ML20148T353
| ML20148T353 | |
| Person / Time | |
|---|---|
| Issue date: | 01/28/1988 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1638, NUDOCS 8802030284 | |
| Download: ML20148T353 (198) | |
Text
A C 2 5 t--l 6 5 F l
1 O UNITED STATES NUCLEAR REGULATORY COMMISSION
- ...........=..........=.=........=.........=..............=....................=.==
IN THE MATTER OF: DOCKET NO:
DECAY HEAT REMOVAL SYSTEMS MEETING
'\ A _
O LOCATION: WASHINGTON DC PAGES: 1 _ 149 DATE: JANUARY 28, 1988
- ..........................................................................=...=..=
,;l)[f7II'7{D!i"DGB31.
u w. . . 'i - .. . c:
.. .';m Lq w,a ,a ,
30 Not itemove from ACRSOffia Heritage Reporting Corporation Official Reporters 12 L Strat. N.W.
O Wunsngton. D.C. 20005 (2021 628-4488 8802030284 880128 PDR ACRS T-1638 PDR
em 1 PUBLIC NOTICE BY THE J 2 UNITED STATES NUCLEAR REGULATORY COMMISSION'S 3 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 4
5 6
7 The contents of this stenographic transcript of the 8 proceedings of the United States Nuclear Regulatory 9 Commission's Advisory Committee on Reactor Safeguards (ACRS),
10 as reported herein, is an uncorrected record of the discussions 11 recorded at the meeting held on the above date. l 12 No member of the ACRS Staff and no participant at 13 this meeting accepts any responsibility for errors or 14 inaccuracies of statement or data contained in this transcript.
15 16 17 18 I l
19 20 21 22 i l
23 24 25 i
() Heritage Reporting Corporation (202) 628-4888 C
1 t%
U-1 UNITED STATES NUCLEAR REGULATORY COMMISSION 2 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 3
)
4 In the Matter of: )
)
5 MEETING OF THE SUBCOMMITTEE )-
ON DECAY HEAT REMOVAL SYSTEMS )
6 )
7 Thursday, January 28, 1988 8
Room-1046 9 1717 H Street, N.W.
Washington, D.C. .20555 10 The above-entitled matter came on for hearings 11 pursuant to notice, at 8:30 a.m.
12 BEFORE: MR. DAVID A. WARD 13 Research Manager on Special Assignment
("T s/ E.I. du Point de Nemours & Company 14 Savannah River Laboratory Aiken, South Carolina 15 ACRS MEMB_ERS PRESENT:
16 MR. CHARLES J. WYLIE 17 Retired Chief Engineer Electrical Division 18 Duke Power Company Charlotte, North Carolina 19 DR. CHESTER P. SIESS 20 Professor Emeritus of Civil Engineering University of Illinois 21 Urbana, Illinois 22 23 24 25 O
l HERITAGE REPORTING CORPORATION -- (202)-628-4888 l k _. - )
2 l
f' -
i O 1 MR. CARLYLE MICHELSON Retired Principal Nuclear Engineer f 2 Tennessee Valley Authority l
Knoxville, Tennessee
! 3 and Retired Director, Office for Analysis and Evaluaton of Operational Data 4 U.S. Nuclear Regulatory Commission Washington, D.C.
5 l_
j DR. CARSON MARK
! 6 Retired Division Leader
) Los Alamos Scientific Laboratory 7 Los Alamos, New Mexico l 8 MR. JESSE C. EBERSOLE Retired Head Nuclear Engineer 9 Division of Engineering Design Tennessee Valley Authority 10 Knoxville,. Tennessee f 11 CONSULTANTS: I l 12 P. Davis I. Catton 13 p'd ACRS COGNIZANT STAFF MEMBER:
14 Paul Boehnert 15 NRC STAFF PRESENTERS:
16 Jerry Mazetis 17 E. Chelliah 18 19 20 21 22 23 24
{
25 O j HERITAGE REPORTING CORPORATION -- (202)-628-4888
I 3^
i-1 INDEX f
2 . Items Discussed Page 3 Status of USI A-45 Resolution 8 4 Review of EPRI/WOG DHR Risk Analysis Report for Point Beach 5 Plant 67 6 Review of NSAC-113 by Sandia Lab 103 7
8 i ,
9 i t-i 10 !
11 12 f i
9 13 l
r 14 .t 15 l
[
j 16 l 17 l.
18 19 :
t i
> 20 .
21 .
t l
22 ;
4 23 -
1 t
- 25 I t
HERITAGE REPORTING CORPORATION -- (202)628-4888 (
--__~--m.-___.-+---m____ ---e---- - _ _ _ _ __.___---.e*
4 j
\
1 P R O C E E D ,I-?J G S l
_ 2 CHAIRMAN WARD: The meeting will now come to order.
3 This is a meeting of the Advisory Committee on Reactor 4 Safeguards Subcoms.ittee on the Decay Heat Removal Systems.
5 I am David Ward, the Subcommittee Chairman. The 6 other ACRS members in attendance are Mr. Ebersole, and Mr.
7 Mark, Mr. Michelson, and Mr. Siess will be here I believe.
8 Mr. Wylie is here, Charlie.
9 Also in attendance are two ACRS consultants, Mr.
10 Catton and Mr. Davis.
11 The purpose of the meeting is t'o continue our review 12 of the NRC staff resolution for USI A-45. Paul Boehnert is
(} 13 the cognizant ACRS staff member for the meeting.
14 Rules for participation in today's meeting have been 15 announced as part of the notice of the meeting, previously 16 published in the Feceral Register on January 13th, 1988. A 17 transcript of the meeting is being kept and will be made 18 available as stated in the Federal Register notice. It is 19 requested that each speaker first identify herself or himself 20 and speak with sufficient clarity and volume so that she or he 21 can be readi]y heard.
22 Before we go to the advertised speakers, I want to 23 make a couple of comments to the members. A-45, of course, 24 has been proceeding along at its measured pace for many years.
25 We seem to be approaching a resolution. I had hoped that in HERITAGE REPORTING CORPORATION -- (202)628-4888'
_ _. . . _ _ _ __ . . ~ _, .__ _ _
5 1 advance of this meeting, we would get some documentation,
's_/ 2 draft documentation, from.the NRC staff with some more 3 specific information on the resolution position they are-4 proposing.
5 We haven't gotten that. We have had indications of 6 approximately what the position is going to be, and'I think we 7 will hear more about that today.
8 We don't intend to go to the Full Committee with ,
9 this in February since we didn't have written material ~and 10 time is kind of short, but I am kind of reluctant to see the 11 thing drag out too far and so depending on what we hear today, 12 I may propose to you that we go to the Full Committee in March
() 13 even though we may or may not have any further formal 14 documentation from the staff in the interim.
15 If we hear enough today so that we think it would be 16 useful for the Full Committee to have to form some sort of 17 consensus our own position on resolution of the A-45, then I 18 will propose that we go ahead and try to do that in or at 19 least begin to do it in the March Full Committee meeting.
20 One other thing, and we are going to hear about this 21 today, but I wanted to call your attention to something.
22 There was a, and I think we saw this sometime ago, an industry 23 group I guess under the auspices of EPRI and'the Westinghouse 24 owners group and through NSAC, the NUCLEAR SAFETY ANALYSIS
(~)g 25 CENTER, published a report, NSACl-113, which was their review HERITAGE REPORTING CORPORATION -- (202)628-4888
6'
~
1 of'the risk analysis of decay heat removal that had been done 2 for the Point Beach plan by the NRC staff contractors. dsing 3 essentially the same inputs that the staff' contractors had 4 used, they came to some rather different conclusions.
5 I asked our consultant, Mr. Davis, to review the 6 report and make some sort of account'.ng and assessment of the 7 differences, and he has done that in a report which I think 8 has been available to you. His report was issued January 9 lith. You probably got that at home. If not, it was attached 10 to the meeting notice that Mr. Boehnert put out I believe, so-f 11 we may be discussing that and raay have a f ew ques tions ~ f or Mr.
12 Davis also in that part of the meating.
() 13 Any other members bave something they would like to 14 say before we start out?
15 MR. MICHELSON- Yes, I had one small comment. I 16 think we use the term dedicated decay heat removal, and I 17 think that when we use the term each of us has in mind- j 18 something maybe a little different than the next person as to 19 what do you mean by dedicated heat removal?
20 It is not clear to me that the Committee has a 21 position on what they mean. It is not clear to me that the 22 staff does, and maybe some time during the day today we could 23 get a clarification of exactly what kind of a system does the 24 ste.lf have in mind? It is covered a little bit in the NSAC 25 document. It is covered a little bit in the Point Beach HERITAGE REPORTING CORPORATION -- (202)628-4888
7 1 document. I would like to know what does the staff really-('T k/ 2 have in mind when they are thinking about evaluating dedicated 3 decay heat removal?
4 CHAIRMAN WARD: Well, I think, you know, since 5 dedicated system is part of one or two of the options, I think 6 that it would be appropriate when they describe those for 7 you--
8 MR. MICHELSON: Just to clarify for sure exactly 9 what they have in mind because it is a little different than I 10 think the Committee has, along the lines the Committee has 11 been thinking, although I am not sure of that, either. We 12 never really got into it, we just kept using the term without
() 13 necessarily defining it.
14 MR. EBERSOLE: Since that came up, I would like to 15 add to it one of the most difficult problems is to identify 16 the conditional state at the time of entry, decay heat removal 17 system into service, because I know quite well, for instance, 18 accommodate a large LOCA, I don't think it will, yet it might 19 accommodate a small one or certain other transients or upseta, 20 but it is not going to be all-encompassing in its 21 capabilities.
22 CHAIRMAN WARD: Dedicated system concept.
23 MR. EBERSOLE: Will have-boundaries of performance 24 that we need to identify.
25 MR. WYLIE: In that regard, maybe we need to HERITAGE REPORTING CORPORATION -- (202)628-4888
8 1 ' redefine what the objective to dedicated heat removal system
[~)
(s 2 is. I mean what are you trying to accomplish?
3 MR. EBERSOLE: Right.
4 HR. WYLIE: Maybe they could comment on that. I 5 think I've got my idea. Certainly mine doesn't include LOCAs.
6 MR. EBERSOLE: Not any at all?
7 MR. WYLIE: Not at all.
8 MR. MICHELSON: This is something we certainly do 9 have a mixed mind on in the Committee and I'm not sure the 10 staff has one mind on it, and that's what I am kind of curious 11 to find out.
12 CHAIRMAN WARD: I assume they know what they mean by 13 it 14 MR. MICHELSON: I assume so.
15 CHAIRMAN WARD: The system.
16 MR. MICHELSON: I just want to make sure that I 17 understand where they are coming from.
18 CHAIRMAN WARD: Okay. Thank you. Let's--the first 19 speaker for the staff will be Warren Minners.
20 MR. MINNERS: Well, I just want to make a few 21 introductory remarks. Jerry Mazetis will be giving you a l
22 presentation which will be a summary I guess of the regulatory I l
23 analysis which we are now working on and have in draft form i
24 and will try to provide you, but in this measured case of
)
25 A-45, I guess we have gone along more than is usual in sharing HERITAGE REPORTING CORPORATION -- (202)628-4888
9 1 with the ACRS the resolution of the issue before the staff has
{/ \
2 all of its positions lined up, and I have no problem with 3 that. In fact, I think it is a good procedure. .I just want 4 to remind the ACRS that what you are hearing is, you know, 5 preliminary stuff, and don't expect-us to stand up and defend' I
6 it to our last breath.
7 We also will later on have a discussion of the l 8 EPRI-Westinghouse owners group analysis of Point Beach which I 9 think is an important question of is decay heat removal a 10 problem? And industry has asked that question and I think we 11 are obligated to answer it, and the purpose or one purpose of 12 the case studies was to answer that question--what is the risk 13 to failure of the decay heat removal function? And once
(])
14 again, what you are going to hear is going to be very 15 preliminary. We haven't discussed this stuff in detail with 16 our contractor. We haven't had a meeting with the people who.
17 did the report, so this is all again very preliminary stuff.
18 I would, since Carl brought it up, I would like to 19 address this point. I'm not sure that you have to define very 20 well what the decay heat removal system would be to address 21 the issues, but at this point we are trying to address the 22 issue, not trying to address fine technical detail. The 23 question comes that any reasonable dedicated system is going 24 to cost'a lot of money, and is it worth it? And that's the 25 question I think we have to be focusing at this point rather HERITAGE REPORTING CORPORATION -- (202)628-4888
10 l l
1 than any' technical design details of decay heat removal. l
' _)
( 2 Dedicated systems are described in the case studies and those 3 are the systems that were used for making a cost estimate. So 4 I think we do have a very specific idea of at least one-5 system. That doesn't mean that has to be the system, but 6 that's the one we used as an example for doing.the 7 cross-studies, and I think they are described in fairly good-8 detail in the case study.
9 Unfortunately, we didn't bring any of that material ,
10 with us so I can't flash it on the screen, but maybe later on, 11 Jerry could give a description of the system, so with that 12 cautionary note that this is all preliminary information, I
{} 13 would like to discuss it with you free and technical 14 discussion, and why doesn't Jerry go ahead? <PHAZ) <PHAZ).
15 CHAIRMAN WARD: Thank you. Jerry Mazetus.
16 DR. MAZETIS: My name is Jerry Mazetis, with the 17 Reactor and Safety Issues Branch in Research, and I'm sorry to 18 say that the new task manager, Dr. Woods, has reported sick 19 call this morning, so I will fill in for him, and forgive the 20 rough edges on the slides.
21 (Slide) l l
22 DR. MAZETIS: This slide is intended to refresh cur 1
23 memory on some of the background on how we got where we are, 24 and it goes way back. It actually seems like yesterday to me, 25 but even before TMI, there were concerns of decay heat HERITAGE REPORTING CORPORATION -- (202)628-4888
. 111 1 removal, concerns being discussed 1and debated-in-house; and
.p'
- v 2 TMI essentially served as a' springboard to formalizing a ~ '
3 variety of decay heat removal issues, some of which were O 4- implemented post-TMI with some corrective-. actions related'to.
5 decay heat removal components.-
6 However, there was still as a result of discussions ,.
7 post-TMI in-house chrough ACRS various tack forces and-8 commissions, still' recommendations that high priority be given 9 to further study of improving the shutdown decay heat removal 10 function.
11 The studies the past seven years conducted by Sandia l 12 focusing on six power plants were included to cover a special 13 emergency such as fire,-flood, seismic ~and sabotage.
14 Not too long after THI, when a lot of these concerns 7
15 had been documented, USI A-45 was. formally identified'and 16 approved in 1980, and the key questions focused on A-45 was do 17 current designs provide the reliability needed to meet the 18 safety goals, core damage frequency guidelines of safety 19 goals, and are there improvements to the. decay heat removal 20 function in operating plants which aren't cost effective?
21 (Slide) 22 MR. MICHELSON: Question--if you don't really define 23 what you mean by dedicated decay heat removal very explicitly, 24 how do you do the evaluation of the improvements that, and 25 risks that come from adding a decay heat removal of undefined-HERITAGE REPORTING CORPORATION - '(202)628-4888 s
12-1 . state? How'do you go about doingithat?
>-)
k/ 2 DR. MAZETIS: Sandia years lago:when--the plants 1 were 3 ' selected and some ground' rules were made,-Sandia~and the. staff-4 agreed.on a certain. level of detail to one.ofLtheir 5 alternatives.- As.I will show later,-there'are six 6 alternatives that were. costed out in.the.past seven yearsLin 7 the case studies, one'of which is a so-called dedicated heat 8 removal system where certain system' assumptions were madelin 9 order to cost the components out, and essentially the 10 system--and perhaps lateroon this morning Sandia can get'into 11 more details--but the system is a, an aux feedwater type 12 system, separate from the normal auxiliary system in a
() 13 separate structure along with a certain amount of makeup' 14 capability, and the makeup capability is limited to the small.
15 breaks that were analyzed in the studies.
16 MR. MICHELSON: Wha *. I didn't find but.what I was-17 somewhat interested in is well, how does high pressure decay 18 heat removal show up? If I were to build high pressure decay 19 heat removal dedicated system, if I were to put in inside j 20 containment versus outside containment, how does that show up 21 and so forth?
22 Was that kind of work done, or did you just pick one
- 23 system which was really nothing more than an additional l 24 auxiliary feedwater and additional high pressure makeup?
I 25 DR. MAZETIS: The makeup, the assumptions, as you HERITAGE' REPORTING CORPORATION -- (202)628-4888
13-1 know, for small breaks, my recollection is the system was
(~)
\- 2 capable of handling small breaks.
3 MR. MICHELSON: Two hundred gallons; it is capable 4 of 200 gallons a minute.
5 CHAIRMAN WARD: You are talking about small, about 6 a--
7 MR. MICHELSON: High pressure makeup for whatever 8 reason.
9 CHAIRMAN WARD: High pressure makeup.
l 10 MR. MICHELSON: That's what they show. And another 11 auxiliary feedwater pump, 200 GPM on high pressure and 1200 on f 12 auxiliary. I think that was it. And you know, is that what 1
() 13 we have in mind, or is there other systems? And I wanted to
'14 see a discussion of other kinds of systems and how they would i
15 appear. This may not be the best option. j l
16 DR. MAZETIS: Sandia may have a comment. Dave? Dr.
17 Ericson from Sandia.
18 DR. ERICSON: There was a comparison of high percent 19 RHR, these kinds of shutdown systems done, and that was 20 published in the '83 studies on decay heat removal.
21 MR. MICHELSON: That was '83, though, and I was--
22 DR. ERICSON: This was selection as a result, this 23 approach was selected as the way to, as the result of those
(~g 24 studies.
\/
25 MR. MICHELSON: Showed this to be the cheapest.
HERITAGE REPORTING CORPORATION -- (202)628-4888
14 1 DR. ERICSON: This was a, based on those studies,
\m-) 2 this was a reasonable way to go.
3 MR. EBERSOLE: I would like to ask a question. You 4 are obligated by law in fact to accept a hodgepodge of 5 designs. And you consider them here, six of them. I would 6 expect in the course of your investigation you would' find 7 substantial differences in dealing with decay heat removal 8 from plant to plant, from type.of plant to type of plant.
9 I was somewhat enthusiastic about the GSAR 2 new PPS 10 which approached in my view pretty close to what might be 11 called the decay heat removal system of extreme simplistic 12 character.
() 13 In your course of assessment, did you characterize 14 these differences or degrees of difficulty as you wing through 15 the plants, and ascribe to them whether, you know, this one or 16 that one needed it more or less?
17 DR. MAZETIS: As you will see in subsequent slides, 18 one of the reasons, maybe the main reason being considered for 19 not backfitting the dedicated heat removal system or for that 20 matter any hardware right now is the f a :t that there was this 21 variety of plant-specific differences that was significant in i
22 the results of the study. l i
23 As you will see in two or three slides, that it was, 24 a major part of the risk vulnerabilities for decay heat 25 removal was seen to be in the balance of plant area.
HERITAGE REPORTING CORPORATION -- (202)628-4888
i 15 1 MR. EBERSOLE: What bothers me there is that because n
(_f 2 you have this composite assignment, you find, you may find 3 that if you can't find it is applicable to all~of them, you 4 won't apply it to any of them. And I think that's a mistake.
5 DR. MAZETIS: Yes, I agree, and we think we proposed 6 a path that will try to avoid the situation of ending up with 7 nothing.
8 In this slide, we go over the objectives of A-45 9 which was to--this bullet--determine the adequacy of decay 10 heat removal in existing plants for achieving both hot 11 shutdown and cold shutdown conditions; developing and 12 evaluating again alternative method for improving the i
(} 13 reliability of the decay heat removal function; assessing the 14 value and impact of alternative methods following the backfit )
I 15 rule; and establishing new requirements, if needed, to reach l 16 the decay heat reliable compatible with our interpretation of 17 the core damage frequency guidance goals.
18 (Slide) 19 DR. SIESS: What is your interpretation?
20 DR. MAZETIS: I think what I was referring to there 21 was the discussion of avoidance of significant release of 22 radioactivity of ten to the minus 6 and backing off to produce 23 a, an interpretation of core damage frequency in the area of 24 ten to the minus 5th.
25 DR. SIESS: I'm sorry. Ten to the minus 6 for large HERITAGE REPORTING CORPORATION -- (202)628-4888
16 1 release and then you back that down to ten to the minus 5 for r's V 2 core damage frequency?
3 DR. MAZETIS: Right, because you have to.
4 DR. SIESS: Core damage means core on the floor?
f 5 DR. MAZETIS: Core damage, core melt frequency, core
! 6 damage frequency because on top of that, you have to add the l l l l 7 containment failure probability. l l
l l 8 DR. SIESS: Core damage means what? Core on the l l i I 9 floor? Or TMI? j 10 DR. MAZETIS: I don't think we have identified 11 I l either way. I'm not sure.
12 MR. MINNERS: I think it means core on the floor, 13 Dr. Siess, but as you know, we can't differentiate in a PRA 14 between those, but I think, I mean when you get to the minus 6 15 we take it a factor of ten for the containment and that means 16 core on the floor. l 17 DR. SIESS: Core on the floor and you say we don't 18 know enough from our research to establish any relationship f 19 between severe core damage like TMI and core on the floor?
20 MR. MINNERS: Well, I didn't mean to be that l 21 extreme. I think in the PRAs that are done, we don't make the 22 distinction, but no, I think there is in the research, in fact f
{
23 part of the accident management program is to try to have some f
24 procedures for controlling the phenomena between core damage 25 and core on the floor, t
HERITAGE REPORTING CORPORATION -- (202)628-4888
17 1 DR. SIESS: Yet you think: we kr.ow enough to' state
, () 2 that the additional probability of containment failure given 3 core on the floor is approximately one 10th, which is what you 4 are using to get from the ten to the minus 6 back to the ten 5 to the minus 5, but that you are not able to put a conditional-6 probability of core on the floor given some severe core 7 damage.
8 MR. MINNERS: Well, these are very rough numbers. I 9 mean there is not a lot of precision --
10 DR. SIESS: Factor of ten is a guess, but it is 11 just, it is, just seems interesting that we know enough to 12 apply that factor on containment failure; with all the s' 13 research that we have done, we can't put any kind of a factor 14 at all on the conditional probability between core damage and 15 core on the floor.
16 MR. MINNERS: Well, I think you probably could. If 17 people were pushed into it, people would be willing to make 18 some guesses, but it doesn't seem worthwhile because the PRA l 19 methodology document doesn't do that. You do a PRA, you just J 20 say you lose cooling and don't really say that this results ir.
21 core damage or core melt or whatever. l 22 DR SIESS: Why are we doing all that research?
23 MR. MINNERS: To try to get some accident management i
24 procedures that we can go and, and give guidance to operators !
(
25 that when they do start to get core damage they will be able HERITAGE REPORTING CORPORATION -- (202)628-4888
- .. - ~. -- -
_ ~ .-
18 t
1 co do something about it.. I mean-the purpose if the research 2 is not to do PRAs, although that's a useful' function, but what l
3 you want to try to do is have something-at the plant and 4 that's my view of why we are doing research is accident 5 menagement.
6' DR. SIESS. You think that view is shared by the 7 people doing the research?. That's a very interesting
}
8 statement. I am inclined to agree with it. But--
9 MR. MINNERS: Since we have one of the research '
10 laboratories here, I won't embarrass him by asking him,why he 11 is doing the research. I don't know. I can't speculate why.
12 I presume that they understand that part of the purpose of the 13 research is for accident management. I mean certainly there 14 is a lot of effort now to.try to have accident management. We ,
15 had one lab write up a report on I insight for severe core 16 damage accidents with a very purpose of trying to make people s 17 and operators aware of what the phenomena are; maybe something 18 happens, that they will be able to respond more intelligently I {
19 to it having the background information. t 20 DR. SIESS: I am not disagreeing. I think that's 21 great, but you did state that the purpose wasn't to do PRAs.
l 22 HR. MINNERS: Not completely to do PRAs. That's one )
l 23 purpose, but I think from our view if you really want to l
l p 24 increase safety, that research has to be applied in accident 1 %
l 25 management procedures.
l HERITAGE REPORTING CORPORATION -- (202)628-4888
19 1" DR. SIESS: :Okay, r
kh) 2 s CHAIRMAN WARD: Jerry, before you go on, you know, 3 on the last chart, you talked about the, being compatible with 4 .the core damage frequency goals specifically.
5 DR. MAZETIS: Right.
6' Cf. AIRMAN WARD: You seem to come up with a number of 7 i ten to the minus 5th for that.
8 DR. MAZETIS: That's the number being discussed now 9 that may appear in the reg analysis.
10 CHAIRMAN WARD: That's the number that you are using 11 in the reg analysis. Do you intend that to be--but that's not 12 necessarily the number that is being discussed in the safety
(} 13 goal policy implementation?
14 DR. MAZETIS: No. I am saying that was an 15 interpretation of the guidance in the safety goal. That's 16 being discussed in the staff to use as a screening tool.
17 CHAIRMAN WARD: Another interpretation of the 18 guidance in the safety goal which is yet not completely formed 19 is ten to the minus 4th core damage frequency. So why, I 20 don't know how important this distinction is in what you do, 21 but why have you chosen to, seemingly to hang your hat on ten 22 to the minus 5th instead of ten to the minus 4th?
23 DR. MAZETIS: All I can say is that in our 24 discussions with the--as you will see later, we are working
~}
,J 25 closely with the IPE severe accident group, with the HERITAGE REPORTING CORPORATION -- (202)628-4888
20 1 individual plant examination evaluation program, and that's (G,/ 2 compatible with their thinking right now. I don't know. I
- 3. don't know why the, I can't explain the difference between 4 what the current thinking is compared to previous thinking at 5 ten to the minus 4th. All I am saying is-right now, with A-45 6 and the thinking on screening tool, it's going to be, it is 7 compatible with the thinking in the IPE severe accident group, 8 and it looks like the numbers being discussed are ten to the 9 minus 5th.
10 DR. SIESS: It bothers me that the decisions that 11 can cost hundreds of millions of dollars are going to be made 12 on an interpretation, rather fuzzy one, of the safety goal 13 that presumably was promulgated by the Commission to give us
({}
14 some guidance what is safe and what isn't.
15 Are we ever going to get better guidance on what the
't 6 safety goal really is in terms of things that are meaningful 17 in a plant, hardware, or procedural or whatever? Whether it 18 is ten to the minus 5 or te.1 to the minus 4, we make a 19 tremendous difference in the amount of money that is spent to 20 meet it.
21 DR. MAZETIS: I agree. I hope the Commission reads 22 your statement in the transcript.
23 DR. SIESS: You know, we are still trying to f- 24 interpret a large release. We have got a fair amount of
, \ ,
25 fuzziness around it. Ten to the minus 6th was pretty explicit HERITAGE REPORTING CORPORATION -- (202)628-4888
i 21 1 1
1 but the large' release, we have had at least four different 2 definitions of it that I know plus a couple more maybe. And I
3 then the factor of ten, this Committee suggested as a. goal for
]
4 containment reliability, but I don't know of any' data that l
5 tells us overall it is a factor of ten or that it is a factor {
6 of ten for all reactors or that is just an average.
7 CHAIRMAN WARD: This--well. Chet, there is a meeting l l 8 with, another subcommittee is meeting with another branch of-9 the staff on February 9th to talk about this, but you know, in 10 the meantime, these user programs of the safety goal are 11 having to go ahead without safety goal implementation being 12 finished so it presents a problem and I think it could be a 13 serious one.
14 DR. SIESS: We are setting ad hoc safety goals as we 15 go along.
16 CHAIRMAN WARD: Neither this program nor the 7'E 17 program should be setting its own safety goals because there 18 is a program which is intended to do that for the full agency.
19 MR. MINNERS: The staff operates in the vacuum.
20 CHAIRMAN WARD: Right. Pete, you wanted to say 21 something?
i 22 MR. DAVIS: I guess I have problems with trying to )
23 translate tne probability of a serious release into a core 24 damage frequency. I don't know of any PRA that has come out i
l 25 with a core damage frequency less than ten to the minus 5th, l l
HERITAGE REPORTING CORPORATION -- (202)628-4888 i i
22 1 and furthermore, if one examines the PRAs that are available,
() 2 you find a.very wide range of conditional containment failures 3 given a core damage frequency. For example, Seabrook, 4 Seabrook PRA has a core damage frequency in excess of ten to !
I 5 the minus 4th, but the conditional probability of containment )
i 6 failure is less than ten to the minus forth, so that plant j 7 meets the safety goals with considerable margin but certainly; 8 would not even come clase to your ten to the minus 5th core 9 damage frequency.
10 Other plants, some PWR, do I believe approach a 11 factor of ten conditional probability of containment failure.
12 The point is I guess you can't, I don't think you can
(} 13 establish a ten to the minus 5th core damage frequency and 14 apply it to all plants as appears to be what you are trying to 15 do here. I realize there is more work that needs to be done 16 in this area, but I think that it is going to be a problem to 17 establish ten to the minus 5th as the, as at least an informal 18 criteria for core damage frequency.
19 MR. MINNERS: What you are suggesting, Pete, is that 20 we have different goals for different plant designs?
21 MR. DAVIS: Well, I think you about have to if you i
22 are going to use core damage frequency. You need also to look l l
23 at the capability of the containment. I don't have an answer, l 24 really, but I think that--
25 MR. MINNERS: I think the discussion is not focused, HERITAGE REPORTING CORPORATION -- (202)628-4888
23 1 and the other question is that core damage frequency by itself 2 you have, can have a number. The Commission has stated in the 3 safety goal policy statement that they don't want to have any 4 more core melts, so that also gives some kind of guidance to 5 the staff.
6 CHAIRMAN WARD: Ten to the minus 5th satisfy that?
7 MR. MINNERS: I think so.
l 8 MR. DAVIS: I don't know of a plant out there --
9 CHAIRMAN WARD: That is interesting.
10 MR. DAVIS: If that in fact is the requirement, I ;
11 think most plants are going to need some kind of modifications ]
i 12 because if one lcoks at the PRAs, as I said, I don't think
( ; 13 that there is a plant out there that meets ten to the minus 14 5th or less.
15 MR. MINNERS: I agree with you. If--we had the same 16 problem in, ATWS okay. The unavailability of the SCRAM system 17 hasn't changed over ten years. The only thing that changed 18 this ATWS was the goal. Start at ten to the minus 7th, and j 19 when it finally got to ten to the minus 5 we declared victory.
20 MR. DAVIS: The initial safety goals had a core melt
) 21 frequency requirement of ten to the minus 4, and I thought 22 that was based in part on consideration of the existing 23 population and how long they were going to operate, and then
() 24 looking at what the probability should be to gain some
] v 25 assurance that you wouldn't have one over the remaining HERITAGE REPORTING CORPORATION -- (202)628-4888
_ _ _ _ . _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . __ I
24 1 lifetime.
} 7 s
j 2 HR. MINNERS: That got deleted for some reason, so
[ 3 somebody didn't like it.
I 4 MR. DAVIS: Now you are using --
l 5 CHAIRMAN WARD: No, it might be coming back in in l (
l 6 the implementation policy.
)
l 7 MR. MINNERS: It is a very useful discussion because 8 the staff is at sea as to what is, the goals to be attained, f 9 so it is a question that really has to be answered and 10 discussed and some decs.sions made.
11 MR. DAVIS: I think there should be a core damage 12 frequency goal, and I think that ten to the minus 4th is a 13 reasonable value for it, but I don't think other people agree
(']/
w f
l 14 with that.
15 CHAIRMMI WARD: I wouldn't say that. I think that f 16 the ACRS letter of when was it, last April I guess, which the 17 commissioners have told the staff to use in large measure for 18 development of the implementation plan, has that sort of a 19 number in it.
20 MR. DAVIS: All right.
21 DR. MAZETIS: Yes, sir.
22 MR. EBERSOLE: After you get done with the core f
- 23 damage frequency or the severe release probability, it is g, 24 unclear to me now how you get to the money question because
()
25 there is two kinds. There is the physical damage to the HERITAGE REPORTING CORPORATION -- (202)628-4888
25 1 public, the man rim, there is the loss of_ resource which is
/ 2 substantial.
3 I, in my own view, it would take a view of both of 4 these, and a view as to how this loss will be borne. It is 5 going to be borne by the public, the rate payers, and we have 6 a public consideration and loss of resource.
7 If it is borne by the corporate entities, that's 8 another picture. They can bear it if they wish. I think you 9 should look at both of these, whether it is borne by the 10 public or borne by the corporate structure, and come to grips 11 with whether you are dealing with a public problem there or a 12 private one.
13 MR. MINNERS: We have done that, Jesse. We' don't I'_T
\/
14 feel that we can tell people how to make decisions, how they 15 do it, but we have tried to present all of the facts, so we 16 have all of the different costs cut every which way you want 17 to. In fact, maybe it is too complicated. If you want to ask t
18 about a cost, we have got it.
19 MR. EBERSOLE: Great. That will come to the front 20 then.
21 MR. MICHELSON: Question -- you said in your second 22 bullet that you evaluated such systems for vulnerability to 23 fire, flood, seismic, and sabotage.
24 Are we going to discuss later today in a little more 1
O-25 detail what might be some serious shortcomings about the way HERITAGE REPORTING CORPORATION -- (202)628-4888
1
- l. .' l I 26 l l
l 1 in which you do that evaluation?
2 DR. MAZETIS: I suggest -- .
l 3 MR. MICHEL3ON: The effect upon perhaps the l l
l l 4 usefulness even of your proposed changes?
I 5 DR. MAZETIS: ~Let me respond to that in two parts.
I j 6 The direct answer is we didn't plan on it, and no, because as f 7 Warren indicated, you don't have the benefit.of our regulatory
~l
. 8 analysis yet, what discusses the details of these areas, and 1
9 that it probably is more appropriate to give you a chance to l 10 take a look at the details and at that time, we can come down 11 again sometime within the next month or so and discuss the l
12 details with the Committee. {
j O 13 MR. MICHELSON: Is that the intent, David, to have )
14 another subcommittee meeting when we get the regulatory
}
15 analysis?
16 CHAIRMAN WARD: Sure. :
f 17 MR. MICHELSON: Then I will just hold until then. -
l 18 MR. MINNERS: I am not sure.
19 CHAIRMAN WARD: You have seen a lot of that stuff in l 20 the past.
f 21 MR. MICHELSON: I have seen it also in this NSAC-13, l 22 and I have a number of comments about the, the superficial way
{
l 23 in which we treat fire in a PRA and to a great extent, flood, ;
l (
l r 24 internal flooding and so forth. 1 1
25 MR. MINNERS: If the Committee would like to have a {
I J
HERITAGE REPORTING CORPORATION -- (202)628-4888 {
}
27 1 meeting on document studies and discuss the methodology, I A
k) 2 guess that could be arranged. I mean that's all laid out for 3 you, Carl.
4 MR. MICHELSON: I have no problem with that. I just 5 want to go on record as do we believe our fire PRAs that we 6 have been doing? If you do, then there is something wrong.
7 There is something wrong with your head.
8 DR. SIESS: There is something wrong with you if you 9 believe them.
10 MR. MICHELSON: And if we don't believe it, then is 11 what you are proposing going to solve the problem of fire 12 vulnerability and how do you know it when your analyses are so
() 13 superficial? I just have a problem knowing that we really are 14 improving the situation at all with your proposed dedicated 15 decay heat. I am not saying you are proposing, but if you did 16 go to the dedicated aux feedwater, for instance, how do we 1
17 know that is any less vulnerable from fire than what we ,
1 18 presently have when we do such a superficial analysis on our 19 present situation? Just an inquiring question.
20 CHAIRMAN WARD: Carl, maybe--excuse me just a 21 minute, Jerry. Maybe at the end of the meeting today we as a 22 committee need to discuss whether, I guess because of the long 23 time laps that is necessary, to go back over some of the 24 ground that we have -- went over.
25 MR. MICHELSON: What is really happening is we are HERITAGE REPORTING CORPORATION -- (202)628-4888
a 28 1 b] ginning to understand the problems a little better I think,
(~h s/ 2 and that what we knew in 1983 is quite a bit different than 3 what we know today, and it might'even be more interesting in i
4 another month or so after we get the results of the current 1 5 look at the vulnerabilities to fire which are coming up in a 6 March meeting.
7 CHAIRMAN WARD: I see. Let's--Paul, remind me, 8 let's talk specifically about that at the end of the meeting.
9 DR. MAZETIS: Maybe I should mention, remind the 10 agenda plan today really has three speakers. Dr. woods had 11 planned to give the overview leading into where we are into 12 the current reg analysis, and how that relates to the
() 13 submittal by NUMARC of the EPRI owners group study for Point 14 Beach. We had two speakers to talk about the details of their 15 review of that report, E. Chelliah and Dave Ericson from 16 Sandia laboratories, and as Warren also indicated, there has 17 not been really any give and take yet amongst these 18 individuals, so there may be some redundancy and overlap in 19 what you hear from these gentlemen. Also in the audience we 20 have some other staff members that may be familiar with the 21 reports that they can respond to some questions perhaps.
22 (Slide) 23 DR. MAZETIS: This slide goes over scope of the A-45 24 program to evaluate decay heat removal systems needed to f5 respond to transients and small break LOCAs, and as we have HERITAGE REPORTING CORPORATION -- (202)628-4888-
29 1 said over the years, this study did not include large break
( 2 LOCAs or ATWS, and several other. areas such as pressurized 3 thermal shock. Also included were an assessment of systems 4 vulnerability to fire, flood, seismic, and insider sabotage 5 where sabotage again was, vulnerability of decay heat removal 6 systems to the insider, and which explicitly did not cover 7 physical security method or the outside or attack from 8 external areas.
9 MR. MICHELSON: What did it cover then on the 10 question of sanotage? How did you evaluate insider 11 vulnerability in some kind of a quantitative way so that you 12 included it in the risk / benefit evaluation which is what you
{} 13 were doing?
14 MR. MINNERS: I don't know how to answer that. I 15 don't think there is a short answer to that question.
16 MR. MICHELSON: I don't think there is sn answer to 17 it.
18 CHAIRMAN WARD: Short answer, it wasn't evaluated 19 quantitatively, but it was qualitatively.
20 MR. MICHELSON: What does that do for risk / benefit?
21 CUAIRMAN WARD: They don't use that directly. As I i
22 recall, you don't use that directly in the risk / benefit 23 computation.
24 MR. MICHELSON: When you get the risk / benefit you 25 ignore sabotage?
HERITAGE REPORTING CORPORATION -- (202)628-4888
1 i
30 l
l 1 DR. MAZETIS: 7.t wasn't ignored. I 2 MR. MINNERS: Excludes sabotage.
I 3 CHAIRMAN WARD: On a page alongside -- \
4 MR. MICHELSON: The inference here is somehow you {
5 evaluated it, and I just don't think you did much on it l I l f 6 really. l l l 7 CHAIRMAN WARD: I don't think --
8 MR. MICHELSON: If you did, I didn't find a f 1
9 discussion of it anywhere.
10 MR. EBERSOLE: You say consider. That's a very- l l
11 large word. To me, I hope it means you noticed the i I
f 12 vulnerability notably of certain systems and you then, at l l !
13 least you conceived something like crash buttons that would l 14 block out the potential for that vulnerability to close in on l l 15 a safe shutdown. I don't know whether you did that or not.
I 16 DR. MAZETIS: Perhaps Dave could amplify, but it 1
l 17 was, there was, in order to cost out certain aspects of the j i
18 different alternatives, there were certain assumptions made l 19 about insider sabotage probability, and there was no in-depth j l 20 studies on likely types of probabilities given certain l 1
21 configurations. For costing out. There were certain gross l
l 22 assumptions made.
l l
23 Dave, do you want to amplify on that?
24 DR. ERICSON: I think that's the key point. As we 25 did with several other things, and as the staff has done in HERITAGE REPORTING CORFORATION -- (202)628-4888
-a 31 1 the regulatory analysis, some things are presented as a range
() 2 for the decision-maker to make his own choice. What 3 quantification we did in the sabotage analysis, where we felt 4 systems were particularly vulnerable, we can discuss in a 5 closed meeting, not in this meeting. We did some conditional 6 analyses, Mr. Michelson. We did not say the probability is.
7 We said if you believe it is--
8 MR. MICHELSON: Is that written up somewhere?
9 DR. ERICSON: Those reports are now being, they are 10 appendices to the individual case studies and they are at the 11 printer's now.
12 MR. MICHELSON: We will see those shortly.
(} 13 CHAIRMAN WARD: We have had drafts of those, haven't 14 we?
i 15 DR. ERICSON: Several of them you have had 16 CHAIRMAN WARD: Maybe not for each of the ones, but 17 you may have not participated in those subcommittee meetings, 18 Carl.
19 MR. MICHELSON: Most of them, but not all of them.
20 CHAIRMAN WARD: We heard about it. I would like to 21 point out that everyone else in the audience has to use a 22 microphone when they speak from the audience, but Dr. Ericson 23 doesn't because we can hear him.
24 DR. ERICSON: Good!
25 DR. MAZETIS: Okay. Refresh our memory with the HERITAGE REPORTING CORPORATION -- (202)628-4888
32-1 A-45 studies, considered several plants, excuse me, six O 2 plants, and the idea at the time years ago was to select a 3 group that represented each NSSS vender to try to cover a 4 range of decay heat removal configuration, and 5 architect / engineers, and the, of course, as always when you-6 are looking for plant-specific data, it depends on the 7 cooperation and the interest and how busy the utility is at -
8 the time of the data needs.
9 Document studies themselves are all completed, as 10 you know, and have been published this past year in the form 11 of six NUREGs for each of the six plants. The more recent 12 being last August, St. Lucie, and the next report from the
() 13 contractor is going to be a compilation or consolidation of 14 the findings for each of those six plants sometime early this 15 year when we get that published so I would expect that most 16 people wanting to know the details of the studies would focus 17 on the last report, and to use these as references as needed 18 rather than going 1ight into reading six reports that are 19 three inches thick each.
20 CHAIRMAN WARD: What about the sabotage analyses Il that are in appendices? What is the status of tnose?
22 DR. MAZETIS: There is some discussions of the 23 methodology in each of these reports and Appendix I, as you '
, 24 have seen, is left out of the report. I think Dave just 25 mentioned that he is pulling together--
l HERITAGE REPORTING CORPORATION -- (202)628-4888
33 1 DR. ERICSON: Those are in the final sign-off and r~s
(-) 2 printing process now, Dave. We should see those within the 3 next month or so much.
4 CHAIRMAN WARD: So they will be coming out as six 5 separate reports?
6 DR. ERICSON: They are a safeguard information 7 appendix to each of the case studies, that is correct. They 8 have to be handled separately in our view.
9 CHAIRMAN WARD: How will that be dealt with in the 10 summary?
11 DR. ERICSON: Some of the detail of why it is 12 particular things are vulnerable are left out.
() 13 MR. DAVIS: I have a related question. On your 14 slide under the selection criteria, you didn't list what I 15 thought was the primary selection criteria, and that is i
16 determination of a potential vulnerability to decay heat 17 removal.
18 DR. MAZETIS: Yes. You are right. I should have 19 mentioned that that, the criteria up on the slide was after a 20 group of maybe 80 or 90 plants that were initially looked at.
21 There was a larger group that was felt to be candidates of 22 trending toward the more vulnerable plants in decay heat 23 removal and out of that group, this was the other criteria.
24 This slide is, goes over a summary of the findings 25 that you will see in the six case studies where the l
HERITAGE REPORTING CORPORATION -- (202)628-4888 -
.. ~ .- - -- .- ..- -- .- , - .. - - - - _ _ _
34-1 probability--the first bullet. The probability of core melt-2 due to decay heat removal function failure'is between two and 3 three, averages between two_and three times ten'to the minus 4 4th per reactor year total for decay, heat removal,: including 5 internal, external causes.
6 The second bullet indicates that, an inside-7 observation that support system failures turned out to be a 8 significant contributor to the core melt probability, l I
9 particularly emergency power, service water, component cooling I 10 water areas.
11 MR. MICHELSON: Excuse me. On the support system 12 failures, were those related to just simple equipment failures
() 13 or were those related to the external event problem?
14 DR. MAZETIS: Both.
15 HR. MICHELSON: Both, yes, but can you give me some 16 kind of feel? Was it predominantly equipment or was it 17 predominantly external event effects or was it about half and 18 half or something?
19 DR. MAZETIS: Depended on the plant. It varied from 20 plant to plant.
i 21 MR. MINNERS: I think, as a very rough thing, Carl, 22 half of the core melt frequency in decay heat removal was 23 internal and half was external, what are called external 24 events.
25 MR. MICHELSON: The support systems, those are much l
i HERITAGE REPORTING CORPORATION -- (302)628-4888 !
35 1 more vulnerable from external events perhaps than some of the 2 other equipment. I don't know. Generally from experience it 3 seems that they are, if you do a analysis. I just wondered if 4 that is how you came out. So it is 50/50 roughly?
5 MR. MINNERS: It varies from plant to plant a lot 6 of--
7 CHAIRMAN WARD: As I recall on the six, it is a 8 range from 20 to 80 percent.
9 MR. MICHELSON: It never did tell me whether it was t
L 10 coming from the support system effects or whether it was 11 coming from the main line aquipment effects. Didn't tell me 12 whether it was the decay heat removal system or was it the
() 13 electrical power to the decay heat removal that was causing 14 that, s
15 DR. MAZETIS: In each of the caso study reports and 16 also to be repeated in the summary for each plant, there is a 17 list of the vulnerabilities, the specific system 18 vulnerabilities. For example, for Point Beach, I know there 19 were about 11, and that's a reference where you could get that 20 information. We can maybe the next meeting where we get into .
21 a discussion of some of the specific numbers for each case 22 study, we could get into that in as much detail as you like.
23 The observation of, again from insights from the study, i
24 redundancy, concerns or lack of redundancy, sharing of 25 systems, support system level, contributed to the risk HERITAGE REPORTING CORPORATION -- (202)628-4888
36 1 significantly for some of plants, including independence,
() 2 separation and physical protection of redundant safeguard 3 trains. Risk from fire, flood, seismic, and sabotage again 4 depending on the plant, was significant. And so that the 5 bottom line you will see that a large number of these areas--I 6 want to make this point now because it leads into the next 7 couple of slides--did vary quite a bit from plant to plant, 8 and was plant-specific.
9 MR. EBERSOLE: Every time you use the term 10 redundancy it brings up the vision of two carbon copies of 11 some function that have common mode potential failures. You 12 don't mention diversity and SCRAMS or sub-systems or even the 13 word at all. Is that deliberate? You don't think diversity
/}
14 is significant?
15 DR. MAZETIS: No. Diversity is significant, but I 16 think that the redundancy, vulnerability comes from 17 observation of walk-throughs at the site, for example, which 18 did see sharing of components and systems, and certain sy. stems 19 that are supposed to be redundant trains are located in the 20 same space side by side, that kind of thing.
21 MR. EBERSOLE: Another aspect of failure, you are 22 talking about functional plant, not to deliver some sort of 23 service in general, you know, step function to zero service of ,
24 water pressure, air pressure, voltage, whatever; do you O 25 consider failure to embody alsc interactive effects like one HERITAGE REPORTING CORPORATION -- (202)628-4888
37 1 pump, for instance, stalls another pump because they can't 2 operate in parallel, or excessive pressure of voltage or 3 frequency as modes of failure or for that metter, a very 4 typical one is failure in the sense of oscillatory performance 5 rather than total step failure?
6 DR. MAZETIS: There was a--maybe Dave can amplify, 7 but there was, I consider what you just went through in the 8 area of common mode concept where the common causes were 9 considered in the studies, and the specific failure rates 10 depended on the system. I don't know what they are, but there 11 is a discussion on what our assumptions were for common cause 12 failures.
13 MR. MICHELSON: Could you clear up for me a general
(])
14 question? For instance, we went through a great deal of 15 efforts under Appendix R to affix spreading rules because of 16 fire vulnerabilities, and we performed certain fixes that were 17 reviewed by the regulators and so forth, and now cfter having 18 done all of that, we go with you and we do this analysis now, 19 these various plants under this A-45 program, and find some 20 plants have an extremely high vulnerability to fire.
21 Does that mean that Appendix R fixes weren't good 22 enough or because the numbers are also fairly high, the 23 probabilities are fairly high? Does that mean that we just 24 didn't do a good enough job or we didn't do the analysis right 25 to begin with or what happened that says that after all the i
HERITAGE REPORTING CORPORATION -- (202)628-4888
38
> r j
1 looking and fixing and reviewing and inspection, and whatever, O
(/ 2 that we still ended up with high vulnerability on some plants?
3 What went wrong?
4 MR. MINNERS: I think that's your characterization.
5 That is high vulnerability. I mean--
6 MR. MICHELSON: It is fairly high. Ten _to the minus 7 5 range is fairly high, and 60 percent of the risk 8 contribution in that plant was fire and spreading, but that 9 percentage doesn't mean much. It is the actual magnitude that 10 counts. It is a fairly high magnitude. Some way on pipe i
11 breaks, you know, fire water header, well, that was supposed 12 to have been looked at way back in 1973 when the Gene Busso
() 13 letter went out and so forth. No energy pipe breaks are 14 supposed to be looked at. They shouldn't be high contributors 15 if they hsd been looked at carefully and properly reviewed and 16 fixed.
17 HR. MINNERS: I don't agree with that assessment. I 18 think on the fire thing people looked at it; it was, said 4
19 well, okay, we have given an exception to the full 20 requirements and not have to have spray sistem. We realized i l
21 that entails some extra risk, we find it acceptable and when l
22 you do the PRA, that quar 'fies the number.
(
23 Now I guess J- Jan look again and say is that 1
4 24 whatever it is times ten to the minus 5 acceptable or not?
25 MR. MICHELSON: If it is acceptable, then why are we 1
HERITAGE REPORTING CORPORATION -- (202)628-4888
39 1 even thinking of adding any additional hardware or whatever to 2 correct the problem? You would think that we already buy off f 3 on the numbers, there is nc reason to worry about the problem 4 further.
l 5 MR. MINNERS: Well, one of-the problems is, Carl, is
}
6 that when you add up all of the vulnerabilities in the plant, I
l 7 okay, that may be something that you want to fix. That's a 8 question you have to answer.. Whether you want to go at these l 9 individual things and fix them or whether you want to take the 10 big fix and take care of it all.
11 MR. MICHELSON: What bothers me a little bit is that 12 it isn't just a fix. It is a fix again proposition. We have 13 already fixed it once I thought. Didn't the Appendix R work?
l 14 And now we go back and fix it again because it is a big 15 contributor yet to risk? If it isn't, then we don't do 16 anything I would think. ;
17 MR. MINNERS: I think it is prudent to go back and 18 look at what you did in the past and say hey, did I do it f
19 right? I know more now. I have methods of analyzing it, had 20 more operating experience, should I do something about it. I 21 don't think just because the case study came cut with two 22 times ten to the minua 5th for vulnerability that says that it 23 should be fixed. I certainly think it puts it up as a 24 candidate that people ought to think about.
25 MR. MICHELSON: What puzzles me a lictie bit is I HERITAGE REPORTING CORPORATION -- (202)628-4888
) 40 1 guess we just'didn't quite understand the problem well enough O
(-) 2 to fix it right the first time which was presumably the 3 Appendix R work, which is fairly recent work.
4 DR. MAZETIS: This slide gets into the current-5 summary of where we are in the staff regulatory analysis, 6 which is in essence, of course,- a translation of the completed 7 contractor activity, Sandia, into a staff position, and you 8 may recall that in, on September 26th, 1986, we provided a 9 pre-decisional rough draft of the case studies in a regulatory 10 analysis format, and we presented it to you in a closed 11 session, and we described the six alternatives that were being 12 considered. At the moment, we have a, an updated version,
() 13 revision that is being prepared that will again describe, 14 describe the same six alternatives, and as you heard a couple 15 of months ago from my boss, Brian Sharon, that the thinking l
16 still is to propose a resolution focusing on the thrust of 17 alternative 2 that is described in, described in the 18 regulatory analysis, which in essence is a plant-specific 19 studies, PRA studies.
20 And just to refresh your memory here, the six 21 alternatives were the first one, no action. There was a 22 discussion of, there will be a discussion of no action, and 23 obviously if it is determined that the NRC, NRC/Sandia studien 24 are too conservative, then that may be justified. In other 25 words, if the EPRI owners group type of study is concluded to HERITAGE REPORTING CORPORATION -- (202)628-4888
41 1 be a more appropriate or more correct, then probably you will
() 2 end up in this arena.
3 MR. MINNERS: Like Pete said, that is also a problem 4 of goals. Depends what goals are selected. You pick two 5 times ten to the minus 4, I agree with you, you don't have to- 5 6 fix anything.
7 DR. MAZETIS: Alternative 2, of course, we are 8 focusing on as a proposal would direct the resolution and i.-
9 merit to marry it with the severe accident program, in 10 particular individual plant evaluations.
11 Alternative 3 --
12 CHAIRMAN WARD: Could I ask you a question about 13 that? When this was--let's see. In September '86 when this
) i 14 list was, first alternatives described, I don't recall that, I 15 mean I recall alternative 2 was some sort of limited scope 16 PRA, but I don't recall that it was explicitly tied to IPE at 17 that time, because that wasn't very well defined.
18 DR. !!AZETIS : You are right, and this next draft is 19 being updated to reflect the marriage of A-45 or actually not ,
20 A-45 specifically. It is going to be decay heat removal per 1 21 se, to the severe accident program. '
22 Alternative 3 was a, a list or composite of 23 modifications that were focused on the cost-out that were ,
l 24 common, more or less common vulnerability seen in the six s
25 plants.
HERITAGE REPORTING CORPORATION -- (202)628-4888 L . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ __ _ - - - - - _ -
~
B 42-1 Alternative 4 was focused on two areas for PWRs. It 2 was feed and bleed, by cost / benefit on feed and bleed and for 3 BWRs venting the containment.
4 MR. EBERSOLE: Call out a non-parallelism or in 5 those two aspects, feed and bleed.
6 DR. MAZ3 TIS: These two?
7 MR. EBERSOLE: Right. There is a large mount of 8 homogeneity in those being in the same place. Feed and bleed 9 does not accommodate containment heat removal, and I, whereas 10 containment venting does.
11 Do you mean containment feed and bleed and still 12 removing heat from the containment via heat exchangers as is 13 normally done? Of course you do. And containment venting,
(])
14 doesn't that always incorporate open cycle boiling out of the 15 primary vessel that is blow-down? So to really make that--
16 CHAIRMAN WARD: Feed and bleed.
17 MR. EBERSOLE: To feed and bleed is open boiling out 18 of the vessel for PWR, but it does not get rid of containment 19 heat whereas the BWR with open boiling contained venting does 20 the whole thing. And I think that ought to be, you know, the 21 differences ought to be stressed. Do you follow me?
22 DR. MAZETIS: Yes. I think they are identified in 23 each case study, a description of the process that was costed 24 out.
25 HR. EBERSOLE: The former is not as comprehensive as t
HERITAGE REPORTING CORPORATION -- (202)628-4888
43 1 the latter.
( 2 MR. MINNERS: PWRs don't need heat removal with 3 their big containments.
4 MR. EBERSOLE: That's a statement that I find very 5 interesting.
6 MR. MICHELSON: That is.
7 MR. MINNERS: They don't need it for a long time.
8 They are built to take a large break LOCA, okay, so they can 9 go a long time before you have bit of a problem, 10 CHAIRMAN WARD: Pressure pool can, too.
f 11 MR. EBERSOLE: Sure.
12 CHAIRMAN WARD: It is the same capacity.
() 13 MR. MINNERS: I agree with you. .They are different l 14 systems. They are not the same system, but to say that one is 15 bad and one is good --
! 16 CHAIRMAN WARD: No. I think that Mr. Ebersole is 17 pointing out that, I mean the BWR already has the feed and 18 bleed, and under this you are adding to it a means of getting i
19 the heat out of the, you know, abnormal situation out of
]
i 20 containment by containment venting. For the PWRs you just are 21 introducing feed and bleed and there in no provision for the 22 second step.
23 MR. EBERSOLE: Exactly. They are not homogeneous 24 objectives.
25 MR. MICHELSON: Could you clear up for me then a HERITAGE REPORTING CORPORATION -- (202)628-4888
44 1 point? Don't most PWRs have the inherent capability now of
() 2 feed and bleed? A few do not I know, but don't most'have it 3 already?
4 MR. MINNERS: No. Most, yes, there are some CE 5 plants that don't have PORVs.
MR. MICHELSON: Don't the majority already have the 6
7 hardware and the capability? Maybe not yet the instruction or 8 whatever?
9 MR. MINNERS: They all are supposed to have the 10 procedures for feed and bleed, but they have--
11 MR. MICHELSON: Most of them already have feed and 12 bleed.
(} 13 MR. MINNERS: their capability is different degrees.
14 They ha,e sized PORVs.
15 CHAIRMAN WARD: Reliability of the capacity is 3 j 16 different.
17 HR. MICHELSON: That's what I am trying to get down 18 to. Is the reliability differences a matter that you can't ,
19 get removal of the decay heat by feed and bleed or what? ;
20 CHAIRMAN WARD: Both.
21 MR. MICHELSON: Depends on what point in time then i
22 you are talking about. You are talking about five minutes !
23 after the event or talking about an hour after the event or :
24 whatever? Most of them have feed and bleed capability within ;
O.
25 reasonable times.
l HERITAGE REPORTING CORPORATION -- (202)628-4888
m 1:
45 ,
1 DR. MAZETIS: The importance--it may be of interest,
( 2 Carl, to take a look at the importance if you were focusing on 3 feed and bleed, for example, to each of the case studies. It
- 4 varied, of course, with each plant as to its contribution to 5 reducing core melt frequency because the individual accident
, 6 sequence varied from plant to plant and its contribution to 7 overall core melt frequency varied'from plant to plant, so for 8 one plant, this may not have been very important. For another .
9 plant, it may have been very important in reducing core melt i
10 frequency.
11 Again, the last two alternatives were the two 12 versions of dedicated decay heat or hot shutdown capability,
(} 13 and cold shutdown capability where the last one included both 1 14 hot and cold shutdown capability.
15 MR. EBERSOLE: Again, there is a disparity. You 16 look at the designs. I don't know what you mean by hot, but 17 in the case of the boiler, at 220 degrees you can zimmer 18 forever if you vent the containment. And I don't know whether 19 you call that hot or not, but it certainly is not hot to, you 20 know, pressurized context, the depressurized boiling system, j 21 with the body containment pressure.
1 22 Ed, is that hot shutdown?
4 23 DR. MAZETIS: I guess this terminology is probably t y!
r 24 more appropriate for the PWR where, where the intent would be i !
25 to--this was a system that did not have the capability, this ;
HERITAGE REPORTING CORPORATION -- (202)628-4888 :
--- - - -_ ~ -
t i-l 46 1
1 dedicated system in a separate structure, did not have the V 2 capability to continue the plant into cold shutdown. It was l 3 in essence the aux feedwater capability with a limited makeup l
l 4 for small breaks, and it was only after you added the shutdown i
f 5 heat removal heat exchanger type pump, heat exchanger f 6 structure that we costed.out the additional dedicated cold 7 shutdown heat removal.
i 8 MR. EBERSOLE: One system, of course, has a l
9 capability prior to core damage, just direct boiling. The l 10 other requires exchangers because of the coolant activity.
11 And it requires, you know, multi-step path to get to
{
l 12 atmosphere.
l 13 You don't develop these differences in these l
14 alternative 5 and 6?
15 DR. MAZETIS: tio . tio . There wasn't really much 16 detail. You may recall that initially we had planned to just i
17 cost out five alternatives where the A-45 studies did not f 18 intend to cost out the additional capability for long-term 19 decay heat removal, and it wasn't until toward the end of the 20 exercise that it was decided to do our best"to tack on the 21 costs associated with adding a cold shutdown capability, so l
)
22 there were some gross assumptions made.
23 MR. EBERSOLE: Cold shutdown capabilities involve p 24 sensible heat removal with heat exchangers and pumps.
V 25 DR. MAZETIS: Yes.
HERITAGE REPORTIliG CORPORATIOli -- (202)628-4888
.- .yy a
,47 1 CHAIRMAN WARD: Jerry, could I go back? Leave that, 1
("T l i
\-) 2 the feed and bleed for PWRs again. As I recall the six case 3 studies, well, none of them showed feed and bleed to be as 4 important as perhaps was expected at least by some people 5 before the studies, but there was variability in the 6 importance.
7 DR. MAZETIS: Yes, there was variability.
8 CHAIRMAN WARD: My question, was that variability 9 due to the fact that, the variability and capability for feed 10 and bleed in the various plants, or was it due, just due to 11 the fact that the need for feed and bleed didn't enter into 12 any actual sequences? Do you see what I mean?
I~ 13 DR. MAZETIS: My recollection, it is probably more
'v) 14 the latter because where you had a plant that maybe had a poor 15 aux feedwater system, with poor reliability, the numbers were 16 not too good, feed and bleed really came out looking great and 17 the contribution of feed and bleed to reducing overall core 18 melt frequency looked more significant than the other plant.
19 You had a great aux feedwater system, so it came into 20 sequences like that. That's why it varied raainly f rom plant 21 to plant.
22 CHAIRMAN WARD: It could be in those two plants, the 23 actual capability to feed and bleed could be essentially the 24 same. You would still --
(~J) :
25 DR. MAZETIS: Sure.
HERITAGE REPORTING CORPORATION -- (202)628-4888
< 4 8^'
'l CHAIRMAN WARD: Okay.
( 2 (Slide) 3 DR..MAZETIS: This againLfrom in the reg anal'ysis 4 that you will see in the near future,' emphasizes that for the
-5 decision-maker, that it gives him. calculations using three-6 methods, and it actually may confuse the: decision-maker more 7 than clarify. Anyway, the methods that were presented --
8 CHAIRMAN WARD: Might confuse them. It still might 9 lead to a better decision.
- 10 DR. MAZETIS
- If he understands the three methods;.
11 as you will.see in the' reg analysis or you ha're seen in 12 earlier versions, that the three' methods.are A, avert a,
()
13 considering averted off-site costs only; B,'considering 14 averted off-site plus on-site costs; and C, all-the above plus 15 the effect.of after-decisional special considerations such as 16 the possibility after an accident of reactors being shut down 17 in moratorium or other unquantifiables'and.using these three 18 methods, you will see and you have seen in reg analysis that, 19 that depending on which method you use,1you may come out with 20 a different answer.
21 You use method A, it is.possible that alternatives 22 2, 3 and 4 from the previous slide may be cost effective.for 23 some plants. If you use method B, of ' course, now the same 24 alternatives may be cost effective-but they would be more cost i 25 effective. And of course, if you use the additional, it is
, HERITAGE REPORTING CORPORATION -- (202)628-4888 L .
. .. .-- ~. . -. - - . - . -~ .
- 1 49
- 1. .very possible averted costs oof -these 'special considerations in
( 2 method C, even alternatives 5'and 6 may_be cost-effective for.
3 some plants.
4 Mk. EBERSOLE: In the evolution of these costs,.did' 5 you give consideration that in the application of say: ,
6 alternatives 5 and 6, you could' cut.back on the extremely, 7 complex and substantial investment in protective features'that-8 you could either soft pedal or eliminate which is put-there-9 now to cover the extended vulnsrability of the plants?
10 DR. MAZETIS: That is certainly a possibility, but
~
11 the ground rules had to be' laid in order to cost things'out.
12 MR. EBERSOLE: So you didn't go back and say if you
() 13 did this then you.didnt have to do fourteen other' things?-
. 14 DR. MAZETIS: No.
15 MR. EBERSOLE: Why isn't that a substantial--
16 HR. MINNERS: I thought it was surveillance or 17 something we thought could be cut back.
18 MR. EBERSOLE: .What about QA, and safety category, 19 categorization, et-cetera, in perioheral systems which are now=
20 all over the place?
21 DR. MAZETIS: For example, for alternate 5 and 6 the i 22 assumption was made. It was installed as safety equipment.
(
23 MR. EBERSOLE: I know, but you didn't, you.didn't.
! p 24 take advantage of going back and including up all the l v 25 peripheral equipment which had to be safety categorized HERITAGE REPORTING CORPORATIOli - " (202) 628-4'888
50 1 because they formed a part of the overa]1 vulnerability of the n
'(>' 2 plant?
3 DR. MAZETIS: No. No.
4 MR. EBERSOLE: Those savings would be substantial.
5 DR. MAZETIS: Well, in essence that's alternative 2 6 now. Remember alternate 2 were, the rationale for choosing 7 alternate 2 is that here is six plants, and here is for each 8 plant what we see as ten or 11 vulnerabilities to their 9 existing plant, and so we costed out for alternate 2 value 10 impact, a backfit if that plant had backfitted those 11 vulnerabilities, so in that sense, we considered improving the 12 vulnerabilities we found, so it is covered in the scope of
() 13 alternatives.
14 Yes, sir?
15 DR. MARK: Value impact analysis is a numerical 16 thing. It is arbitrary as may be. You still put in a 17 thousand dollars as if it had anything to do with the man rem 18 and so on.
19 But how you can even, how you can allow yourself to 20 talk about a value impact analysis based on unquantifiables 21 escapes my imagination.
22 That's just a judgment to me. It may be that the 23 method C says we should do it. It will be important to do it.
24 It has got a big point or something, but you can't possibly 25 put it in the bracket of saying it is cost effective.
HERITAGE REPORTING CORPORATION -- (202)628-4888
51 1 You can put a number on sabotage. It is either one 2 or zero or it is something in between, and the whole list 3 there. It is a value judgment, and I'm not. objecting. It,may-4 be worth doing, but it can't be called cost effective.
5 DR. MAZFTIS: My response to that is yes, there is a 6 lot of uncertainty and vagueness to these areas that I!have 7 listed in method C, and we included them because when'we. asked 8 ourselves the question should we ignore them, the answer was 9 no. That's the decision-maker has to cover, include the 10 spectrum of considerations, and although we may not have as 11 good a handle on all these areas as we have other areas, it's 12 something that shouldn't be ignored, so that's why I was
() 13 including that.
14 MR. MICHELSON: Question -- the proposed dedicated 15 aux feedwater plus high pressure makeup arrangement, was that 16 considered alternative 5?
17 DR. MAZETIS: Yes.
18 MR. MICHELSON: Thank you.
19 MR. KNIEL: Jerry, could you list which one, which 20 ones were quantified in the alternative C moratorium, 21 quantified in three different ways as I recall, or four 22 different ways?
23 DR. ERICSON: That is correct.
fs 24 MR. KNIEL: Put the slide back up again on V
25 the--which of the ones were quantified and which weren't?
HERITAGE REPORTING CORPORATION -- (202)628-4888
-._ r
52 1 DR. ERICSON: Okay. Sabotage in the' conditional-
/~T
(_/ 2 sense allowing you to make the judgment of what likelihood is, 3 but a' range was included. The moratorium was quantified in 4 several ways. The resolution or cut was made at the economic 5 impact of the resolution of the other generic issues. Both of 6 those had dollar values assigned. We did not do those. The 7 staff did those.
8 DR. MARK: What you are saying is that tney are all 9 subjectively filled in?
10 MR. MINNERS: Sure.
11 MR. KNIEL: We did actually make some dollar values.
12 DR. MARK: You invented numbers to go with them.
(} 13 DR. ERICSON: I wouldn't say we invented them, Dr.
14 Mark.
15 MR. KNIEL: I think the moratorium thing was treated 16 fairly cleverly because we handled it in one, was it three 17 different ways or three different ways, right, so you had your 18 choice of how you wanted to look at it.
19 DR. MARK: That's what I have been saying.
20 MR. KNIEL: Fine.
21 MR. MICHELSON: How did you treat the generic issues 22 then, and there are quite a few of them, and in A-47, A-17 and 23 so forth?
24 DR. MAZETIS: Excuse me?
25 MR. MICHELSON: Proceed. You are going to tell me.
HERITAGE REPORTING CORPORATION -- (202)628-4888
53 1 DR. MAZETIS: What that means is wo looked at the 50 (m
(_) 2 or 60 or however many there are_ generic issues, and using
, 3 judgment, for example, one-of them was A-44, Station-Blackout, 4 conditioned on the, a backfit of alternative 6 which was the 5 dedicated heat.
6 MR. MICHELSON: Why don't'you take.a more difficult 7 one like A-17, which you haven't even resolved yet? That, was 8 that included in this or was that ignored at this point? I 9 DR. MAZETIS: I don't recall the specific ones that 10 were included. There may have been about a dozen that are 11 discussed or were considered and--
12 MR. MICHELSON: Did you look at the unresolved ones
/~' 13 or only the resolved ones? Maybe that would help.
V) 14 DR. MAZETIS: I think we looked at all of them, but 15 we concentrated in order to come up with a reasonable 16 judgment, as to what the outcome would be, concentrated on 17 those, that some resolution was already perceived on paper.
18 MR. MICHELSON: One of the big ones that keeps 19 coming up as the system interaction A-17, I don't know how you 20 would possibly handle that at this time, 21 MR. MINNERS: Didn't quantify it.
22 MR. MICHELSON: Certain of the generic issues were 23 left out of the quantification?
24 DR. MAZETIS: That's right.
25 MR. MICHELSON: Okay. Thank you.
HERITAGE REPORTING CORPORATION -- (202)628-4888
.54 1 DR. MAZETIS: As I indicated-earlier, the r'
(~)s 2 alternative 2 is what the. reg analysis will focus for future 3 . potential improvement'of decay heat removal systems where.the 4 basis or the reasons that we came up with this proposal is 5 that the studies, the case studies again showed that a large 6 number of the risk contributors are plant-specific because 7 they lie in support systems, balance of plant areas, not in 8 NSSS specific areas.
9 Also use of method C which perhaps could justify 10 even alternative 5 or 6, goes beyond value impact methods 11 previously documented for unresolved safety issues and generic 12 safety issues, and also the insights gained from the six case 13 studies and the EPRI owners group analysis is planned or
(])
14 expected to become guidance to licensees as the severe 15 accident program progresses, and licensees, if what is 16 expected is implemented, turn to their plants for examining 17 for severe accident vulnerabilities.
18 CHAIRMAN WARD: Now we have seen your reg, the 19 results of much of the regulatory analysis, but I'm, I have 20 lost track of what form or when. Did we see a rough draft 21 report or did we just see your earlier --
22 DR. MAZETIS: You saw a rough draft report that we 23 submitted to you prior to the closed session in '86.
24 CHAIRMAN WARD: Okay. Have you changed or have you 25 done any new work since then or--
HERITAGE REPORTING CORPORATION -- (202)628-4888
55 1 DR. MAZETIS: 'There has been no change to the-
'- (q ss/ 2 contractor's findings, no, but the change is, is going to be 3 in the draft proposed resolution in terms of more definitive 4 as to what direction we feel.the A-45 program should take, 5 that you didn't have at that time.
6 CHAIRMAN WARD: Right. Okay.
7 MR. MINNERS: The numbers'have changed because we 8 have made errors, and we have made little modifications, but 9 the general idea is the same thing. It'is only minor changes.
10 DR. MAZETIS: I think at that time most of the 11 results of the studies had been completed but they hadn't, 12 they hadn't been published yet in final form, and it could
(} 13 have been there was some finetuning.
14 CHAIRMAN WARD: Yes. Well, they are still not 15 published in final form, are they?
16 DR. MAZETIS: Yes. The six case studies from 17 earlier form are in six NUREGs that have been published for 18 the plants.
19 CHAIRMAN WARD: The case studies are?
20 DR. MAZETIS: The case studies.
21 CHAIRMAN WARD: The regulatory analysis is what I am ,
l 22 talking about.
23 MR. KNIEL: No. l r~g 24 CHAIRMAN WARD: That's what I was talking about. We V
25 saw before the September meeting, a rough draft report I HERITAGE REPORTING CORPORATION -- (202)628-4888
56 1 thought of the regulatory analysis. Or am I dreaming?_
( 2 DR. MAZETIS: No. You did, and I agree that the 3 major part of that draft remains in tact in' terms of the 4 technical findings. I mean they haven't changed. There may 5 have been some finetuning of the numbers, but--
6 CHAIRMAN WARD; What is the status of that report?
7 DR. MAZETIS: Okay. The status -- did you add 8 something to that question?
4 9 CHAIRMAN WARD: No.
10 MR. MINNERS: Wanted to know what the status was of 11 the regulatory analysis.
12 DR. MAZETIS: As I indicated earlier, that the next
(} 13 revisien is essentially in the typewriter, and our plan is 14 sometime within the next couple of weeks to get it distributed 15 internally so that within the next month or two, at the 16 outside, ACRS probably would be a contributor to commenting on 17 that report. .
18 DR. SIESS: May I ask a question? What guidance has 19 the staff had from the Commission on the fact that it should 20 be included in the value impact statements?
21 DR. MAZETIS: Well, there is a backfit rule that we 22 have tried to follow with its references to various value 23 impact NUREG guidance that have been published by the staff I
r^g 24 the past several years, and so the format of this regulatory
(/
25 analysis will track with other recently, other recently l
HERITAGE REPORTING CORPORATION -- (202)628-4888
57 1 . submitted backfit analyses. For example, more recently A-4; 1 .
2 it follows-in general the same backfit format as the A-4 3 format.
l 4 DR. SIESS: Is that an answer to my question?
l 5 DR. MAZETIS: Anyone have a better answer?
I l 6 HR. MINNERS: The only explicit guidance, any 7 written guidance that I know you have we have.gotten from the 8 Commission, is the backfit rule which lists half a dozen or so 9 specific factors that should be considered in factor analyses, 10 and also has an indirect statement something like in any other 11 relevant factors or something like that.
12 DR. SIESS: Does that give you guidance on off-site-f 1 13 versus on-site?
l 14 MR. MINNERS: No. t l
l l 15 DR. SIESS: Moratorium?
l l 16 MR. MINNERS: No.
l 1 l
17 DR. SIESS: So we are operating in an area where we 18 don't know what the safety goal is and don't know what the 19 values and impacts should be? Making common law.
20 MR. MINNERS: Well, I guess the direction that I 21 would take is one of reasonableness. I think you ought to 22 discuss all of the relevant significant factors, and I don't 23 think anybody can disagree with that. I don't see any reason 24 for excluding any information except on the basis that it is 25 irrelevant or insignificant.
HERITAGE REPORTING CORPORATION - (202)628-4888 f
)
58 1 DR. SIESS: So you would you say include everything?
2 MR. MINNERS: Then the decision-maker is-going to 3 make his decision on whatever basis he wishes. I have no-way 4 of controlling how the decision-maker is going to make--
5 DR. SIESS: Who is the decision-maker?
6 MR. MINNERS: The Commission I presume on the USI.
7 CHAIRMAN WARD: Could we take a--
8 DR. SIESS: You would propose to give the Commission 9 a regulatory analysis that covered the range of strictly.
10 off-site up through on-site and moratoria and let them decide?
11 MR. MINNERS: I think they can handle that.
12 DR. SIESS: You said it. I didn't!
13 CHAIRMAN WARD: So in other words--
(')T
~
14 MR. MINNERS: We have had an opinion from OGC on 15 on-site costs, but I haven't heard anything from the 16 Commission on it.
17 CHAIRMAN WARD: Well, if you look at your three A, B 18 and C, methods for value impact analysis, I mean isn't it 19 clear that C, isn't C clearly outside of what would be 20 intended under the backfit rule?
21 MR. MINNERS: No.
22 CHAIRMAN WARD: That isn't clear to you?
23 MR. MINNERS: There is introductory statement in the e- 24 backfit rule which says you should consider all relevant bg 25 factors.
8 HERITAGE REPORTING CORPORATION -- (202)628-4888
l 59.
1 1 CHAIRMAN WARD: You don't think, I mean you-have l
(
'El 2 made the determination I guess that neither A, B nor C is- I I
3 particularly favored for compliance with the backfit rule? I 4 mean I guess I -- ,
5 MR. MINNERS: If I am making the decision, okay, I 6 would look at method C and I would weight the different 7 factors as to how much I believed them. I would 8 consider--everybody thinks about moratorium. It is ridiculous 9 not to say that. It is a possibility.
10 CHAIRMAN WARD: What woUld the backfit rule require?
11 I mean what would the lawyers say?
12 MR. MINNERS: What it would require, it has six
(} 13 factors which are basically implementation costs and averted 14 off-site costs, and occupational exposure.
15 CHAIRMAN WARD: So that would be B?
16 MR. MINNERS: That would be A. That would be 17 required. You have to at least address those things. Now in 18 addition to that, the backfit rule as I read it let's you go 19 beyond that and consider other, what I would say is relevant 20 significant factors.
21 CHAIRMAN WARD: Okay. And you say that could then 22 go as far as B and C or C?
23 MR. MINNERS: To me, it goes as far as C, sure.
24 Moratorium is a serious consideration and all we have done is 25 done some number manipulation to try and give you a better i
HERITAGE REPORTING CORPORATION -- (202)628-4888
._ . . .. . . .~ . . - . . -- .
~6 0 - ;
~
l
.1 perspective on it. No intent to say that our calculations of' I) 2 cost'of a moratorium'are any estimate of;what the actual'. cost
)
3 'would'be. Just trying to give some perspective.
4 Sabotage is the same question. It is a matterfof
- 5. belief. I agree with Dr. Mark. There is no way of knowing 6 what the probability for sabotage with them is,Ebutif you are 7 going to make some estimates'you are going to-havefthe balance 8 of cost of sabotage from, against your assessment'of risk. It' 9 is going to be.a very subjective assessment, but it'is still 10 going to have to be balanced.
11 CHAIRMAN WARD: Thank you.
l 12 DR. MAZETIS: This slide is really a review of some )
{) 13 of the considerations that are now being given. Obviously it 14 is not cast in concrete yet, but-this will give you a feel for 15 the implementation, and as we try to implement the.thrustHof 16 alternative 2, how we are going to do that specifically.
17 Again, the idea is to require plant-specific 18 reliability analyses to try to focus on plant-specific l
l 19 vulnerabilities, and two questions that are being debated are j l
20 should, should it be part of IPE or should it be separate PRA l 21 requirements concurrent with the IPE program? And there are 22 pros and cons to each as you can see here. i l
23 One question we are wresting with now, I really-24 don't have the answer for, is if we marry it to'IPE, does it l
( l 25 mean we have to require a more rigorous or more thorough PRA .
HERITAGE REPORTING CORPORATION -- (202)628-4888 (
61 1 methodology than what is being currently contemplated for the L( 2 utilities to use on IPE?
3 Also not a small consideration is, is somewhere down I 4 the road should the staff get involved in either some kind of f limited audit efforts or in overall review of these analyses?
- 5 l
l 6 So that I guess up front here we are thinking there should be l 7 some commitment,-if we expect to get something, some safety l
l 8 improvement out of the idea of marrying to IPE A-45, there l 9 should be some kind of a commitment to review downstream these l 10 analyses as a minimum on an audited basis.
1 11 MR. MICHELSON: Question--if you elect the IPE I
12 route, will you be expecting them to look in sufficient detail 13 to understand the effect of fire, floed, external, internal,
(}
14 wind storm, et cetera, which is a part of the question? Would 15 that be now brought--I know IPE can't. I didn't look too 16 closely at those areas. Maybe I am mistaken.
17 MR. MINNERS: If you mean the IDCOR method, that 18 doesn't include external events, but my understanding of the
-1 l
19 severe accident policy statement is that it says explicitly i l
1 20 that external events should be considered. 1 I
21 MR. MICHELSON: So you are thinking IPE will include i
22 a sufficiently careful examination of external events to 23 satisfy this criteria?
gg 24 MR. MINNERS: We have some people looking at what we U
25 should consider in the external events.
HERITAGE REPORTING CORPORATION -- (202)628-4888
62 1 MR. MICHELSON: From what I have heard today, that,
(> 2 would have to be sort of a ground rule if you are going to 3 cover these events by the IPE approach?
4 HR. MINNERS: I said we are investigating how much 5 external events should be included in IPE, and I guess a 6 decision hasn't been made on what is going to be in IPE.
7 MR. DAVIS: On the previous slide, that item under 8 the separate requirement?
9 MR. MINNERS: Where are we, Pete? Slide 8?
10 MR. DAVIS: Nine.
11 DR. MAZETIS: Here?
12 MR. DAVIS: Yes. It indicates you would require, if
() 13 it is separate, A-45 resolution is separate, this would 14 require each licensee to perform two separate analyses, and I 15 guess that's true.
16 However, I don't see that that is a significant 17 problem because I don't think anything they would do under 18 A-45 could not also be used in the IPE. Is there something 19 more to it than that?
20 DR. MAZETIS: No. I guess the idea was that if they 21 were required under the IPE program to go into their plant and i 22 employ a certain methodology to look at the total plant i 23 vulnerabilities and that they finished that study and l
l l s, 24 concurrent with that there is a requirement to do the same
~)
25 thing but focusing on just decay heat removal in a more HERITAGE REPORTING CORPORATION -- (202)628-4888
63 1 rigorous PRA methodology, it may be an additional burden, and
-) 2 redundant to what they did under IPE, but what you say is 3 true. They may be able to use something they learned from IPE 4 or vice-versa, but it would be two separate burdens on the 5 licensee to respond.
6 HR. DAVIS: . Well, they might be separate, but they 7 wouldn't be independent.
8 MR. MINNERS: At least double the paperwork.
9 MR. DAVIS: Okay. Well, to me, that doesn't seem 10 like a significant disadvantage.
11 MR. MINNERS: It is quite possible, too, depending 12 on what order you did it, you seem to be suggesting if we did
() 13 A-45 first that could be used in IPE or vice-versa.
14 MR. MINNERS: That might not be the case. If IPE
, didn't include external events, and that's not a decision made 16 yet, then it might not be enough.
17 MR. DAVIS: No. I am not suggesting it wouldn't be 18 enough, but what they would do would be applicable to at least 19 internal events part of of A-45.
20 CHAIRMAN WARD: At least if a licensee knows he is 21 going to have to do both of these things, whatever they are, 22 he should be able to organize his effort.
23 MR. DAVIS: Avoid duplication.
24 CHAIRMAN WARD: Yes. But if you come out with 25 instructions for doing an IPE next month and eight months HERITAGE REPORTING CORPORATION -- (202)628-4888
64 1 later come out with instructions for doing something under
(')s
(_ 2 this, that could cause licensees a lot of problems it would 3 seem to me.
4 DR. MAZETIS: Okay, and the last two slides kind of 5 shifts gears into the next couple of speakers, where maybe now 6 we could point to the submittal by EPRI and the owners group 7 on Point Bea.h which is the same plant, one of the plants that 8 NRC/Sandia studies concentrated on, and you might ask the 9 question, well, if we are going the way that I just described 10 in our reg analysis, why bother? I mean this document may be 11 a nice piece of information to have, but why bother with 12 resources of reviewing this document?
() 13 I guess that's a good question, and our current 14 thinking is that when the IPE generic letter is published, we 15 would not lose what we have learned on A-45 the past seven 16 years. There is planned and explicit address of, focuses on 17 decay heat removal in the IPE exercise, and we perceive it 18 perhaps as an additional appendix on what we have learned on 19 A-45 through the case study analysis plus in this particular 20 case, the whole breadth of the picture on some additional 21 insights provided from others in the industry.and hopefully by 22 the time it is published, some staff comments in the IPE i
23 program as to where, what agreements we have had with NUMARC l
24 and where our disagreements still lie so that the utility as 25 he goes through this severe accident program and starts HERITAGE REPORTING CORPORATION -~ (202)628-4888
65 t
1 looking in his plant can get some specific guidance for decay-3
/ 2 heat removal-vulnerabilities en where to focus his resources 3 and where the areas perhaps for his plent would contain the 4 more vulnerabilities, where the soft spots are and no forth.
5 So we think there is something to be gained by 6 continuing to the dialogue which is planned with the NUMARC 7 sometime within the next one or two months to interface with N 8 NUMARC to discuss in essence as you will see a need for 9 additional understanding as to why certain assumptions were 10 made, so that we could at some point down the road form an 11 opinion as to what we feel are the better assumptions. And as 12 I indicated earlier, you might ask is the alternative 1 or the
(} 13 no action, do nothing, justifiable? And again, if the EPRI 14 PRA is, turns out to be correct, and if ye if you like it, 15 and you think that Point Beach was a bounding plant, then 16 again you may think, may end up in this box where doing 17 nothing is justified, out I don't think as far as the staff is 18 concerned that obviously this early that we are convinced that 19 either of these are correct. We have this dialogue to go 20 through on the NUMARC studies, better understand the basis for 21 a lot of their assumptions.
22 MR. MICHELSON: Clarification--when you talk about 23 the EPRI PRA, do you mean the NSAC-113?
24 DR. MAZETIS: Yes.
25 MR. MICHELSON: Okay. Thank you.
HERITAGE REPORTING CORPORATION -- (202)628-4888
A M -m L 4 & p. p "Am.-.
, &,w & m 4 3-4 &*,.:: A f - sk& k A -.A ,-
66
- i. 1- CHAIRMAN WARD: Jerry, it'seems,to me that the first 2 one is a,' reasonably: determinable, that is, the differences .1 i
l 3 between the Sandia PRA and the EPRI revision of1that.-'You
'l 4' know, that's what you are going to-discuss. Maybe you1can 5 resolve those.
6 DR. MAZETIS: 'Yes.
7 CHAIRMAN WARD: But the second one, how are you ever.
8 going to figure out whether Point Beach is a bounding plant 1or 9 not?
10 DR. MAZETIS: We'are not intending to do that. 'All 11 I am saying is if someone is asking himself and is. convinced .
12 that doing nothing is justifiable, as far as the staff is 13 concerned, this may be the bases for them' believing that. All
(])
14 I am saying is we disagree. We are not convinced yet that 15 that rationale has a basis.
16 HR. WARD: Okay. You are-saying your position is
~
17 alternative 1 is not justifiable?
a 18 DR. MAZETIS: That's right. Exactly.
19 CHAIRMAN WARD: Maybe they can convince you on.the 20 first part there, but it looks tough on the second one.
21 DR. MAZETIS: I doubt it on the first part, too, 22 after preliminary interface with some of the analysts, but you L
23 will hear that in a couple of minutes. ,
24 (Slide) 25 DR. MAZETIS: And my last slide is--you have seen i
HERITAGE REPORTING CORPORATION'- (202)628-4888
67J l this table before. It is just a lead-in to the next
[
t 2 presenter, from the EPRI owners group document. You can-see 3 the difference for the up-front end of the document, the core 4 melt frequency, where the major differences are identified and 5 focus more on the first four areas where you can see the, they 6 are in some cases a large difference between NRC/Sandia 7 reports and the EPRI owners group report.
8 So that completes my presentation, and Mr. Chelliah 9 is the next presenter with 15, 20 minutes discussion of what 10 his observations are so far, and again I remind you that he 11 has not yet closed. There have been no give and take with his 12 observations with Sandia, nor with NUMARC, so there may be a tT 13 little overlap here.
Q 14 CHAIRMAN WARD: Thank you very much. Let's take a 15 break until 10:30.
16 (A brief recess was taken.)
17 CHAIRMAN WARD: Our next speaker is Mr. Chelliah of i
18 the Research Office.
i 19 DR. MAZETIS: If I can just make a brief comment, I l
20 notice from your agenda that you had an additional gentleman, !
21 Dave Pyatt, scheduled, and we would like to have his area, 22 which is the back end of the calculation, deferred as far as I
23 an actual presentation, but have him available if questions s 24 come up, so really we have got two speakers left, Mr. Chelliah l
.)
25 and Dr. Ericson, who probably we could get through before ;
l HERITAGE REPORTING CORPORATION -- (202)628-4888
68 1 lunch.
(-) 2 MR. CHELLIAH: Good morning. My name is Chelliah.
3 I am from the Risk Analysis from Research.
4 (Slide)
~
5 MR. CHELLIAH: As the, as part of this' decay heat 6 removal study, the lab, Sandia Laboratory, they performed the 7 decay heat removal type of'PRA on the Point Beach facility on 8 the power in cooperation the EPRI. They performed the 9 reanalysis of this particular Sandia study with the EPRI, 10 phone company, of the Point Beach facility in cooperation wich 11 the EPRI; they performed the limited review of this Sandia 12 study. They have provided us ti.e reanalysis the staff looked
() 13 into, and we have the preliminary comments here.
14 In summary, the EPRI report claims that their core ;
15 damage frequency estimate just allocated to decay heat removal 16 is lower than the Sandia estimate. What we have done here in 17 the presentation is to provide you with a summary differences 18 on all the accident sequences. Here I would like to put the 19 ACRS members to read the sequence. There are a couple of 20 pages on the back side. I have described the sequence with, 21 what this means. Also I will explain to you as I go through 22 here.
23 Here we have two estimates, that is Sandia National 24 Lab, EPRI estimate. And we have here three major bullets I 25 would like to explain one at a time, and I will also try to HERITAGE REPORTING CORPORATION -- (202)628-4888
69 1 say whether we concur or don't concur.
1
() 2 First the sequence, the frequency of failures. You 3 can see the difference in the frequency estimate. The first
~
l 4 difference comes from the frequency in the EPRI study.
5 Basically here what they did, they looked into the Sandia 6 frequency, the smaller frequency; it covers wide range of 7 break, varying from .3 inches all the way up to 1.6 inch, 2-8 inches. It includes all the failures, all those things.
9 What the EPRI did, they did look into the 10 experience, the operating experience. They have decided that i
11 those breaks involves the smaller ones, so what they did, they j 12 assigned the small LOCA frequency about 1.five inch, which
{} 13 would--accordingly they are just doing this frequency.
14 MR. MICHELSON: Question--what data base did you use i 15 to determine small break frequency?
16 MR. MICHELSON: Or what did Westinghouse use or 17 whoever did the work?
18 MR. CHELLIAH: Yes. In this particular report, 19 there is also one, in the report it is indicated that they l l
20 used Westinghouse operating experience I assume. l l
21 MR. MICHELSON: Westinghouse, every time some l
22 utility has a small break, does Westinghouse somehow get l 23 informed or become aware of it and put it down in their data 24 base? Is that the way it works? Or are they just pulling out 25 some numbers and playing around with them? They are real HERITAGE REPORTING CORPORATION -- (202)628-4888
70 1 legitimate scientific data base that.they maintain small break r
(_3) 2 LOCA?
3 DR. MAZETIS: I think you are asking some of the 4 questions that when we get.with NUMARC that we would expect to 5 gain a better understanding of what the basis for their 6 numbers are.
7 MR. MICHELSON: You haven't done that yet?
8 DR. MAZETIS: However, Sandia could address the 9 basis for the numbers in the six case studies if you want to 10 get into that.
11 MR. MICHELSON: That's a good place. What.at 12 Sandia--
I' 13 DR. ERICSON: Generic data base published in the V) 14 procedures guide.
15 MR. MICHELSON: Where did it come from?
16 DR. ERICSON: Accrued by the staff and by the 17 community over a number of years.
18 MR. MICHELSON: I have seen some of those numbers.
19 I am no PRA expert, and the first to admit it, but a lot of 20 numbers I have seen in the past seem to not have a good 21 scientific base, but rather some, early on they were 22 assumptions because they didn't have good numbers and they 23 began to become accepted even though they weren't necessarily 24 scientifically based. I just wondered, you are using the same 25 numbers everybody else is using? Is what you are saying?
HERITAGE REPORTIliG CORPORATION -- (202)628-4888
71 1 DR. ERICSON: That is correct.
O k) m 2 MR. MICHELSON: And how do our PRA people feel about 3 that? Are they satisfied those are good numbers, or just the 4 numbers everybody uses?
5 MR. DAVIS: According to the EPRI report', the small 6 LOCA frequency that was used by Sandia was dominated by some 7 pump seal LOCA data thatHis in an internal NRC memo which I 8 haven't seen, and according to the EPRI report, most of those 9 breaks were pieces of small data that would not require 10 recirculation, and that at least my understanding is the basis 11 for the large difference in initiated frequency.
12 CHAIRMAN WARD: Mr. Ericson, is that right? The
(} 13 data you used--
14 DR. ERICSON: He says what I understand to be in the 15 report. I disagree with him. And the S2 data that is 16 published includes all of small break LOCAs below one and a 17 half inch, doesn't throw some of th'em out. In fact, if you 18 take what is in the EPRI report, the five additional events 19 cited, I think that is correct. Then you come up with l 20 something like three times ten to the minus 2 if you add those l l
21 into the data base they talk about.
22 MR. CHELLIAH: One point, Dr. Michelson--the EPRI--
23 MR. MICHELSON: I am not quite finished with the 24 question yet.
25 MR. CHELLIAH: Okay. I'm sorry.
HERITAGE REPORTING CORPORATION ---(202)628-4888
72 1 MR. MICHELSON: If the pump seal LOCA.is a dominant
(~ . .
(_.) 2 contributor, which I think is what you are'saying, then the 3 NRC is aware of that, it is their data base, is that what you 1
4 are saying? You said it was an internal NRC memo, so really. l l
5 the NRC is the custodian of small break LOCA predominant I 6 contributors at least. And so you people are the ones that 7 are the experts and have *he good data base on it I guess.
8 That's what I thought I heard.
9 MR. DAVIS: According to the report, the data that 10 was in the memo was based on a search of LERs.
11 MR. MICHELSON: Now we are getting a little closer.
12 And does NRC maintain a small break LOCA data base which they
("T 13 keep current on all of these? They find, of course, the pump V
14 seal is made to contributor. I don't think I necessarily 15 disagree, but you people are the custodian and I can go and 16 find out who is maintaining this data base and see how they do 17 it?
18 DR. MAZETIS: Joe--
19 MR. MINNERS: Joe, you guys--
20 MR. MURPHY: I am unaware of formal data base.
21 MR. BOEHNERT: Would you identify yourself, sir?
22 MR. MURPHY: Joe Murphy from Research; I am unaware 23 of a formal data base in terms of someone surveying the OERs 24 once a month and adding that in.
25 MR. MICHELSON: I am unaware of it, too. That's why HERITAGE REPORTING CORPORATION -- (202)628-4888
73 1 I was a little surprised. I am, I am'not going to argue if 2 that's what everybody thinks is the source.
3 MR. MURPHY: There was~a search done a couple of 4 years ago, on the pump seal LOCAs, and that information was in 5 the memo that was referred to and that has been used in a lot 6 of PRAs sponsored by the staff.
7 Now there is a standard policy in doing a PRA that 8 the analysts usually do surveys of EORs. If they'see an 9 unusual pattern of behavior, they go back and analyze the data 10 more explicitly, but AEOD may do some formal data base, the 11 basic organization for gathering and maintaining the data base 12 for the agency. They may have something explicit in that 13 area, but I'm not aware of it.
14 HR. MICH'4LSON: I -m not either, but I will go back 15 and ask them again for a current answer. Thank you.
16 MR. EBERSOLE: Ask you a question? First I want to 17 thank ycu for characterizing these sequences because it blows 18 my mind to try to remember them. So in the first one there, e
19 you have in the explanation it is a small LOCA followed by 20 failure of low and high pressure recirculation systems.
21 HR. CHELLIAH: Yes.
22 MR. EBERSOLE: I thought a PWR, for that matter BWR, 23 really has no problem with recirculation since it has natural 24 convection and the problem is one of makeup to hold pressure 25 so you wou3d have a thermal differential to carry the heat to HERITAGE REPORTING CORPORATION -- (202)628-4888 i _
_. J
74-
) 1 the secondary side. 'o
} (o) 2 I don't know.why '. say fallure'of low and high 3 pressure recirculation system. You have got one in natural f
4 convection. I can see a problem in actually supplying the
{
5 pressure to make up the leak to hold the pressure up so_you l
6 can transfer heat.
}.
7 DR. MAZETIS: I think--this is a different 8 terminology, recirculation.
f 9 MR. EBERSOLE: You are going on out to recirculating =
i 10 the water. I am talking about the presence of the 11 secondaries. s
( 12 MR. MICHELSON: Are you defining small LOCA for BWRs l
q 13 diffecently than for PWRs? By defini, ion it is the makeup of V
14 the normal, beyond the makeup of the normal system I think is 15 the definition of small break LOCA, and it~is'very large s
16 break.
17 MR. uTNNERS: What difference does it make if we 18 define it dif., .ely?
19 MR. MICHELSON: Makes quite a difference in what
~t 20 kind of systems are going to address it. <
l 21 MR. MINNERS: Small break LOCA medium, and we do i
+,
22 large break I.OCAs.
f 23 MR. MICHELSON: You said 2 inches was the rancs of 24 your small break LOCA. I think it is considerably bigger than 25 that for the boiling water reactor? I
{
s.
\
HERITAGE REPORTING CORPORATION -- (203)628-4888
75 1 DR. MAZETIS: My recollection.is that was about the
(~h s/ 2 size for the PWR case study. Perhaps Dave could correct me if 3 I'm wrong.
4 DR. ERICSON: I can't pull it out of the top of my 5 head, e 6 MR. EBERSOLE: You are taking heat out-through the 7 exchangers. ' Iou have got for even the boilers, I mean the l l
8 steam generators--
9 MR. CHELLIAH: Yes.
10 MR. EBERSOLE: Degraded state in the fir'st place; i 11 you have lost the secondary.
12 HR. CHELLIAH: The point is the EPRI analysis assume i
() 13 that is the, is needed for breaks more than 1.five inch 14 breaks. Accordingly, they are just small LOCA frequency, so 15 that is one of the difference here.
16 Dr. Michelson, going back to your question, the PRA 17 does provide a document of all the reported small LOCA events, t
18 including the Zion and the Robinson event, and some of the 19 small leaks varying from 20 GPM all the way up to hundred 20 GPMs. It does provide some documentation. If you are 21 interested, you can look into the NSAC-60 report.
21: The second difference is the modified common cooling 23 water success criteria. Basically what it is here, the EPRI, h 24 they did look into the cooling system configuration and 25 associated cooling by common cooling water system, but their s
HERITAGE REPORTING CORPORATION -- (202)628-4888
76 1 claim is this. The CCW provides seals rather than it provides
(~%
'V 2 the cooling, so the claim is this particular system is not 3 really needed during the injection phase of this particular 4 sequence up to one hours. Accordingly, they have, the 5 frequency here has accounted for some removal dependency due 6 to the CCW system. We do concur with the plant-specific 7 details.
S We need some additional information on some of the 9 test data on this seal performance related to why it doesn't 10 fail during injection, why it would fail only after one hours 11 because we have more than two-hour time sequence from the time 12 so basically we do have that, but we need some additional 13 information here, u
14 The third one is really--
15 CHAIRMAN WARD: How much would that change the 16 numbers?
17 HR. CHELLIAH: I don't have it here. The 18 difference, added to the other thing they have here.
f 19 DR. ERICSON: Eighty-six percent of the difference 20 in the small break LOCA frequency, in the two numbers; above 21 the CCW and the service water balances are very minor 22 contributors. CCW is like 6 or 7 percent.
23 HR. CHELLIAH: This is the load balancing, j (s
\_]
24 Basically this particular sequence used as the hour, there are 25 three series pumping each unit at Point Beach. What EPRI HERITAGE REPORTING CORPORATION -- (202)628-4888
77 1 claims based on data, review of their procedures,-related to nss 2 this system, there are two, it is no*. meant to--operator ~has 3 reduced some of the heat load. It is not needed so that they 4 can live with the one hour satisfactory, so accordingly, there 5 are some redundancies. That is also one of the reasons we 6 went down.
7 We do have some other differences which I did not 8 list. They did look into some human ratio. One of the things 9 that this frequency involves is the operator failure to 10 establish recirculation following the event. -You have two 11 hours1.273148e-4 days <br />0.00306 hours <br />1.818783e-5 weeks <br />4.1855e-6 months <br /> to go. The Sandia, they used something like three minus 12 per demand. EPRI claims that there are, 1 E minus 4 would be
(~T 13 more appropriate based on the procedures. They have not U
14 provided any human analysis. Right now we do, we need more 15 analysis on that particular operator failures.
16 There is one more on the sequence--the tank, because 17 two hour, you are going to dump the water. What they are 18 saying, you bleed from the tank; this particular area you 19 bleed through the tank from spent fuel or you can bleed 20 through the tank, okay, bleed the tank spent fuel or use the 4
21 boric acid storage tank, about a hundred thousand gallons.
22 EPRI basically claims that the Point Beach facility, 23 they do have procedures which dictates in case of water is 60 24 percent, 30 percent, you can do all these accidents, so based 25 on the plant-specific procedures and training which are in HERITAGE REPORTING CORPORATION -- (202)628-4888 -
1
18' 1 place, they claim, they have taken some--that's also a i (^h 2 contributing factor. -Other thing, too, you~can see_the
(_/
3 sequence.
4 The next sequence is-- j 5 MR. DAVIS: Excuse me. A question--have you come to i
6 a conclusion about whose frequency of small break LOCA is more 7 valid for this case? I don't believe you said whether you 8 agree or not.
9 MR. CHELLIAH: We do contend the EPRI's frequency 10 because we did look into it--you don't need these below 1.5 11 inch break. They are right. It is right. We do agree with 12 that.
13 Now the second sequence is really that loss of Ou 14 off-site power sequence, basically Station Blackout sequence.
15 DR. MAZETIS: Excuse me. Again, if you ask the same 16 question of Dave when he gets up here, you may or may not.get 17 the same answer, and I just want to point out that the staff 18 has yet to close the loop on interfacing back and forth.
19 MR. DAVIS: I understand. Thank you.
20 MR. MICHELSON: Question--if you have a safety 21 relief valve go through an operation and stick, is that 22 considered a small break LOCA under this business?
23 DR. MAZETIS: Yes. PORV open air, sticking open.
24 MR. MICHELSON: Those kind of events do not occur as 25 frequently as pump seal failures? Or if not, why does pump HERITAGE REPORTING CORPORATION -- (202)628-4888
l 79-
'l seal failure dominate the data. base?
O)
(_ 2 MR. CHELLIAH: Dr. Michelson, we do have the 3 sequence. We will come to that. We do.have a sequence to 4 answer your question.
5 DR. ERICSON: That is not the initiating event. The 6 stuck open PORV is not the initiating event; subsequent event 7 after transient.
F 8 MR. MICHELSON: That's why I asked the question. Is 9 it considered a small break LOCA? What you are saying is that 10 only if the pipe breaks first is it in this table in the way 11 it is presented, but a small break LOCA may be involved in 12 some of the other events. The fact is it was a fairly high
[} 13 probability.
14 MR. MINNERS: It is event T30D1D21 15 MR. MICHELSON: I really think it is there. That's 16 why I asked the question. I thought it would have been 17 considered a small break LOCA, but I see why such logic is you 18 are looking only at the initiator, finding it that way.
19 MR. CHELLIAH: The second sequence is basically loss 20 of off-site power, the failure of--the L is the failure of aux 21 cooling. I'm sorry--secondary cooling, and E is the, really 22 the failure of feed and bleed, so this sequence is really 23 dominated by some of the Station Blackout concerns, failure 24 combinations. There is some difference you can see close O 25 to--there are two major differences here. Updated initiator HERITAGE REPORTING CORPORATION -- (202)628-4888
~
,801 1 frequency, loss of off-site power frequency. The Sandia 2 study, they use something like about .zero 8 per site year 3- based on one of the reports, HUREG 1032. At EPRI they did ,
4 look into the plant-specific operating experience. We did 5 have one event over there in the sixteen-year operating 6 history, so they are just going accordingly.
7 The second one is really a very important one I 8 thought in my view. 'They added, like they said in the report, 9 they added one of the third time class battery system, 10 seismically qualified. Basically, some start and control 11 failure probabilities, safety equipment, because this 12 particular sequence frequency is really dominated by the
() 13 battery common mode failure, so this would help now, so they 14 have not provided us--we do concur with this one. We don't 15 have much detail what they did, how they did all these things.
16 We need some information from EPRI, but basically we do 17 concur.
18 There is some other difference also other than these 19 two bullets. EPRI, what they did, they did look into the 20 common mode failure analysis of the D cells and batteries 21 reported in the Sandia. They used a somewhat different :
22 method. For example, in the Sandia study we used factor for 23 the common cause failure of common mode failure of the 24 diesels. EPRI, they started using one of the multiple ;
25 failures. There is a difference, about *hree difference on HERITAGE REPORTING CORPORATION -- (202)628-4888 I
81-1 the diesel common mode failure. So that comes into this
() 2 sequence frequency.
3 MR. EBERSOLE: Did they add an engine-driven DC 4 charger by any chance?
5 HR. DAVIS: Just batteries; site qualified.
6 HR. EBERSOLE: They just added more battery 7 capacity?
8 MR. DAVIS: It is an independent battery.
9 MR. EBERSOLE: Rather than putting a small 10 engine-driven charger which has indefinite capacity, storage 11 life; well, okay.
12 MR. CHELLIAH: We need more detail on what the
{} 13 battery does exactly, you know, where does it hook up, all of 14 these things. I think Sandia is going to discuss the data we 15 have on this particular item.
16 So basically, we need some information on this 17 particular bullet.
18 The third and fourth sequence is really transient 19 where you have main feedwater is available. The Q is the 20 stuck open, the relief valve, and these are the recent 21 failures and here is the injection failures.
22 EPRI, they look into this sequence carefully. Their 23 claim is they did look into some of the Westinghouse, some of 24 the operations performed by Westinghouse. They came to the 25 conclusion as long as you have main feedwater available, you HERITAGE REPORTING CORPORATION -- (202)628-4888
82 1 don't get into this situation because--I'mLsorry--because you
, ) 2 don't reach first to the relief' valve set point. Therefore, 3 you go on to something, so they are, according to their 4- analysis, the sequence shouldn't exist, so they did not 5 comment.
t 6 MR. EBERSOLE: That exit is a transient involving a i
7 stuck open PORV and failure of recirculation systems.
8 The main feedwater system is assumed to be. initially j 9 available. Well, and that says no need for recirculation, and 10 that's right, but there is a distinct need for pressurization 11 and supply capability to deepen overpressure on the primary 12 side against the secondary system. 'Are you with me?
() 13 MR. CHELLIAH: Yes. ,
14 MR. EBERSOLE: I have both to have a thermal ,
15 gradient--maybe you can pull secondary down to 29 inches of 16 vacuum. Does it involve that? You have got to run downhill.
17 Is that what is contemplated, that you will run the secondary 18 pressure down as the primary pressure declines?
19 MR. CHELLIAH: Okay. Answering your question, Dr. !
20 Ebersole, we don't have much detail because if you look at the 21 EPRI analysis, they did not even draw a branch point when you 22 come to this particular event.
23 MR. EBERSOLE: It implies they are going to transfer 24 to the secondary somehow, when they say no need for 25 recirculation.
HERITAGE REPORTING CORPORATION -- (202)628-4888
e
-83 1 MR. CHELLIAH: But their-claim is this valve is, 2 won't reach that set point'as long as you have initiating 3 feedwater available, so they say you will be able to cool down 4 with the existing, the other ejection syrtems.
5 MR. EBERSOLE: Feedwater is not worth anything 6 unless you can get heat from the primary to the secondary, and 7 that requires a gradient in temperature and that requires a 8 primary system be at higher pressure than the secondary.
9 MR. ERICSON: If you accept the argument that Q 10 doesn't occur, there is no loss in primary, primary stays i 11 pressurized.
12 MR. EBERSOLE: Fine if it does that.
(} 13 DR. ERICSON: That's their argument. '
14 MR. CHELLIAH: The question that we are taking 15 basically, here is the similar situation as well as here. The 16 injection is failing, falling this particular event, so you 17 are in the same argument. If the valve doesn't open, you 18 don't bleed the water from the primary system. Basically it 19 is a logical sequence according to them.
20 Right now the question of what position we are 21 taking, we do agree with them, but I think they have to show 22 that the specific analysis for this facility, falling back to 23 some of the analyses or typical Westinghouse analyses to some i 24 raising pressure, we haven't seen yet in the report. This is l
25 part of the verification of the criteria. Basically we do, we 4
HERITAGE REPORTING CORPORATION -- (202)628-4888
84 1 agree with them.
2 CHAIRMAN WARD: You are saying you agree on both of 3 those cases?
4 HR. CHELLIAH: Yes. Then this sequence, the fifth 5 sequence, is very important sequencew. The T is really loss 6 of feedwater event.
7 HR. MINNERS: We have a littic.
8 MR. VANDERMOLEN: Harold Vandermolen from Risk 9 Analysis Branch--the ruling means we agree with them 10 contingent upon seeing the thermohydraulic calculation. We 2 11 are not accepting this unconditionally.
12 MR. CHELLIAH: That's right.
(} 13 MR. VANDERMOLEN: I want to make that clear. When 14 we say we agree, we agree contingent upon some type of 15 receiving thermohydraulic calculation verifying that 16 assumption. Follow me?
17 CHAIRMAN WARD: Right. Thank you for the 18 clarification.
19 MR. CHELLIAH: As part of PRA review we have to 20 verify the criteria so we have to look into those 21 calculations.
22 This fifth sequence is a feedwater event followed by 23 the stuck open, the relief valve.
p 24 MR. MICHELSON: Which relief valve? That has got a V '
25 Q in it also.
1 HERITAGE REPORTING CORPORATION -- (202)628-4888 :
85 1 HR. CHELLIAH: I will explain that. Okay. -Such 2 sequence also exists in EPRI. They do have some ectimate.
3 Have, sometimes have difference. The major difference is the 4 revised PORV probability. I think here it needs a little 5 explanation. When there is, Sandia did, performed ,their 6 study, based on the information available to then, what they 7 assumed here, the safety valve will open because the PORV is 8 the block valve.
I 9 Also they made assumption that 7 percent of the time l
10 they do open, and they are just also, they gave critically the 11 blocked valve failure. Probably accordingly they computed Q l
12 in, but the EPRI, they did look into this particular event l
(} 13 under the associated calculation performed by Westinghouse.
14 What their claim is, if you get the feedwater event, 15 the only way you can reach the set point, the PRA set point f
16 the PRA set point is with the number 35 pressure, PSIG. This 17 is based on the flow mismatch betweem the primary and 18 secondary system so they say that there is only 1 percent of
- 19 the time the valve will challenge because of the rising 20 pressure. So accordingly, they computed this Q. They did '
21 also give credit to the blocked valve failure on the PORV.
22 According to them, it will not reach the set point which is 23 cver 2485 PSIG. That's a big difference over there, so 24 basically some sense here so they assumed 7 percent of the 25 time it will open. But what the licensee claims based on HERITAGE REPORTING CORPORATION -- (202)628-48,88
86 1 their operating experience and the Westinghouse with
() 2 calculations, is only 1 percent of the time it will challenge 3 because of the flow mismatch.
4 We do agree with the EPRI position but we need some 5 additional bases under the assumption 1 percent, how they 6 got--we need some bases on that, but we do concur with that 7 position.
8 Then the last sequence, this is companion sequence 9 here, that's the failure here, the small LOCA followed by the 10 injection failure, so you have this particular bullet, place 11 it here; also this one because this affects it as well, the 12 LOCA frequency. And also doing the injection phase, the CCW 13 is not needed because it prevents the seal rather than the
[}
14 cooling, so that bullet goes into this risk we estimate.
15 MR. EBERSOLE: Let me ask you a question. You have 16 got some pretty low numbers up there, like ten to the minus 8.
17 Isn't there a school of thought that says as you go up--
18 MR. CHELLIAH: Hine he is 79.5.
19 MR. EBERSOLE: As you go up, rather go down in 20 probability like this, that there is a progressive 21 deterioration of the reliability cf these numbers because of .
22 common mode especially in complex systems, and so as you run 23 down this scale of probability, your accuracy really declines 24 and everything gets rather fuzzy at about ten to the minus 25 5th, or 4th.
HERITAGE REPORTING CORPORATION -- (202)628-4888
87 1 MR. CHELLIAH: Well, yes. We_do agree. When you go
) 2 to minus 7, you do have substantial amountHof uncertainty. In 3 fact, right there the-S2 frequency is the frequency itself 4 because here we are talking about break and 1.five inch, the 5 sequence, because the same initiating frequency is assigned 6 here, so in absence of data, you know, light water reactor, 7 you do have uncertainty.
8 MR. EBERSOLE: A cry for super-simple systems that 9 don't have so many elements to have unreliability.
10 MR. CHELLIAH: The D1, D2 here, we are talking about 11 the LOCA rejection. These systems are ECCS system. Sandia 12 has modeled extensively common mode failures. We need to look
{} 13 into that. That's factored in here. .
14 MR. MICHELSON: I am still trying to sort out the 15 small break as opposed to a relief PORF safety valve being 16 open.
17 The revised PORV probability you show there is 18 really the probability of that having experienced a transient.
19 The relief, the PORV rather will also open. Does it mean it 20 sticks open or remains open indefinitely? Is that what that 21 probability is supposed to be, or just the probability it 22 opens?
23 MR. CHELLIAH: No. This Q is, it is the joint 24 probability of two events. Not only this will fail open, but I 25 also this includes some critical for unblocking the blocked HERITAGE REPORTING CORPORATION -- (202)628-4888 l -
88 1 valve.
2 HR. MICHELSON: It is the probability that it opens 3 and the flow path remains open for some reason, is that right?
4 M... CHELLIAH: Yes.
5 MR. MICHELSON: -When does the flow path: finally 6 terminate, or is that yet another. calculation? See, relief 7 valves have been known to open at high pressure but reclose.
8 In fact, many of them will reclose at low pressure. The LOCA 9 does not go on forever. In fact, it gives you a worse J 10 situation than if the thing remained open.
11 Do you have any probability for experiencing an l 12 isolated small break LOCA which might be a worse situation by
(} 13 far than continuous one?
14 MR. CHELLIAH: Maybe Dave-probably?
15 DR. ERICSON: The argument is event Q is open, and 16 sticks open. It doesn't reclose.
17 MR. MICHELSON: Never recloses; the real world 18 history is that it is a rarity to ever have a relief safety r
19 valve stick open.
20 DR. ERICSON: When you take that into account, the 21 numbers get very small.
22 MR. MICHELSON: But the event changes significantly 23 when, for instance, you have a small break LOCA terminated at
- -w 24 the time when you are now using it as heat removal, it i
25 suddenly ceases to be the heat removal, and you have to HERITAGE REPORTING CORPORATION -- (202)628-4888 .
- _~ r__
_l 89 1 establish communication with the' steam generator again. 'And 2 now it gets real sticky. Now you have got a refuel problem 3 and a natural convection problem, and so forth, regenerated
)
4 which wasn't there as'long as the'small break LOCA procedure.
5 That's what TMI did really. They had a small break LOCA for a 6 .w hile in terms of open PORV and later they terminated the l
l l 7 break, and it creates a whole new situation at the wrong time.
l 8 I just wanted to make sure I understood what your probability 9 you plugged in. Thank you.
10 HR. CHELLIAH: Okay.
11 (Slide) 12 MR. CHELLIAH: So the previous slide says some of 13 the sequences. The Sandia study, it is very detailed study.
14 That's my thinking with respect to the other, comparable to 15 some of the industry performed historical analysis on the 16 flooding and seismic events and fire events, and also they did 17 some on the wind analysis and they don't dominate here. So 18 the EPRI, they did also review some of the Sandia study 19 sequences. First one is flood sequences. This is basically, 20 the sequence here is really there is fire in the main pipe.
21 It is running on the top of the room, so basically they have 22 postulated the pipe break and releasing something like about 23 2,000 GPM type flood, what they call moderate flood, and so it 24 is going to result in the failure of the pump. The failure 25 mode was really the spray failure mode.
HERITAGE REPORTING CORPORATION -- (202)628-4888
90 1 Sometime back, Dr. Michelson, you asked me a
() 2 question is there the system failure, how this failure as 3 opposed to random failures? Here the spray failure mode, so 4 basically, if you lose the water system, the water does 5 provide cooling to the PCS. Basically you will see a 6 transient now. Also water cools some of the safety equipment 7 like the pump and HVA pump su basically they postulated 8 sequence frequency of this, 9 Now EPRI frequency you can see is almost more than 10 three. Some situations, there are two bullets. First bullet 11 is very, very important. In the estimating this frequency, j 12 Sandia, they assigned something like 2 E minus 2 of frequency 9
(} 13 for the flood frequency for the fire in the pipe. Basically 14 that frequency is not really a pipe frequency. That. frequency
) 15 is what they did look into generic experience. They have J
l 16 extracted what they call moderate flood, 2,000 GPM flood.
17 minus 2 frequency. What the EPRI claims, no, that's not the 18 way to calculate.
19 MR. MICHELSON: Excuse me. From any kind of a break i
, 20 that would give 2,000 GPM? Is that what you are were saying?
21 They look for 2.000 GPM breaks of moderate energy pipes I 22 assume? l l
23 MR. CHELLIAH: yes.
24 MR. MICHELSON: That is correct.
25 MR. CHELLIAH: Which is beyond the isolation l l l
HERITAGE REPORTING CORPORATION -- (202)628-4888 l
91 1 capability in that room, flood isolation capability in, like a
() 2 draining all those things. So they did some careful review on 3 that.
4 CHAIRMAN WARD: That's a lot of floods, isn't~it?
5 MR. CHELLIAH: I haven't seen that facility. Water 6 pumps sitting there are going to be affected according to 7 them. They will tell more explanation in the latter part of 8 the presentation.
9 What EPRI claims is that 2 E minus 2 frequency is 10 not appropriate. What they have suggested is this. Postulate '
11 the pipe break frequency of the pipe failure, and then 12 postulate the sequence, so the EPRI data turned integral i
13 formula which uses not only the pipe failure rates, also some 4
14 of the quality factor, dynamic load factor, all those things.
15 Eobething like frequency of, flood frequency, something like 4
16 about 4 E minus 5 type frequency. So right there they are the 17 magnitude difference on the frequency.
18 We did look into some of the Point Beach review, 19 water failure in pump room, 1E minus 4 type frequency, at the 20 point we are taking. We do the EPRI matter, but we need more 21 data to assess the reasonableness of the flood frequency 22 because it is still meeting that frequency a little lower.
23 Also we believe the other frequency is a little higher.
24 That's our position right now.
25 MR. EBERSOLE: There is a confusion factor in the 4
HERITAGE REPORTING CORPORATION -- (202)628-4888
92 1 matter of identifying floods. Does--as the maximum credible
() 2 flood or possible flood. There is a flood that overruns 3 grade, which is much higher probability.
4 DR. ERICSON: We are talking internal, Mr. Ebersole.
5 MR. EBERSOLE: It is all internal? Sorry.
6 MR. CHELLIAH: Pump room, we are talking about 7 internal flood here. In the pump room we are talking about.
8 Did I say sitting on the room? We are talking internal 9 flooding.
10 MR. MICHELSON: When people talk about flood, there l
11 are various ways that one type of person talks about flood on l 12 the floor, what the elevation rises to and whether the
() 13 equipment is affected. Other people talk about a flood in an 14 upper floor and the water cascading, running, dribbling and 15 ':hatever down the lower floors. ,
16 Which kind of flood are we talking about here that 17 was evaluated?
! 18 DR. ERICSON: This is essentially a spray problem, 19 not submergence. This main is above--the pumps have a certain 20 credibility credited to them for existing even outside. We 21 took that into account. The break we postulated was 2
1 22 sufficient to cascade water on to the pumps, not to--
23 MR. MICHELSON: All of them? :
24 DR. ERICSON: Any one of them or any combination of 25 them.
HERITAGE REPORTING CORPORATION -- (202)628-4888
93 l' Mk. EBERSOLE: Why don't we say something besides ;
( 2 flood because--
3 DR. ERICSON: We did-call it spray study.
4 MR. MICHELSON: That particular problem took a 5 unique break I think.
6 DR. ERICSON: Take a break about those pumps, that's 7 right.
8 MR. MICHELSON: And that narrowed the probability 9 quite a bit because it is much lower probability of unique 10 break than general break.
11 DR. ERICSON: The point is all six of those plants 12 see at least 15 feet of the header, and now you are into part
(} 13 of it. That the point. i 14 MR. MICHELSON: We will get to it later. We will 15 get to it. ,
16 HR. EBERSOLE: The more general context of flood, do ,
17 you find plants that large have capability to drain the 18 various areas compatible with the largest source they might 19 see from a sewer pipe? Has that been designed into the plants 20 or not? r 21 DR. ERICSON: We do not do detailed analyses of 22 drains. We looked to see if ruptures in specific rooms could 23 submerge or affect that pump in general, particularly Point l . 24 Beach. It just goes away, returns out doors, and this sort of I
i 25 thing. You just can't pond it.
HERITAGE REPORTING CORPORATION -- (202)628-4888
94 P
1 MR. EBERSOLE: Great. I am not sure that is always
() 2 the case.
3 DR. ERICSON: In this case it is. You can't find 4 any place you can contain water inside.
5 MR. CHELLIAH: So going back to the previous 6 question, Dr. Michelson, you realize your question is whether-7 any support system failure other than random failures is due 8 to external event? Yes, they did identify the' failure, spray 9 failure mode, but I think right now we think that 2E minus 2 10 frequency is a little higher, the EPRI estimate of 3E minus 5 11 on the pipe break frequencies a little lower. We need more 12 data from them right now EPRI. THAT's the position we are 13 taking right now.
O(~N 14 on the seismic events, the Sandia, they did perform 15 very detailed study on seismic events. Basically what they ;
i 16 did, they made use of some of the program. Also they doubled 17 up detail of, for various structures and components, and this 18 is estimate. Three types of sequences; I have listed out on 19 the fourth page some loss of feedwater sequences and some 20 small break LOCA type sequences.
21 EPRI did look into--when I say look into, they sort 22 of did the scant type analysis because they have not performed 23 detailed analysis of what they did like Sandia, but there are
, 24 three points I would like to mention. One is the RWST tanks i
25 here, they are refilling the tank during a seismic event. Now i
HERITAGE REPORTING CORPORATION -- (202)628-4888
95 1 this particular tank is called very low DBR in plant,-
( 2 something like .192 median G as opposed to some other kind 3 like .8 or more than 1 G, so that claim is here that according 4 to EPRI, that tank will not catastrophically fail. Take about 5 30 minutes to deplete the water, so they started giving some i 6 credit to refill within the time,'you know. At the same time, 7 also refilling the CST, the condensate storage tank, for'the 8 secondary site cooling, they took some credit. Here we do, we 9 agree with the EPRI's plant-specific process for seismic 10 events, but I think here we are going to look into very 11 carefully this particular tank. We are finding very unusual, 12 somewhat unique design feature here in this plant. We need 13 some more additional information i i this.
(]}
14 Of course the third plan, second, third plan, lE 15 system, we do agree with the EPRI, some of the yearly blocker !
16 type sequence frequency comes along here.
17 Sandia, they have some similar ones there on this
]
18 particular one.
19 Here the EPRI, what they claim, they have doubled up 20 more site specific results using some improved technique.
21 They have not provided the curves, so we have to make some 22 provision here. We need some management information looking 23 to the curve, how they double up on all those things, but 24 basically we do agree, plant-specific, to estimate the 25 plant-specific seismic frequency on looking at those decay HERITAGE REPORTING CORPORATION -- (202)628-4888
96 1 heat removal problems. We are doing that, but we need more
) 2 data on that.
J 3 The final one on the fire events, Sandia,-they did 4 perform some limited study on this looking into various, 5 looking into fires in various critical zones. Here the fire 6 is limited to only transient combustible fires. We are not 7 talking about panel fire or the cable fires like what you see 8 in typical PRAs, the first code PRAs. So within that 9 limitation, they came up with this 3E minus 5 frequency. EPRI 10 did look into--you can see this. It needs some explanation.
3 11 One of the bullets I thought, the difference here, how you IN 12 estimate the initiating frequency for fire, Sandia, what they I'T V
13 did, the Sandia is consistent with the state of art PRA 14 matter, basically looking at the historical experience, report 15 on fires, and then looking into the amount of available 16 materials in each critical zone and they estimated the fire 17 frequency. That's what Sandia did.
18 Now EPRI, they started deviating from their practice 19 which they followed in other sequences. That is, they didn't 20 make use of the historical experience. What they did, they 21 went to each room like the LW pump room. They postulated just 22 one fire aven though the plant doesn't, may have any specific 23 operating experience, and they started estimating the fire 24 frequency based on operational experience, and based on so 25 many critical zones, so they came up with substantial lower !
i HERITAGE REPORTING CORPORATION -- (202)628-4888 i i
l l
.97 l 1
l 1 frequency.
O) s_ 2 We do not agree with that estimate. We agree with 3 the original estimate. This is based on the consistency with 4 the other review of the PRA review practices.
5 Then the second bullet, second and third and fourth 6 item here, the EPRI, they did look into plant-specific, they 7 identify one of the second train, halon system. I guess the 8 Sandia, they do also agree. They just have probability 9 according to here. And then here, the revised halon system 10 failure probability; Sandia, they looked into some experience, 11 some general experience. Here the EPRI, they did look into 12 some non-nuclear facilities, some facilities like the Rocky
(} 13 Flat, all those things. They came up with something like a 14 one in 17 halon failures, so they had halon system failure 15 probability as one of the automatic systems went down so 16 that's taken care of herr..
17 Also there is some human failure probability. They 18 took some additional credit, additional credit for the AWL 19 plant. There is, I realized it is not switch gear room here; 20 the AMWP pump room at Sandia. They assigned something like 4 21 one in ten chance that you can affect the turbine-driven pump 22 by opening the steam visible valves following the fire at the ;
23 LW pump. We don't know how far they are located. I'm sure
. 24 Dave could tell.
25 What Sandia is, what EPRI is saying, these pumps are HERITAGE REPORTING CORPORATION -- (202)628-4888
98 1 located far apart. The pump is located far apart from the
() 2 motor-driven. They have cut down that probability something 3 like .zero 3 by a factor of three. We do not go with that 4 kind of assumption. This is primarily because this, there is 5 some core melt timing involved. Depending on the fire damage, 6 if the core damage has to dry out the steam generator, dry 7 out, and core damage has occurred within about 30 or 40 8 minutes, probably the original estimate is probably .1, which 9 you see in other PRAs also.
10 In other words, if there is a slow fire, if you take 11 about two, three hours, then probably we can say that the EPRI 12 estimate could be justified, a little different situation 13 here. It needs little more detail from EPRI in this case, so
[}
14 basically here what we are saying the staff did look into 15 various contributing factors to the difference of the factor i
16 of 30 in EPRI estimate.
17 What we think right now that the Sandia estimate is
> 18 higher, but it is not lower like what EPRI says. It could be 19 somewhere--we are not quantified. We need additional -
20 information. We have asked them for a meeting. After that ,
21 we will be able to requalify and tell where we stand. ,
22 That concludes my presentation.
23 HR. MICHELSON: Question--on the fire protection of 24 Point Beach, is it all manual, or is it automatic sprinklers?
t 25 And if so, what type or what--it wasn't citar in the reports HERITAGE REPORTING CORPORATION -- (202)628-4888 I
a -
)
i 99 i 1 when they talked about manual mitigation, so forth, whether or {
2 not it was automatic backed by manual or just what it was.
)
3 Could you tell me? l 4 MR. CHELLIAH: The way the PRA analysis was done 5 using--
6 MR. MICHELSON: What is the plant? I don't care 7 about the PRA analysis. What is the provisions in the plant )
8 for fire protection and particularly say in the area of the 1 9 auxiliary feedwater pumps? l 10 MR. CHELLIAH: It has got both, manual as well as 11 automatic. l l
12 MR. MICHELSON: Does have automatic sprinklers?
13 MR. CHELLIAH: Halon system.
l 14 MR. MICHELSON: They are using Halon for fire 15 protection around the aux feedwater pump? l 16 DR. ERICSON: The pump room, that is one of their i
17 main cable routing rooms. That's why it is an issue. l 18 MR. MICHELSON: It is a cable routing, okay.
19 MR. CHELLIAH: Yes.
20 !!R . MICHELSON: They back the halon w.ith hand-held l 21 hoses?
22 DR. ERICSON: If I recall, that's true.
23 HR. MICHELSON: Each pump is in its own separate 24 compartment.
25 MR. ERICSON: It has a wall, but it is a common HERITAGE REPORTING CORPORATION -- (202)628-4888
100-1 room.
1
() 2 MR. MICHELSON: Ten foot wall?
3 DR. ERICSON: Eight foot, six to eight foot. ,
I 4 MR. MICHELSON: There is a very similar' arrangement 5 in number of other plants except a lot of plants use water i ;
6 sprinklers over those pumps.
7 DR. ERICSON: In this case, there is considerable 8 cabling on either end of the room runntng through'the room.
9 MR. MICHELSON: This was the case, too, that that's 10 why the staff made them put sprinklers in-the room and 11 sprinklers are a real problem like you talk about fire 12 anywhere in that area because the heat and smoke propagate to
(} 13 the sprinkler system and it thinks there is a fire there and ,
14 it will spray all the aux feedwater pumps and it is no 15 different than that header break that you talked about 16 earlier. You have seen a pretty good example of that lately j 17 on something else.
.i 18 MR. CHELLIAH: Thank you.
19 CHAIRMAN WARD: Okay. Any other questions? Have 20 you got any comments you want to make now?
21 MR. DAVIS: No, I don't.
22 CHAIRMAN WARD: Your plans now, you have left us, 23 you know, hanging on an opinion here on a number of these, so 24 your plans are to work with who in resolving these i
O 25 DR. MAZETIS: Again, our near-term goal is, as I i
l i
I HERITAGE REPORTING CORPORATION -- (202)628-4888
'101 1 have discussed with Jerry Nyles yesterday, of NUMARC, is to
() 2 sometime over the next month or so, meet with them, the-staff 3 meet with them, with Sandia, and discuss these various areas 4 of differing views or need for additional information and i
5 understand the basis for th Jr numbers.
4 6 Now it is not obvious to me that we necessarily are 7 headed in the direction of coming up with a new number per se.
8 Right now, I think our 7bjective is just to understand the 9 reasons for the differences. That's our primary objective, so 10 that eventually, any insights that we gain from this interface 11 with the industry can be disseminated to the rest of the 12 utilities, i
[} 13 HR. MICHELSON: Did the industry make any real 14 attempt to look for financials that you might have-left out of 15 your analysis that parhaps 3hould have been included? ,
16 DR. MAZETIS: I noticed there were a couple of !
17 places in the report where they explicitly pointed out several 18 areas of possible non-conservatism in the Sandia study. Maybe
~
19 Dave has some comments on that later, but in answer to your 20 question directly, my observation is yes, they did.
21 HR. MICHELSON: Are you picking up on those to go 22 back to see if perhaps you should revise your numbers upward l F
23 to reflect these, this lack of conservatism? !
24 DR. MAZETIS: Again, I'm not sure our objective is 25 to change any numbers in the future. .
l l
HERITAGE REPORTING CORPORATION -- (202)628-4888 i
)
s .
l[
1# ' .
qw
.if 3 ;n.
o s
.q 4 * >- gf .
a :.ic 3~
' _, ' 4
. ; prbh .
l -[ ,IyY l 1- MR. MICHELSON: If you vere tos any, numbers, q ;.y changg N % s y;y *,T VL 2 you would then re-evaluate sche.o'f thhsdpeinapsjinadedate.
e , r . -
.o.
' ..' {; ,4 y - , . , ,
-3 conservatisms' as well as where. they poini(1 outl..maybe you iwer:ed -
. >~/ ,
- dg '
,. :P 44
?
y .M 4 too' conservative? N" ' :
y, f JiwM"V S Not necessarily to m.
rehevalbst4~them*,.
,.>. ./
'DR. MAZETIS: -
4 c.m :y1 6 but to identify for the b'enefit~of whoever wants.to use tho ,
e s, ,a 7 guidance'from.these studiesc to identify.that'this'is an area 8 that they should.eith'er explore urther, or use new: numbers.
c 's.~ h' Parallel qudstic'n to-Mr;e "
9 CHAIRMAN WARD:
. _yl ./ ~ J.
10 Michelson--what about ' areas of incomplet5neds tha,t. Micig: . .
. . 3 /- .t.
- 11. study--did their study suggest any, you knos,.new. sequences or 12 things that the Sandia. study had: missed?
7s 13 DR. MAZETIS: Other than.the couple of areas,where
, x.
14 they mentioned because a sequence.may have bestnDoW rlookesi or' ^
.% e s , 4 ,,..,
15 an assumption,may have been' overlocked the sequence 'iinplied, .
16 maybe not conservative, I don't remember.anything that=was 17 significant that sticks out in my mind from the. report. Does, 18 that answer your question? , .
19 CHAIRMAN WARD: Yes, I think so.. Okay. Thank yc0 4
20 Mr. Chelliah. Now let's see. Mr. Ericson, so you'are going, 21 to kind of cover the same?
22 LR. ERICSON: Mine will be very short. It hasfall
- 23 been covered.
24 CHAIRMAN WARD: Let me poiiU: out to the t
O . m. s
- 25 Subcommittee, I think what we have.to consider here-is not HERITAGE REPORTING-CORPORATION - - -
(202)628-4888 m
-; . i,': .
. . _ _ ._,_ . - - . - ~ ._ _ ._ _ ~ --. _
k 103 .
- 1. ,.just the specifics of each case,. figuring-fout whether a' 1
, 2' . particulartnumber3ought to.be adjusted one-way-or.the,other,'
3- Lut.rather that.you know',Lpart of the. basis for the
~
kh 4 iecis' ion-making in this:whole USI'is the understanding-of. risk
-5' -in these six plants,'to the extentithey represent,lyou know,
~
6 the total population of plants outethere, to the extent;that-7 they'can. 'And so I think'wechave got to be,-think about.the r
8 process here as a meansof providing information'for the
- 9 essential decision-making as wellias the, just the' specifics s' ,.
10 of this particular; case.
11 DR. ERICSON: .Well, I said that partly.inLjest.and 4
12 partly seriously. I have a number of these introductory ,
s . .
/ /*) 13 vugraphs I am just going to flip through because we' talked
'V 14 about much of it before. ,
15 (Slide) 16 DR. ERICSON: For those of you'who don't know,'I am ,
17 no longer with Sandia, but I-am still involved.with the A-45 18 study. Mr. Adams, who is here, has been very much involved in 19 looking at the EPRI work and comparison with what we did 20 originally.
21 Just to reiterate, the objectives we set out to do 22 were to set some--and the point I would like to make, we'were 23 interested in the industry perspective in these objectives, to 24 try to get.as broad a view as we could during-this study, and 25 therefore, from day one, we tended to use generic data in the
? ?
g; 1r HERITAGE REPORTING CORPORATION -- (202)628-4888
.m .. 4- , . , _ . . - .., _ . _ . . _ . . . . . . _ - . . , . . . . . _ , . _ . , _ . . _ _ . _ . . _ , , , _ . _ , . - _ . _- n. . . . .;,,
104 1 available data bases unless there was some overriding reason
(_) 2 at that particular point.in time to use plant-specific data..
3 The limits were, as had been stated, to--we only 4 looked essentially at the small LOCA and'the' transients. We 5 were looking only at DHR, and although again we had specific 6 plants, we were trying to get some broader perspective.
7 Now the EPRI report deals specifically in the 8 analysis with Point Beach. Stated objective is to get-the 9 best estimate of Point Beach risk from an industry 10 perspective. And I think that's, that's the important thing.
11 It is from their perspective.
12 I think we should mention that they did no original
/"N 13 fault tree analysis. Started with the fault trees that were L.)
14 reported in case study. They did not go back and redo all of 15 those. They did use this risk management query system which 16 is a way of, developed by SAI, of storing event trees, fault 17 trees, system cut sets, and data, and it allows you to ask 18 questions. If I change the quantification, what is the 19 result? And these can be very useful tools. One has to be 20 very careful in coing this. I feel that you have to be 21 careful how much you change, because when the original fault 22 trees are solved, when the reductions are made, you make 23 certain assumptions or review assigned certain values, and if 24 you go back and completely eliminate systems or make gross 3
~J 25 changes, gross changes in values, you may be led astray.
HERITAGE REPORTING CORPORATION -- (202)628-4888
105 1 I don't think that has happened here. I think in)
(')s
(- 2 are all right and that they are all r'ight.in using it in this 3 particular application, but it is.a caution again in the 4 broader perspective as we look toward what industry may be 5 asked to do as-part of the resolution package.
6 (Slide) 7 DR. ERICSON: All of this has been mentioned before.
8 Let me put some numbers with this and I will come back to them 9 again, but new, the internal core melt which includes the 10 long-term Station Blackout, the reduction is a. factor of 50, 11 but if you look at this, the big issue or the biggest 12 contributor in this reduction is that small LOCA frequency 13 that we have talked about. The small LOCA reduction takes
( ']
14 about 50 percent of the frequency, and then if you add to it 15 the changing relief valve, about 80 percent of the differences 16 is accounted for by those two, so while these other iscues 17 were addressed and are e.ddressed in their, in the EPRI report, 18 there are one or two things that dominate.
19 Same way here, and I will just skip through this one 20 and come back to some of these comments on the specific 21 issues. In particular, for example, if you look at the 22 details of the analysis, the biggest difference in the fire is 23 caused by the credit given for suppression reliability. We 24 had a per train of .2. They have something on the order of 25 ten to the minus.3, so when you start multiplying those HERITAGE REPORTING CORPORATION -- (202)628-4888
an 106 1 through, that's your big player. I will say some more about-(~)%
(_ 2 that.
~
3 MR. EBERSOLE: May I ask you a question back.on the 4 other slide? You had in your accident, you. dismissed large 5 LOCAs and other things.but you did include small ones.
6 DR. ERICSON: That's right.
7 MR. EBERSOLE: Now implicit in that, although I 8 haven't heard anything about how you are going to design a 9 decay heat removal system, is presumably that it will'have to 10 have a replenishment system for PWRs, but it really won't have 11 to have any for boilers because they are going to cool by 12 evaporation anyway.
[} 13 DR. ERICSON: We did include one that has the 14 capability of that.
15 MR. EBERSOLE: They add I am sure quite a bit to the 16 cost as you have to cope with a small LOCA. In other words, 17 you have got to provide makeup to get transfer in a PWR 18 whereas it is only a part of the normal evaporative process in 19 the boiler.
20 DR. ERICSON: The capital cost for the high pressure 21 pumping equipment are not the major factor.
22 MR. EBERSOLE: It is not?
23 DR. ERICSON: It is the civil costs associated with l i
g 24 building those separate buildings and tank and water supplies.
J 1 25 if you look at the details, in fact, as an instance, the BWR, j i
1 1
I HERITAGE REPORTING CORPORATION -- (202)628-4888 l
. - . . ~ . ...- . - -
't ze
~
1107-1 we looked-at itiwith'ut o some high pressure makeup,1and'I1think' 2 the cost'was-like 80~some;million'and it wasLanother million1 3 and a halk to put a high pressure pumpLin one. The'BWR'had
'4 ,been designed, much of the civilicostsEthattjust: gobble?youi
'5 ~ alive on the a'dd-ons.
6 HMR . EBERSOLE: .All right.
7 DR. ERICSON: The results, these have been mentioned 8 before; overall the factor of 30. Additiondlly, they'relooked 9 at the core melt or at the containment perform'ance, and the
(
10 consequence sometimes another' factor _of-seven, and.some 11 increases in costs, and we will.come back to touch just 12 briefly on those today.
13 (Slide)
~
14 DR. ERICSON: I will sk'ip the-next one. We have 15 talked about the general forms of the accident sequences, 16 transients and small LOCAs, and we have although not 17 specifically listed it as such, the real. reasons for the 18 differences are this removal of the CCW in'the objection 19 phase; the changing frequencies ~for various transients, very 20 low numbers for, lower numbers for Point Beach, the.
21 cross-connect capabilities and then operator or human error 22 factors in which there is some changes, some areas of 23 disagreement. q
- 24 CHAIRMAN WARD: The second one dominate, .right?
25 DR. ERICSON: In the internal, that's a biggerg the
' HERITAGE REPORTING. CORPORATION -- (202)628-4888
108 1 initiating event frequencies are multiplied. When when you
'( ) 2 change those by-order of magnitude, you have a- tremendous 3 impact on the obvious.
- 4. MR. DAVIS: Excuse me, James. Put that slide back 5 up.
6 DR. ERICSON: Yes.
7 MR. DAVIS: The biggest single difference in the two 8 studies was the flooding sequence. Now you are talking about 9 internal only?
10 DR. ERICSON: This is pretty much internal.
11 CHAIRMAN WARD: Yes. Bu'. what is--this is all 12 internal, isn't it?
/~'s 13 DR. ERICSON: These are internal, yes; initiating, V
14 yes.
15 (Slide) 16 DR. ERICSON: Let me--some things I think we need to 17 talk about, some of which we have already talked _about, and 18 those of you who have heard me for four or five years know I 19 am always willing to talk about something again.
20 First, the timeliness of data; in this review, and 21 in several other outside reviews that have taken place over 22 the, for the case studies, we have frequently been chastised 23 for not using data that was available in 1986. Well, I'm 24 sorry. We did the study in 1984 and 1985. It took us a year p)
\-
l 25 to get it printed, but that's life in this business, as you I
l HERITAGE REPORTING CORPORATION -- (202)628-4888 !
-l
109 1 all know. So we are dealing with snapshots. .You can't l fT A_/ 2 . continually; update just because the new piece of data came out l 3 yesterday. It just, you just can't do it.
4 Plant characteristics, in my view, to some extent, 5 this study deals with a different plant than we dealt with.
6 There are seismic one, new seismic one batteries. Now it was 7 said they were added. I don't know whether they added or 8 replaced what was there. I can't tell from what they say in 9 the record, quite frankly. That's an issue that we have to 10 discuss with them, as has been pointed out.
11 The other area that we will talk about, human 12 effectiveness and recovery actions--we have been' accused from Il V
13 day one of being very conservative with our recovery. I would 14 suggest to you that the EPRI analysis is very optimistic. Let 15 me give you some specifics. Question of whether a plant will 16 feed and bleed? Now as has been pointed out in the bottom 17 line, whether they do or do not doesn't make a big difference 18 in the core melt frequency when you talk about what operator 19 will or will not do. Davis-Besse, they could have, but they 20 didn't. This report says we talked to a couple of operators 21 and they said they would have no hesitancy. I'm sorry, 22 gentlemen, I have had technicians for many years working for 23 me and I also taught school, and people who have been through 24 training know what the school solution is. If you have asked O.
25 them in the classroom what are you going to do, they will give HERITAGE REPORTING CORPORATION -- (202)628-4888
j 110 1 me, better give me the right answer or they flunk.
() 2 What they will-do at 3:00 in the morning with, as 3 our former colleague Mr. Reed used to say,'the world turning 4 brown,- I don't know. I am more sceptical than those who did 5 the study. The times is available. Those are things that are 6 discussable and we will be discussing those.
7 Modification costs--I am surprised that there is 8 such a difference in the cost as I will come back to that in 9 in a minute. So in general, to that extent it was a different 10 plant. I think this is a key issue, and I heard the word 11 "correct" used here a couple of times this morning. I don't 12 think anybody in the PRA business, and I hope Mr. Davis will
(~T 13 agree with me, knows whether they have got quote, the correct U
14 answer.
15 CHAIRMAN WARD: That is a good place to pause.
16 (A discussion was held off the record.)
17 DR. ERICSON: Wouldn't you agree, Mr. Davis, this is 18 the crux of some of our discussion, what is enrrect or what is 19 not correct? I wish we had, could go to the grade book and 20 get the right answer!
21 At times, I think even with the new emergency 22 response guidelines, though, we agree and understand they are 23 very specific. At this point I am not sure we have enough
- 24 experience with them to really know what performance will be.
s 25 And in our review, and again I think it was--
HERITAGE REPORTING CORPORATION -- (202)628-4888
111 I l
l 1 MR. MINNERS: The new guidelines for this plant? ]
/"% l
(,) 2 DR. ERICSON: In general, but certainly here,'and in 3 this particular review, and again as we said preliminarily we 4 haven't had a chance to interact, because you have got to 5 study a report a'1cng time before you'know what questions you 6 want to ask or you are going to waste a lot of time. There 7 are times we are not sure exactly what data was changed and 8 where. That is we know pretty much, but in the details. This 9 is then success criteria, this has been commented upon. The 10 EPRI document, EPRI owners group document, says you don't need 11 it for cooling, but in almost all the other PRAs that 12 generally have been used, so this is a point we netd to 13 discuss.
14 The balancing of service water, what they have 15 presented certainly sounds like a reasonable thing to do and 16 if the procedures exist and'it can be done, we would expect 17 that it does get done.
18 Human reliability quantification, they make the 19 point they have reduced the failure to initiate feed and bleed 20 one times ten to the minus 3. One times ten to the minus 3, 21 that is because they believe we have, the individual event 22 activity, or criteria in our study, was one times ten to the 23 minus 3, but they argue that we have. summed those and summed gS 24 it erroneously as a point we need to discuss, several times, U
25 and this is one place they make the point that there are many HERITAGE REPORTING CORPORATION -- (202)628-4888
112 1- people in, many people in the control room following the event O
(.) 2 so something is, so.the action has been sure to be taken.
3 Does the mere presence of lots of people mean something will 4 get done? That's something we can discuss.
5 HR. MINNERS: And here we are!
6 DR. ERICSON: I can show you lots of places where 7 lots of people were there and nothing happened.
8 MR. EBERSOLE: 'Like the Browns Ferry fire, it took 9 them six hours to put water on it.
10 DR. ERICSON: The reduced failure to initiate sump
~
11 recirculation, as a matter of fact we acknowledge-that in the 12 study because we originally were going to require some
(} 13 redundancy and went back and said no, we don't need to. We 14 were assured the training cases and we have suggested maybe 15 another alarm, but then did not make any changes.
16 The reduced failure to depressurize, the big issue 17 here is not in the baseline values, but the application of 18 stress factors, and I would not claim to be a human 19 reliability analyst or specialist. We multiplied and you can 20 see the difference is a factor of five that we had there for a 21 stressful situation.
22 MR. EBERSOLE: Has the operator, did you look at the 23 matter of what the operator can see, validate, that he should 24 depressurize?
- 25 DR. ERICSON
- We did not, not in our study. Again l
I HERITAGE REPORTING CORPORATION -- (202)628-4888
113 1 remember we did not do a detailed HRA. We used the---
2 MR. EBERSOLE: I can't find, see does.the 3 information flow to him that says that's the last bit I need,-
4 now blow it?
5 DR. ERICSON: We certainly did not look at that.
6 MR. EBERSOLE: I think it is hard to find.
7 DR. ERICSON: I think they have, I'think in the EPRI 8 as I recall, the way they worded this thing, they have looked l
9 at that somebody said A, B, C and D, not in the report, but I 10 think in the procedures that now exist in the ERGS, in fact 11 one of the EPRI people may comment, it gives you a specific 12 point as to what to look for in the, in those new procedures,
(} 13 because they are symptom oriented. And then they have used 14 the new--not new, but SAI approach to the time reliability 15 correlation and in some cases additional actions which we 16 might comment on later.
17 Just to refresh your memory, though, we have been 18 repeatedly accused of only using one recovery. We do in fact 19 have two before two hours. If you include loss of feedwater, 20 either recovering the off-site power or loss of feedwater plus 21 another after two hours, we have had that plus two other 22 things.
23 The study, owners group study looked at the number 24 of options available, the time that they felt required to tO 25 perform any given action, whether it existed in the procedures HERITAGE REPORTING CORPORATION -- (202)628-4888 I
114-1 or not, and whether there had been training on it,-so their
/~'s
() 2 application of recovery I would say is certainly more 3 plant-specific and allows for more flexibility.
4 MR. EBERSOLE: May I ask you in the particular case 5 of the boiler, the staff says that loss of containment heat 6 removal, they believe every boiler has a means to vent the 7 containment to prevent disastrous overpressurization before 8 the core is hurt.
9 Are you with me?
10 DR. ERICSON: Yes.
11 MR. EBERSOLE: You find that to be true or not, in 12 terms of equipment and procedures? I don't think it i s '.
13 DR. ERICSON: I am shifting gears, Mr. Ebersole, (a~}
14 because we were not--you keep jumping to the BWRs and I 15 haven't looked at those in a while.
16 MR. EBERSOLE: That has been a curtain that is 17 lifted. Oh, yes, you can dismiss the--
18 DR. MAZETIS: Of the six case studies, two of them 19 were BWRs.
20 CHAIRMAN WARD: Sample is only two of--
21 MR. EBERSOLE: Okay. You only had the two.
22 DR. ERICSON: We gave credit for limited venting.
23 DR. MAZETIS: Limited examination of procedures
. 24 observing the credit could be given for that.
J 25 MR. EBERSOLE: Can you, did you find the equipment HERITAGE REPORTING CORPORATION -- (202)628-4888
115 1 to be present and the procedures there?
b 4/ 2 DR. MAZETIS: The procedures were there, but I just 3- don't remember about the details-of the equipment.
4 MR. EBERSOLE: Conflict is whether in fact cven the 5 equipment is there or not.
6 MR. MICHELSON: Or if the equipment will function 7 under the conditions that will likely exist at the time 8 venting is required.
9 DR. MAZETIS: I know that wasn't.
10 MR. MICHELSON: If they won't, then the procedures 11 and equipment are worthless because these can't function.
12 DR. ERICSON: The next slide just recaps those
, ("i w/
13 recovery actions that we had in our, in the case study we did.
14 We cove. red off-site power, main feedwater, battery common 15 cause, battery faults, diesel, and then other things that one 16 can do from the control room or things one can do locally 17 without spelling those out specifically.
18 In the EPRI owners group analysis, again, as was 19 mentioned earlier, the refill of RWST which would allow you to 20 stay in injection mode because you have additional water, I 21 think for internal events if indeed that plumbing system 22 allows you to do that, I would certainly accept that. I have 23 a little trouble with someone telling me that refilling the 24 RWST is a recovery action for a seismic event. We will come 25 back to that in a minute. Cross-connecting buses, those l HERITAGE REPORTING CORPORATION -- (202)628-4888
116 l
l 1 cross-connects do exist. Manual control turbine drive aux i
( 2 feedwater pump, they take considerable credit for this now-3 because of the new batteries, allowing, giving the supply even :
4 though it has to be done locally.
5 MR. MICHELSON: Have they ever operated those pumps 6 under that condition? Have as a test?
7 DR. ERICSON: I do not know.
8 MR. MICHELSON: Does anybody know? It would be a 9 fairly--if it is, it is purported to be not that difficult. I 10 have talked to some people on the side as to how difficult it 11 may be, but I just wondered if does it--have the staff ever 12 known of anybody that has actually demonstrated?
() 13 DR. ERICSON: There is one. I can't think of which 14 plant it was. I have read a report where they, indeed the 15 fact went in and ran some of the stuff manually.
16 MR. MICHELSON: That is the way to-put-it to bed in 17 a hurry. Otherwise I hear conflicting stories, maybe very 18 equipment specific. The real problem is control, of course.
19 It is all done manually, they can't feed back fast enough to 20 vent the trips.
21 DR. ERICSON: I guess the other one, recovery from i
22 common mode failures, they relied heavily on. AEOD study that '
23 said significant numbers could be recovered and so they r~g 24 applied, reduced some of the common mode stuff apparently V
25 somewhat arbitrarily.
HERITAGE REPORTING CORPORATION -- (202)628-4888
117 1 (Slide)
/~T,
(/ 2 DR. ERICSON: I'think we have beat this one to death 3 at this point. The small break LOCA frequency as to what you 4 count or don't count, we will be discussing that more with 5 them. As has been pointed out, it does make a significant 6 difference.
7 The PORV, the relief valve PORV LOCA situation, I 8 guess I would have to agree that the arguments they present, 9 as the staf f has pointed ot.t, when we ourselves take a look at 10 the thermohydraulic analysis, we made, certainly accept that Q 11 or that event as not having occurred given that feedwater is 12 present so that you have got a transfer mechanism for the heat
[} 13 to circulate, reasonable argument.
14 MR. MINNERS: What is RCMF?
15 DR. ERICSON: Reduction core melt frequency.
16 MR. MICHELSON: This statement doesn't pertain I 17 guess to boiling water reactors? Is that what you are saying?
18 DR. ERICSON: We are dealing specifically in our 19 comments here with Point Beach, Point Beach only.
20 MR. MICHELSON: Okay. Thank you.
21 (Slide) 22 DR. ERICSON: I might, I would make the comment in 23 the small break LOCA frequency, this study uses a lot of 24 Westinghouse data and Westinghouse topical reports that J
25 so-called WCAPS, to make their point, but the small break HERITAGE REPORTING CORPORATION -- (202)628-4888
118 1 frequency argument is not for a Westinghouse plant. They use
() 2 Oconee, and I don't think that's a Westinghouse plant, but it 3 is a smaller number.
4 (Slide) 5 DR. ERICSON: Common cause failure rates, they have 6 used a number significantly less than.ours, taking exception 7 to our use of beta factors in some places. They talk about a 8 design review and that certainly is the method that Carl 9 Flemming has argued for, still that you really need to.look at 10 those things, and yet I don't believe based on the way I read 11 this, the EPRI record, that they really did those detailed 12 design reviews on this plant but have used similar sorts of
'N 13 things, and certainly they did use significant amounts of data (O
14 from the Millstone 3 PRA.
15 (Slide) 16 DR. ERICSON: You can see very quickly that 17 significant differences come in in the batteries. For 18 example, other pumps are down by about a factor of three. The 19 combination of motor driven and turbine driven aux feedwater 20 pump is down order of magnitude and those do have their 21 effects promulgated through the system, and I think this is an 22 area where we will want to have considerable discussion with 23 those folks as to how their numbers were arrived at.
24 MR. MICHELSON: Clarification on common cause--are 25 you dealing here with the common cause perhaps being seismic HERITAGE REPORTING CORPORATION -- (202)628-4888 l
119 1 or flood or a fire, et cetera?
(~N
(/ 2 DR. ERICSON: In those numbers, no. Those were--
3 MR. MICHELSON: Purely equipment?
4 DR. ERICSON: Equipment failures; in the seismic 5 analysis we did a correlation, though, so that if the pump.was 6 in the same room, they both fail, that sort of thing. We 7 certainly understand that they have installed new batteries 8 now and where batteries exist in our seismic sequences, they 9 may have an effect, but it does not appear in the report as we l
10 have analyzed it thus far where they looked at these new 11 batteries at the two, three and four times SSE levels where i
12 there are significant seismic risks.
13 The refueling water storage tank failure and
[}
14 recovery, I think it has already been mentioned, but I will 15 mention it again. The RWST at Point Beach is unique. I can 16 best describe it as a smokestack. It is very tall, and very 17 short in diameter. Our people have done the fragility 18 analysis. They are convinced that tank will fail relatively 19 catastrophically when it goes. When it buckles, it will 20 buckle. So I am not sure you can refill it.
21 I would also suggest somewhat humorously but very 22 seriously when you talk about recovery actions in seismic 23 events in 30 minutes, I don't think you have ever been in an l
24 earthquake. Things are pretty chaotic in those first 30 l gs ,
%] l 25 minutes, and-- I HERITAGE REPORTING CORPORATION -- (202)628-4888
120 1 MR. MICHELSON: When dealing with the RWST failure,
(~
(_)\ 2 catastrophic, potential catastrophic failure,lare they looking 3 also at the flooding effects of the nater release?
4 DR. ERICSON: They did not, but that thing will dump 5 a lot of water in places you don't want it.
6 MR. MICHELSON: The piping arrangement coming into 7 it, there may be a cavern underneath that tank with the piping 8 in it.
9 DR. ERICSON: This tank could get back up in aux 10 building.
11 MR. MICHELSON: Runs into potentially interesting 12 situations.
() 13 DR. ERICSON: This tank could get water back to the 14 aux building.
15 EPRI has argued seismic hazard curve conservative 16 for factor of two for low accelerations, three to five for 17 high accelerations. We disagree. They have not reanalyzed 18 this particular site. Our approach was to normalize curves to 19 the frequency at the SSE, look at local attenuation factors, 20 local soil, column effects, to get the curve. In talking with 21 our seismic people, they feel that our numbers will be, would 22 be very consistent with the complete site specific curve if 23 one were to take the techniques that now exist and do it, 24 which would take you probably three or four man months in time 25 to generate that analysis.
HERITAGE REPORTING CORPORATION -- (202)628-4888
121 1 MR. MINNERS: Low accelerations below SSE?
/3
() 2 DR. ERICSON: They said conservative at one to, 3 factor of two at 2 to 3 SSE and factor of five above as I 4 recall.
-5 MR. EBERSOLE: That tank is the source of highly 6 borated water. To simply shut down the--
7 DR. ERICSON: There are, the reactor makeup water is 8 also available on site.
9 MR. EBERSOLE: Is it better off than this?
10 DR. ERICSON: As I recall, it is a squatter tank.
11 Again, the question of how much your recovery action in early 12 on in seismic; we were asked earlier did they identify things
(; 13 that we had? Yes. And this is a specific example. Turns out 14 the way Point Beach opens the CSTs intertied, so if you lose 15 one, you have lost them both because the valving is locked 16 open between them.
17 MR. EBERSOLE: Did you find out by any chance as we 18 did at Diablo Canyon the emergency manuals' fall over the 19 floor, you couldn't find them in seismic events?
20 DR. ERICSON: Well, I haven't looked at the book 21 racks in the control room, so I don't know the answer to that 22 question.
23 Internal flood, as has been pointed out earlier, the r~g 24 significant differences in the, occurs in the frequency that U
25 is used. We used a generic flooding frequency room by room HERITAGE REPORTING CORPORATION -- (202)628-4888
.. , , ,__ , . _.~. , - - . . , - -. -
122 1 for aux buildings. EPRI has used a pipe frequency, break gs
(_) 2 frequency correlation by Thomas, which they use and then ,
1 3 recompute, or estimate the pipe break frequency for the pipe, 4 service water pipe, over those fire, water pipe over those 5 service water pumps.
6 MR. MICHELSON: Do you know what that number is 7 approximately? Pipe break frequency I guess per foot of 8 length or something?
9 DR. ERICSON: Oh, P--no. This one is a computed 10 number, and when you use their numbers, this is the 3.7 times 11 ten to the minus 5 number.
12 MR. MICHELSON: Per unit length of pipe?
(} 13 DR. ERICSON: Isn't unit length. It is on a--here 14 it is, right here--1.8, one ten to the minus 8 per year per 15 cube. These are the quality factors I would call them.
- 2. 6 MR. MICHELSON: What is cube?
17 DR. ERICSON: Relates to how thick and how big the 18 pipe is.
19 MR. MICHELSON: So you, somehow that is a per foot 20 of pipe, isn't it? And then you multiply by--
21 DR. ERICSON: Q is ratio of--you look at this 22 product to pipe diameter in feet, timas length in feet, 23 divided by thickness squared.
(^g 24 MR. MICHELSON: If I have a hundred such lengths, it V
25 is a different probability than if I only have one such HERITAGE REPORTING CORPORATION -- (202)628-4888
123 i 1 length.
(_/ 2 DR. ERICSON: Q is dimensionless parameter.
3 MR. MICHELSON: How do you convert it into, to 4 reflect the number of total length of pipe that you have in 5 your plant?
6 DR. ERICSON: That Q is based on this three-foot 7 lengths. See, EPRI argued that there was only one three-foot 8 section of pipe that could affect you in this particular 9 situation.
10 MR. MICHELSON: They are doing--okay.
11 DR. ERICSON: For that specific location, one 12 three-foot.
(~% 13 MR. MICHELSON: That's a strange--
q.)
14 DR. ERICSON: We don't understand that because we i
15 feel that the pumps can see-anywhere from ten to 15 foot.
16 That's a point of discussion.
17 The other problem I had, quite frankly, when I 18 looked at this correlation and went back to the reference, I 19 find that what EPRI reports in the document is not what I 20 found in the original reference. There has been modifications 21 made to it that are not documented. The original work by 22 Thomas doesn't use a dynamic loading factor but uses factors 23 related to age and design learning.
rw 24 They have used a restricted pipe length only three
()
25 feet which we need we have to discuss that with them why they 1
HERITAGE REPORTING CORPORATION -- (202)628-4888 i - _
i i
124 1 feel it is only that three-foot length, and these two are (s~)% 2 related and the definition in' terms in'the equation, so we 3 have, some discussions have to take place on that correlation.
4 MR. MICHELSON: How did they treat the non-seismic 5 piping that could produce floods versus the seismic that could 6 produce floods? Did they treat those as any different than 7 probability of failure?
8 DR. ERICSON: .The only one that was computed was the 9 one we argued, fire main water. They used this as an 10 empirical relationship based on data, on industry-wide data.
11 MR. MICHELSON: Only pipe in_the whole plant that 12 could give them a problem if it broke? Is that the idea?
/~} 13 DR. ERICSON: Our argument was that given, looking V
14 for internal flood sources, this was the one we found.
15 Whether they went and looked for others, I don't know, one of 16 the reasons being, for example, there are significant rooms at 17 93 int Beach which have big service water pipes but they are in 18 guard pipes, so if the pipe ruptured, it would go outside the 19 room.
20 MR. EBERSOLE: Tell me, there are three high 21 pressure pipes at least, in a PWR, there is more in others.
22 That have a substantial different potential in others. It is 23 the, I will pick--well, the first one is the steam line that fs 24 feeds to, the steam to the steam-driven aux feed-pump. That O
25 is service area either pressurized up to the valve or right up HERITAGE REPORTING CORPORATION -- (202)628-4888 L
125 1 to the stop valve. I don't know. Two other reactor waier 2 clean-up systems; if_those pipes break, the effect of break is 3 regressive; if the flow continues on out into the machinery 4 area, disables the whole damn plant.
5 One must place dependence on valves' operating in an 6 unusual and untested and not in a statistically verified mode.
7 Did you go to that level detail?
8 DR. ERICSON: For clean-up, for example, we did not.
9 MR. EBERSOLE: What about steam on the aux feeds?
10 DR. ERICSON: Steam on aux feed can't cause a No.
11 spray or flooding; causes steam, so we did not look at it.
12 MR. EBERSOLE: Saturates the whole electrical
/~T 13 network in the plant. It is a Turkish bath. Everything O
14 shorts out.
15 MR. ERICSON: If it is in the right place; no, we 16 did not look at the steam lines, that steam line.
17 MR. EBERSOLE: Those happen to be focal points of 18 severe potential.
19 DR. ERICSON: In the area of containment, success 20 criteria, EPRI argued and the owners group argues that primary 21 system leakage in small LOCAs do not place a significant load 22 on containment, and I think Mr. Minners made that same comment 23 a while ago. They look at recovery. The report says we used
- 7. 24 critoria for large LOCA. To some extent, that's right. We U
25 said originally in the studies that we would use the RESMAP HERITAGE REPORTING CORPORATION -- (202)628-4888
126 1 ' approach because we did not have the resources at the time to p)
(, 2 go into detailed containment performance.
3 We did, however, modify that as the NUREG 0956 and 4 other reports were coming out, the 1150 work, and reduced the 5 failure probabilities; we have'something like a 55 percent 6 non-failure probability for the large containment, so we did ,
7 not do it sequence by sequence the way EPRI has, and I cannot 8 comment on theirc because at this point we have not had the i 9 opportunity to get back through all of those yet.
10 They provide a comparison of using'95 and BMI 2104, 11 the IDCOR data, their judgment, and using our baseline case of 12 .3 times WASH-1400 for the source terms, and get this range of
(} 13 values. You may recall that we, we had.an estimated person 14 reactor year of something like 38 I think is what we report in 15 our report based on our analysis and the 3.3 times WASH-1400 16 numbers.
17 (Slide) 18 DR. ERICSON: We have to do some more comparison and 19 work on that one. We have not done that yet.
20 Cost estimate differences, failure to consider all 21 design requirements for specific aspects of modification, I 22 think this is, refers primarily to the add-on system.
23 Failure to account for existing structures and/or rs 24 buried piping in cable, frankly I find this one very puzzling b
25 that they raise this issue since after we did our conceptual HERITAGE REPORTING CORPORATION -- (202)628-4888
., - . - . . ~ . _ _ _._ -
127' 1 designs we went to the plant, walked through'the plant with a
() 2 plant staff, asked if we could do this, can we do that, will 3 this fit there, and what are your local labor costs.and what 4 is your experience? So to tell us that we haven't done that 5 and we are way off, I don't understand.
6 Failure to account for iteration between ' initial 7 design and final installation, and then they argue cost 8 cxperience and installing new batteries, quite frankly, we 9 have not explored all of this. I am very, very puzzled by it 10 since the architect / engineer who did this work does-this kind 11 of thing for the industry all the time, and I know they were 12 very conscientious about checking their costs to make sure 13 they weren't optimistic because somebody might ask them to bid
} ,
14 on them, and so it is puzzling just why we have such a mtrked 15 difference.
16 DR. SIESS: Was there a time difference?
17 DR. ERICSON: The EPRI report doesn't say anything 18 about time, but we were used, we used 1985 dollars and even if 19 you use 1987, I don't think labor rates and that sort of thing 20 have changed that much in the last two years.
21 CHAIRMAN WARD: Is that what you mean? Okay.
22 DR. ERICSON: In fact, some areas, it is cheaper now 23 than it was before.
24 One other thing we have looked at, though, is this 25 question of human reliability and emergency procedures, and i
HERITAGE REPORTING CORPORATION -- (202)628-4888
128-1 there is a NUREG that came out last year that suggests that e
k_r) 2 emergency operating procedures can markedly improve,'and 3 perhaps reduce core melt by as much as 12 percent and that 4 there are synergisms. I.have no reason to argue with that, 5 and--
6 CHAIRMAN WARD: -What NUREG-is that? Who wrote that?
8 DR. ERICSON: Are you familiar with that one?
9 DR. SIESS: Did you mean what you said by factor of 10 1 percent? This says factor of one eighth up there.
11 DR. ERICSON: Isn't that 1 percent?
12 DR. SIESS: That would be 88 percent if I reduced it
(} 13 by--
14 DR. ERICSON: Reduced it by tight, that is 12 15 percent.
16 DR. SIESS: By an 8th.
t 17 DR. ERICSON: That's what it said.
18 DR. SIESS: That's practically nothing.
19 DR. ERICSON: Well, I have learned one has to be 20 very careful when people talk about reduced by factors of, and 21 that sort of thing.
22 DR. SIESS: I would multiply it by an 8th if I just
, 23 read it and hadn't heard what you said.
24 MR. ADAMS: I think you multiply it by one 8th, 25 Dave.
HERITAGE REPORTING CORPORATION -- (202)628-4888
129 l
1 DR. ERICSON. Eighty-eight percent, okay;.I haven't
~ l
/ )T
\_ 2 gone through the whole report-yet. We have to be careful of- 1 3 that.
4 DR. SIESS: Twelve percent, you forget about it.
5 Wouldn't even make the slide.
6 CHAIRMAN WARD: Okay.
7 DR. ERICSON: There is one slide.
8 CHAIRMAN WARD: Also, well, whether it is, if it is 9 12 percent, I don't want to say this. If it is 88 percent, I 10 guess I do want to say it, so I don't know. I don't want to 11 take the time. But I don't know what sort of study this was, 12 but you know, reports we have had about the efficacy of EOPs
/~3 13 is that it is, maybe it is good in a theoretical study that V
14 INEL might be doing, but as far as EOPs, that is actually in 15 the hundred and some plants, there is some question about 16 whether they are doing any, much good or how uniform that good 17 is spread around the plant.
18 DR. ERICSON: I think your latter point is certainly 19 valid. I think it will depend very much on staff, on 20 management attitudes, that sort of thing. Even with the best 21 of procedures, I think, you know, we can talk about simulator I 22 experience, but somehow that just isn't quite the same. You 23 know; if you make a mistake on a simulater, can always go do fg 24 it again.
U 25 HR. ADAMS: That EG and G report at Idaho, HERITAGE REPORTING CORPORATION -- (202)628-4888
s i. ;
.130 1 high-level report on the on-site assessment of the -
' :,+ -
s d 2 ef f ectiveness and impacts .of upgraded emergency operaying'-
3 procedures. ,
4 MR. MICHELSON: What is that name again?
5 MR. ADAMS: On-site assessment of the.offectiveness{."'
6 and impacts of upgraded emergency' operating procedures, s 7 CHAIRMAN WARD: Carl,. this is another subject, but I 8 think--
9 MR. MICHELSON: What is the NUREG number?
10 MR. ADAMS: 4617. '
11 CHAIRMAN WARD: It is en the sheet.
12 MR. MICHELSON: I missed it. I'm sorry. I'm o sorry.
(} 13 CHAIRMAN WARD: This is kind of another subject, but 14 your subcommittee was going to look at this matter.
15 MR. MICHELSON: Yes. I have got it.
16 DR. ERICSON: Let me close with the following 17 additional general comments that aren't in.your package.
18 Quite frankly, in the review process, this one and others, one 19 of the problems we have had is people misquoting what we did.
20 This report says we considered only trensient combustible 21 fires. That is incorrect. We have responded at least to four 22 different times when people have said that.
23 We only used a calculation of a trmnsient '
24 combustible to bound the amount of energy in fires. The 25 frequency, locations are all done by separate analysis. I get HERITAGE REPORTING CORPORATION -- (202)628-4888
,,i ~ 1 131 1 frustrated-and I know our analysts do,-too, when that keeps
/~T.
(j -2 happening.
3 MR. CATTON: What about the comment about acetone 4 being everywhere?
5 DR. ERICSON: That is not what we assumed. The i- 6 prior frequency in a room is a function of the histori' cal' data-7 for those kinds of rooms. The acetone fire was only looked to 8 see if a fire of certain intensity could propagate to cables-9 and propagate to something else, and I see Mr. Busley smiling' 10 - over there. He knows what I am saying. The best estimate 11 analysis, EPRI claims to have a best estimate analysis. There 12 is no discussions of uncertainties.
13 DR. SIESS: Uncertainties in the best estimate or
.(/ r')
14 uncertainties in the--
15 DR. ERICSON: Uncertainties, how good is that best 16 estimate? We have used a conservative number. If you look at 17 the results coming out of, that have come out of NUREG 1150, 18 and every other PRA where a detailed propagation of 19 uncertainty is taking place, the 95 ard five are a long way 20 from the best estimates or means.
21 CHAIRMAN WARD: So you are saying that your PRA 22 numbers are not what you would characterize as best estimate 23 or--
fg 24 DR. ERICSON: We called them point estimates based N-)
25 on available generic data. We also gave you some discussion i
j HERITAGE REPORTING CORPORATION -- (202)628-4888 i
-132
~
1 of the potential effects of uncertainties, but we did not O
(m/ 2 propagate uncertainty through our analysis.-
3 CHAIRMAN WARD: But did you cover uncertainty by 4 using conservative, purposefully conservative assumptions?
- 5. DR. ERICSON: No. We did not say now let's take the 6 top number. What I am saying is in our analysis we have used 7 the generic data which has some conservatism in it I will 8 argue, and we have presented it as a point estimate using 9 generic data; no more or no less, 10 DR. SIESS: What is the difference between that and 11 what EPRI did now?
12 DR. ERICSON: Their argument is best estimate, I
(} 13 interpret to mean they have looked at all the data they can 14 get their hands on and then selected or arrived--
15 DR. SIESS: You don't mean best estimate the way you 16 say point estimate?
17 DR. ERICSON: No, because in the point estimates we i
18 did not dig into all potential possible data bases. We did 19 not get deeply into the human, into the EOPs as they did.
20 DR. SIESS: Maybe I don't know what best estimate 21 means. I thought best estimate was the statistical term that 22 gave you the most likely value out of a universe of values.
23 DR. ERICSON: That would be a definition.
24 DR. SIESS: Is that what they mean by best sometime?
! 25 DR. ERICSON: I assume so.
I HERITAGE REPORTING CORPORATION -- (202)628-4888
1133 1 DR. SIESS: In other words, value like a mean or
() 2 median or something.
3 CHAIRMAN WARD: Some attempt.at a central estimate.
4 DR. ERICSON: Some sort of-central estimate.
5 CHAIRMAN WARD: My impression is that is what most 6 PRAs are intended to try to do.
7 DR. ERICSON: We treated them as a means whenever we 8 could. l l
l 9 DR. SIESS: And your best, your point estimate l l
10 quantitatively, how would you describe.it?
11 DR. ERICSON: We used mean values from the generic 12 data bases.
13 DR. SIESS: Incidentally--
14 CHAIRMAN WARD: So why is there a difference? I 15 mean I don't understand.
16 DR. ERICSON: Because they haven't used mean values 17 from published generic data bases. They have reworked those 18 values and massaged them to get what they consider as the best 19 estimate for this particular plant.
20 HR. MINNERS: Didn't we use median values for the 21 seismic part?
22 DR. ERICSON: Yes. There is a', the problem there is 23 it is off by about a factor of three as a result.
24 HR. MINNERS: Mean value is accepted.
25 DR. ERICSON: That's right. That's a good point.
HERITAGE REPORTING CORPORATION -- (202)628-4888 I
134 1 We owe you a piece of paper that is somewhere in the--you will
( 2 get in.
3 CHAIRMAN WARD: The difference isn't so much that l'
4 their analysis characteristically used, you know, central ~
5 best, median, mean, values, and yours didn't, but rather they 6 used data bases or.whatever that were specific for this 7 particular plant and you didn't in some cases?
! 8 DR. ERICSON: Well, in some cases they have chosen l
! 9 to use other data, to change the generic data, and the small l
l 10 break frequency is an example. The values they use are not 11 Point Beach specific. They were derived for Oconee, and 12 that's the number of the Oconee report. They adopted that
(} 13 argument and so used that number, so--
14 DR. SIESS: Mr. Chairman, I think it is very I 15 important that we get a copylof that slide because it is the l
l 16 only slide he has used that has the word "uncertainty" on it.
17 I just didn't believe anybody could get through any discussion 1 \
l 18 of PRA FOR that long without the word uncertainty.
I 19 MR. MINNERS: You almost made it! l l
l 20 DR. ERICSON: Schucks! I should have left this one '
l 21 off and I would have set a record!
l l 22 And the other thing that I think needs to be I
l 23 commented about the report is that this report deals with more 24
(^)
\J than just the Point Beach study. Significant comments are l
25 made about what is presumed to be the NRC decision on this l
HERITAGE REPORTING CORPORATION -- (202)628-4888
r 135 1 matter and not the Point. Beach specific report.
I 2 As Mr. Mazetis has said, we will be in an iterative, 3 not necessarily iterative process, but. interactive process .
4 with the EPRI owr s group people to understand the source-of 5 numbers, and argue and fuss about why we use different 6 numbers.
7 MR. EBERSOLE: Can you talk a little bit about'a 8 matter that is somewhat in detail? Do you recall the GDC-19 9 required early on the establishment of points of control from 10 the plant at places outside the containment. That resulted in 11 a rash of extension cords stuck out all over the place from 12 the main control room into these discrete other areas, thus
{} 13 extending the fire vulnerability.
14 It is recognized now that those are regressive 15 circuits and we ought to cut them all out because now we-can 16 take care of damage in the control room rathor than just a 17 stink in it.
18 Did you look at vulnerabilities from fire, including 19 those spread-out circuits all over the place?
l 20 DR. ERICSON: We looked at the plant with Appendix R i
21 in place. We used Appendix R submittals. j 22 MR. EBERS0LE: I understand that, but did you look 23 at increment in risk presented by those old circuits that ara )
gS 24 still there? i V
25 DR. ERICSON: Specifically, I don't know. I can't i
HERITAGE REPORTING CORPORATION -- (202)628-4888
, --. -. .- . - , . . . . . -, ~ . . . .
- -,=2-136 1 answer that question.
) 2 MR. EBERSOLE: They are still there.
F 3 CHAIRMAN WARD: Any other questions'for Mr. Ericson?
4 Okay. I think as a Subcommittee, we need to discuss what we 5 think about what we have heard, where we want to go 6 particularly. We haven't heard directly on the issue which we 7 can spend some time on this, that the Point Beach analysis, we
)
8 haven't heard directly from the EPRI group or whomever it is 9 now that is responsible for that and_we might want to do that 10 in the future, but I think we need~to put that in the context 11 of what we are trying to do.
12 So before we go into that discussion, though, and I
(} 13 am going to suggest that we have that after lunch, I-would 14 like to hear from Mr. Minners a little bit more specifically, 15 you know, what is your schedule? When are you going to do 16 what, toward, you know, final resolution of this? What do you 17 need to do, and what would you like to have from the 18 Committee, from the ACRS? What do you see that you need, if 19 anything, from the Committee. Isn't necessarily what you are 20 going to get, but just like to know what you think you would 21 like to get. Could you tell us that?
22 DR. MAZETIS: Let me start and just--Jerry l
23 Mazetis--as far as schedule in the near future, I may have i
24 indicated in recent conversation with NUMARC, Jerry Nyles l 4 s / I 25 yesterday, led to our tentative initial dialogue with, the i
HERITAGE REPORTING CORPORATION -- (202)628-4888
137 1 interface with them perhaps around the first or second week in
(~% .
\-)
2 March, and before then, of course, we would continue amongst 3 ourselves in-house interfa6ing with completion of the review, ,
4 and preparing ourselves for that meeting.
5 I would anticipate concurrent with that wo'uld be the 6 availability as I indicated earlier of the draft or the 7 updated regulatory analysis distribution, at the moment 8 unknown, but it would be my guess available for the ACRS 9 within that timeframe, the next month or two.
10 CHAIRMAN WARD: Okay.
11 DR. MAZETIS: Let's see. You had another question?
12 Your other question was?
{} 13 MR. HIriNERS: What do we want out of the Committee?
14 I think the usual word is that you know, we have brought this 15 down to you. We value any comments that you had. I don't 16 know whether you want to give it your blessing, but we are 17 going to present a proposed resolution and traditionally the 18 ACRS has commented on it, and sometimes said yea or nay, so I 19 guess it is more up to the Committee exactly what you want to 20 provide.
21 CHAIRMAN WARD: I realize that, but I just wanted to 22 know what your perspective was.when you say the draft 23 analysis, regulatory analysis, that's going to present your 24 proposed resolution, and right now you think that is going to 25 be item number whatever it was.
HERITAGE REPORTING CORPORATION -- (202)628-4888
138 1 MR. MINNERS: Two.
() 2 CHAIRMAN WARD: And then present the arguments for 3 that; that's what you mean by the regulatory analysis?
4 MR. MINNERS: Well, the purpose of the regulatory 5 analysis is not to present the arguments just for alternative 6 2. The purpose is to discuss all six of the alternatives, 7 okay, so that people can see what basis the decision was made 8 on.
9 CHAIRMAN WARD: Okay.
10 MR. MINNERS: I think--
11 CHAIRMAN WARD: Okay. That's all?
12 MR. MICHELSON: Warren?
(} 13 DR. SIESS: Warren, they have got to do some kind of 14 PRA, right?
15 MR. MINNERS: To do what?
16 DR. SIESS: The individual plants.
17 MR. MINNERS: Yes.
18 DR. SIESS: How is the staff going to handle this?
19 I don't think there has ever been an instance of a licensee ,
20 making a PRA that the staff didn't think the probability was 21 one to two orders of magnitude higher.
i 22 Is the staff going to makeup, do a PRA to go along 23 with every one the license does, accept the licensee's value, 24 or are you going to apply a factor of say of ten on it..
25 Conservatism or at empirically ten to 15 seems, ten to 50 l
HERITAGE REPORTING CORPORATION -- (202)628-4888 I
139 1 seems about right.
() 2_ MR. MINNERS: I don't know the answer to your ,
3 question. It is perfectly possible that the staff might not 4 even review the PRAs'done by the licensees. That's a 5 possibility.
6 DR. SIESS: You really think that's a possibility?
7 MR. MINNERS: Anything is possible. .Now Vhat is 8 probable is something else again, and I don't think we know 9 the answer to your question because it takes a-lot of staff 10 resources. And I guess that's going to be, I mean this is ,
il something that is being worked out basically in the IPE 12 process. If we look on to those guys, we come out and it is 13 their worry, but we are, you know, it is something that has
[}
14 been being discussed as we say.
I 15 One of the purposes of reviewing this 16 EPRI-Westinghouse owners group analysis was because we think 17 it is kind of--I don't know how to say it--one of the 18 precursors to an IPE submittal, okay, and maybe some precedent 19 will be set or at least indicated by our evaluation of their 20 analysis, but I would point out to you is that yes, you do, ;
21 you get a factor of ten lower, but here is a particular 22 incident that if, a reason you have got some of the difference 23 is because they installed some new equipment, and that's fine 24 by me. j 25 LA. SIESS: I saw the Susquehanna IPE evaluation HERITAGE REPORTING CORPORATION -- (202)628-4888
140- q 1 with their ten to the_minus 6, ten to'the minus 7, and they m
. 2 say it is, all comes not~from installing any new equipment but 3 from taking advantage of accident management procedures in 4 _ site training, based on their PRA.
5 MR. MINNERS: Okay.
6 DR. SIESS: I remember when you had the six choices, 7 one of the objections to No. 2 was opens everything to 8 argument, and No. 3 or 4, whichever it was, prescribed fixes, 9 at least was clear cut, somewhat reminiscent of ATWS. ATWS 10 ended up being prescribed figures, right? Were negotiated-or 11 developed let's say in cooperation with the venders.
12 MR. MINNERS: We are going the opposite way here.
() 13 ATWS, that's right. We aaid we weren't going to have 14 individual evaluation models. We will paint'everybody with 15 one brush. And we are going, seemingly going the other way.
16 I don't know what to do, Dr. Seiss. I think it is 17 recognized that when you get into PRA, you have got this 18 problem. You have got many knobs and switches that you can 19 turn on these analyses, and people have different views of 20 them. And you can--and take different views. I guess that's a
21 a problem.
22 The, I guess what you have to look at is your 23 alternatives. Your alternative is a very expensive I
r~s 24 prescriptive system, or a less expensive prescriptive system
'd 25 which I'm not sure that no matter how you train the operator, l
HERITAGE REPORTING CORPORATION -- (202)628-4838
141 1 he will ever turn it on.
( 2 DR. SIESS: Well, even on that ATWS you had to have 3 different systems for different plants.
4 MR. MINNERS: For feed and bleed you are talking 5 about?
6 DR. SIESS: Hardware fixes for ATWS, for each 7 vender, the hardware fixes for DHR are likely to be different 8 for every darned plant.
9 MR. MINNERS: That is correct.
10 DR. SIESS: Going to be more variations that affect 11 the SCRAM system.
12 MR. MINNERS: Well, you know, we work with, for the
(} 13 Commission, and I think the Commission set out a severe 14 accident policy, and I think we should be guided by that, and 15 their decision was to do it in a PRA way, and I guess we are 16 going to work out the details of how that gets reviewed, and I 17 think that A-45 is more of a severe accident issue than 18 anything else. I mean kind of falls into that category 19 naturally, so I think, you know, I think there is, to my view, 20 some Commission guidance that's the way they would like to go.
21 Now whether policy statements and that kind of stuff 22 is the proper thing to use for severe accidents is, that was 23 their decision. I think that's what your question is. I 24 think you are more questioning the IPE process than you are 25 our A-45 because that's going to cover more accidents than we HERITAGE REPORTING CORPORATION -- (202)628-4888
. 142 l
1 have covered. If we went out with a prescriptive solution and I (N
(,,) 2 fixed decay heat removal, there could still be ATWS, or .i 3 interfacing LOCAs, or any other, or other sequences that could 4 dominate and all our fixes would be for naught because it 5 would be overridden by these other possible more dominant 6 vulnerabilities at particular plants.
7 MR. DAVIS: Mr. Chairman, I had a question. I think 8 I agree with.the alternative you have chosen. It seems to me 9 that your studies have shown that there are significant plant 10 differences, and therefore, we need to look at each one 11 individually.
12 But I guess I am still troubled by how we decide if
(} 13 we have got a problem or not. Both your study and the EPRI 14 study indicated that there were no fixes justified on a 15 cost / benefit basis I thought for Point Beach, and yet you seem 16 to, you seem to have a high core melt frequency that you now 17 think should demand some sort of effect. Am I missing 18 something here?
19 MR. MINNERS: Depends what method. I don't agree 20 with you that there is, that on the cost / benefit method they 21 are not justified. I mean I don't know what you, I mean we 22 presented three and there is probably any number of 23 combination you want to use. My view is that the, you should s 24 include on-site costs, and that's in my view, a kind of a 4
25 negative impact. You subtract that off of the limitation cost HERITAGE REPORTING CORPORATION -- (202)628-4888
. _ _ _ ~ _ _ _ _ _ _ _ _ - _ _ _ _ _ - - - _
8 143 1 and then do_your ratio bit, and in most cases that comes-out n)(_ 2 it is worthwhile.
3 Now I don't think you should be regulating on the 4 basis that you are going to reduce on-site costs. That should 5 not be the impetus for the regulation. The impetus should be 6 you are going to have off-site safety improvement, but when 7 you consider the balance and what you put on each side of the 8 billion as you have to put on to the cust, and the affected 9 on-site costs are real costs, somebody is going to pay those 10 costs. Okay. And they have happened and somebody is now 11 paying for them. Okay. So I think they ought to be put in 12 the balance.
{} 13 MR. EBERSOLE: Don't you think it is important who
! 14 pays those costs?
15 MR. MINNERS: No, I don't care who pays those costs.
16 MR. EBERSOLE: Let me tell you why I bring this up.
17 It is, if it is a corporate cost, there are forces that will 18 come in the direction of saving corporate costs. If they are 19 public costs, that's different. I am putting incentives to 20 fix the incentive, the pocketbook as ever, and the public has 21 no way of reaching in and influencing that incentive, as you 22 know, except the intervenine, process.
23 MR. MINNERS: I don't agree with that. I think PUCs 24 have very strong inf'.uence on--
< O 25 MR. EBERSOLE: Get down into this?
HERITAGE REPCRTING CORPORATION -- (202)628-4888
144
- 1. MR. MINNERS: They sure have.
O lss/ 2 DR. SIESS: If we assume that the off-site costs are 3 an industry concern, a business-type concern, is it fair to 4 make the industry consider the on-site costs with your PRA 5 results, which:is two orders of magnitude higher than their 6 PRA? They look and say it is on. ten to the minus 6 7 probability that it is going to cost me a billion dollars, and 8 you say that is a ten to the minus 4 probability that it is 9 something happened to the public. Make them use the ten to 10 the minus 4, with their billion dollars.
11 MR. MINNERS: I don't think there is any such thing 12 as industry cost, Dr. Seiss. The industry has no way of
{} 13 paying for anything. Only way they get their money is out of 14 people. I mean we are the source of all of their funds, and 15 so eventually, the country pays for those costs. t 16 HR. EBERSOLE: There is a--
l 17 DR. SEISS: That is not NRC s concern.
18 MR. MINNERS: I think it is NRC's concern in making 19 the balance. l t
20 HR. EBERSOLE: The corporate stockholders can fire 21 the president, chairman of the Board, everybody else, if they 22 don't have a good design. The public can't do that. And I 23 think there is a difference in the incentive. If the cost is 24 seen as corporate costs, management will rise up and pay
,)
25 attention to whatever the plant is designed.
HERITAGE REPORTING CORPORATION -- (202)628-4888
145 1 MR. MINNERS: J don't think I have any regulatory
( 2 authority to say who will pay the cost, the Public Utility 3 Commissions will decide in their prudence hearings, prudence 4 hearings who pays what costs. I have no influence on that.
5 MR. EBERSOLE: You believe that can be shunted off 6 to the PUCs?
7 MR. MINNERS: Believe it can be? It is. I don't 8 read anything in the Atomic Energy Act that let's me regulate 9 their costs.
10 MR. EBERSOLE: I am not impressed by PUCs as I dig 11 down this far for details.
12 DR. SIESS: If you were advising the PUC to look at
{} 13 this sort of thing on a cost / benefit basis, what PRA would you ,
14 advise them to use, the licensee's, or the NRC's or the quote, 15 correct one?
16 MR. MINNERS: I would advise them to use the best 17 PRA they can get.
18 DR. SIESS: But in other words, should they--let's 19 face it. The NRC PRA or Sandia or whoever is always a higher 20 probability than the one that comes from the licensee. I 21 don't think I have ever seen a case where they weren't in that 22 usually by a factor of one to two orders of magnitude. There l
23 are conservatisms in the regulatory PRAs. Should the PUC be !
24 using a conservative--
25 MR. MINNERS: No.
HERITAGE REPORTING CORPORATION -- (202)628-4888
146-1 DR. SIESS: Value even by-one order of' magnitude or
~( ) 2 be using the most rea'listic thing they can get?
3 MR. EBERSOLE: Generating their own.
4 MR. MINNERS: For. economic-calculations, it is clear 1
5 you should be_using the'mean value-for making any of your -l
.. i 6 economic calculations, so they should try to give a PRA that 7 gives them the mean core melt frequency.
l 8 DR. SIESS: You wouldn't offer _ opinion as to which
)
9 of the two we have looked at is the closest to that?
10 HR. MINNERS: No. I think that this is an honest j 11 effort, and I appreciate it, and industry has their new and -
12 they tend to take a more optimistic view. Fine.- I think
- 13 that's good because I think we need some balance in lookiing
[}
i 14 at that. I think some elements of ours are conservative, but
.?
15 then again, there is always the question that we have missed ;
16 things in PRAs. It is a very--I tried in the beginning to get 17 people to sit down and write all the assumptions that went 18 into these things, and then from that, make a qualitative
! 19 assessment on the conservatism or non-conservatism of the 20 PRAs, and we were unable to do that. It is too difficult a i 21 task to do I think. It is sounds good ideally, but I think f q 22 it, as a practical matter, it is not good. But nevertheless, 23 I think decision-makers are going to have to do that. It is '
24 going to have to come to some judgment of who has got_the "O
1 25 appropriate PRA. Some of the stuff can be resolved on i
HERITAGE REPORTING CORPORATION -- (202)628-4888 i
147 1 -technical bases I am sure, go out, get a consensus this is the l
) 2 right number, but I think, on human error rates, what do ycu 3 pick? That's anybody's guess. !
\
4 DR. SIESS: You think they are both biased?
5 MR. MINNERS: Yes. I do think both, do' agree with 6 that, and don't imply from that that I think there is a gross 7 bias, okay?- I think that when Sandia came to a point of 8 decision they would tend to go to the conservative. Okay.
9 When they had to make a decision I think when--and disagree 10 with me. I expressed my opinion. I think when industry came 11 to that same decision, they go more the optimistic route.
12 DR. ERICSON: I agree with you, Warren. The answer 13 is that every study that is done is biased, without putting
[}
14 bias in a derogatory or negative fashion. We all are biased 15 by our own experience base, and by our own perception of the 16 problem. That doesn't mean that it is bad or negative or l 17 wrong, but we are biased.
18 MR. MINNERS: Well, see I think maybe the point that i u
19 ought to be made here, I guess it is said all the time with l 20 PRAs, don't bottom line it. This is a very interesting set of 21 numbers, okay, but don't get mesmerized by the bottom line 22 numbers, okay. Look more ir.to what the PRA tells you and then !
23 start to use your engineering experience and say hey, is that
, 24 a reasonable thing?
]
25 CHAIRMAN WARD: I think the bias that we are l 4
HERITAGE REPORTING CORPORATION -- (202)628-4888 l
148 1 concerned about is not the bias resulting from your personal l
() 2 experience, but the bias which comes from being interested in i
3 the result, having some particular interest in the result.
4 That's the bias you are concerned-about, we have to deal with.
5 DR. SIESS: It is conceivable--
6 MR. MINNERS: They both produce bias.
7 DR. SIESS: It is conceivable that the normal 8 regulatory bias, you know, we know it is there. I think we 9 always think a little bit on the conservative side, that the 10 regulatory bias itself could account for a 54 difference.
11 Maybe it only accounts for about 30 and the other 20 is some 12 optimistic bias on the other side.
13 MR. MINDERS: Factor of 50 is well within 1150s
[}
14 feather diagram.
15 DR. SIESS: Everything is within that 16 HR. WARD: Everything is, yes, 17 DR. SEISS: Zero and one, you know, we have about 18 covered that 1150.
19 CHAIRMAN WARD: Okay. Well, thank you very much, 20 gentlemen. Before we--let's see. We will take a break for 21 lunch, but I would like the Subcommittee to come back and we 22 will be in Executive Session, which means we don't need a 23 record, at 1:45. I want to spend about an hour or so in a 24 more general discussion. i 25 (Whereupon, at 12:40, the recorded portion of the 4 ,
HERITAGE REPORTING CORPORATION -- (202)628-4888
. ~ _ _ . _ _ _ _ . . _ _ _ _ . _ _ _ _ _ _ _ . _ . _ _ _ - . . . _ _ _ _
149 1 i f l' meeting was adjourned.)
t 4
I )
i b i
3 :
i 5 5 1 i 4 j 1 .l 5 (
6 i
- i i
+
7 i I
, 8 ;
i 9
i i 10 i i !
) I
, 11 '
l 12
. i
[ l 13
)
i 14 <
) l 1
i 15 I
4 I
16 J
/ 17
' l 1
18
(
l f 19 l l
{
20 i
21 l l l 22 l
l l 23 i
r i
l 25 HERITAGE. REPORTING CORPORATION -- (202)628-4888 f
1
1 1 REPORTER'S CERTIFICATE 2 ,1 s 1
,) 3 COCKET NUMBER:
-- 4f CASE TITLE: ACRS--Subcommittee on Decay Heat Removal 5 HEARING DATE: January 28, 1988 i I.
', Washington, D.C.
6 LOCATICN: l' 6
7 4l
, I hereby certify that the proceedings and evidence 8l '
are contained fully and accurately on the tapes and notes I 9
s reported by me at the hearing in the above case before the !
10 !
Nuclear Regulatory Commission.
11 : I 12
~\ 13 i Date: January 28, 1988 g
kl 14 I
1 15l, l
16
[.
17 Official Reporter 18 A E REPORTING CORPORATION 1220 L Street, N.W.
19 ashington, D.C. 20005 20 l
l 21 cs 22 -
h1 23 24
,O
\s'i I 25 HERITAGE REPORTING CORPORATION '
(202)628-4888
m -
(.
- u. -
i LO ..
i 1 .
1 I
1 i
I STATUS REPORT ON USI A-45
\
i "SHUECWN DECAY HEAT REMNAL REQUIRENNTS"~ -l
\
l
. l l
i
.l l
1 PRESENTED TO THE 9
ACRS SUBCOBNITTEE FOR DECAY HEAT REMOVAL JAN'JARY 28,'1987 0 .
i 1
ROY WOODS, SENIOR TASK IMNAGER l DIVISION OF PEACTOR AND PLANT SYSTEFE OFFICE OF NUCLEAR REGULATORY RESEARCH O
4 O
BACKGROU_N_D_ ,
i CONCERNS ON RELIABILITY OF DHR FUNCTION:
t PAISED BY TNI
-I
- EXTENDED TO SPECIAL EERGENCIES (E.G., FIRE, FLOOD, SEISMIC, '
SAB0TAGE) -
e f ACRS, TASK FORCES, AND SPECIAL C0ftllSSIONS REC 0ftENDED THAT HIGH l
PRIORITY BE GIVEN TO STUDIES OF IffR0VING SlUIDOWN DHR FUNCTION' :
O (POST-TMI EFFORTS)
UNRESOLVED SAFETY ISSUE (USI A-45) APPROVED DEC. 24, 1980 (SECY-00-325) !
KEYQUESTIONS:
I D0 CURRENT DESIGNS PROVIDE THE RELIABILITY NEEDED TO T ET CDF G0AL?
ARE THERE IffROVEEN13 TO DHR FUNCTION IN OPERATING PLAffTS WHICH ARE COST-BENEFICIAL?
O 1 !
1
O USI A 145 SECIFIC CBJECTIVES
- DETERMINE THE SAFETY ADEQUACY OF DECAY HEAT REMOVAL IN EXISTING POWER PLANTS FOR AQilEVING BOTli HOT SHUTIX7nN AND COLD SHUTDOWN CONDITIONS ,
- DEVELOP AND EVALUATE ALTERNATIVE METliODS FOR IFPROVING RELIABILITY OF DECAY HEAT REMOVAL FUNCTION, O
- ASSESS TliE VALUE AND IMPACT OF ALTERNATIVE E1110DS
l l
I o
2
Ng
[
L)
USI A-45 SCOPE EVALUATED DHR SYSTEMS NEEDED TO RESPOND TO TRANSIENTS ABS SMALL-BREAK LOCAs (DID NOT INCLUDE LARGE BREAK LOCAs OR ATWS)
ALSOEVALUATEDSUCHSYSTEMS'VULNERABILITYTOFIRE, FLOOD, SEISMIC, INSIDER SABOTAGE t
SABOTAGE CONSIDERATIONS:
(] -
CONSIDERED VULNERABILITY OF DHR SYSTEMS TO "lNSIDER" SAB0TAGE
- DID NOT CONSIDER "PHYSICAL SECURITY" MEnl0DS ("OUTSIDER")
l l
l 1
( \
G' 3
1 O
A-45 STUDIES DHR SPECIFIC STUDIES OF SEVERAL PLANTS PLANTS SELECTION CRITERIA: .
REPRESENT EACH NSSS VENDOR
) -
COVER "RANGE" 0F DHR CONFIGURATIONS, AEs UTILIT/ INTEREST IN STUDIES O
PRA CASE STUDIES FOR DECAY HEAT REMOVAL (ALL COMPLETED)
PLANT TYPE REPORT NO, D_ ATE (NUREG/CR-)
POINT BEACH W2LP 4458 3/87 TURKEY POINT W3LP 4762 3/87 ST. LUCIE CE 4710 8/87 ANO-1 BaW 4713 3/87 QUAD CITIES BhB3 4448 3/87 COOPER BWR4 4767 7/87 (Stfi%RY) (ALL) -
(EARLY '88)
O 4
1 I
T l'
1 o
i SLWARY OF FINDINGS _FROM SIX CASE STUDIES PROBABILITY OF CORE ELT DUE TO DHR FUNCTION FAILURE [P(CM)DHRI AVERAGES 2 TO 3 x 10-4 PER R-YR (INCLUDES INTERNAL AND EXTEFNAL- ,
CAUSES) l SUPPORT SYSTEM FAILURES (E.G., EERGENCY POWER, SERVICE WATER, !
CGPONENT COOLING) CONTRIBlITE SIGNIFICANTLY TO P(CM)DHR LACK 0FREDUNDANCYANDSHARING0FSYSTEMS,PARTICULARLYATSUPPORT SYSTEM LEVEL, CONTRIBUTE SIGNIFICANT RISK FOR SO E PLANTS ,
LACK OF INDEPENDENCE, SEPARATION AND PHYSICAL PROTECTION OF REDUNDANT SAFEGUARD TRAINS CONTRIBlJTE SIGNIFICANT RISK i
RISK FROM FIRE, FLOOD, SEISMIC, SAB0TAGE IS SIGNIFICANT BOTT m LINE IS TllAT RELATIVE IlPORTANCE OF VULNERABILITIES IS PLANT-SPECIFIC O
5
i O REGULATORY ANALYSIS
- SEPTEMBER 26, 1986 PRE-DECISIONAL DRAFT PRESENTED TO CLOSED SESSION 0F ACRS SUBC00NITTEE DESCRIBING SIX ALTERNATIVES
- DRAFT REVISION BEING PREPARED, WILL DESCRIBE SAE SIX ALTERNATIVES.
- lE WILL PROPOSE RESOLUTION WITH ALT. # 2 (PLANT-SKCIFIC ANALYSIS)
ALTERNATIVE 1 - NO ACTION WOULD BE ACCEPTABLE JE NRC ANALYSIS RESULTS ARE ;
OVERLY CONSERVATIVE (EPR150G: POINT BEA00 l
4 O ALTERNATIVE 2 - LIMITED SC0K PRAs I
SEVERE ACCIDENT PROGRN'i IEs ALTERNATIVE 3 - S KCIFIED SYSTEMS I'0DIFICATIONS !
USIs AND Gls 4
g ALTEPNATIVE 4 - DEPRESSURIZATION AND COOLING l l
PWR - FEED AND BLEED BWR - CONTAlffENT VENTING ALTEPNATIVE 5 - DEDICATED HOT SHUTDOWN CAPABILITY ALTEPNATIVE 6 - DEDICATED COLD SHUTDOWN CAPABILITY
> O
)1 6
\
'I O
l REG. ANALYSIS (Cottr'D) /
DIFFERENT APPROACHES TO VALUE-lMPACT ANALYSIS U VALUE-IMPACT ANALYSIS PERFORWD 3 WAYS:-
A. AVERTED OFFSITE COSTS ONLY ,
B. AVERTED OFFSITE PLUS ONSITE COSTS ,
C. AB0VE PLUS EFFECTS OF SPECIAL CONSIDERATIONS O (E.G., SABOTAGE, MORATORIUM, RESOLUTION OF OTHER GENERIC ISSUES, UNCUANTIFIABLES)
RESULTS:
ETHOD A - ALTERNATIVES 2, 3 & 4 PAY BE COST-EFFECTIVE E THOD B - ALTERNATIVES 2, 3 & 4 MAY BE FORE COST-EFFECTIVE I
METHOD C - ALTERNATIVES 5 & 6 MAY BE COST-EFFECTIVE O
7
?O REG ANALYSIS'- (Corn'D)'
l
- RES ENDORSES ALTERNATIVE 2, PLANT-SPECIFIC ANALYSES AND.IMPROV9ENT OF DECAY HEAT R90 VAL SYS191S, BASIS: ,
A-45 CASE STUDIES'SHOWED MOST RISK CONTRIBlITORS TO BE ,
PLANT-SPECIFIC.
O USE OF "ETHOD C"-(CREDIT FOR "f0RATORIUM AVOIDANCE") G0ES BEYOND VALUE/IfPACT ETHODS PREVIOUSLY USED FOR USIs/GSIs i AND THEREFORE ALTERNATIVES FIVE AND SIX CAN NOT BE JUSTIFIED ON THE BASIS OF CONVENTIONAL VALUE/IffACT-INSIGHTS GAINED FROM SIX CASE STUDIES AND'EPRI-WOG ANALYSIS (PLL'S NRC/SANDIA REVIEW) WILL BECOE GUIDANCE TO LICENSEES.
1 cO 8
-t **J,'y
- - -,- ry-.--%-- r*,e v -v. -+-vv v=.y*.eu , --- =,rw+ -me.,-- ---..- w - < -.,.4. , ,-%v.w-e.--.-- r++. * - --N - W 3. m-r**vw
l l
l O '
IffLEENTATION DETAILS ;
(ALTERNATIVE 2)
- SHOULD BE PART OF THE IPE? ,
- GREATER EFFICIENCY, BUT
- WOULD REQUIRE MORE lliOROUGH PRA
- WOULD OBLIGATE LARGER NRR REVIEW EFFORT FOR-IPE ANALYSES O
- SHOULD BE SEPARATELY REQUIRED?
- SIGNIFICANT REDUNDANCY WITH'IPE, WOULD REQUIRE EACH LICENSEE TO PERFORM TWO SEPARATE ANALYSES
- COULD PERFORM MORE THOROUGH PRA FOCUSED ON DHR ONLY
- C0FNITENT TO REVIEW ANALYSES
- LARGE EFFORT NEEDED, NOT f1ECESSARILY AVAILABLE
- SAFETY IMPROVEENT DEPENDENT UPON REVIEW C0FNITENT O
9
-- . _ . . . - _ _ . t _ _ _ . _ _ _
l 4 l
l EPRI/WOG POINT BEACH PRA l
- NRC REVIEWING FOR 2 REASONS:
l
- 1) GUIDANCE FOR UTILITIES PERFORMING A-45 PRAs i
- WHAT ElliODS ACCEPTABLE
- WHAT NUTRICAL ASSUMPTIONS ACCEPTABLE
- 2) IS ALTERNATIVE 1 JUSTIFIABLE?
O JE THEIR PRA CORRECT, AND
- JE PB IS A "BOUf0 LNG" PLANT (THEN ALT. 1 WOULD BE ACCEPTABLE) i WE ARE NOT CONVINCED THAT EITHER OF TliE ABOVE ARE CORRECT
- THUS THE REVIEW RESULTS URGBITLY NEEDED
- WILL BE DISCUSSED BY NRC AND SANDIA O
10.
.r,9 y, ..
w... _ , , _ . , ..%- ,,..c,- - ~ r ---H* " "
O COWARISON OF NRC/SNL AND EF.d/WOG RESULTS CORE IELT FREQUENCY PER YEAR REDUCTION-SOURCE OF RISK NRC/SNL EPRI/WOG FACTOR INTERNAL 1,4E-4 T.,6E-6 54 SEISMIC 6,1E-5~ 7,4E-6 8 FIRE 3,2E-5 6,3E-8 500 0 INTERNAL FLOOD 7,7E-5 1,0E-8 7700 EXTERNAL FLOOD 1,9E-8 1,0E-8 ( 2)
WIND 4.0E-6 1.CE-8 ( 400)
LIGHTNING 5,8E-8 1.0E-8_ ( 6)
TOTAL 3,1E-4 1,0E 31 0
11
[
L 1
0 TA3 A-45 SHJTDOWN DECAY HEAT REMOVAL ANALYSIS COPAR S0s 07 Ss _/sRC As) E3R /WOG R ESU _~S TO R 301s" B EAC-ERU A33A S. C- E__ A- l O
s RC/RES/)RAA/3RAB ACRS SUBCOMM ~ TEE Os JECAY -EA~ REMOVAL JAsUARY 28, ' 988 l l
O
, ,. _ _ , , . _ , - . - _ _ - , , _ . . . . - _ - ~ , , . - - - - , . - - - , ,
0 i
O POINT BEACH 2HR/3RA SEQUENCE FREQUENCY /RY'
SUMMARY
DIFFERENCE l
- REVISED SMALL LOCA FREQUENCY l
(
5.8E-7 EPRI: e MODIFIED CCW-SUCESS CRITERIA i e CREDIT FOR SW BALANCING T1MLE 6.7E-6 'SNL
- UPDATED INITIATOR FREQUENCY i
7.7E-7 EPRI:
- NO NEED FOR RECIRCULATION T30H1'H2' 'SNL NA [EPRl]
T30D1D2 4.6E-6 [SNL'
- N0 NEED FOR RECIRCULATION NA EPRI:
T2MQH1'H2' 3.5E-6 SNL'
- REVISED PORV PROBABluTY 1.9E-7 EPRI' S2MDID2 8.7E-6 SNL' e REVISED SMALL LOCA FREQUENCY 9.5E-8 [EPRI
- MODIFIED CCW SUCESS CRITERIA O
l POINT BEACH DHR/RA l
ll l'
l 11 SEQUENCE FREQUENCY
/RY:
SUMMARY
DIFFERENCE ll ll l
- NEW METHOD FOR FLOOD FREQUENCY l 1.0E-8 [EPRl]
- UPDATED HPIS PERFORMANCE SEISMIC 6.1E-5 'SNL]
- ADDED CLASS 1E DC SYSTEM l O . USED MODiflED HAZARD CURVES l
FIRE 3.2E-5 [SNL]
- CREDIT FOR SECOND~ TRAIN l i
EVENTS 6.3E-8 [EPRl] HALON SYSTEM
- REVISED HALON SYSTEM FAILURE !
PROBA81UTY l
e CREDIT FOR RECOVERY OF AFWS l FOR SWITCHGEAR ROOM FIRES i i
i i
O I
i I
_-_._a----_-.-.
. u i
1 O l POINT BEACH DHl/3RA '
SEQUENCE DESCRIPTION I
S2MH1'H2' - SMALL LOCA FOLLOWED BY THE FAILURE OF LOW AND HIGH PRESSURE RECIRCULATION SYSTEMS TIMLE - A LOSS OF 0FFSITE POWER EVENT FOLLOWED BY THE COMMON WODE FAILURE OF THE DIESELS, RESULTING IN EARLY CORE WELT O T30H1'H2' - A TRANSIENT INVOLVING A STUCK OPEN PORY AND FAILURE OF RECIRCULATION SYSTEMS. THE MFW SYSTEM IS ASSUMED TO BE INITIALLY AVAILABLE.
T30DID2 - A TRANSIENT INVOLVING A STUCK OPEN PORY AND FAILURE OF BOTH HIGH AND LOW PRESSURE INJECTION SYSTEMS. THE WFW IS ASSUMED l TO BE INITIALLY AVAILADLE.
T2M0H1'H2' - A LOSS OF FEEDWATER EVENT FOLLOWED BY A STUCK OPEN PORY AND FAILURE OF RECIRCULATION SYSTEM.
S2MDID2 - A SMALL LOCA EVENT FOLLOWED BY THE FAILURE OF BOTH HIGH AND LOW PRESSURE INJECTION SYSTEMS. !
O a
o 30lNT BEACH 2HR/PRA SW FLOOD - A FLOODING EVENT IN SERVICE WATER PUMP ROOM CAUSING IN0PERABILITY OF ALL SW PUMPS, RESULTING IN LOSS OF CORE COOLING SYSTEMS.
THE FLOOD IS ASSUMED TO RESULT IN A TRANSIENT.
SEISMIC EVENTS - SEISMIC EVENTS OF LOW TO MODERATE INTENSITY RESULTING IN:
(A) LOSS OF FEEDWATER SEQUENCES INVOLVING O INJECTION FAILURES, (B) SMALL LOCA SEQUENCES INVOLVING INJECTION FAILURES, AND ,
(C) TRANSIENTS INVOLVING THE FAILURE OF BOTH l PRIMARY AND SECONDARY COOLING SYSTEMS.
FIRES - TRANSIENT COMBUSTIBLE FIRES WERE CONSIDERED.
(A) A FIRE IN AFW PUMP ROOM FOLLOWED BY PRIMARY COOLING SYSTEM FAILURES.
(B) A FIRE IN 4.16 KV SWITCHGEAR ROOM FOLLOWED BY PRIMARY AND SECONDARY COOLING SYSTEM FAILURES.
O
TAP A-45 Shutdown Decay Heat Removal Analysis Comparison of SNL/NRC and EPRl/WOG Results for Point Beach ,
David M. Ericson, Jr ERCI incorporated -
Kenneth G. Adams Sandia National Laboratories ACRS Subcommittee on Decay Heat Removal January 28,1988
- O e
a . .
- USl A-45 (TAP A-45) Program Objectives 1
l Determine adequacy of DHR in existing NPP E valuate feasibility of methods for improving DHR, including diverse dedicated systems l
Assess the value and impact of potential l
2 alternatives i
1 2
i
Limits on Scope of PRA 4 initiating events considered did not include Large LOCA, reactor vessel ruptures, Interfacing system LOCAs, ATWS Only plant systems contributing to DHR were analyzed . ,
Although specific plants were identified for j analysis, the analysis was to provide j representative results for similar plants i
) -
9
EPRl/WOG Analysis Best estimate of Point Beach risk from industry perspective -
l Same analysis scope as USI A-45
- Primarily used RMOS l
l l
l
~
Risk Management Query System (RMOS)
Store principle components of a PRA initiators, Accident sequences, System cut sets, Basic event (component) probabilities, Risk management factors Ask questions of the PRA investigate effect of parameter changes initiator frequencies, recovery events, failure probabilities, new loops
i Internal Core Melt Frequency NRC 1.4E-4 WOG 2.6E-6 Reduction Factor 54 New Batteries .
Small LOCA freqQency
! CCW success criteria i Operator action
- Refill of CST Diesel generator failure rates i
i Handling of relief valve LOCA i
h g
External Core Melt Frequency NRC WOG R/F 6.1 E- 5 7.4E-6 8 Seismic Recovery, new batteries, seismic hazard curve 3.2E-5 6.3 E-8 500 Fire Recovery, new batteries, initiator frequencies
>7700 internal Flood 7.7E-5 <1.0E-8 Flood frequency, success criteria for cooling of HPl Wind 4.0E- 6 <1.OE-8 >> 40 0 Diesel generator exhaust supporta strengthened i
j t-_ - -- - _ - - - - - _ - - - e O
Results of Reanalysis Factor of 30 reduction in core melt l
frequency Additional factor of 7 in offsite consequences 50% to 400% increase in cost estimates 1
O 9 g
General Form of Accident Sequences Transient with failure of secondary heat removal and feed-and-bleed Transient with failure of secondary heat removal and successful feed-
-and-bleed in injection mode, but failure of ECCS recirculation 0fPR mode)
Smal LOCA with fature of ECCS recirculation Smal LOCA with failure of ECCS injection arid failure of either secondary heat removal or t.PI Transient-Induced LOCA and falure of ECCS recirculation Transient-induced LOCA and failure of ECCS injection (and failure of either secondary heat removal or LPI)
- O e
Reasons for Differences HPI does not require CCW 1
SBLOCA and transient-induced LOCA frequency l Low Point Beach-specific transient frequencies l
{ Miscellanecus operator recovery actions or human error in performing normal actions 4
I 1
i
- # O O 4
1 Areas for Discussion i
Timeliness of data (snapshot)
Plant characteristics (interties / interconnects)
EPRI/WOG included new batteries Human effectiveness (recovery) .
Times available for recovery Initiator frequencies Modification costs 1
i l
- e e
h General Comments Different plant based on new batteries and success criteria Current methodology does not provide a "correct" answer .
Difficult to establish effectiveness of human j
performance (particularly under new ERG)
Difficult to ascertain exactly what data, l
models and assumption were used
! O 9 g
-- - - - - - - - - - - - - - - - - - _ -- - _ _ - _ _ - --_ _- -_ a
Success Criteria for Support Systems i
i CCW provides seal rather than bearing cooling and is therefore not required for HPl operation in injection mode SW not required for cooling of CCW and
. recovery possible by operator when less .
than three of six pumps available (balance load) i i
i 4
- e .
l Changes in HRA Quantification Reduced failure to initiate feed and bleed 9.E-3 to 1.E-3 Reduced failure to initiate sump recirculation 3.E-3 to 1.E-4 Reduced failure to depressurize
! 1.5E-2 to 3.E-3 Change the human recovery time reliability l
l correlation, reduction by up to a factor j of 50 4
\ Additional Recovery actions l
i i
e
4 4
Criteria for Recovery i
~
USI A-45 or Two actions before two hours (RA-1
- RA-2 + another)
Three after two hours (two + RA-1 or RA-2) -
i WOG Number of options available l
i Time required to perform each action Presence in procedures and training l
(Emergency Response Guidelines or plant l
I specific procedures)
O 9 g
1
'I i
! USI A-45 Recovery RA-1 Loss of offsite power
- RA-2 Loss of main feedwater i RA-8 Battery common cause
! RA-9 Battery fault
! RA-10 Dieset common cause RA-11 Diesel fault RA-6 Other failures from control room l - RA-7 Other failures locally
! # e e
l 4
I EPRl/WOG Recovery i
I RWST refill using spent fuel pool or chemical and volume control system Cross-connect AC or DC buses Manual control of turbine-driven AFW pump i Provision of backup supply of feedwater
! Use of charging system for loss of feedwater
! Balancing loads on the service water system Recovery from common mode failures l
l I
l l
! O 9 g
! Small LOCA Frequency and Modeling 2.OE-2 to 3.0E-3 eliminate seal leaks Small small break LOCA 0.5 to 1.5 inches Significantly less injection for makeup
. AFW can delay core uncovery, more time .
for recovery actions Small enough for recovery by the alternate 3
- sources of water for the RWST
)
4 e , ,
0 Relief Valve /PORV LOCA '
1.4E- 3 to 1.1 E-4 RCMF 2.9E-5 i Survey of operational Westinghouse plants conducted in 1981 showed no failures in PORVs. Included 163 operational openings 1
i i e -
e e i
Common-Cause Failure Rates Reasons for Changes Design review for determining beta factors The design review process has yielded a factor of two lower in EPRrs application than the beta factors calculated using all l industry experience. (i.e., Factor of 3 on pump failure)
Recovery from common-cause failures l A review (AEOD report C504) indicates that roughly 60% of i both human and hardware failures were recovered within one l hour
! Used data from Millstone-3 PRA i
(
Common-Cause Failure Rates USI A-45 WOG Batteries 9.6 E-4 4.0E-4
- MDPs & TDP 2.0E-4 3.OE-5 2 running SWPs & 1 standby 2.OE-5 4.OE-6 l
i Running & standby-CCWPs 8.0E- 5 3.0E' S Other pumps 1.0E- 4 3.0E-5 1 MOVs 4.0E-4 8.OE-5 Diesel generators 1.5 E-3 5.0E-4 l
i l
- e e L -- - - -- -
Seismic New Seismic Category I batteries RWST tank failure recovery Seismic hazard curve conservative by factor of two for . low accelerations, and by a factor of five for high accelerations l
Frequency of small break LOCAs l
l AFW recoveries implemented from control room I
Failure of one CST will cause loss of inventory L* * .
Internal Flood (Service Water Pump Room)
USl A-45 study used generic room flood frequency of auxiliary building moderate flood
- 2.2E-2 WOG used frequency based on relationship of break to pipe -characteristics length of pipe pipe diameter pipe thickness I weld quality I dynamic loading factor
! 3.7 3 E-5 i
Pipe Break Frequency Correlation (EPRl/WOG) 4 Po = (Po/Pi) * (Q, + A
- S
- Qw)
- BF
- P Po = probability of break over the pipe length Pc/Pi = percentage of breaks out of leaks (.06)
Qp = ratio of product of pipe diameter with length of pipe to square of thickness A, S = factors related to weld quality (50,1)
Qw = same as Op but for the welds BF = dynamic loading factor (2)
- P = Global pipe failure rate per Q (f.E-8/yr/Q)
Length =3 ft. diameter-10 in. thickness =.5 in.
i
- O e
Issues Related to EPRl/WOG Approach (Internal Flood)
Correlation modified without explanation J
I Restricted pipe length considered Definition of terms i
' e ,
l l
Containment Success Criteria EPRI/WOG RCS leakage and small LOCAs do not place a significant load on containment Recovery considered USI A-45 Used criteria for a large LOCA design based accident
- O e
i Source Term Comparisons Using EPRl/WOG Core Melt Frequencies expected dose person-rem / reactor yr EPRl/WOG 1 0.08 BMI-2104 and draft NUREG-0956 EPRl/WOG 2 .
0.16 .
DCOR program EPRl/WOG 3 0.7 erg.;e.2g Judgement USI A-45 1.2
.3 time WASH-1400
- O e L____________ _ _ _ _ _ _ _ _ _ _ _ _ __ ___ ____._____________ __. __________ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
i i
Cost Estimate Differences
)
Failure to consider all design requirements (seismic) for specific aspects of modification
! Failure to account for existing structures and/or buried piping or cabling ,
Failure to account for iteration between initial design and final installation i Cost experienced in installing new batteries i
l 1
L* -
1 Upgraded Emergency Operating Procedures There is evidence NUREG/CR-4617, March 1987 that EOPs markedly improve operator performance (CMF coservatively estimated to be reduced by a factor of 1/8) l '
Synergisms amoung transient analysis, upgraded training, and management overview l
1 1
e O O L__ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ________ _. ____.-_ __ _ _____ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ______ _______