feel for the relative points of various isotopes 9 and the assumptions are right up here. 10 DR. OKRENT: Okay. 11 Are there comments, further comments or questions 12 that the Subcommittee has or that the Staff has7 13 DR. KERR: I firstly want to commend the people ) 14 who were responsible for what I thought was a very well 15 organized report. I found it quite informative. I applaud 16 this moving of direction which, it seems to me, we need to 17 go. 18 DR. REMICK: I concur. 19 DR. OKRENT: Well there is something wrong with 20 this Subcommittee meeting. 21 (Laughter.) 22 MR. ERNST: It is a first. 23 DR. OKRENT: We might even end early. 24 But in any event, I have nothing to add at the () 25 moment. We will be interested in seeing what you have ready ACE. FEDERAL REPORTERS, INC. 202 347 3700 Nadonwide Coveraos No.33HM6

5970 07 05 78 1 for our next meeting. I think the issues will become -- g MPBagb 2, that are hard to deal with will become more clear at that i 3I time. l 4 In any event, if there are no further comments, I 5! will thank the Staff for this prosentation and adjourn the 6 meeting. I 7 (Whereupon, at 5:00 p.m., the meeting of the 8 Subcommittee was adjourned.) 9 10 11, 12 13 g 14l 15; i 16i 17' i 18' 19l 20 21 22' 23 24 h 25 ACE. FEDERAL REPORTERS, INC. 202147-)*fX) Nation *tde Coverage MO 13WW:

CERTIFICATE OF OFFICIAL REPORTER O ~ This is to certify that the attached proceedings before the UNITED STATES NUCLEAR REGULATORY COMMISSION in the matter of: NAME OF PROCEEDING: ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON SAFETY PiiILOSOPIIY, TECIINOLOGY AND CRITERIA DOCKET NO.: PLACE: Washington, D. C. O DATE: ^ Wednesday, October 9, 1985 were held as herein appears, and that this is the original transcript thereof for the file of the United States Nuclear Regulatory Commission. (sigt) J (TYPED) Madolon P. Bloom Official Reporter Rep $$[kNf f1TNfn"> '"'- ^ O

'O PLAN TO DEVELOP CONTAINMENT PERFORMANCE DESIGN OBJECTIVE O ACRS SUBCOMMITTEE MEETING OCTOBER 9, 1985 O n, t. eRNST

'O PURPOSE OF PLAN o DEVELOP PROPOSED DESIGN OBJECTIVE AND GUIDANCE o OBTAIN PUBLIC COMMENT o PROPOSE TO COMMISSION AS ADDITION TO SAFETY G0ALS O o EVALUATE THROUGH TRIAL USE o ASSESS TRIAL USE AND MAKE FINAL RECOMMENDATION TO COMMISSION O

ATTRIBUTES OF DESIGN OBJECTIVE ggg o PROTECTION OF PUBLIC HEALTH AND SAFETY NEEDED MINIMUM LEVEL OF SAFETY IMPROVE OR MAINTAIN PERFORMANCE COST / BENEFIT OF POTENTIAL IMPROVEMENTS CONSIDERATION OF UNCERTAINTIES CONSIDERATION OF EXTERNAL EVENTS MAINTAIN NEEDED DEFENSE-IN-DEPTH o INTEGRATED PERFORMANCE OF CONTAINMENT l STRUCTURAL INTEGRITY ESF TO REDUCE CONTAINMENT LOADS ESF TO REDUCE FISSION PRODUCTS ggg EMPHASIS ON TIMING 0F CONTAINMENT FAILURE o FLEXIBILITY PERMIT TRADEOFFS BETWEEN ACCIDENT PREVENTION AND CONSEQUENCE MITIGATION GREATER EMPHASIS ON MORE T0XIC FISSION PRODUCTS SMALL VERSUS POTENTIALLY LARGE RELEASES (E G., VENTING) UNCERTAINTIES o EFFICACY UNDERSTANDABLE lll IMPLEMENTABLE COSTS

4 SPECIAL ISSUES CONTAINMENT BYPASS EVENTS EXTENT OF RELIANCE ON PREDICTIONS OF ACCIDENT PHENOMEN0 LOGY O WHETHER THERE SHOULD BE A DIFFERENT DESIGN OBJECTIVE OR IMPLEMENTATION SCHEME FOR NEW VERSUS OLD DESIGNS WHETHER THE DESIGN OBJECTIVE IS AIMED BASICALLY AT MAINTAINING VERSUS SIGNIFICANTLY IMPROVING CONTAINMENT PERFORMANCE O

1 O CONCEPTUAL ALTERNATIVES 1. ACRS PROPOSAL (NUREG-0739)--NEW DESIGNS AND CONSTRUCTION PERMITS l <0,1 (CONDITIONAL) 0PERATING LIMIT <0,01 (CONDITIONAL) GOAL N 2. ATTENUATION 10 FOR IS0 TOPES (OTHER THAN NOBLE GASES) WITH HALF LIVES GREATER THAN X DAYS, WITH PROBABILITY OF Y OF (]) ACHIEVING ATTENUATION i 3. ESTIMATED REDUCTION IN RADI0 ACTIVE RELEASES OVER A SPECIFIED PERIOD OF TIME AFTER LARGE SCALE CORE MELT ( 1 4, ESTABLISH PROBABILITY FOR UNCONDITIONAL LARGE SCALE RELEASE, BUT PROVIDE A MINIMUM LEVEL OF CONTAINMENT PERFORMANCE AND i THEN ALLOW FOR SOME FLEXIBILITY IN TRADEOFFS BETWEEN CORE MELT FREQUENCY AND LIKELIll00D OF CONTAINMENT FAILURE i

i 1 i OVERALL PLAN

U n

ESTABLISH INTEROFFICE WORKING GROUP AND A SENIOR REVIEW GROUP i i ) DEVELOP DESIGN OBJECTIVE BASED ON EXTANT INFORMATION AND DRAWING I HEAVILY ON REFERENCE PLANT ANALYSES i DEVELOP GUIDANCE REGARDING IMPLEMENTATION: I i o SELECT SEQUENCES I l 0 MODELING ASSUMPTIONS .,i i O ACCIDENT SEQUENCES j I ACCIDENT PROGRESSION SOURCE TERM i l CONTAINMENT BEHAVIOR t CONSEQUENCE CALCULATIONS s o UNCERTAINTY ANALYSES i I EXTERNAL VERSUS INTERNAL ACCIDENT INITIATORS o DEVELOP DECISION ANALYSIS O 1

~. - _ l SCHEDULE O l MEET WITH ACRS OCTOBER 1985 FINAL PROGRAM PLAN I I NOVEMBER 1985 COMMISSION INFORMATION PAPER MEET WITH ACRS FEBRUARY 1986 3 j MARCH 1986 DRAFT DESIGN OBJECTIVE AND GUIDANCE FOR PUBLIC COMMENT i MEET WITH ACRS APRIL 1986 MAY 1986 WORKSHOP lO MEET WITH ACRS JUNE 1986 I AUGUST 1986 REDRAFT MEET WITH ACRS I DRAFT POLICY STATEMENT TO OCTOBER 1986 COMMISSION WITH GUIDELINES PUBLISH FOR PUBLIC COMMENT NOVEMBER 1986 FINAL TO COMMISSION FOR l SEPTEMBER 1987 INTERIM USE t TRIAL EVALUATION FY 88 AND FY 89 I ,l 1

i NRR STAFF PRESENTATION TO THE l O ACRS l

gr CoSTNRMERT 9625cEMAG omECTM5

SUBJECT:

fe>g. 55Ne2.E ACC.tDE.MTS DATE: Oq t'E. 3 y 19 65 PRESENTER: ThCU 6-NS -O PRESENTER'S TITLE / BRANCH /DIV: 6EC.TlDH LtM@ EthCIbR.SN6TEMS BEM b\\\\Ml0W CA' SN5~tEK':2 ItEE@ PRESENTER'S NRC TEL. NO.: 46)2 o)Q4~.1 b M @Y 9 @LDSD? W,Tec4htom W SUBCOMMITTEE: A % CE\\'tE9lh o. f

MAKING COMPARISONS WITH CONTAINMENT PERFORMANCE OBJECTIVES AND WITH THE SAFETY GOALS - A PROPOSAL COMPARISONS WITH CONTAINMENT-PERFORMANCE COMPARISONS WITH OBJECTIVES (PROPOSED) SAFETY GOALS RETENTION OF APPLICABLE APPLICABLE FISSION PRODUCTS IN-VESSEL RETENTION OF ~ ' FISSION PRODUCTS EX-VESSEL RESULATED CONTAINENT ALL SYSTEMS AND IN CONTAINMENT AND ESF STRUCTURES SITING N/A APPLICABLE EMERGENCY APPLICABLE plate 4ING N/A Safety Goal 7 CPO f h f f f f ( ( { e e

-=we RELEASE FRACTIONS PWR 3 PWR 4 BWR 3 SSTI SST2 ~ Xs, Kr .8 .6 1.0 1.0 .9~ I .' 2 .09 .1 .4S .003 Cs-Rb .2 .04 .1 .67 .009 Te-Sb .3 .03 .3 .64 .03 s. Ba-Sr .02 .00S .01 .07 .001 Ru .03 .003 .02 .0S .002 Lo .003 .0004 .003 .009 .0003 O O O

• 4

RELEASE FRACTIONS S.LARGE" TRAWMAN RELEASE PWR 3 PWR 4 BWR 3 SST1 SST2 FRACTIONS Xe, Kr .8 .6 1.0 1.0 .9' 1.0 I .' 2 .09 .1 .4S .'003 .1 Cs-Rb .2 .04 .1 .67 .009 1 Ta-Sb .3 .03 .3 .64 .03 .1 .g. Ba-Sr .02 .00S .01 .07 .001 .01 Ru .03 .003 .02 .0S .002 .01 Lo .003 .0004 .003 .009 .0003 .001 e e 4 i e i e e e e

l DIFFERENCES IN PROGRAM EMPHASIS FOR VARIOUS ASSESSMENTS OF CONTAINMENT PERFORMANCE DURING SEVERE ACCIDENTS l FEDERAL REPUBLIC STATION BLACKOUT 0F GERMANY NRC CONTAINMENT LOADS SPECIAL PROBLEMS # 1-5 : STATION BLACKOUT WORKING GROUP SPECIAL PROBLEM # 6 : POWERED IGNITERS - 's NRC CONTAlttENT PERFORMANCE OBJECTIVES SPECTRUM OF DOMINANT CONTRIBUTORS e.g., STATION BLACKOUT, EVENT V, SGTR, SEISMIC G 9 O

RELATIVE IMPORTANCE OF ACCIDENT SEQUENCE FREQUENCY CONDITIONAL CONTAINMENT RESPONSE FOR A BWR SHOWS A PARADOX CONCERNING THE PERCEIVED CORE MELT VECTOR. s ACCIDENT CORE E LT FRACTION RELATIVE SEQUENCE FREQ. ( CW) 0F CW CONSEQUENCES Ts PERSPECTIVE 1 TQUV 1XE-4 99% LESSER RELEASE ~ TC IXE-6 1% GREATER RELEASE PERSPECTIVE 2 TOUV IXE-S 90% LESSER RELEASE TC IXE-6 10% GREATER RELEASE TQUV = STATION BLACKOUT WITH LOSS OF HEAT REMOVAL TC = ATWS e O ~ G A

f f , Table A-3 Containment Failure Modes e ~ LARGE DRY / ICE MARKI/ MARK III $URATMDSPHERIC CONDDtSER MARK II STEM EXPLD510NS ~~ very unilkaly to be sufficiently energetic to fall containment i 3 Unlikely. Design dependent Failure to isolate -,--- Design dependent .Subatmospherte Interwediate less likely then time frame. large dry seal failure may be a ? problem I i Ilydrogen burn / ~ Unlikely. Dominamit Unlikely Dominant failure detonatten Intermediate failure sede. mode. Wetwell failure time-frame failure may be more probable than - [ i a problem in drywell failure i e same contalements and less risk significance. i j Overpressurizatten Early failure Late failure Dominent late failure uniftely. Intermediate mode of lesser Dominant late, e, . time-frame risk significance. i failure mode but failure mode 1 lesser risk significant-to risk.- 1 significance. s Pessible to unlikaiy, of lesser risk significance j Sesenetpenetration l t Design dependent, u risk significant Centalement typass. i l

i l I f Q Z m;k o >- -,.. g ..., ss s,. } N: ' 3I

• {$k

,,,j ;3 [q l ggggg w ' ap:'- . M&R Wm m za p-4u 5 9 a% IQ & ~% A,g p $,,8 $2,%,g;i ~, w;.- ??:._ i ....,.,,' ?~f,:w -d ,4

&i A.,'

,D)x l ) 'Iy W sur.m a s ex e nt: t.a s a;.m ca:num;wwx anwcq:t:mm %9: i ' MMWW ;i;tt;!;n ' ~ 7tvC?r%D t me yo ky" 31gs; _3 m em > a wris riMMM 5:4;:S;t:1;U 2$ Q, E I$. D!?f ^:dM.221h beiT y 'S - m;7 ;ca ks %2O?t!:WC.&, F -5 C-7 / Ment'sanac eme'!.SG h h20:010tY(A; '"m mue x- '2i i:8 L " ;ET - A ,; :. 7 ^ M ~ ~ -; w

nQg:??;t@$!C' d2Y r:' Cc;;;.%;;q" * :'d;:12":'

7 J

f'9 o$

2:Rin :: t*%:::::L ?? T:;V Ls i misgamaw dywww m w-4 q

1. -symmy%

y;'a;;;::T:-;;;S,W ' m,a-VT UW;",,G. G..y? J ya n. - %s L c.+.. _. g, 3 L;- U a?]$ O y _I g i ;f: W.-. vc h p

p 3

._I o,y g- '3o ci . z a e c E -. [s;'; '-' U) ry 1 p I.d [c c: >H z_M Z F Om jc gg n 03 ZA 't p p U i gr._, L.gy.- ^ k' o_ C e: a m a, - -.. _ u> c.,, L

l l Ik'$$k L_ L ~ . _J .m x. yp. yW o N..' I" 'l jZ z a_ -'s, i._ U) ' g g) Cf.3 x% + [ W .D h [3Q e, 's, .r:' fh?A{/// / 'Y 47 :- y ./ .s .r $'//A(l'$': .s. >-0 m .- -hh u.

5.

I 4^ h,b hhb h: 5 ! Ihh hhik'. h .I y!N$)Dt,:: E h. ( f.k..- f 1, k k-c 1 .h.ik.:.I.h..: *,;;n;:S$hfh. 'I [Sh c;.Ub..iiY 9N.A:fh.;.s'-;-'I:$h,N $hk. ) $N v.:wAV5:h .y.nn><.x .v 3m>+:%.>:./..->...> I

e n

i 5: I' b f IfUh'5, !; M. '>. f ".*d;f03 O!!!NSy, CT;;iit.!?id

ER

7-X71difstfCMS:!.

ie;

?

??2f:!$P^G ;'.;nnhfCO. 3dssRai!PA%31 lEs!EM6ks%Ciddesid%.. indii;Z%5;33 %2f%U22#n%s%2t%%2:st-. $!2!$1?2Ef%d2 bl.%s%M20$sififi2diT%l$ d ( t!id%d%2h3 (f)lyACH:i%t?#Cf3n!ML3%fl5Mbd%Ol'!%32%$ste!!n #E% t sEma WissWe n:n L!aisisited %i!% 4MtsS!!M22l1 7J Et$$titinHi!%$2 int.14Mili? E t%d%21 6 h!W W2M Os#!2MiC 4"%"gjS%2.&sanf wdRyp.?m!$?:5@!Rt DM*%O s n ss au :nw:an $.$jij 25!!.$.; $.9F?[}. A..M$.,.IN..y4;j.f_((g^. . Y[fj [ + a g4 g g~, .....2

sN.N$I3M7Ndeess V+.c*s#Hs*Disd@d'WN' M5hf[-EbM4 I

-M$Nf5'MN:*/N5IkY$3[$h

' ENi - :.[.Mb$!!5 )$$NNNh!N d@2222Mti%s%ME%%%!m$ K!* !% !nntitis2tinLMN2!!S20 2nO!$Es$sEijW" SW2P4 msy,s %~ W._w.19331+020!GW312sW' p -m. -:<...m.e. d-l m us-mm:.,+ ? M V.WAM-

i Table 4.1 Summary of Containment Matrix and Conseqeence' Bin Assignments for Surry Contaltnent Containment Late Failure Due failure Due Overpressure to Early to Late Fallure of Containment Overpressure Hydrogen Burn Containment Failure Containment Containment te Sequence Precedes Spray Spray Spray Spray Cont. Cont. Basemat Isolation Containment Containment Evaluated Core Nelt Falls Operates falls Operates Rupture Leak Melt-through failure Bypass Failure TMLB' O.005* 0.13 0.38 0.20 0.02 0.09 0.002 0.18 1 9 8 9 10 14 6 15 50 0.004 0.04 0.005 0.05 0.15 0.02 0.16 0.002 0.58 8 3 1 2 9 8 9 10 13 7 ,15 5 , s. e 50 0.001 0.01 0.006 0.06 0.15 0.02 0.28 0.002 0.48 2 3 4 9 8 9 10 13 7 15 AD 0.17 0.02 0.30 0.002 0.51 9 10 13 7 15 AF 1.0 5 V 0.75. 0.25

11. 12**

• For a.

15 the conditional probability of the containment failure given a core melt condition, and b. 1s the assigned bin. 0*ll assines the break occurs under water, while 12 assunes otherwise. O O 9

Table 4.2 Source Term Bins for the Surry Plant Ilme of Release Warning Release Energy of Fraction of Core Inventory Released Release Duration Time Elevation Release 10' 8tn (hr) (hr) (hr) (m) (Blu/hr) Xe-Kr Cs-Rb Te-5b Ba-5r Ru La 1 2.5 10 .5 10 2 1.0 7.5x10

• 5.8x10
• 5.5x10 3 1.0x10 '

l.3x10 ' l.7 10 ' 2 1.5 1 .5 10 2 1.0 7.5x10

• 5.8x10
• 4.2x10 8 5.1x10 '

1.3x10 ' l.4x10 5 3 2.5 3 .5 10 11 1.0 5.1x10 ' l.1x10 " 1.1x10 2 3.2x10 8 1.tx10 ' 9.4x10 ' 4 2.5 1 .5 10 0.7 1.0 5.0x10 8 4.810 5 2.5x10 ' 6.8x10 ' 2.6x10 ' l.3x10 ' 5 12 10 2 10 2 1.0 3.9410 1 3.8x10 I 2.0x10 ' 8.6x10 ' l.2x10

• 1.7x10 "

$g 6 2 10 0 10 2 1.0 2.7x10

• 1.3x10**

1.2x10 I 6.2x10 ' l.6x10 3 2.6x10 ' 7 1 10 0 10 0 1.0 5.3x10 3 2.6x10 ' 1.8x10 ' l.8x10 ' 2.7x10 5 -,5.8x10 # 8 8 1 6 10 0.7 1.0 5.0x10 ' (1.10 ' 1.5x10 ' l.2x10 5 4.3x10

• 2.1x10 '

9 12 1 10 10 ' 1.1 1.0 7.8x10 3 3.9410 ' 8.5x10 ' 1.8x10

• 3.3x10 '

8.1x10 5 10 2.5 10 .5 10 0.2 1.0 6.9x10 ' 1.1x10 8 1.3x10

• 5.8x10 3 1.7x10 '

2.4x10 ' 11 1 2 .8 0 3.4 1.0 8.4x10

• 7.3x10
• 2.5x10 8 2.2x10
• 4.5x10 '

1.1x10 8 12 1 2 .8 0 3.4 1.0 4.1x10 ' 4.0x10 ' l.2x10 ' 1.3x10 3 2.7x10

• 6.4x10 8 13 24 1

22 0 0 1.0 - 0 0 0 0 0 0 14 24 1 22 10 0 3x10 2 1.5x10 ' 0 0 0 0 0 15 cegitgtble A

CORE MELT FREQUENCY AND RISK VS. PLANT DAMAGE STATE SHCP D&M PLANT CORE RISK DUE CORE RISK DUR DAMAGE FAILURE MELT .TO EARLY MELT TO EARLY STATE NODE FREQ. FATALITIES FREQ. FATALITIES SE LATE GE-S 1E-4 GE-6 SE-6. AE LATE GE-6 IE-5 5E-7 SE-7 V3 EARLY 4E-6 IE-3 IE-7' 2E-5 TE LATE 9E-7 IE-6 6E-8 SE-8 IE-4 IE-3 7E-6 2E-5 Totals [ SHCP = Seismic Hazard Characterization Program D&M = Dames and Moore. O O 9 a

ax,r Cot 4\\)OLtd.IlOE O sgep of set 6M1C i a m ou I O T fut4cTton RAtkeb d i 5 I' [ p@tettxty I D i x

1. o i

O m tL & / x, + ~ 6 i A=.9 C.b i ds I g> Fe%ttny 4 i t h= i.e1 %' 6 k i i a i ~ .4 4 g 4 i y \\ s J 3 1 3 2 3 4 i I ll-36 4 \\ \\ r\\ 1 s \\

• i L

\\ \\ N O x .h .k .i, oh 1.0 12. l.4 141.02.02.12.& O 6Els Mtc. Gr LE\\JE L (

Table 1 CRAC II Sensitivity Study Ground Release 50 New' York Weather Sequences with Probability of Release = 1.0, Time of Release = 4.0 hours, Ouration of Release = 1.0 hours, Warning Time = 0.5 hours, Heat of Release = 0.0 BTil/hr, Height of Release = 10.0 meters, No evacuation but relocation after 5 hours. Shielding factors: Ground = 0.33, Cloud = 0.75 -(average activity), Population = uniform distribution of 100 people per square mile. Base Case Isotopic Grouping Release Fractions Release Fractions Xe 1 I C C Te Te Ba Ba Ru Ru La la La 2 2 3 3 Xe

1. 0

1. 0 1

0.1

1. 0 0.1 2

C 0.1 1.0 0.1 3 Te 0.1

1. 0 0.1 Ba 0.01 1.0 0.1 Ru 0.01

1. 0 0.1 La 0.001 1.0 0.1 0.001 Acute Results Fatalities 8

0 46 1 7 0 26 1 66 3 229 18 1410 107 0 Injuries 20 1 115 8 3 0 63 4 22 2 732 54 4180 356 1 Latent Results Initial 29 2 24 3 42 4 56 5 105 11 999 94 4850 472 4 Total 482 2 74 9 1800 55 56 6 3170 251 1980 233 8120 986 12 Groupings Xe = Xe,R g ; != I ; Cs = Cs, Rt ; Te = Te, Sb; Ba = Ba, Sy ; Ru = Ru, Co, Mo, Tc, Rh ; La = La, Y, Zr, Nk, Ce, Pr, Nd, Np, Pu, Am, Cn O O O

23 ] Table 5.1 Comparison of Release Characteristic Bins and Frequencies i 1985 Surry RSS-$urry IP233 IP311 Zion 3* M111 stone-3 9 (NUREG-0956)2 (Rebaselined)12 1*" PWR-1 8 8 Z-1 M2, M3 4E.8 1E-7 3E-8 3E-8 1E-8 2 F F 2E-7 SE-6 6E-6 3 0** D** 3E.8 IE-6 1E-6 4 3E-7 5 5C 3E-7 2x$0-s 6 PWR-4 G" G** 4 M4**, 7E-9 3E-7 3E-7 2E-8* 7 PWR-5 G** G** 4 M4** PiiR-8 7E-8 SE-8, 4E-5 3E-7 3E-7 2E-8 8 3E-6 2-5, 83 M65 SE-7 5E-6 9 PWR-3" C C 2R 321 M7" 7E-6 3E-7 1E-5 1E-5 1E-7 9E-6 10 7E-7 11 3E-6 12 PWR-2 8 8 2, 4, 81 M1A 1E-6 7E-6 4E-7 4E.7 1E.7 8E-7 13 PWR-6 M** M** 6 M10 3E.7 6E-7 6E-5 6E-5 14 PWR-7 M** M** 7 M9 SE-6 3E 5 6E-5 6E-5 4E-5 15 PWR-9 I I 8A, 88 M12 . 2E 5 4E 4 2E-4 2E-4 3.5E-4 IE-4

• Frequency < 10-8 yr-1
  • Release category is not readily assignable to NUREG-0956 bins.

i

• " 4 : is the source term bin, and b: is the frequency (yr-3)

[ O i M'--"-----

O Table 4.3 Assignment of Reactor Safety Study Frequencies to NUREG-0956 Source Term Bins and C-Matrix (Per Reactor Year) Bin Frequency (yr 1) 1 4.1 x 10 8 2.4 x 10 7 2 3.1 x 10 7 3 3.1 x 10 7 4 2.7 x 10 7 i 5 i 6 6.6 x 10 ' t.O 6.8 x 10, 7 ,i 2.7 x 10 8 8 6.6 x 10 ' 9 7.5 x 10 7 10 3 x l0 ' 11 1 x 10 ' i 12 3 x 10 7 l 13 8.5 x 10 ' 14 2 x 10 8 j 15 0 0 ,,r O f \\.

.=,,, - l O Table 1 PLANT DAMAGE STATE VERSUS EARTHQUAKE ACCELERATION LEVEL Median Fragility Plant Damage State Equipment Failure 0.20 g SE (1) offsite power and emergency generator failure due to masonry wall collapse and/or .88 g oil cooler bolt f ailure and/or .91 g control building collapse and/or 1.00 g service water failure 1.30 g or O (2) offsite power and 0.20 g failure to scram due to core geometry distortion or O.87 g CRDS f ailure and 0.88 g RWST footing failure 0.08 g i 1.59 g AE RCS pipe break and RWST footing failure and 0.88 g (see above) station blackout V3 crane wall collapse 1.82 g e ESFbuildingcolkspse 1.70 g TE --- _ _ _.}}