ML20235N045
| ML20235N045 | |
| Person / Time | |
|---|---|
| Issue date: | 09/29/1987 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1612, NUDOCS 8710060375 | |
| Download: ML20235N045 (372) | |
Text
_ . _ .. . _ ...___ .. ~ .._ _ . .. . L . _ . . . _ _ L .
m 'm: \
\ ) 17 -
TRANSCRIPT OF PROCEEDINGS UNITED STATES NUCLEAR REGULATORY COMMISSION ADVISORY COMMITTEE ON REACTOR SAFETY l
In the Matter of:
ACRS SUBCOMMITTEE ON
)
EXTREME EXTERNAL PHENOMENA
) ,
'l 0
l RM8%RgEgyy M91Bemov.e' rom ACRS3? ice
- Pages: 1 through 292 Place: Washington, D.C. l Date: September 29, 1907 .j l
. Heritage Reporting Corporation Official Reporters e' 1220 L Street. N.W.
Washingtes:. D.C. 20005 ,
I (202) 625-1888 8710060375 870'729 PDR ACRS T,1612 PDR e
F l(,) .
1-- PUBLIC NOTICE BY THE 2 = UNITED STATES NUCLEAR REGULATORY-COMMISSION'S 3 ADVISORY' COMMITTEE ON REACTOR SAFEGUARDS 4-5 . ,
f 6
I 7 The contents of this stenographic. transcript of the 8- . proceedings of the United. States Nuclear. Regulatory-1 l 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.
12 No member of the ACRS Staff and no participant at r-j 13 this meeting accepts any responsibility for errors or
.Q 14 inaccuracies of statement or data contained in this transcript.
15 16 17 18 19 20 21 22 i l 23 24 L
25-T Heritage Reporting Corporation (202) 620-4099 <
a _ _ - - -
1 UNITED STATES NUCLEAR REGULATORY COliMISSION 7 _. _
'^'
2 ADVISORY COf1MITTEE ON REAC'IOR SAFEGUARDS i
3 f--~ y l kJ -
4 In the Matter of: ) )
5 ACRS SUBCOMMITTEE ON
)
9 !
)
s 6 ! EXTREME EXTERNAL PHENOMENA !
1717 H St. NW l 7 Washington, D.C.
Tuesday, J 8 September 29, 1987 )
9 The above-entitled matter came on for hearing, l
10 pursuant to notice, at 9:00 a.m.
11 BEFORE: DR. CIIESTER R. SlESS Chairman 12 Professor Emeritus of Civil Engineering University of Illinois, 13 Urbana, Illinois 14 ACRS MEMBERS PRESENT:
15 MR. CHARLES J. WYLIE j Retired Chief Engineer, !
16 Electrical Division Duke Power Company j 17 Charlotte, North Carolina 18 ACRS COGNIZANT STAFF MEMBER:
19 Dr. Savio.
O !
PANEL MEMBERS:
21 A. MURPilY D. GUZY 22 l f] R. MURRAY C> D. MOORE 23 M, K. RAVINDRA r3
,/ 24 I
25 Heritogo Reporting Corporation (Ifl2) 428-4800 b________.____.______._
t 2
I p PANEL MEMBERS (Cont'd:)
U R. EUDNITZ !
D. GUZY T. SPEIS 4
)
5 i j
6 j
1 7
l l 8 i i
9 !
I i 10 j l In 1
12 13 14 15 16 17 18 l 19 l
20 21 1
2 ~' l O 23 l
i 24 25 . I j l l Heritogo Reporting Corporation ;
< maa..... l l
t 1
l l a 1 PROCEED I NGS i 2 DH. SIE5S The meeting will come to order. This is 3 a meeting on the ACRS subcommittee on Extreme Extra phenomena.
l 4 My name is Ken Siess, and it says here that I s,m subcommittee 1
5 Chairman, but I am not. I am Acting subcommittee Chairman for 6 .this meeting. The Chairman of the Extreme External phenomena l
7 subcommittee has for a number of years has been.Dr. Okun,'and 8 Dr. Okun is about to leave ACRS a fter 24 years of service, and 9 I will not be succeeding him as Chairman of this subcommittee, j l
10 I will be leaving ACRS in March after 20 years of service. So i i
11 those of you who are interested in the activities of th<is 12 subcommittee can lovk forward to a new Chairman some time, 12 assuming the Committee continues.
() 14 And we have one other.ACRS member in attendance 15 today, Mr. Charlie Riley on my left, and the cognizant ACRS l
16 st af f rumber, Dick Savio on my right, i
17 The purpose of the meeting is to discuss the NRC 18 seismic design margin program and in particular the application l 19 of the methodology as to the Maine Yankee plant. There was 20 participation in a meeting that had been announced prior to the l l
~!
21 Federal Reaister notice on September 14, 1987. Please note 22 that a transcript is beinD kept and will be made available l
, I 23 within five working days,and we do have microphones through the
]
24 Reporter, so'picase identify yourself when you . speak with your
- 1 25 name, and then speak with sufficient clarity and voiame.so thct {
4 Heritage Reporting Corporation (202) 628-4888 y L_____-____ - _ - _ _ - - _
Ti l
4
- . .1 you can be heard by the Reporter, 2 We received no written comments or requests for time 3- to make oral statementsf f ror members of the public. And by way 4 of introduction, we have a very full' agenda. We are planning 5 to go, I guess, on the agenda until some time around 5:30. .I 6 don' t think that there is much' chance'that we wi'11 finish, 7 If.there is anybody that'has an earlyl plane for; earlier. )
1 8 anywhere, please check where you come on the agenda and if i
' l, 9 things tend to run~a little late, let me know'and we will do
10 what'we.can to schedule things your way, i
11 VOICE: We can' t hear you.
12 DR. SIESS: You can' t - hear me? I' m sorry. We' ve got-13 to turn up the mikes. I have a cold and I' ll have trouble j
() 14 talking any louder. Testing? Is'it on? Is it~on? If you 15 can' t near me you' re going to have trouble hearinD anybody. In 16 that any better?
17 I am looking forward'to the presentations at this i 18 meeting. The question of seismic margins has been with us'for i
~
19 a number of years. I am not quite sure that I know' exact 1y 20 what we mean by " seismic margins," and I.use the "we" ;
21 advisedly. I think I know what I mean by it and.I know what j 22 some other people mean by it, but I am not sure we have ever 23 had couplete agreement.. I don' t recall where' exactly the issue 24 originated. Somewhere it originated within-the ACRS, concerned 25 with the possible increases.in design bases for plants in the'
\
'I l
Heritage Reporting Corporation-1202) 628-48E8 i
l o
i
l
, 1 5
1 . Eastern U.S. even before the U S.G.S. unleashed Charleston,-
2 'there hadIbeen some' cases of? increasing design bases and 3 questions of whether the margin was adequate. !
4- There was the more important question raised by:some- t 5 people about whether we were on the edge of a cliff, if'we had
- . J 6 exceeded the design basis, the SSE, was everything going to-7 fall to pieces? The plan failed all at once.
l l ~8 The other thing that. drove it to some extent Was the i
1 .
L 9 introduction-of probablistic approaches. I think in the very.
10 early days there'were some people who thought that the SSE was j 11 the maximum possible earthquake and there was a time that we 12 took the worst earthquake that ever occurred in the region and 13 added a one to the intensity to get a design, basis, and then
() 14 thought, well, nothing could be worse than that.
15 Then war began to get more knowledge, in quotes, about 16 seismic hazards, and it turned out that a lot of people thought 17 that the SSE-had a probability-of between 1 and 1000.-- one in j
}
18 ten thousand -- per year. i l
19 And if everything was going to fail when we had the 20 SSE, or exceeded the SSE, we weren' t too happy >about it,-
21 especially those who thought we had a one in a millior, chance 22 of exceeding it.
23 So the question about seismic margin started to be 24 raised. All the engineers said, "we' ve Dot margins. We' ve 25 dpsi gned conservative stressesi conservative analysesi Heritage Reporting Corpor.ation 1 (202) 628-4888- !
1 l
1 I
.l 1
i
-i u 6
[
~
1 conservative' assumptions' conservative dampingi.et; cetera,':but 2 nobody kisew how much. And,there was a time when mechanical' 3 engineers, structural engineers, used.to.come in and take the 4 calculated stress and made very simplistic calculations of what 5 the margin was, how much the earthquake could % multiplied and 6 still not exceed the allowable stress. About hal f 'the' t ikne 7 they did that wronD. And that was about as simplistic as you j 9 could get;.
1 3 So everybody began to have some concerns about I i i i 10 margirn, which engineers always knew were there, There is a l il feeling we' d like to know a little better how big they were. j i
12 And the result was, I think, first the SSMRp, and the seismic 13 safety reargins were asserted, which was almost entirely'a 1
() 14 probablistic approach to both hazard and er.sistance, and I 15 guess that evolved into the present program. I will expect the 16 staff maybe to trace out that line of. thinking a little bit 17 better for me. And what we'are now callinD the " seismic design 18 margin" program is almost the antithesis of the SSMRp, in other 19 words, the hazard design of it, the hazard is one of the main 20 thrusts of SSMRP -- thn hazard is the main uncertainty. . 8 21 Uncertainty there is measured in orders of magnitude. ,
22 The seismic design margin program concentrates on the .
23 resistance or capacity of the plant. The.e are still j 24 uncertainties,but these uncertaint ies are not measured in 2D orders of magnitudei they are measured in fact as -- engineers l
1 1
Heritage Reporting Corporation i (202) G28-4888
7 1 l
1 'are more used to th!nking of.like, 1.5 or.2.or 3, o~r.something
.O 2 like that. There'are still probablistic aspects, certainly, 3 present in the approach to the seismic design margans, as you :
l 4 will hear.
5 I mention this for the benefit of'the speakers, 6 because this is my concept of sort of how we got.to where we - j 7 are. I aiw not sure it is correct in all of its aspects or'any 1
8 aspect, ano I am hoping that what I hear frore you will make it - i a
9 a lot clearer to me where we are, and a little bit no to how'we 10 got there. I think we need to remember how we-got where we i 11 c e, and the ob,)ect, I think, then, is what we' ve Dot,- what I
12 we' re doing, why we are doing it, and what are we going to j I
13 learn from it about either making plants safer against
() 14 earthquakes, or deciding whether they are safe enough?
15 " Safe enough," I hate to mention because it always ;
i 16 sounds like cafety goals, and there is nothing more confused 1 17 right now than safety goals. ..j iB Charlie have you got any preliminary comments.you 19 would like to make? :
1 20 M R. WYLIE: No. l
)
21 DR. SIESS:
~
Mr. Wylie hasn' t been in on this es-long .
1 22 as I have so I am sure he is going to have a lot of questions-42 3 as we go along.
1 24 With that i nt r~od uct i on,- Andy, would you 1 A ke tu sort !
25 of summari::e things op to where we've been and where we' are r
Heritage Reporting C0rporation !
(202) G28-4888 l
m
)
j g . 1 going?-
, 1 2 M R. MURDHY: All right, I'will take a ch&nce at it..
3 I am Andrew Murphy. I am the new Chief of.the New 4- Structural and Seismic Engineering branch in. Engineering. I am 5 a seismologist by training, and today I will make a brief l
l 6 presentation on the status of the seismic safety rascarch !
)
7 program.
8 I intend to follow the outline of the program as we 9 presented it in the recent'NUREG 1147 (rev.1). This' document I 10 is simply entitled, "The Seismic Safety Research Program Plan." .
11 It has been reviewed by NRR price to publication. It was 4
1E published in May of this year. This is the first revision of f i
13 this p11.n. And in this program plan,' we wi.11 examine very ')
() 14 broadly the regulatory issues that are being addressedt.the 15 research needs associated with'these issuesi and the1research 1
16 program that we have put'together to address these issues and 17 needs.
18 The program is divided into four programmatic
-]
j 19 elements that are shown here on the first viewgraph: ' earth j
- l. l 20 sciences 4 the seismic margin program 1 the fragility / response j
)
21 programi and our valid 3 tion effort. !
i 22 In looking, as I said, very broadly at this program, i 23 the three issues that we have ioentified are " seismic hazard," l I
24 wh ai, we call "seri smi c margi n,, " and the adequacy of current f 25 criteria.- What we are concerned about with seismic hazards are O* HeritaDe Reporting Corporation l
(202)-'628-4888 l
l l
l 1
9 1 the uncertainties that do exist in estimating or assessing the e seismic hazard that we use for licensing. purposes. .)
l 3 tjexb would be the seismic margin. What fr.argins,'as 4 . Jack has indicated, what margins do retually Exist in the 5 plants as-built' A good bit of the emphasi.e for the nurrent.
6' seismic design margins program came-from a couple of. letters:
7' from the ACRS written to the Chairman, one in January 1983 and 8 t h e r, again in Jarsuary 1984, calling into quest ion the act ual 9 quantification of the margin that had been quoted in a number 10 of licensing hearings. The (,CRS specifically asking for this 11 margin to be quantified.
]
12 The next item, or next .i s s u e , Has the adequacy.of 1 l
13 current criteria. Are the assumptions or judgments that have 14 been made in laying out the current criteria, are they i
13 adequate? Are they fully-verified or validated? I 16 DR. SIESS: Andy?
l 1
17 MR. MURPHY: Yes, sir" !
18 D R. S1bSS: A thouDht just came to me. The first and l 1
19 last of those items would apply equally to any. future designs I 20 as to exist 2ng plants.
21 MR. MURPHY: Yes, s i r.
22 lh% SIESS: The second ene, as it is stated, is 4
23 concerned chiefly with existing plants, am I not correct?
24 MR. MURPHY: Yes sir, that would bx a fair 25 interpretation.
O- Heritage Reporting Corporation .
(202) 626-4888
-i
.i' 4 !
t '10 1 DR. SIESS: .Well, the major mission of NRC has-been
~
l 2 stated by the Commissioner to be operat ing plants. Some ever q i
3 so optimistic that maybe we will.see some. future plants or j i
4 future designs or standard designs, although coming up -- is j 4
5 that Detting any significant ettention in our program now? Are l
l 6 you thinking much about future, or mainly concentrating on l l
7 existing plants? , l l
' 8 MR. MURPHY: We are concentrating mainly on existing i l
9 plants, part iculary through the seismic margins or psi-margiris 10 program.
11 DR. SIESS: Well, that clearly 19 existing.
12 MR. MURPHY: But the uncertainties that exist in the !
13 earth sciences that we are addressing today and that we are. l l
() 14 addressing principally for the benefit of the operating plants, 15 certainly will have applicability to any future plants.
IB DR. SIESS: If the research that is beiMD done on I 17 seismic. hazard does not increase the uncertainties,-do we know 18 enough now to come up with a basis for standardization, l 19 standard plants? We have been seeing standard designs that are
-l '
20 proposing a, let's say a .3 G input, which they claim would 21 meet X percent of the sites east of the Rocky Mountains.
22 MR. MURPHYe We probably know enough now. He don' t 23 have any further curprises, that if everything is within the 24 uncertainty bands that we see now, we' re probably pretty close 25 to knowinD enough. But the thing that 61 ways turns around and i
Heritage Reporting Corporation (202) 628-4888
, ~_ _; .
I 1
I 11 -l 1 kicks us in the teeth, is that the next set of data may change
~O' 2 those uncertainty boundee. ]
3 DR. SIESS: Which way?
4- MR. MURPHYs What does "Murpicy's' Lay" , way? It ~ won' t 5 go down.
6 But in order ior us to be prudent s. bout what we are 7 doing, or what we wil'a do with the future, take it frcr,a a 8 seismologist, riDht now we have to have some handle on 9 mord toring what is Eoing on, on understanding about this. -And i
1 10 there are certain ways about the way'we regulate and apply' _
j 11 Appendix A. How we say what ground motions are going to be-at 12 the plants. They appear to be reasonably conservative and 13 rational. Mother Nature may surprise us once in a while, ,
14 DR. SIESS: Is somebody came=in tomorrow -- or maybe 15 yesterday - with a standard plant design .3 G defined in some 3
~
16 meaningful way, could the staf f tell them in any definitive ~way- l 17 which areas in the Eastern U.S. they could not site that plant 18 at?
19 MR. MURPHY: I think they could give them a i i
20 reasonably good handle on that, yes, E i r.
21 DR. SIESS: How far from Charleston? How far from -)
'l 22 New Madrid? ]l 23 MR. MURPHY: New Madrid is the-easy orte. Charleston'
' 24 is more diffieult. And that is'you want to say a' question ed l 1
25 how far we let the Charleston earthquake run? Even there.I
]
i HeritaDe Reporting Corporation (202) 628-4899 i
r 1 t
- - _ . - - 2.a A. . - - - - _ - - _ . -
y
'lj 12 1 th;nk we could put some handle on wh0re that earthquake could 2 be located. But personally I think we could give them a very 3 good handle in a reasonably 'la'rge area unere they could build .
4 those plants, using the good .3 G.
5 DR. SIESS: Would it be any Easier if they came in 6 and said, "instead of .3 it.Will be .4?"
l 7 MR. MURPHY: A little bit. I don' t think it would a
8 make tha,t much difference. The uncertainty ebout things like' ]
i 9 the Charlesten earthquake are such that changing the l
10 acceleration of that margin.just woul dn' t effect it that ;
11 significantly.
12 This next viewgraph identifies what we have referred l
13 to as our "research needs." ' Which, I guess, is Just another l l
() 14 way of saying what tools we needi what handles are we trying to l
15 develop to address the issues, thy regulatory. issues of our
!6 previous slide? j 1
I 17 Here we are talking -- we improve our estiraates of ]
18 the hazard to assess the dif ferent ways that th.is hazard 19 affects the power plant, through the soil response, capability f 20 of the structuresi we are also interested in validatxng what we 21 cali " current criteria," but this is also the way we make our j l
) 22 measurements and operate the codeg that are used in the desi-gn 1
1 P1 practice itself. i l
24 The short-term research is part icular . it ems, 25 projects, compinent response program, that looks at the-O. Keritage Reporting Corporation l
l (202) 628-4888 i
_ . _ _ . -_ _ __ -- ______ _ __ O
4 13 f '
1 uncertainties like in the components - what' are the re ' -l 0- 2 fragilities in these devices that are used in the power plant?
1
-j 1
3 What we are-interested in them for in the long term is'their j
~
4 applicability to resolving the' seismic issue as a whole. To I i
5 their input into programs like the design rdrgin approach ---
E the desigr; margin approach earns a great deal of credit because l i
I 7 it has been able to avoid the need'tcw a extremely critigal J B hazard assessment. It has taken the approwh of going' beyored '
l 9 the nazardous estimate. Wh erer, what acceleration: levels can i 10 these plants meet in general? 3
(
11 Gimply, like your question about the Charleston )
12 earthquake. It doesn' t make a difference bet weer . 2 and . 22.
13 You go enough above that to the HCLPF value where we are j O 14 ceer1eem+ that en1e n1amt cam enerate eareir et th:e particeier 1 l
15 acce1eration. i 16 DR. SIESS: In the short term resecrch your second 17 bullet, is " remove conservat ishes where they are ur,necessary. " i 18 What does that mean in reaction to exist ing . plants? Does that j 29 mean " remove conservatism in the assessment?" '
i 20 M R. MURPHY: " Rem 0ve conservatisrc, " associated with 21 things like the leak before break, where we have got snubbers, 22 pipe lift constraints, and so-forth, that are in there becaue*
23 of t he way we have denigned power plants.
)-
24 DR. SIESS: In assessing the plant, how does 25 licensing? You are not talking about existing plants?
(3 seri a,e neportin, corpor. tion (202) 626-4888
i 14' i
1 MR. MURPHY: -We are talking about existing plants. !
2 SIESS: Oh. For future plants, for' example, a.
C R.
3 conservatism in damping, specifically, and some of that- )
4 conservatism in being removed in evaluations of~ existing i
5 plants. If it is also removed in future designs, which does 6 not seem unlikely, then aren' t we at ' the se. roe t ime reducing .
7 margins? Conservatism is Just one source'of our margins. If.
j i
8 we design these things without any kind of conservatism, we 1 3 wouldn' t have any margin. I 10 MR. MURPHY Thab's right. That's one part ly- fi l 163 l
l 11 guess here, i 12 Now, for the future plants, if we remove the 13 conservatism, there is the potential for removing some of tha 14 margin. The ucenc-io would bre -- hopef ully be, that we are 15 intelligent enough about what conservat isms to remo /s, t'h at 16 there in still an adequate margin in the piant to withstand the 17 uncertainties that1ar2 still going to be-there about the 1G seismic challenge. We are simply attempt ing to do the. best. Job li 19 se can at the time.
P0 DR. SIESS: I think you said something I'd- like to -l L 1 j 21 tha11enge some other time, but what I thought you said is "we i
]
L 22 can put enough ruargin in the struct ure to offset the I 1
.)
21 uncertainties in that h a c Grd , and I don't believe you con, j 24 But that is beside the point. )
25 If all the future pl ants were desipned to . 3 G, that:
( Heritage Reporting- ' Corporation l (202) G28-4388 1
.f.'
l .
l L
. -1
'?
_ ,, ,Q \
15-
'1- is not the same'as having a HCLPF high pr.obability of:.3 G.
-O '2 .That'would be some whatLlarger.. But the design for .J.G with J
3 modified damping. factors, modified this, modified that and so-4 ~forth, would be~different that what we were doing?
5 MR. MURPHY: Yes, it would.
6 DR. SIESS: More realistic?
7 MR. MURPHY: That would be the. intent.
8 DR. SIESS: Okay.
3 M R. MURPHY.: This next figure is taken directly out.
10 of the NUREG I mentioned.~ And bepefully everybody has got.a 11 copy of the handoi>t --
12 OR. SIESS: 7 didn' t find it very helpf sti. It says, 13 "everything affects everything else." Most of the arrows go
() 14 two ways. .
15 MR. MURPHY: Yes, they do. A coup 1? of them v- three 16 of them -- don' t . )
17 DR. SIESS: They go around in cire)es.
18 NR. MURPHY. But it is a problem that does feed on 19 itself. The principal intent of this slide at this time is to l
20 show you what the program eleuents are, to give you a quick l 21 look et the number of regulatory appiteatione of each of the 22 program elements $ and then Just to show over h e re on the left 23 what we rall "related oregrams," to indicate that the program-S4 is to some extent a cooperative program 5 that we are 25 interacting with a number of apancies and statistician, to l
O l
Heritage Reporting Corporation
} (202) 6.28-4888 1
~ ,
16 i
1 expand the ef festiveness of the " hang .for the buck, " of this i i
2- program. We wi21 come back to this chart again a little bit i 1
1 3 later.
l 4 What I would like to do next is just to put~~up the q j
5 pecbable areas of the elements of the research program. There 6 are a number of things on there you can read -- what I would 7 like to do at this time 15 Just to mention for each of these 8 areas, a couple of the projects that we are carrying out, a-9 pouple of projects that are being looked at by these elementh, 10 I think one of the principal ones for the earth 1
11 sciencen is the -- program with the U.S. Geelogical Survey for i 12 the establishment of the national seisniographic network' for the 13 eastern 2/3rds of the nation. That interagency agreement wat l
() 14 signed off last November. >The folks at the Geological Syrvey 15 are reasonably well-along on some of the design parameters: we- .j j
16 anticipate that very shortly a couple of pectotype stations
]
i l
17 will be installed and we will begin to just sit down and decide l
l 18 exactly where the stations that will make up this network will l l 19 be. I 20 The current plan is for approximately 55 to 60 1.
[ 21 stations in the eastern 2/3rds of the nation, from the Rocky.
22 Mountains east. This compares with approximately 250 stations 03 that are currently operated within the NRC's regional network 24 program. The stations will be --
l 25 DR. SIESS: Excuse me, Just a minute. I think we l
() Heritage Reporting Corporation i
(202) 528-4888 I
l
-______-_-__=_______-__- _ _ _ _ _ _ _ _ _ - _ - . . - _ _ _ - . _-_ .- _ _ .
. . - - . -_--___-_._-__j
17
, l' would befbetter off if.we hooked that mike up to.you therc.
~
- 2. MR. MURPHY; The stat' ions that'wil1~make up the 1 3 network will be three component. stations, a vertical and two 4 horizontalt,i will be broad-band instruments-andLproviding high-5 quality digital date, 'This data will probably be transmitted.
6 real time via satellite to the Geological Su rvey' s" nat iona l 7 earthquake information' center. We anticipate that that will be 8 tu-ned around within a half-hour, which is the design goal, so-9 that if any earthquake has magnituda 3.5 vr larger that has 10 occurred anywhere in the Eastern United States, will be 11 reported to the public within a half-hour of its occurrence.
12 DR. SIESS: What is the advantage of that from a 13 research point of view?
O 14 sR. muRnHv= erem a research geint er view, we w111 15 be in a position for the universities that have it, and the 16 Geological Survey to get portable equipment out to the 17 earthquakes for aftershock studies. Considerable amount of the 18 information about the earthquake's force parameters, fault size 1
19 and orientation., is gathered from the aftershocks. l 20 It also will. allow us rapid interpretation of this 1
21 data, so that we will understand and have a data bank for '
22 decisions within the --- office filed for recurrence, j 23 essentially recurrence of OBEs in the Eastern U.S.
24 DR. SIESS: Now you are talking about 55 locations?
25 MR. MURPHY: Yes, sir.
O Heritage ReportinD Corporation (202) 628-4898
_ _1_ ___n_ ____;
18
. f- _f DR. SIESS: There are how many power plant locations
.\
R that now have sei s,mic inst rument at ions?
3 MR. MURPHV: I don' t know right off the top of-my 4 head.
5 DR. SIESS: All of them was a type-fit into the older l
l 6 plants, I think itLwas.
7 MR. MURPHY: No, it wasn' t to my knowledge. Do you 8 know the' number, Leon?
9 MR. RYDER: Leon Ryder. I just don' t know off'the 10 top of my head.
11 DR. SIESS: Well, it must be on the order of 50 or 12 60.
13 MR. MURPHY: Yes, sir. That would oe'a good number.
() 14 DR. SIESSr How does the instrumentation at those-l 15 nuclear power plants compare with what the U.S.G.S. will be
]
)
16 putting in these other sites? How are they going to be 17 currently utilized? .
18 MR. MURPHY: The instrumentation that is at the power 19 plants sites, is principally what we call " strong motion" 20 instrumentations. And this is a --
21 DR. SIESS: Okay, and this won' t be -- this won' t be 22 --
23 MR. MURPHY: The majority of it will not be. In 24 addition to the three components there will be principally i
25 velocity measurements. There will be a number of the stations "
l l-O HeritaDe Reporting Corporation j 1
(CO2) 628-4888 l
1 l
L ____ _ __ _ _ --___ _ _____
19 L 1 that will have strong motion instrumentation as well. We don' t O- 2 'know exactly what percentage of the.55 will have thab 3 instrumentation.
4 DR. SIESS: .Does.U.S.G.S. . have.any plans for 5 utilizing the nuclear plant information?
1 l
6 MR. URpHY: No, sir, they do'not have specific.
7 plans. At this stage, it is just a matter of a,particular
'8 occurrence of earthquakes, and getting that data from the power-9 plant.
10 That program has been in effect since November. It l
l 11 has got a six year track, so that the entire package will be in l
12 operation by the end of Fiscal'1992. In that time period, we 13 anticipate the NRC will .phece cut it regional networks. The O 14 phase-out will be timed'so that a new network will be ' in ._pl ace i 15 we will have several months to a year of co-existence, so'that 16 we can be certain of the operating characteristics of the new 17 instrumentation t
18 Another important program that we have going on is an i
19 assessment of the ground Motion in the Eastern U.S. What is 20 the source of importance of these high-frequency, high-21 accelerations that have been recorded recently? Are they a 22 characteristic of the source of the earthquake source 23 mechanism? Or are they a function of the travel path? The 24 materials that the earthquake travels through? This program is 25 being addressed on a couple of fronts! one commercially through Heritage Reporting Corporation (202) 628-4888
20' 1 a -- procurementleanother through a granti and then a third in.
2 cooperation with U.S. Geological.
3 DRs SIESS: And you' re talking about things like the 4 Perry?
5 M R .- MURPHY: Yes, sir.
6 DR. SIESS: Now, has a' decision been made'somewhere 7 that these high-feequency,: low-energy earthquakes-have:any 8 significance to a nuclear' power plant?
9 MR. MURPHY: No,. sir. I do not believe a regulatory 10 decision has been made. There is a tendency to discount their 11 importance.
12 DR. SIESS: It seems to me what would be important 13 before we spend a lot of money is its standing.
4
() 14 MR. MURPHY: Yes, it is very definitely important.
15 But I mention the program because of its importance in here.
16 It is not a high-cost item. We are getting cooperation from ,
l 17 the Geological Survey and the National Science Foundation on it 18 so to address this question is not expensive. i l
19 DR. SIESS: Okay.
20 MR. MURPHY: We also have a couple of programs 21 looking at soil liquefaction, recurrence rates of earthquakes 22 again, they are in the Geological Survey and then a competitive 23 procurement with the commercial contractor where we are 24 examining soils in the Eastern,. Southeastern United States, to
.25 again get a handle on the recurrence rate of the Charleston Heritage Reporting Corporation 2 (202) 628-4888
J l
21 <
l 1 earthquake. !
() ' 2 preliminary indications are that the Charleston 3' earthquake has' recurred, in'the.same vicinity, in the order of 4 1500 years.
5 DR. SIESS: Of sand --
i 6 MR. MURPHY: Yes, sir. Those are a couple of the 7 projects within the earth sciences. Dan is going to cover 8 seismic design margins in. considerable detail shortly so I will 9 skip over that element.
.10' The next two elements, component fragility and.
11 response program, two of the important programs here. First is-l 12 the seismic category I structures program that-is being carried' i
1
[ 13 out at Los Alamos. Current status of'that is that we have done
() 14 two tests -- recently completed two tests, and the static mode l 15 to investigate the' frequency shift that was observed ' was --- the 16 frequency shift, to-my mind, the softening that was observed in 17 that dynamic calculation, in the dynamic models. The two 10 static tests that have been finished have confirmed current 19 theory. current codes, and do not show the frequency shift or 20 softening of the models.
21 DR. SIESS: I' m sorry, say that again? They do not-22 show the frequency shift?
23 MR. MURPHY: They do not show the frequency shift. ;
24 DR. SIESS: That is not what I heard.
25 MR. MURPHY: I believe that I am correct in that they Heritage Reporting Corporation (202) 628-4888
y 22 ,,
1 do not'show the softening of the frequency shift'that~ was 4
2 observed in the dynamic model.
3 DR. SIESS: Did they crack the concrete?. j 4 MR'. MURPHY: pardon?
5 DR. SIESS: Did they crack-the concrete? This i s.
..J 6 FRT-4?
7 MR. MURPHY: Yes, sir, i
8 D R. SIESS: Did not show a reduction in stiffness 4
1 9 when the concrete cracked? You must be -- I-Just saw a report. '
10 I got a --
11 MR. MURPHY: Unless I am'misspeaking in..-- )
j 12 DR. SIESS: If they made stat ic tests that don' t . show 13- any change in frequency of a shear wall -- and that would be 1
() 14 cracks -- I would be extremely surprised.
15 Both, do you know something about it?
1 16 MR. KENNEDY: Yes. Bob Kennedy, I am on the, I 17
~
guess, call it " peer review" for those tests. The last;two 18 tests were very carefully handled samples.- The issue was, all i
19 of the previous tests showed a frequency shift at very~1ow I 1
20 input levels, below the cracking levels, and it was not known 21 what that would do to and some of us had concerns. The last 22 two tests were very carefully handled so that they were 23 convergent samples. They were uncracked samples before they 24 were put on for the test, they showed -- under-static loading, 25 showed very low stress levels, well below the cracking stress- .
j j
O Heritage Reporting Corporation (202) 628-4888 l
JJ I
- ]
4 3
23 1 levels, carried on crack theory where it is well for their 2 stiffness. When they got to the cracki'ng levels they did crack i 3 and the frequencies did shift down.
4 The issue is all of the previous. tests showed a' 5 frequency shift at very low stress levels. It is some~of our l
6 suspicion that those previous models;were cracked before they )
7 went on the shake table. )
8 On the other hand, real buildings are cracked, too, 9 before the sial flakes and so on --
10 DR. SIESS: That is exactly the point, Bob.
11 MR. KENNEDY: So Mehta Sozin and I both'believe very 12 strongly that all this really is' indicating is there is a 13 significant variability in the stiffness of concrete
() 14 structures. Both sets of tests are right. One with very 15 controlled samples and the other.with samples that were 16 cracked, l l
17 DR. SIESS: I would like to have the money they put 18 in that project since I told that that three years ago. I 19 mean, cracked concrete is going to have reduced stiffness. If 20 it is a sheer wall it is going to be a tremendously reduced 21 stiffness because everything cracks at once. And if you are 22 going to make any kind of test to try to predict when concrete 23 cracks, you are Just spending money throwing it down the drain.
24 And you don' t have to load concrete for it to cracki.go out and d5 look at the damn buildings. Midland.was cracked from i
O Heritage Reporting Corporation-
.i (202) 628-4888
'l l
I
-_-_ _ _ _ _ _ _ _ _ - - _ _ _ _ _ _ - _ _ _ ___ _ ___________-___m
24 1 set t l erne nt , but you could walk through Midland and you could O 2 find cracks in walls that had never settled. Concrete cracks.
3 Period. God.
4 Okay, go ahead.
5 MR. MURPHY: On that p ro g r arn, the current plan is to 6 exarnine the characterist ics of the dynarnic tests to see waether l
7 or not it is possible to decide what caused the cracking on low 8 levels.
9 DR. SIESS: That is an absolute waste of rnoney. I era 10 going to have to get to sornebody at a higher level in this 11 out fit and tell them that. You' re not going to learn anything 12 f rorn any research in the laboratory on why concrete cracks or 13 when concrete cracksi just Do out and look at your buildings.
O 1* 't'e o 1"o * "e crecueo- ame t"ex e"oht
- oet " it" *"e
. 15 original purpose of that prog rarn, and that i s, applying 16 analyses to coraplex shear wall type structures with shear lag 17 and al1 these other phenornena that need to be 1ooked at.
18 And if they are worried about changes in frequency
~'
and core response, then get out and find out what is i rnport a nt .
20 Okay, let us get off of that one. Because that is 21 secondary to rnargins.
22 MR. MURPHY: The next thing I wanted to bring up Wa 23 Torn Bryan's cornponent f ra g i l i t y prograrn. That prograrn has 24 looked at the nurnber of cornponent s, panel boards, rnot or control 25 units, and established a fragility data base for that pro g rarn.
Heritage Reporting Corporation (202) 628-4888
i
+l 25 l
)'
1 This year he is looking at some additional O 2 ' components.
3 DR. SIESS: Who is doing this, now?
4 MR. MURPHY: John Bryan is the project manager. It 5 is being -- part of it is being carried out principally at i
6 Brookhaven.
7 DR. SIESS: How does it relate to the EpRI-generic --
8 I forget what'they call it, that whole series of. qualification 9 tests and developing --
10 MR. MURPHY: There was a serious, in the earlier 11 stages in this program, where there was considerable 12 cooperation and trading of data back and forth, this particular i 13 -- at this particular time there is a number of componentsi the ;
() 14 NRC is looking at, independent of the EPRI program, 15 independent, in that we are effectively "doing our own thing,"
16 completely aware of what is going on with them.
17 DR. SIESS: You are checking on the EpRI results of--
l 18 MR. MURPHY: Yes, sir. Our results were also feeding 19 into the A-46,-the seismic qualification effort in the program, :
'l 20 and coordinating that with Knute Anderson. I 21 D R. SIESS: Will this program involve actual tests or 22 looking at results of previous qualification tests?
23 MR. MURPHY: Both, sir. There is some-actual testing 24 going on of components, and then there.is simply collection of i 25 data. ;
O Heritage Reporting Corporation (202) 628-4888
i l
{ 26 l
1 D R. SIESS. Okay..
O. 2 MR. MURPHY: That program at this stage we anticipate
/
- 3. winding that up with the work that is being done, this year.
4 The next element of the program is our validation 5 program. Here we have three projects underway. The first is 1 6 the HDR program with'the Germans. On that, the so-called 7 " shad" series of tests have been-finished. That was the large,
)
B exciter high up in the structure.
9 This year we are doing what is called the " sham" 10 series, which is exciters on a piping system, and then varying 11 the types of energy absorbers, snubbers and other energy 12 absorbers, ' hat have been applied to that piping system.
13 The second program is the Tadotsu shaking of about a
() 14 1/2 scale PWR pipe loop. We anticipate that that experiment 15 will physically be carried out in. March or April of this year.
16 The other program is the on going sales structure 17 interaction program with EPRI and Thai Power at Lotung. There 18 they' ve got a 1/4 scale containment building with some 19 internals sitting there being shaken by a natural earthquake )
i 20 ground motion. I 21 DR. SIESS: I recently saw in the form of some tests 22 on piping with three shakers attached to it, and so forth. In 23 that NRC-sponsored work?
24 MR. MURPHY: That was some of Dan's work.
25 DR. SIESS: Why is that not listed? Th6t's I guess Heritage Reporting Corporation
.(202) 628-4888 i
1 l
- - - - - - - - - _ . _ . - - - _ . - -___.__._m___a-_-___.m___._a_m_...a__._
i 27 1 bccause we did pipe research also? -
(:) 2 DR. SIESS: We co piping research also. That is-
]
i 3 Jooked at under the fragility response program. I just. picked ~1 i
'l 4 particular ittims to talk about today. Anything else?
-]3 5 DR. SIESS: I have'a problem with this validation 1
6 being pulled out as a separate item. 'You say'thosa' piping- ,)
l 7 tests are actually built in responce that HDR is under {
i 8 valication? Simply.becauso:HDR is in funny-looking oower 9 plant, end the others were on the shake table? This validetion {
l 10 category, I guess, bothers me. Just why it is separated out. ?
11 MR. MURPHYr I guess I Dot to say I don' t .know 100 12 percent sure why at was originally separated out. This was a 13 programming decision tnat was made and it works in some senses
() 14 and there is valid criticism of it in other senses. It ic 15 tote.lly artificiali it is just something that was a management 16 tool to work with --
17 LR. SIESS: In the other three categories that you 18 had the research divided up i n, has a response, has it that the ;
i 19 fragility has been putting those together to get some kind of l 20 margin, but not the same kind of margin we' re talking about 21 now. I guess it made more sense, i
22 Then I found validation _and I wasn' t quite sure what 23 we ware valideting. ;
24 MR. MURPHY: Okay, there we are i nt eres't ed in )
25 validating codes and assumptions on the fragilities and !
O Heritage Reporting Corporation.
(202) 628-4888 s 1
28' 1 response, and we' re . principally looking at their pathering of 2 data, whether it is new data from experirr,ents theuselves, or
. data that had been carried out by vendors.
4 L R. SIEGS: That sounds. good when you say it fast, 5 but when I s t e-rt trying to break it down as to uhat data 6 constitutes validatier, and what cata constitutes something 7 else, I think it breaks down. It is a strange cattgery but I B will accept' it and go ahead.
~
~
9 It doesn' t mean- it's' validat ion and codes in the Q/A l 10 sense.
11 MR. MURPHY: No, sir. It doen not. Okay, that'is la this presentation i have for this morning.
13 DR. SIEGS: Questions, gentismen? Well thank you
() 14 very much. It was a ' nice overview. I' m going to declare a 15 break at this point in time. About ten rdinut es. [Off the 16 Recor d. 3 17 EBack on the Record.: Okay, Don, you have got the 16 floor.
19 MR. GUZY: Okay. My name is Dan Guzy and I am-20 heading up this presentation er,the seismic design reargins 21 program. In thf.e session, I think we are prepared to follow 1
22 the agenda-pretty much as it ic here. One suggestion I'would 23 like to make is that the comments for Maine Yankee would come 24 before my future plans part. Bec se it is more logical. I 25 cover other material other than Maine Yankee's in the D-start Heritage Reporting Corporation (202) 628-4888 L____-______ . - _ _ - _ _ - - _ _ _ _ _ - _ .
29
.q 1 review.. So I would like to switch that.
%)
- 2. ,
D R. SIESS: Good.
3 MR. GUZY: Also,.I am on twice'today, and what I 4 would like to do is to do part A of the' status and the 5 discussion .about the methodology first and leave both future 6 plans and information about on-going and then recently finished l
l 7 projects until the second session,.so a quick status. discussion 8 about the methodology and run right into Maine Yankee, and I 9 will cover the other task later on, if it is all right.
10 Okay, a little information.here'about where we have 11 been, and as I noted before, the seismic design margin program 12 began in the 1984 time frame. This is in response to a number 13 of requests and beefs, such as the Charleston issue questions
() 14 and ACRS concerns about -- that were addressed to NTOLs and 15 generic letters. written to EDO.
16 The EDO, in response to these ACRS requests and 17 concerns of his own, I guess, develops a response plan of 18 action which included that the establishment of an NRC working 1 1
19 group to look 't a the margins issue, and also at that time we i 20 realized we needed outside help and we formed an expert panel 1
21 on the quantification of seismic margins, j i
22 The first year was mainly a planning year and mainly )
23 through the. advice of the expert panel, we developed a program ]
)
24 plan. I also should note that Lawrence Livermore.
25 subcontractors helped'in developing this plan.
\ Heritage Reporting Corporation (202) 628-4888 ;
i 1
1 30 i
)
1 The plan itself'is a little bit. obsolete, but the.
2 general guidance in it is being followed. The fact is.that1 we 3 developed a new science review methodology and we have 4 implemented it in two trial pl ant s. .The first' trial plant is, 5 Maine Yankee, which we will be talking'more about today.
6 In the 1985 and 1986 period they developed this 7 methodology and it was developed looking at current seismic 8 information, earthquake experience datal test datal in 9 particular, work that would be done by the Scrub A-46 effort, )
l 10 plus information of all'the pRAs1that were available at the 1
11 time. We have developed two documents on -the - approach and the 12 guidance for using this approach that were-published.in that 1
13 year.
() 14 That was about where we were when we had a last 15 briefing last year with the ACRS -- we discussed the l
\
'l 16 methodology and the planning for future work and.the program ,
1 17 then. The Maine Yankee review was ongoing at the time but we .
18 requested that it not be discussed because it was all the l
l 19 results were interim resultsi the program wasn' t completed. j 20 And today we will hit it in its full glory, I guess.
21 The. Maine Yankee. review is complete. It was ,
22 initiaced in the spring of 1986. The real'-- and that ' s . t 23 initiated beginning which the contract is put in place. .The 24 real kick-off. meeting was June and I guess it started in 25 earnest in July 1986. It ran all through the fall and the, l O Heritage Reporting Corporation-(202) 628-4888 L
i
- -.---_.N_--.--_-__ --------.A.---
31 1 winter and was; completed in March 1987 with'the publication of.
-2 the three volumes on' Maine Yank'ee. NUREG ' came i out' in Mar ch 3 1987 which essentially closed this'out.
4- These NUREGS and the draft of these NUREGS were 'used 5 in developing a SER or. Maine Yank.ee wh'ich: was also . issued in G . March 1987 Just before the v3 organization. . And especially~this
'7 closed up the main ceismie licensing' questions-t%at were'being
~
8 asked at Maine Yankee, so we finished the program and .i t ' wan -
9 used in an actual licensing case. You will hear much more 10 about this today.
11 Another study which har been a long time in coming,:
12 was the PWR systems study. At the time of the expert panel, 13 there was enly one seismic pRA, but since.then, there has been
() 14 five more that have been made available, and we initiated a 15 program to do the saine kind of systems evaluations and develop 16 recordations that was done for pWRs, which had a lot more pRAs 17 available.
18 This program was initiated in FY1986. We have only 19 recently have got the draft report on. The review of this 20 draft was ur:derway,.and we hope to get thel final report 21 reviewed and out within the next few months. I will. talk about 22 this later on this afternoon probably briefly.
23 Another. prograrn project that is on-going.is a 24 comparison of the two HCLpF calculational methods. I will 25 discuss those a little bit later this morning. But this was a Heritage Reporting 1 Corporation (202) .628-4888 1
m - .. . . . . . . . . . . , - .
h i 1
'1 1
- 32. l j
' 'L 'l task that was called for by'the expert panel way back when the )
2 approach document wa written. It has taken a while'to first
)
3 develop.a work scope to which'everybody agreed'that would be '
4 the best thing'to do. plus, one of the reethods, the CDF 5 method, was not developed at the time that we recommendedEthis y I
6 st udy.. And since then, UBI han further developed it %nd has .!
l 7 given us a bench mark so that we can make a comparison with the j
~
8 fragility analysis methi>d with, q
9 So this has been a plan for a-long time, but the Lj l
10 actual kick-off meeting was not unti1 ~ July of.last year. This: 1 I
11 project is ur.derway. It' plans to be completed'in Mcrch 1988 e 12 with a report. -i
.i 15 The third item here as I want to point out is that i 1
() 14 there will be signif-icant findinDs. I g'uess'I should say, in 10 the conclusion, well, the significant findings hopefully.by- .
16 these are in the NovemLer meeti.ng. There is another November 17 meeting, and ,f'chere iG any sf gnificarit differences or if 18 thi n a. wer*+ some of + he significant findings that came out, it 19 probably will be in the November time frame that we'will know 20 about it if there is something lacking or different --
greatly
- 2) different, in both methods.
22 DR. SIESS: Who is doing that ?
23 MR. GUZY: I will get into that. But we have a 24 panol. We have a penel for everythinD. We have a panel work' 25 group for this.
O Heritage ReportinD Corporation (202) 628-4888
33 '
1 Another. project that is part of our overall program
<-)
2 is something that came through the Small Business. Incentive 3 Reaearch program, and this was' initially a smal1~ task that 4 Future Resources Asc., LTD, Bob Bennis' company, had proposed l
j 5 on and we accepted, that essentially looked at'the seismiciline 1 i l, 6 margins approoch and see ways that we can improve'its use and L 7 to extend its.use for more knowledge. I will: discuss this l.
l 8 .again this afternoon.
9 _But'that has.been completed. The s' coping phase.one ;I
.j 10 aspect has been completed, and we have awarded-the' phase two 11 part of that and that will on going next' year. I willidiscuss 12 this more this afternoon.
'13 Okay, now, the major activity now currently on going
() 14 of the program is to review the EPRI seismic margins 15 methodology. We are planning to go cooperative with them in a 16 BWR trial plant review. One of the steps is to give them a 17 good peer review of EPRI methodology -- if something existu, 18 NRC will look at it. So the first step in 'this cooperation, we )
19 received draft reports of methodology in the spring 1987. We 20 received both the draft of their' methodology: report before and 21 results of their . Catawba st udy, which they used tbnir L 22 methodology in doing a margins review in Catawba study.-
l 23 DR. SIESS: Has the Catawba study been released? i 24 MR. GUZY: It has not been released to the public but l 25 it has been released for our review. As far as I know. Bob O Heritego Reporting Corporation (202) C28-4888 :
l a
34:
^
1 Castelara of EPRI is here. He-can give vou a status on that..
Bob?
R 1
3 MR. CASTELARA: This is Bob Castelara here from EPRI.
4 The Catawba review has been completed and'the report is in .
5 draft forrn. It has not been. released as an'EPRI report as l
l l 6 yet.
! 7 MR. GUZY: It has.been well-rev,iewed by:both the NRC 8 -- well, by the.NRC Work Group and this panel, as a 9 demonstration of how the EPRI. methodology could be used.
10 DR. SIESS: But there are no licensing issues 11 involved in Catawba.
12 MR. GUZY: No, there was no licensing. Unlike Maine 13 Yankee, there was no licensing issues there to solve, j
() 14 As part of our review we established another panel to-i 15 review the EPRI methodology, and this was an intense two- o'r' {
'j 16 three-month effort which was completed in July 1987. We issued. l 17 a draft report and it has had some impact on the FORI l
18 methodology in that, at least'their report, they have-revisadL l
19 their methodology to clarify items and incorporated comments.
l l 20 This, again, is in a draft form, and thit latter' report is now 1
21 being reviewed by the NRC Working Group. We plan _ to have a. ;
i l
22 comp?ste area evaluation next month. I' ll . talk a little bit -
l l 23 more about this this afternoon also.
l 24 And at the end of all this review at'least now we 25 think we will have a NUREG report that will tie all of this l Heritage Reporting Corporation (202) 628-4888 l
35 l
-1
.1 together, both the ' panel's report i, the EPRI's comments in that i O 2 report ano their revisions and.the NRC Working Groups, in 1
3 par t i c u l ar.- Basically it la favorable, but I will get toithat )
4 later on.
5 Ttue next big thing that we are playing will'be the
. .I 6 trial BWR plant review. And as I said, we plan to do this in- i l
l- 7 cooperation with EPRI and we have a working agreement between l l
8 EPRI and Research, which I will discuss later,.and we have a 9 proposed. plant to look at, which is Hatch I.
10 There was o meeting in late June with Georgia Power, 11 EPRI, NOR and Research to discuss the possible use of Hatch in 12 the margin review in a cooperative effort, and negotiations are j 13 on-going now and I will discuss that later this afternoon. But
() 14 if all goes well, we will probably begin this review in early ,
l 15 1988 -- maybe as early as January.
16 DR. SIESS: Don, an extraneous' question. Regarding I 17 BWRs, the last time we talked about the SSMRP there was a 18 seismic PRA to be done on PWR -- LaSalle, I-believe. Did that 19 ever get done?
20' MR. GUZY: The report has not been published, but I ,
l 21 understand that there has been a final draft and it is due out 22 any day now,.right? That LaSalle report has been used by the 1 l . .
l 23 Martin program as one of the PRAs that we have looked at in.the 24 BWR systems study. That was a draft of that. .My understanding i
I 25 is the draft will -- the final report is coming out any time
! Heritage Reporting Corporation (202) 628-4888
36 1 now.
! Chet, that 2 CUMMINGS: Garth Cummings, Livermore.
M R.
i 3 draft is being monographred and typed right now,.so we:can-4 assure them -- l i
5 DR. SIESGL Thank you.
6 MR. GUZY: Up to here is more or less our firm plans ,
1 7 for the seismic design margin program. There may be a few.
8 wrap-up tasks. I don' t' foresee' anything . big, other than what I 9 we have listed up to this last bullet herei in +he: current 10 context of the seismic design margin approach, 11 However, there are severe accidents, which we will 12 discuss this-afternoon, which may have impact on what' we do in I 13 our division, the Division of Eng'ineering. Surely it has an 14 impact an what people in Brian Shearing's division will 'do, but 15 we haven' t worked it out yet. Right now, we don' t have 'uro/
l 16 concrete plans to do something specifically to address. severe 17 accidents other than what we have done already. There is I
18 separate research being done by the other division and it is I 19 being funded by them.
1 i
20 So I am saying, this in an " iffy" thing for us. We 21 don' t have any concrete plans. We may'do something in this ~;
22 area.
23 D R. SIESS: All right. To'take that last item.and 24 turn it around, the role of severe accidents in seismic design 4
. 25 margins program, when the-margins program.has.been in terms of ,
l 1 Heritage Reporting Corporation (202) 628-4888 1
_ ___m___-.____._-_m_ . - _ . _ _ . _ . . _ _ _ _ _ _ _ _ u _ - . _ _
1 07 1- severe accidents, has it not,.not in' terms of design basis 2 = accidents?
3 MR. GUZY: In termsDof.the global sense of a isevere 4 accident, yes. But in terms of what the policy statement says 5 on how you. implement it are internal events.
~
6 DR. SIEGS The Maine Yankee seismic desion margin 7 pr.agram at Catawba have all been looking at the. role.of margins 8 against severe accidents,.not against design basis accidents.-
9 MR. GUZY: Specifically,-it is looking at all the l
l 10 margins mitigating or preventing core melt.
- 11. DR. SIESS: Okay, that is a severe accident.
12 MR. GUZY: Just a quick run-through. of- the . stat us. I l
l 13 have been asked to tal'k about methodology, and-I will do that 1
() 14 next, but first I would like to recap ,what our objectives are.
15 There is two objectives we currently cite now. The 16 first one is to assess the inherentIcapabilities of' nuclear-17 power plants to withstand earthquakes to develop.the design 18 level. The second is to arovide an effective and. efficient 19 means to ident ify seismic vulnerabilities.
L 20 The points I woulo like to make about these two 21 objectives are: first, the first one is probably better 22 understood when we have a bettte feeling both about the NRC-23 prograto and the EpRI program, of meeting thic objective. I 24 think people being involved with the program, feel fairly 25- comfortable saying, "yes, we can use this methodology and meet Heritage Reporting Corporation (2E2) 628-4888 l
/
___.___.__m__...m._-_m_____.____._.____1 . _ . - _
I i
, 1
(
'6
, 38 this objective."
l 1 g l
2 'But it should be pointed out in writing this' 3 objective and understanding this objective,-we.weht through a 4 lot of iterations defining what a margin;wasi what margins'are 5 importanti what failure isi what failures are'important -- so 6 it took.us a lot -
you know, this is a modification.of the s
7 original request to quantify. seismic margins and we[ sort of put,
-l 8 it in our own terms and defined the objective in ter4as of what .)
j 9 was reasonable and correct to do.
10 Tha second objective is not as well-developed in my
~
11 opinion. The book warranty programs, the NRC program and the 12 EPRI programs do find vulnerabilities, But I don' t think we 13 have a firm handle on what vulnerabilities we really need to (f 14 look for and how this f'ts i into our general policymal:ing here 15 within the NRC.
16 DR. SIESS: How would you define vulnerability? l l
17 MR. GUZY: That is -- I would define'it one way and 18 somebody else would oefine it another way. It is -- t he i
19 vulnerability would be determined -- in the context of the' '
j 20 margin programi would be a weak link or a system or component 21 that would +- whose failure would lead to core melt-and would j E2 be identifiable below a certain seismic margin earthquake E3 level.
- 24 DR. SIESS: In other words, if you were using .6 G as 25 your seismic margin of earthquake, the vulnerability would be I O. Heritage Reporting Corporation !
(202) 628-4888
_________-_2___-_____.- _ _ _ _ _ _ _ _ - - _ _ - _ _ _ - - - _ _ _ - - _ _ . _ _ _ _ _ _ _ - _ _ _ _ _ _ - - - _ _ _ - = _ _ - .
y
, ,.~~-- = -
, e .
4 1 3
.i i .l 39 i fm 1 .anything that would' lead to a core melt below .3'G?. )
\wh .
l 2 MR. GUZY: Anything-of the functions you identify are 1 3 important to mitigating core melt, a. system or component thht 4 -led to a HCLpF that was below .3 Gs would be a vulnerability.
5 DR. SIESS: Why wouldn' t you-accept my definition?:
6 MR. GUZY: I tried to ekpand'it to clarify my terms. >
'l 7 Vulnerabilities could mean -- I ;nean, you could have a 1 8 vulnerability in terens of radioact ive release - -
9 DR. SIESS: I' d put it in terms of core melt.. And-10 you qualified it-by the HCLpFs and' trains needed.
.1 11 MR. GUZY: Because you could have vulnsetbilities --
\
IS you could have things that would fail or not haveLa'HCLPF apply l I
13 at the highest .3-Os might not bS important to core melt. I J
t( ) 14 guess'you said.that. Maybe you covered it? Bob?
J 15 DR. SIESS: But that definition. hinges on picking an j i
16 earthquake to begin with. )
17 MR. BUDNITZ: I'am Bob Budnitz. There is a crucia'l l 1
18 point here, and that is, somebody'inLthe-regulatory crowd, j l
19 presumably it is the Commissioners, ultimately -- will have to
]
1 20 tell us whether a very strong plan has any vulnerability. That l l
21 is, suppose you screen a p2.an for .3 G and it makes it. And- !
~
22 then somebody is not happy and they say, " great, we' re going to 23 screen it for .5 Gs," and it still makes it. '
i 24 In other words, if it pl ans- o'ut my way then I hope- .
i 1
23 will make it at . 5 G. Of cotrse that is a different hazard, i l
i
/~'. -
Heritage Reporting Corporation !
(202) 628-4888 I
i l
t .!
'0 4
1 But suppos2 an Eastern plant was designed well so that they.did
- i 2 & ,S G HULpF with'high confidence? ~Does that plant have j i
3 vulnerabilities? Presumably.
')
4 DR. SIESS: Suppose you do a pRA and.the core melt j 5 prchability is 10-6, Do ycu care whuther it is caused by an 1
4 .
.l G' earthquake or not? , J 7 MR. BUDNITZ: Well, I was asking that question. New,,
l .
l 8 I can' t answer it On3y the Commissioners and their 3,000 9 staffers, can answer that question. Put'it has to beianswered j
10 because what one does in this program depends critically on the !
I 11 answer to it.
1 12 I have a personal view, but then again, I.am only a j 13 citizen. And my personal view is, once they decide to screen, '
1
() 14 and they screen it so there are no vulnerabilities as.thab 15 plant. That is what I would think. I would .like to think of l
l 16 it that it is therefore otronD enough'for us, and &ns can move .j 17 on and worry about floods or LOCAs or something elso, a j
18 But thet question is sti.11 floating here at the end 19 of 1987, even though we started in with this program in 20 December 1984. And I want to let everybody know that that open- I i
el question ought to be answered by somebody. Policy level -- and ..
l ~ 22 it is not Dan'"a fault 1 it has got to'come frofa thra Ticensing _ -
4 23 and the Contaissioner level.
24 DR. SIESS: Where do the safety goals coms.in? j
' l 25 MR. GUZY: I think we will be revisitinD l 1
1 O. Heritage Reporting Corporation
($02) 628-4869 I
s
's-... .-_ ,_ _
l l
4.1
]
+ 1 vulnerabilities this norning and a lot rnore this afte.*nnon s so 1
2 we taove cyn. j i
3 I would like to give you Just a brief run-through on i
4 scrne of these groups 14e' vo 'han. I think sorne people are 5 conf used about what partirular groups are and what'they do. l l
6 The rnain point is that we have a lot of. di f ferent i 7 review groups, work groups, and testing. panels, and the Way we 8 have been working at is'that we develop these panals on a te.sk-1 l i
9 specific basis, se that sorne cases the sarne . people e,re on I l
!O offferent panels and sornet irne s we have new people in and it 11 gets very confusing end we put the whole list of everybody 12 here, i 13 The next group panel was essentially was the group. ,
i O
14 that devel ped the rnath d logy of the approach and the 15 guidelines. We had a separate review group that served as a )
16 sort of O/A function on the Maine Yankee review that tried to
, 17 keep the reviewers honest and your own inLight or input into j l l l
18 activity into the review itself. We plan to have a sitnilar 19 peer review group for.BWR plant review.
20 Wn also havelthis panel to review the seisrn;i c rnar gin, l
21 the EpRI seisrhic rnargin rnodality, and this was a very short i 22 effort -- two or three rnonths. Other groups conforrn here for l
l 23 the HCLPF cornparison, t h t> CD&M vereus FA HCLpF cornpari son --
24 and we have sort of forrced groups on an ad hoc basis depending l
25 on what we r.eed these people for. A 30t of t irnes it looks like l
1' l
Heritage Reporting Corporation (202) 628-4888 l
1 l
' '~
.v O
e 42- J 1
w 1 a pick-up game here. Who's on the playgrodnd and, who we need.
i 2 Lawr.ence Livormore has been the main contrant with us
~
1 13 on the seismic me.rgin. program. :The Chief project.Nanage,r is- 1 4 Bob Murray, who:will be t, peaking ch the Maine Yankee a bit !
f 5 later, and Pete Cassanot 'nas been th:2 main people at Livermore' l
6 on the program. , -
- 1. f 7 There are also a lot o* consultants in'this program. j
- i b .i 6 I think it<was one tirm ne figured that about o0-percent of t h e. j l 9 ccut went to corselting firms. And these are all represented !
l l
10 here today by dif ferent peUple, we have EQEh Future Resources, i l 1 la Asc. , Energy, Inc., and Gpplied Risk Techno1c gias. . Plus I 1
l 12 think there han been some others that we have contracted.
I I l 13 DR. WYLIE: Do you have any left over we could use' ]
() 14 for ACRS?
15 MR. GUZY: I think we hard enough time playing up l 16 industry and research. '
1
. 1 17 Okay, I hava been asked to present some st uf f on. the 1 l
18 methodology. I would like to cover just the main points on j i
19 this slide. First of all, as noted, we don't develop sei.smic !
20 hazard curves. The *.iargins' program han developed and used the I 21 high-confidence -- low probability of faiJure in the HCLPF 22 concept. I will discuss that. limit ver-y soon. ;
23 Also, the use of screening is very integral to the >
24 mcrgin study. We screen both on component. strength and on what -!
L 25 systems or functions are really import ant to safety during the I i 1 Heritage Reporting Corporation (202) 628-4888
.l 1
i i
_. .m_. _ _ _._______..__E_.___-___
= m
'l.
! i V !
i
,. 1 1
'43
.1 1'
1 core _ welt. The program, becauseofthe.peohle-involved, 2 because.of how recent it is,fwe made-use o'f a' lot ~of recenti j
' i 3 earthquake experience data and test data, particularly the data ,
a i
4 that was developud undue.tho A-46 squab.e.ffort'.
D One of the key things ar d one of the key' bene fits of G t hit seismic margins. , methodology, is tieat . it requires two fj 7 extensive plant walk-Uowns. You have-two people in the phant 8 who ere checking thi'nus such as anchorages, systems
~
1 l
S interactions -- they spendla lot of time.there -- I think the j i <
l 10 first review wou2d no;ainally be a week in lengthi.the.second
~1 11 would be shorter.
12 DR. WYLIE: Are you-going to discuss further the 1 13 Scope of the interactions?
(h 14 MR. GUZY: The systems interactions? I think the 15 best demonstration of that would.be here up on theaseismic --
16 the Maine Yankee review to explain.how the tea ^ went cut
, .\
l 17 together and how they selected with what was important. q 18 But basically, cover something like, it could be a 1
1 19 rron-safe -- block wall falling dawni it could be this system is i l i 20 out -- it will be covered by example in the Maine 7ankee review t
1 21 very well.
1
.I 22 But along trose same lines,. though, ~that in these i j
23 walk-downs, she fragility people and the systems people go :
24 together.and they help'eachother and thers is probably mora 25 interaction between thern than there h&s been in the past PRAs, J
)
r, t
Heri t e.]e Reporting. Corporation (202) 62864888
, 4 i
--._.-____.-.__._-._.._..-..-----.__-J
U
- 44
-', <1 certalnky more than~we ever'had in desiDn Gnalysis'or anyt h i'ng .
2- This is critical to a walk-down to have both- the fragility 3 analysts and the syst6ms analysts and also to have.the plant
? operating people. involved in the' walk-downs.
15 The effort itself breaks down into about twice'as 6 inuch ef fort ; for the' fragility evaluati~o'n, the components 7 strength evaluation, as it is fo.* the systems evaluation. In 8 contrast'to the PRA there is no seismic hazard developmen;
- S although there'is a -- you have to soecify an earthquake.- sa l 10 target earthquake.
l l 11 DR. SIESS: Wha *c 6oubles that. distri bution cif ef fort ?
l 12 Scope or the difficulty of the approach?
13 MR. GUIY: Both. I guess what we.are"looking-at
() 14 basically is plant capacity, and the systems part is toove cut-15 ond-dried than the fragilities part, However, this may:be in l'
l-16 context of PWR. In DWR tt may be a different' percentage.
17 D R. CIESS: Now, that systems evaluation effort 10 includes tP,e systems evaluation that'goes with screening?
15 MR. GUZY: There is a screen part of it-that includet 20 the levar-high systems out -- fault array. analysis. Whatey*3r 21 you do for a systems analysis. It'is what leads to a system a
22 HCLpF and a t plant HCLPF.
23 Okay, a few words'about the recults of-the margins 24 program. First of aII, it makes a statement about~seicmic core 25 melt resistance, and I say it is a robust statement, the fact' Hovitage Reporting Corporation-(202's 628-4868 l
l
7, t i' 45 1 'that I thing there is less, perhaps, argument, among the design .i cO .
2 types atrout a HCLPF than there would 'an about some roedian t core -I 1
3 melt probability with sorr,e uncertainty band on it. If the ,
l
- 4. HCLPF is:itself a lower bound number it is not a bent estimate. I 1
5 Margin wtudy does not specif'ically quant'ify core mel t ']
'l
^
6 frequencyi doesn' t give yot a meltdown . prer year pe plai
] l l
1 1 7 raumber, and it does not give you a nurober for a release, a. j 8 'r i s k , or consequences. We cannot call ite l .
9 Anc.ther possible result of a PRA:should he noted tliat 4
I
'10 it'is quite possible that when you did a PRA you may come out 'j t
~
-)
11 wit h a statement that ---
]
12 (Continued on the next page.)
13 O 14 ,
1 15 l
16 1
17 j l
18 l 13 J
20 l i l 21 l
l l 22 1
1 23 24 'l 25 :
Heritage Reporting Corporation (202) 628-4888 l
1 l
y .
'46 T4 -
1 MR. GUZY: (Continued) It seems like'to.me, it 2 seems to.me this could'be.quite likely.for a new plant at:a 3 point 13G livel. That may be your product.- -You may say,the.
4l plant-HCLJP 'Is'above point 3Gs and you,have yto identification ~
5 .of a specific weak linki or you'krrbw, t h a t ' s _ a . --
6 DR. SIESS: Dut s new plant .~ design 'doesn' t say HCLPF
-l 7 above three-tenths. It aust 'says an SSE of. three-tenths.
8 MR. GUZYi But when you design.tolan SEE ofr.3Gs, <
l I h~ .
I 9 automatically your HCLPF is poing to be much greater than that. 1 10 DR. SIESS: If you are good.
11 MR. GU7Y: If you are good. I mean if you design la accounts for all systems interaction and the effect of 13 nonsafety categories and that kind of stuff.
() 14 DR. SIESS: Well,1 now arriving at the HCLPFLypu are 15 looking at all of those' things. Now I don' t' have any ; feel . for, 16 because how many of those things got looked at in,the original q l
l 17 design, and I think if I go back f ar enouDh- a lot . of them i
18 di dn' t , are more impertant? How many of those1 things will be l 19 looked at in a new design. ;
1 20 It seems to me that one of the benefits from this 21 type of review is looking ' forward'. .That.the next "cime somebody i 22- designs a plant if they know the kind of things that came up in:
23 say the Main Yankee ce'the Catawba review, they~coUld take 24 those things into account.
25 MR. GUZY: I think if they were to' design a plant, C4. Heritage Reportiag Corporation (202) 628-4888 a
47 i 1 they would certairily look at; thinks ~ as anchorages, and I think 2 relay' chatter would be'probably.betterl addressed.- I think the 3 things we are finding from the margins program for pRAs should
.i; 1
l l
4 be better addressed.
l S. DR. SIESS: Now is there 6 mechanism for getting l
6 these concerns and thece insights to the designers of: fut ure l '7 plants?
l S MR. GUZY: My understanding is that --
9 DR. SIESS: A lot of your panel members are --~I'know .
10 about designers of future plants'as being both NSSS suppliers-l 11'- and AEs. I don' t . see a 1st 4cf AEs represented.on your' expert..'
l 12 groups, although I see a lot of people that work for AEs, i
13 . consultants.
() 14 Will this word get out?
15 MR. GUZY: I think it will get out maybe in a couple j
-l 16 of different ways, but my understanding is that 'we.are going to 17 start requiring PRAs at the plants. 'At least,- we are. thinking 18 about requiring in the plants.
19 DR. SIESS: Well, now that's a dif ferent . story. I 20 MR. GUZY: I think -- I don' t know. I think if you-21 deal with seismic pRAs, certainly you don' t have to do a 22 seismic margin review. I 1
l 25 DR. SIESS: No. right. That's a point that I' m 1
l 24 a? raid is going to come up, and we.have gotten so involved in 25
~
thinking.in terms of pRAs that every time people look at.the I O' Heritage Reporting Corporation (202) 628-4888
i 48 i
i 1 seismic design. margin thing, they try to fit it-into;to the pRA 2 picture somehow. I' m,not going to name the people, but I' ve 'l L i 3 seen-it h.appen. e l l
4 And what,I have read in some of the reports here the 1 i
\
5 point is being - made you -shouldn' t try to do that. You can not l s
? tie it back to the hazard' curves. You are not going to reduce
~7 - the uncertainty in ha::ards. You are-only talking about what i
8 the plant can resist, and you dord t know th'e probability of it 'l 1
9 ever seeing a three-tenths;g earthquake 10 But if it does see a three-tenths g earthquake.with l i
11' not ;nore than: 14 percent or 16 percent exceeds or something, it-1
)
12 has e high probability of surviving.- And'that's'all yod can 1 1
i 13 say. j
() 14 But the pRA says something different. pRA says it's 15 10 to the minus 5 for a reactor year probability.that any.
1 16 earthquake'will cause a core melt on ,t h i s , pl arit . -
17 M R. GUZY: I think the advantage -- 1 1
18 DR. SIESSr That's a completely dif ferent statelnent.
1 l
19 MR. GUZY: Well, there~are some similarities between 20 a pRA and the margins review'in that they do 'look at the 21 systems aspect and, you know, if you. looked at'Just the --t what
]
22 in the old days called a margins-review, y,ou would look at all- ,'
23 Category 1 structures and notning else, and assess some sor't of I 24 design margin or response margin or something'on that. )
25 I think we have learned that the systems part'is ;
1 Heritage Reporting Corporation ]
l l (202) 628-4888 l
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ l
i 49 1 important in design, and that would be something that I would 2- imagine could be changed, but we have a lot of regulatory. 1 i
3 : inertia to change that, trying to resist changing that. ;
4 DR. SIESS: There are two ' different ways of thinking '
I J
- 5. about safety. I keep seeing the statement, the seismic PRAs.
6 show t at the seismic risk contribut' ion doesn' t really begin 7 until you are at least twice the SSE and sometimes three or 8 four times the SSE.
1 9 Now there is no way'I'can easily relate that 10 statement of say two times the SSE before I have a seis'mic-rink; 2 11 with a WCLPF for a plant, Design is designed for what point 12 two - .17 3 and let's say the PR4 says it.took .34. It was 13 higher than that on design. I think it was closer to 4 on j i
/~* l 14 that. But I don' t know what a -- a HCLPF isn' t a median.
(_)
I 15 MR. GUZY ' I cari --
16 DR. SIESS: It's a low bound. There ain' t no longer 17 bound on a normal curve. )
i 18 MR. GUZY: I think as a working number.I, guess for. I J
19 the HCLPF versus median, it's felt that there's at least a 20 factor of two between the HCLPF and the median vedue. Now it's i 1
1 21 something that you cannot-quantify exactly, and you cannot back i 1
22 a median value out of a HCLPF value.
23 DR. SIESS: No.
24 MR. GUZY: But' conceptually you have a margin of-a 28 median of the HCLPF of a factor of at least two.
's 1 O Heritage Reporting Corporation (202) 628-4888
'I
__._d_____________.____________ _ _ _ _ _ _ _ . .
50 1 ' DR. SIESS: The median is the value that gets
- 2. involved in PRAs. . I' m either going to think medians or I' m. not 3 going to think -- if I. start thinking uncertainties, I don' t '
l '4 get anywhere.
5 MR. REITER: Leon Reiter for the staff.
6 I wanted to point out'something that.you had:
l 7 mentioned before about the relationship between HCLPFs and 8- design values, andlI think'that relationship is unclear.
9 Dr. Ventroe awhile back.-presented a, comparison of 10 HCLPFs for various plants, and: it was very obviously seeking a 11 literal relationship between the HCLPFs and the design values.-
12 In fact Seabrook was .25g, had a' lower HCLPF in some plants 13 which had the desiDn for lower values. That was explained by O 14 *e v-15 One way would be that.the design of the HCLPFs are i
16 controlled by -- in may case by off-the-shelf items which are i l >
17 not dependent upon design values.: And the other argument posed la was that those HCLPFs are not all -- are not all apples. Some 19 'of them were put together by people using more conservative 20 methodologies, some by less conservative methodology, i '
21 So 1 toin4 -- at least?from my perspective the-22 relationship between.HCLPF and design values is' unclear, and 23 they don' t -- t hey don' t fitLthe. scale literally.
l- 24 VR. SIESS: Yes, but HCLPF tends to be governed by-25 the lowest capacity. And-if you make a perfectly good design 1
1 Heritage Reporting Corporation (202) 628-4808 c_____=____ . _ _ _ _ _ _ _ . - _ _ -_ :
1 51 at two-tenths gs, then goof somewhere, and that's in a critical 0-1 l I
2 system, that's going to govern your HCLPF. If you do a design
- 3. at ' two-tenths g and don' t make any mistakes and don' t overlook 4 anything, then your HCLPF ought to be higher.two-tenths. It 5 may or may not be.
6 But I agree. I wouldn' t expect to see much l 7 correlation between a HCLpF and the. design.value for old 8 plants. But every time we find a governing vulnerability for 9 something that wasn' t designed, you had a -- as you'say, it was q l
10 wasn' t designed for. I 11 Mr. Budnitz.
12 MR. BUDNIT7.: I have another comment about your 13 previous question about new plants. I want you to imagine that-() 14 a new GSAR, a new GSAR came in the. door, a new combustion 15 engineering PRA came in the door, and that they claim that.they 16 designed at a . 3g, let's suppose, which by the way is the sort 17 of level that one is hearing about for the vendors for their 18 newest designs.
i 19 Now if you do' a PRS on a GSAR that's never been l 20 built, just a paper plan, you wil1 never' find'that the tanks in q 21 the yard have chairs that are this hiDh instead of this high. )
22 You will never find that they forgot to anchor water control -
23 center No. 6. You will never find that the battery spacers 24 weren' t put in right even though they will. tell you that ~ it's l- 25 designed, and on paper'it should be.
~ Heritage Reporting Corporation (202) 628-4888
I i
j 52 1 i
1 The lesson I' learned from that is.that having'done
~O 2 that pRA and got all.the systems, you know, they have a'certain 3 number of systems, and they are configured in a certain way and' 4 they are separated and all that stuff, that there'has to be.a
'5 commitment'to make it come true in the field when each of those ,
i l
6 standard plants, which one hopes there will be many, when each i 7 Lis built..
8 And that commitment requires diligence to get~ out 9 there and see that it's so. But.that's not.something the-10 regulatory staff. finds'in their review of that standard 11 application. It's just not findable.
12 DR. SIESS: I' m glad you said diligence instead of 13 QA.
() 14 MR. BUDNITZ: I don' t even. know what GA is,. but I 15 know what diligence means.
i 16 MR. KENNEDY: -Yeah, Bob Kennedy.
17 I think I would like to make that same point having 18 participated in two seismic pRAs on nonbuilt plants.
19 DR. SIESS: Those are safe. We know this.
20 MR. KENNEDY: Basically the margins program.has a lot 21 of advantages because you are looking at the plant as it 22 exists. And so you get to physically see. what's khere.- When 23 you do a: margins effort or a pRA before the plant is built, you 24 don' t find weak links, because everything1you think about-you- l 25 fix, so you don' t. find weak links,- and you basically,: at least O
Heritage Reporting Corporation' 'l (202) 628-4888 i I
+
53 1- in my judgment, need to do the seismic walkdown of the. plant 2 after it is built. Whether the walkdown is-part of a margins 3 proDram, part of a pRA or part of anything else you want to ,l 4 call it, it's the walkdown that is important. ]
5 DR. SIESS: But you Wouldn' t refer to the walkdown as 6 a OA or QC?
l 7 MR. KENNEDY: I have never been -- i 8 DR. SIESS: The point has been made -- no, but the l 9 point has been made in a couple of places that the seismic-10 design margin effort is not a GA/QC effort.
11 MR. MENNEDY: The reason that point.is.made in both 12 the NRC seismic margin program and the EPRI one is that we felt 13 it was not cost effective. It was not the efficient thing to
() 14 do to worry a margins program about catching every possible QA 15 or QC problem. j 16 In other words, when you do a margins program, you 17 are looking at the anchorages. But yousare not out there 18 putting bolts -- I mean putting wrenches on the anchorages and 19 testing them, because it Just seems that you could increase the ,
20 program a factor of three or four, and you would lose sight of 1
wh a't your real goal i s,
' ?
21 which is finding potentially weak links j 22 such as lack of spacers between batteries, deficiently anchored 23 tank or something such as this.
2* And so we have always said that-we are not -- we are t
25 not doing a GA/QC program because we~are not doing that ;
O Heritage Reporting Corporation ;
(202) 628-4888 !
4
_____m__ . _ _ _ _ . . _ _ _ _ _ _ ___m.____ . _Em. .__m._m __1___.___..
~1 54 1 detailed level of inspection. You are certainly doing a l
1 2 program.with.your eyes open looking for weaker links. So in, 3 some. respects I guess it is a QA/QC, but.it+s not to the level '
4 of detail.
5 DR. SIESS: QA would be look.at that. They just ~look 6 at paper. & j 7 MR. KENNEDY: That's one of the problems.
8 Your point though on'the HCLpF capacity versus' PRAss 9 the HCLpF level and the approaches to getting it were selected 10 based upon reviewing more than 20 seismic PRAs, and basica11y' 11 it was set as to be.the lower level, to be a level at which 3 12 there was low seismic -- was a low percentage of the seismic i 13 risk came from levels below the HCLpF.
() 14 And so the HCLpF is reall'y the level at which seismic 15 risk to the plant . begins, And it:tends to be maybe a couple 16 percentages of thel seismic risk would come below that. level, l 17 but it tends to be a lower bound to where the seismic risk, the 18 earthquake levels that dominate seismic risk.
19 M R. GUZY: I would sort of like to wrap up this 20 methodology talk, again it's the Maine Yankee review. Just a 21 couple more points I would like to make. ;
22 There are two -- well, first'of all, we-used 23 experience data and test data to develop tables'that help .
I 24 screen components. There are certain components we don' t have 1 25 to look at at all for certain earthquake levels, certain.D O. Heritage Reporting Corporation 1
j (202) 628-4888 !
I
i 1
'l 1 55' ;
1
~
'1 levels.- - There are cert ain cornponents that we only'have to look; 0r 2 at certain things such as acquages or'whatever. So'that thia s
t --
]
3 is part of the component GOeening. 4 4 For instance, a concrete containment you wouldn' t A
5 have to look at for all practical purposes.
]
6 There are two approeches. Doth of them were ,
7 received, or at least endorsed.by the expert panel. The
-i 8 fragility approccbr was developed.mrre fully, and this is the s ,
9 one that wn used ir.cthe Maine-Yankree reviewi Its'requirds a I 4
t
,l 10 pRA type f ragility analysis where you have got yourg wie,11an .;.
- /
11 values fragilities and your betas on yourt random miss.and.j, , .
,t I.
12 uncertainty. And it's a quickfalculasion to get at HCLPF out
\ '
t fx g 13 of that er fcR . Lat me ,3ust gice dou t?]at. T
, t
, c. .
- 'N $
) r ,. ., j T 14 Tr.1s i gyour clasoie pRA fsa' S 1ty type cuqve wqerrt l 15 J
yourmediansaluemay{ beat scor.a nutnt fer, ! riht
$ )
hepe
)' A
. !WGs.
a ' . , },) g r \
16 This slope is the rand:vi miss and tt e spr ead iridt O uncertainty-
- t <. y '3 . k. 4 7
. e
( 'g 17 or competerce limit. ( dq,o ,
'gi /*
t ,
a hji i T9 get 4 ' HC2PF qsji this egnay off curves, yotnl ould go 18
'* l
,j )
[/ ,
Nt, ' ' q) 19 to the 95. percerd compqtence, the[curvei y o dc ch! M 5 percent ,
Q 20 probabilst.yfof failureI pick"up your HCLPF values. So your .,
- t. '!
21 mediarovalhe may be .90ss. HCLPF' for ,tbJ s case wou!d be . 20gs. I
~
((
~
I 22 rhe uncertainty drives ()he p
HLLPF valud: a muchfas ;
the a, l
23 probability of. failura. And in*het in some cases where we i <
.% \
24 don' t br.ve any inforrn9t ion co mit,Swe car.rsot' calcuNje a
.1 , n 25 component's,HCLPF, You(automatically assume i t ' s . v/sy low. And 4
{. -i
~ t 4 O ' , , . , ,
I r' q Heritage fecorttinr,(Corporation A 'p,
'f (204h 62874888 /' /o .
yy} 3' ?s9g I .l k,., a .
I
/ I kp] .f > ?g( e i
.l l.
tl %4 b, , , , ( s i
~
L. u '\ _'
(..
'l 56; i 1
1 1- .you will see'this in the Maine Yankee review. "There was some j 2 lead anomy batteries we didn' t . have ' data on, ' and ' you couldn' t~
- 3. -do.anything. You either had.to assign a'very low HCLPF or -- .
4
~
an'ci i t was al' uncertainty. 'We' didn' t know anythinD about 1
5 those' batteries,1 but because of the. uncertainty. 1 6 DR. GIESC: ,You' re still' talking about uncertainties 7 now. They can be expressed as energies,'.not as powers of 10.
1; 8 MR. GUZY: That's right.
9 DR.'SICSS: When we talk about. hazard, we' re ' ;
10 probabilities.
11 So, you know, without all of.this fancy curves, what 12 you are trying to come out with is a' good low bound,- like on.
Di h- 13 engineering lower bound. think of everything you'can and say I L
I
( f; . 14 don' t really think it could ever be below this. And'if there
- 15 is some uncertainty somewhere that you 'can' t . reach a judgment s
7 16 on, you say zero, take it out or fix it.
s e l 17 ,
MR. GUZY: This is one way.to get.a,HCLPF. TheLother-
- ' i I
18
~
way is the CDFM method which we will talk about this afternoon..
Iv2 '
19'- DR. SIESS: Backing off'on a design.
1 20 MR. GUZY: Right. Based more on design calculation i
21 procedure.
., l3 c i 22 DR. SIESS: And right now the comparison of the two s
I I
r ,( 23 is in front of us.
+
p3 , s, . 24 MH.-GUZY: It's in . progress. There was'a comparison l' C '
\
L 1 25 under Maine. Yankee. We did a-few components there, but we have w j a
O' \
Heritage Reporting Corporation (202) 628-4888 L3 3 'i j
,( ~)
l
- ((
- 57 1 a separate task altogether to work on.
2 DR. SIESS: How close do'you want.the two approaches 3 to come?
4 MR. GUZY: Well, I . think that's what we will find 5 out. The question is is it the approaches or the people that i d
6 are doing the calculations. In the Maine Yankee. review, the 7 limited cases we looked at, the numbers came out surprisingly )
8 very close. {j i
9 In the study we' re going to -- I' ll talk about this 10 afternoon. In the' study we are goinD to run, we are going to l
L 11 have different organizations do line calculations, doing
! I L 12 different methods and we will find out, t
13 DR. SIESS: And if the. difference between the two
() 14 methods is not any greater than the difference between two 15 people using one method, it's probably --
16 MR. GUZY: Right.
17 DR. SIESS: -- okay on either one.
1 1
18 Bob. !
l 19 MR. KENNEDY: Having used both methods on about 25 20 components now, I' ll bet almost any amount.-of money that the 21 differences between two people will be much better than the j
22 di f ferences between the two inethods. The differences between 23 two people are likely to be rather big.
24 DR. SIESS: And that's two people on the same side of 25 the table?
Heritage Reporting Corporation >
(202) 628-4888
-l l.
e-____-__-___--______-__-__________--_--_-- _ _=_ - - _. - - - _ _ _ _ _ _ _ _ .
I 58 f
s l' MR. KENNEDY: Two people evaluating '; se sarne piece of 2 component but with no communication between:them, i.e.,.after ;
l 3 they have had a chance.to communicate and use what is sometimes 4- called the Delfi method,.I suspect-they will get;much closer 5 together.
G But doing the calculations on the.same component'with 7 no communications between the two people is the experiment that-8 Dan is going to tell you.about later here.
9 MR. GUZY: Last, I would just'like to brieflyLtalk. y 10 about the systems approaches. The screening-is based on 11 functions that were identified to be important for core melt
~
12 and PRAs.
13 For PWRs, it comes down to two main functions,
() 14 reactor sub-criticality. You know, to put your control drives 15 in or put more injection in. And early core cooling, an ECCS 16 system.
17 For BWRs, the study we just recently had_ completed, I 18 it's not so simple. There is more functions you have to look l
19 at and actually there is more mitigating events.that you have- 1 20 to look at also.
21 You have to look at both the front line and the 22 suppor' systems. The support systems'are very'important. But 23 once you do the screen you essentially can screen out all.those 24 components and structures that did not have the influence on 25 'these functions. So that if it doesn' t directly and' indirect ly .
1 O Heritage Reporting Corporation s202) 628-4888
)
'59 i 3
- .1 ' affect t hoso .f u'nct i ons, then 'you don' t ' have. to look at i t'.
s./ - .
2 That's a great savings in' terms of the amount.of work'you have 3 to do in a margins. review. . -It prevants you have hhving to look l k
4 atLevery Category l'whatever out there.
- 5. DR. SIEFS Now what'< s the significance. of early core l 6 cooling?
7- MR. GUZY: Early core cooling, you get the_ water.into l
8 .the core as soon as possible.
9 DR. SIESS: But does that mean only early? ,
10 MR. GUZY: 'Well, early in the ARC approach I think i
11 tha -- probabilistically if you have the early core cooling, 18 then you probabilistically you are going to nave the other 13 stuff you need later on. So it's a'--
J
() 14 DR. SIESS: You have enough time to get other
)
15 systems.
16 MR. GUZY: Or you have the same systems'that are used )
I 17 in both. I 18 Okay, just something that will come up this -l 19 afternoon, but the NRC systems analysis once you do an analysis '
20 uses a fault tree analysis' approach, and the difference between 21 that and the EpRI approach will be they use the success path' 22 approach, and that will be discussed in greater detail this
.i 23 afternoon.
24 That wraps up my morning discussion unless you have.
25 any more comments.
O Heritage Reporting Corporation-(202) 628-4888 1.
1 '.
l I
U 1
- )
60 -
J
'i- Okay, where are we? Bob Murray down to 3(b) on
(:)- 2 . applicat ion on Mali)e Yankee. j 1
3 MR.-MURRAY: What'I would like so do for the first a
4 half-hour this morning is give aclittle summary and overview of .
5 what we have d'one on this Mainc Ya nkee'- t ria l review. .
6 My name'is Bob Murray. My-background is in civil-7 structural engineering.from Lawrence Livermore National 8 Laboratory.
l I think we are pretty much on our agenda here. The l
9 10 way I would like to divide up this two-hour time' slot'is to il spend about a half-hour giving you.an-overview of our program, .
12 how we put the project together, the people involved, what 13 their backgrounds were, the process we went through, and then
() 14 Just h.ighlight the results, and I would like to let each;of our 15 ' systems and fragility team leaders spend about'45 minutes i l
16 discussirg the details of what was done..
17 And this will be broken down by first approximately 18 11:15 to 12 by Dave Moore from Energy, Incorporated, discussing 19 the process he went through in the systems' analysis. 'And then 20 followed by Robby, D r. Ranindra from EQE discussing the 21 fragility calculations, and then getting into more detail'about 22 the results and findings of this project when we applied it at i 23 Maine Yenkee.
24 So we have a'two-hour time slot to cover the details
~!
25 of the review. I O Heritag'e Reporting Corporation ]
(202) 628-4888
)
l 1' 1
. I
r N , }
.O f 1
.1 DR.~SIESS Sounds all right. 1 O
~
l a MR. MURRAY: Okay, does that' agenda fit Lyour 3 schedulen j
.4 DR. SIESS: Yeah, we may break;it somewhere'in there j1 5 probably.
6 MR. MURRAY.: Okay, okay,'but it's li ke hal f-bour,, 20 (
7 minutes to begin and then two 45-minute segments.
8 DR. SIESS: Okay.
9 MR. MURRAY: Why did we need this margins review?-
10 The main emphasis, 'I believe it's been summarized- in the j 11 opening ~ comments, the reason.we needed something here was to1 12 try out the methodology. We had spent about.a year.and' half,'
13 two years putting together an overall' program plan'for a' ,
O 1+ raie or or>< , eeve1 aima *8eee1 ear *e eeter iee m raie-15 in existing plants, and then putting together procedures, 16 guidelines, trial applications for carrying this 'out on plants. Jj 1
17 So we really wanted to comp 12te our research: efforts 18 by trying this out, testing it on an actual plant where real 19 problems would be surfaced. >
20 The second application of this was that the problem.
21 or the outstanding licensing issue at the. Maine Yankee plant 22 where the perception of the seismic hazard had changed. The {
23 plant's initial design level was an SSE of .lg. The staff had 24 recently changed that estimate up to .18g earthquake. So this 25 was a, side benefit. If our results, if our project could be-Heritage Reporting Corporation !
(202) 628-4888 1
l- 1
f 4
62 <
1 applied to this l'i censing action, that- wasian- extra , benefit of.
.0;'
2 the program. l1 3 I would like.to send a little time discussinD bcw we 4 put our project together. This was a research project under 5 Dan Guzy. Dan has Just reviewed the methodology. And Lawrence
'6 Livermore Lab was selected as the. central' organization for-7 . putting together the project team and carrying out this 8- analysts.
9 So I put.an_overall organizational. chart here showing 10 how we did this. As the team leader for this process, the 11 first step was to go out to bid'and select some contractors in ;
1 l
12 the systems area and the fragility area. 1 13 So in about the December ' 85 t'ime frame we put j
() 14 together scopes of work for these two broad categories. We.
'i 15 issued requests for proporal to four-firms'in the systems area, 16 and six firms in the fragility area. We received.back all four; ]
q 17 bids here, and five out'of six of the fragility people 18 responded back. I 19 We then put a panel together at Livermore to review 20 the quality of these proposals, the technical makeup of the 21 team, initiative, ideas, approaches, whatever, and.we selected l
.l 22 two teamst. one to do the systems and one to do the fragility l l
23 analysis. !
l 24 In the systems area, we selected Energy, j l
25 Incorporated, and that was led by Dave Moore who will be I O Heritage Reporting Corporation I (202) 628-4888 3
1
1
'{
'63 - {
.1 discussing what he.did in'aLfew minutes.
.(:) 2 .In the fragility area, we selected EDE' led by Robby,- -. ;,t 4
3 and he put togetnur a staff of people to assistfhim:in both the !
4- structural.and the mechanic 31 or' equipment. area.
5 One ingredient"that came with EGE,was a leak.to the 6- Squag Data Base through Sam Swan, and that'organi'zation was '
7 responsibility for assembly of much of the.information .used by.
l 8 Squag. So that provided a nice' ingredient.for our' program.'
9 We had previously set up two other panels.which.' Dan a l
10 highlighted. The expert panel had been in existence for i a
11 approx t:nately two years at this time frame. Bob Budnitz is1the 12 lemder of that panel. This panel was respo'nsible for-the ;
13 .overall methodology development for.the NRC. margins program as l h 14 well as guiding the.-- put t ing together. of the guidance and' '
15 applications document. Paul Amico was also involved in the BWR-16 study and he is in the audience here this morning. :1 1
17 The peer review group was a group especially 18 chartered to examine and wa$ch the ' application of this .
19 methodology specifically to the Maine Yankee, review. And that 20 group was also charted by. Bob Budnitz to provide some 21 continuity with the methodology development, and we added Mike 22 Bond from Sandia where both Budnitz and Bond had. essentially' .
23 systems type backgrounds. We put together John Reed and Loren E4 Wyllie who both had backgrounds in the structural and fragility !
25 area to complement things. Wyllie is also a. member of the' t
O- Heritage Reporting Corporation l
(202) 628-4888' i f
l
- 1 1
d 64 j
~
1 i .
~1 ISRAp organization._ l l
2' Jim Thomas was -added' to this to represent an ' industry J l
3 perspective _ as well as bringing some of the Squag' steering gj
~
4 committee ideas to this overall revlew. {
l 5 'So this Droup staytd in existence during the nine j 1
6 months or so of our review of Naine Yankee, anc we met with 7 this group'on a bi-monthly basis.
i 8 Let's see, we had-one interaction during this process. 'l 9 with the expert panel. That was early in'the project, about a 10 month after its beginning. We had a conference call to explain L
11 our approach, what we were going to do, to the expert panel, 12 how our systems and fragility' people were going to carry out 13 their efforts, and we 9c.t feedback from the expert panel on the j
() 14 application of this_ methodology. f 15 Our teams met essentia11y'on a monthly basis face to 16 face and carried out numerous telephone calls throughout;the 17 process.
18 Some important other links in our project here were 19 the link to the utility, Maine Yankee, and Yankee Atomic 20 through the NRR organization, Pat Sears. We were able to get 21 excellent cooperation out of the utility organizations in
-i 22 carrying out this review. And that's an essential feature in i 23 conducting a timely review, to be able to get the. kind of 24 information you need rapidly and timely. Also provide the kind 25 of question and answer service when we are trying to understand O' Heritage Reporting Corporation' ,
(202) 628-4888- ,)
a
n
.65 l' and put together systems models of the plant, gettingl rapid
.2, response from the utility is essential.
3 The utility had a consultant arrangement with Dr.
4' Kennedy to participate with them and provide support to them as 5 part of this overall review.
6 There was one other group of people, and; this is 7 called the NRC. Seismic Design Margins Working Grcup made up of 8 NRC ctaff members. At this time, at the time of this review, 9 co-chaired by Jim Richard from research and Nute. Anderson from 10 then NRR. So a number of the members of that review group are 11 also in this audience today.
12 Dur role was to keep that review group informed of 13 the process of our project. We met with them approximately
() 14 twice during the Maine Yankee review. The final meeting was 15 after the results had been in to cresent the overall project
)
16 status.
17 Two other interactions not mentioned on here was'our 18 fragility team met with Combustion Engineering. One of the 19 concerns that came up was not having enough information 20 available on the control route drives and' reactor internals.
21 So we negot i ated with Combustion Engineering. They were.able -
22 to get information for our fragility team to make some 23 decisions.
24 Another important interaction which I would say both 25 of the analysis teams as well as Bob Budnitz, representing the Heritage Reporting Corporation (202) 628-4888
66- )
'I 1 ' peer' review group, and-the expert panel, was our' interaction i
.O 2 .with the EpRI: Margins team. We. met with them after their:
I 3 -walkdowns of Catawba.
- We. picked their brains. , We tried'to !
A find out what worked-for.you~st Catawba, what areas did you i 5- have difficulty._in,-what areas started dragging and.taking a.
6 lot of' time. -And so we got;a' lot of insidhts on ways to ~ make . -
.q our project, our application of this project, the Maine.' Yankee, l- 7 1'
l 8 go along smoother. 4 9 In about October, after our first walkdown of Maine i 10 Yankee, we also met with the EPRI group and shared our
]
11 experience with them, and used that exchanue=to better define-12 how we.would do our second walkdown. So we tried to interact i
EndT4 13 with that program throughout. ;
I O 14 <cemtimeee om *"e me
- P oe >
l 15 16 1
'l 17 l
18 19 20 21 22 23 l l
-24 l
'i 25 4 j I
l A Heritage Reporting Corporation (202) 628-4888
_ _ _ . _ . _ _ _.____.______.__________.____j
1
.i 67 q 1 .M R . MURRAY: ~Just to h'ighlight'the'overall process, Q
2 the seismic margin review is a process which~ takes about nine 3 months from the time that you actually start. The starting.
4 process is to Select an eerthquake for1 review.g And this;is a-l 5 review-level earthquake, not a design earthquake that assumes l
~
l 6 that the plant has been designed with conservative design l
7 approaches.
l 1
i 8 The first. major steps are to gather information, and 9 system teams gather informat' ion about the import ant 1 syst erns, 10 the systems that we want to include in our review. . The 11 . fragility team gathers information on broad classes of 12 components so that they can make decisions whether.to include 13 these components in further analyses, whether'they need to gain O 14 "e e ** " *"e- r ~" ether *"ex "oe2e ecreem *"e- "*-
15 So this part of the process takes about a month or 16 two to get as much information as you can before you set foot 17 in the plant.
18 The first walk-down for our plant took place about 19 two to three months after we started our prograro. We coupled 20 this with a walk-around with the peer reviewers. And a five-21 member peer review . group participated in the actual walk-downs 22 of the plant in both the first and second phase. So we felt 23 that it was an important ingredient, an important link-that 24 reviewers were actively looking at the plant and looking at .
25 components, pointing out items, discussing areas.that were'not O Heritage Reporting Corporation (202) 628-4888
<. i .
- ,7 g-t -
68 l 1
1 1 .of-difficulty. ;
(:) <
& And these eteps.went onlduring-the August, September,,
1 3 OctoberLtime-frame where the'systemsLpeople were improvingiand-4 coming up with their event-free, fault-free model of-the ]
5 systems. The fragility ar!alysts-were refining.their. data, ')
-)
6 identifying components that would need additiona1'information, j 7 Much information at this time, was. pained by interaction with !
j 8 the utility.
9 . Finally, we did.a second walk-down'in the November -
10 1986, time-frame and this was a' shorter walk-down and again, 4
l 11 the neer review group participated in this. phase of the H l'
12 project.
l 13 And: finally,.we put everything together and came'up .
l l
() 14 with some overall estimates of the plant level. That is the 4
15 process. We were able to carry this process out in a'9-month- 1 16 time-frame to meet the NRC requirements.
)
17 I am going to discuss some of the limitations and j l
18 assumptions that are specific to the Maine Yankee Review. The 19 first one is that we started with the review level earthquake 20 of 0.3G and we anchored a new rate, 0098 Newmark and Hall-21 spectra of 50% tunnel spectra to that value..
I 22 And we interpreted this review level earthquake as a 23 uniform hazard spectra, at the 84 percentile non-exceedance 24 level. ]
I 25 Throughout this' project, every time we have had a i
i O.N Heritage Reporting Corporation-1202)'628-4888 <
~
u, ,.
69
- l' p' r oject meetir,g,. we discussed' and : argued aboutEthis review P level etrthquake and its meaning. So one of the areas:that we-3 would like to see further guidance, is to.better, clearer 4- uefine what this earthquake means up front, and be done with 5 it.
6 Another important assumption was that we~were using 7 in-structure spectra, and existing seismic structural _models of 8 the Maine Yankee plant. 'The assumption was to assume.those _
i S were adequate. We did spend quite a bit of time reviewing; 10 those models verifying their adequacy and we decided that they--
11 could be used fine in the' evaluation. I l
l 12 The point here was that we did not want tx) go in~and 12 remodel the structure and perform new analysis on the i 14 structure.
15 D R. SIESS: What do you mean by the in-structure 16 spectra? Is the spectra calculated for - - 4 I 17 MR. MURPHY: The floor. spectra,.the spectra at: floor 18 levels used for the analysis review of components at diffe' rent !
19 floor levels.
20 D R. SIESS: You sa'a d, eNisting in-structure spectra? !
21 MR. MURPHY: Right, our asnumption here was that the 1 22 models in the in-structure spectra generated from the Maine
)
23 Yankee existing models, could be used in our probings. ;
24 DR. SIESS: Okay the models for those that ex.isted 25 from the time that the plant was designed --
O Heritage Reporting Corporation (202) 628-4888
y ,
o q
70 1 MR. MURPHY: Actually they had --
O 2 DR. SIESS: Dr .from the ' 82 . review?
4 3 MR. MURPHY: 'Actually they had more recent models: ..
4 from their reviews.
i 5- DR. SIESS: And-the spectra'were usinD those models l
i 6 and the three-tenths earthquake? l 7 MR. MURPHY: The spectra that were generated were at j 8 the -- I believe it was the .18G. number that the Maine Yankee l 9 people were using.
9 10 Is that --
11 DR. SIESS: The I ~ don' t understand how you could-12 use the .18-G spectra with a .3-G earthquake?
13 MR. MURPHY: The spectra.were scaled.
() 14 DR. SIESS: Oh, they were scaled.
15 MR. MURPHY: What we wanted to get was-from the i
16 existing models and the spectra that were generated from those il more recent models -- they have;a shape and we then scaled 18 those shapes up accordingly. But what we wanted to avoid was a 19 complete reenalysis project where we would start from 20 essentially scratch and come up with models of the structure, 21 and analyze them.
22 MR. KENNEDY: I am Bob Kennedy, and these were new.
23 models, they are not the old models. They are new models and 24 the analyses of floor spectra were somewhat more conservative 25 than what you would like to have for a peer margin's review O Heritage Raporting Corporation (202) 628-4888 I
f s
n <
71
_g 1 prograto and so the use.of them introduces some.. slight-
-'d' 2 conservatism to the HCLPF Ostimates but it is slight.
3 <
And furthermore these in-structure' spectra really did
~
4 not end up governing the HCLPF level of the: plant. IfEthey 5 had, it would have led;to some'slightLconservatism.
i But they-6 are' good quality new models.
7 DR. SIEG3: This'is:a rock site?-
8 MR. MURPHY: Yes,,it is v rock site. If we need to 9 re-run this, the' schedule would be impacted and so on.' -
10 Another area was that we did not do any. walk-downs-l l -11 f.nesoe of containment. If we.went inside' containment, we_would 12 have looked at the control rod drive mechanisms, as well as l 13 some of the impulso lines. What we did do WAS'assurrG a small '$
14 break LOCA and we did not spend time incido ofJcontainment at 15 this review.
16 'We used conservative methods l'n our' approach.and the
- i. 17 HCLPF calculations procedure has been discussed. We used the l
l 18 fragility analysis method for this program and the reason'for 19 that was that our fragility contrsetor had' used ' it on a number 20 of previous PRA applications and they felt the most comfortable 21 with that calculation on that. We want to be able'to use both 22 or either CDFM or fragility analysis method.'There is'not'hing 23 uniquely tying the NRC program to the use of the;. fragility 24 approach in calculating a HCLPF.
25 But we would like to think of a HCLPF as if we'put a
+
Heritage Reporting Corporation (202) 628-4888 1
l
4 ' .
i i
, 72
.i f _
1; group of engineers together, instead, we1wouldflike you to come j I
2 up with a consensus estimate of the capacity of this componer/t'. j 3 The'HCLPF is a number.that, those knowledgeable engineers could j 4 reach an agreement. There is'nothinD.in it, that saysibhat.you i
5 must have~a fragility curve, and it mustocome from that. l
~l That is how we look at HCL.PF of lower.' bound estimate
^
6
! 7 of the seismic capacity.
l 8 We based our calculational'results, our fragility j i
9 results on the modified and upgraded components, that came.out.
10 ab c result of this review, rather_than going back and 11 analyzing the plant, prior to this review. i l
12 A number of modifications were deemed -- not deemed, j l
13 were decided on by the utility to be'made'at their next
() 14 refueling outage, so that we took advantage of. that in our
.t 15 overall review, 16 'We did not include aginD as looking at the plant in 7 the "uture. We took a snapshot of this plant in the-1986 time-l 18 frame, so at that po'nt, i the equipment had certain age to.it, 19 but we did not say, okay, in 1990, the'HCLPF would be reduced i
1 20 by this much because of aging in the future. So essentially a 21 look at the plant at the time that of our review, was what was 22 done.
l 23 DR. SIESS? The next ?o the last item, as I I 24 understand it, there were components to which you_could not l 25 assign a HCLPF value?
l l
Heritage Reporting Corporation (P92) 628-4688
73-
'1' MR.' MURPHY: Correct.
O 2- DR. SIESS:- So'there is.really no way that you could 3 tell us what the HCLPF was as found.
4 MR. MURPHY: 'That is correct.
5 DR. $IESS: You are really-giving i t ' as 'l e f t '. -
6 MR. MURPHYr As modified.is what we did.
7 DR. SIESS: As monified.
8 MR. MURPHY: Yes, vne had a' number of components 9 exactly like that, that if they were not modified or identifieo' 10 to be modified, by the utility, we would spend many, many man-11 hours- arguing over what' the HCLPF eapacity' of that item was.
12 The approach by the utility was, let's J ust . upgrade -
13 it, or replace it, and a. void this uncertainty.
l () 14 DR. SIESSi HereIthere any low values as found that 15 there wac_not an argument about? They were just clearly low?
16 MR. MURPHY: Well,, we spent a loticf time arguing.
l 17 about tanks, so th9 ones thab'would come to mindi I would say, 18 is probably this anchorage on the d_tesel-day tank. Once it was 19 identified, tnere was nc argument. The decisien'was made to 22i improve that anchorage, and osv61d the issue.-
21 DR. SlESS: It was 'learly c low?
22 MR. MURPHY: Yect it was the anchoP_ bolts were 23 shorter than the drawingo had indicat e;L 24 DR. SIESS: Was it lower than cne-tenth?
25 MR. MURPHY: We probably would argue about that one.
Heritage Reportir.g Corporation (262) 628-48G8 l
L - _ _ - - _ _ _ _ _ _ _
y _ __ _
j j
74
, I
'0dr'gues9 was po~sibly, but there probably'would'be' argument-1
]
a from some.of the Maine Yankee and'their consultants on that 3' issue.
1 4 DT.. SIESS:_ Well, that.would be a vulnerability _or an 'I 5 outlier.
6 M R. MURPHY: The~other'one in-that category'iu this 7 transformer. This was a floating bus-bar in this transformer, 3 and we traced that back through the General El'ectric ,
9 organization and we found now, for transformers' applied to
]:
10 plants in the west coast they have a bracket attached to that ]
)
11 bus-bar to prevent motion and possibly shorting of that.
12 Our feeling here, is that'our review team felt 13 possibly below . 1, but again, we would have probably sharpened q
() 14 pencils and got into arguments and spent a lot of hours 15 deciding if it was .09 or.08.
16 Again, the decision was made, let's not spend our 17 time that way, let's - j ust modify this component and reduce the I
la uncertainty in that general area. j 19 MR. KENNEDY: Bob Kennedy, and this is the same kind 20 of a thing that has happened in most of the seismic PRA's that 21 have been done. It is the real advantage of the walk-down.
I 22 You find items that you would spend a lot of time 23 arguing about, and for instance, this transformer, has been an l 24 issue of Loring Wylie,,who is on their peer review, and my 25 issue as a part of the SRAP issues of PRA-46, because we have
~A k/ Heritage ReportinD Corporation '
(202) 628-4888
+ 1 l
)
~7 5
-1 .been very concerned about transformers'not.,being anchored.
1 0~ 2- inside of their cabinets, they exist all-over the place.
3: We cannot find any failures of such. transformers, in
'4 the earthquake experience. data; base. We are still. worried.
5 about it. There is no way o'f' knowing what the seismic capacityL i l
6 'of that kind of a transformer-is,.because there is no evi~dence I
.: l 7 of seismic failures of those' transformers in the' data base, !
B But Loring Wylie who was on a peer review on one 9 side, and I, who was a consultant to. Maine Yankee,.both felt 10 rather-strongly that this transformer should be restrained 11 inside the cabinets, and the utilities simply found from GE how 12 to do that and fixed it.
13 And those are the thinDs that a mar. gin review needs I
(}
~
14 to look for and I think that it would bog down if'it Was trying 15 to assess, bog down in tremendous argument'if it was trying'to 16 assess what the capacity was before the fix, even on the diesel 17 fuel day-tank, the expansion bolts were not long;enough.
18 But it would be very argumentative as to what the 19 capacity of that day-tank would be. I personally think that it 20 would have been well above the SSC level with.the expansion-21 bolts the way that they were. But it is a trivial fix and if 22 they had not been fixed, they would have been an important ;
i 23 contributor to a lower HCLpF level for the plant.
24 DR. SIESS: The problem.is, that if you are an i 25 optimist, you look at these things and say, after the walk ;
( Heritage Reporting Corporation (202) 628-4888 4
1
- = _ _ _ _ _ _ _ _ - _ _ _ _ - _ - _ _ - - _ - _
76 3
l' down, after the review,-whether it islPRA'or seismic margins,.
2 we found some things that looked like.they needed fixing and 3 some of them, we knew that they needed fixing and now, we have 4 .got a pretty good plant.
5 lAnd if you are pessimistic, you look at a -- every G one of these things that we looked at, we four.d some things 7 that needed fixing and so'all of the rest of these plants out 8 th'ere, must have things that: need. fixing.
i l 9 MR. KENNEDY: .Every seismic PRA I.have' participated.
10 i n, we have found some things.
i 11 DR. SIESS: Yes, butnyou'see, it is not as clear cut '
12 as saying, they needed fixing. What you said is,' that once you 13 look at it, you know, it is.probably all right, but I cannot
() 14 prove it, it is easier to fix it, let's fix it and get it out 15 of the way.
16 MR. KENNEDY: Yes.
17 DR. SIESS: But we really don' t know how bad those 18 were.
19 MR. KENNEDY: That is right, we do.not know. That is 20 the problem, we don' t know, you know, in -- )
21 DR. SIESS: See, once we have done it, we have a fair 22 amount of confidence at this new level. If we have not done 23 it, we cannot automatically say, gee, that plant actually,that 24 nobody has looked at, has probably got a series of micro-bolts 25 missing underneath a relay cabinet and the damn thing is going.
l l
l-(:) Heritage Reporting Corporation (202) 628-4888 l
a' [
?
f7 1 to' fall down.
- O 2 MR. KENNEDY: Within my experience, the' utilities, i
3' once these issues are pointed out to them, _ Jump right on, j 4 improving them. Maine Yankee was a particularlyngood example 5 that way, in that they were extrernely cooperative and obviously .
I 6 they fixed, at least in ray opinion. some mors things than were i
7 needed to. l 8 DR. SIESS: But for plants that have not been looked ;
I 9 at, we don' t know how bad they are and'how good'they are. 1 10 You know --
11 MR. KENNEDY: Most plants should have some kind of a 12 seismic walk-down.
13 DR. SIESS: I guess we have got some plants where ]
() 14 they did not even design-in anchorages and you could check that 15 by looking at the drawings, but there must be.a lot of plants 16 out there, where the drawings showed anchorages and they are !
'j 17 not there. Like we have found them on more than one case on a J i
18 plant that had been walked-down twice and the third time,.they 4
19 found some bolts mis; sing.
- 20. MR. KENNEDY: Clearly walk-downs have found anchorage 21 missing on important components and it is a very limited 22 number. I mean it is not a prevalent feature in these . plant s, 23 but it does exist. They have been found in walk-downs.
24 DR. SIESS: You have done PRA's, does the PRA get 25 down to that kind of detail that if I tell you that this O Heritage Reporting Corporation (202) 628-4888
- _ j
78 1 electrical cabinet does not have an anchorage, you can tell'me-(:)- 2 what its contribution to risk is?
3 MR. KENNEDY: No, the PRA's that lI have been' involved 4 i n, if you find a cabinet without having an anchorage, the 5 cabinet is anchored and then you do the PRA on the anchored l 6 cabinet. In most'every PRA there has been something done to l
the plant, before the risk numbers were repceted.
7 !
1 i
8 DR. SIESS: But suppose that you were asked to do the l 9 PRA assuming that.it was not anchored,_do you get to that level' 10 of detail in the PRA?
11 MR. KENNEDY: I think that you would be guessing as 12 to what the capacity was. I would be very uncomfortable, in 13 fact, making a fragility. estimate. It would have very large
() 14 uncertainty and there would be the possibility that it was_ low,_
15 but there would also'be the possibility that it was not low.
16 It would be very uncertain.
l 17 DR. SIESS: Now, what comfort can we get, assuming 18 that we want comfort, or deserve comfort, from experience, in 19 non-nuclear plants that have been through earthquakes, with 20 similar kinds of equipment on them?
21 MR. KENNEDY: You can get the comfort from this type 22 of equipment that there has been rather few failures in non-23 nuclear plants. The majority of the failures don' t have an l
24 anchorage failure. And they have been eit'her missina anchorage !
l 25 or deficient anchorage. But there has also been a number.of )
a 1
i i
O Heritage Reporting Corporation ,
l (202) 628-4888 I
i i
l w_-____--_-___
79 1- cases, where you have-h'ad deficient anchorage or missing O 2 anchorage that the equipment.did not fail. So, I mean, it-is a 3 very mixed bag, when you start looking at deficientlyfanchored 4 or unanchored equipment, it is a really mixed bag as to how 'it 5 is going to perform in an earthquake.
I l
6 D R. SIESS: But what you are saying, then,,really is 7 that even with 15 or 20 Seismic PRA's and even if we had 10 or 8 15 seismic margin reviews that it still'would not tell us very 9 much about the other plants.
10 MR. KENNEDY: Not if they did not have a seismic 11 walk-down.
12 DR. SIESS: But did any plant that has been through a 13 seismic PRA, or seismic design margin. review, that prior to
() 14 walk-down, we feel a lot more comfortable about-it, 'l 15 qualitatively if not quantatively? i q
16 MR. KENNEDY: Yes, I think that you could gain.that 1
17 comfort though, just by the. walk-down that you would notineed I 18 to throw the bells and whistles of the margin review or the PRA 19 in .
20 Now, the margins review concentrates more on 21 equipment than the seismic PRA's. A better walk-down has been 22 done as a part of the margin review programs'than was done 23 under most of the PRA's. So that there have.been exceptions 24 where this same level has been done.
25 DR. SIESS: Has anybody thought about howLyou can I
Heritage Reporting Corporation (202) 628-4888 I l
1 1
-1 l
80 1 formalize a walk-down and.a result of'a walk-down, in. terms'of u 0: 2 a report or something, without having it as a part of 'a.pRA.or 3 ,a seismic design module?
4 MR. RAVINDRA: 'The generic'implementationfprocedures.
- 5. for A-46 was an attempt to do that.
6 :MR. KENNEDY: A-46 will' accomplish'that but then, A-i 7 46 is likely:to build up into a very substantial documentation 4
8 effort, and'a very substantial bolt-checking ~ effort'to.where it 9 may become, by the time that it gets finished, it may become-a 10 rather substantial proDram for a p1' ant, which has me' worried.
11 How do y'ou accomplish a good documentation, and. keep 12 the costs down, so that people are highly encouraged'to'have it-13 done on their plants, is, I think the real issue.
() 14 How do'you Det the. cost efficient walk-down on all 15 plants?
~
16 DR. SIESS: I see-your point, but A-46 does.have.a 17 formal walk-down procedure, the walk-down is essentially the 18 point of the thing.
19 MR. KENNEDY: Yes.
20 DR. SIESS: But you don' t 'want to QA it to death.
21 MR. RIVANDRA: Although I share Kennedy's concerns 22 about this degenerating into something much more detailed.
23 DR. SIESS: Try to plan it so.that you could get-24 results without having to put them on a computer.
25 MR. MURPHY: I would like to make two other points on
. Heritage Reporting Corporation (202) 628-4888
j 81 I
- i
~
'l this slide right-here. . One of them . is' that during' our; walk-O.-
2 down we had' broken it'up into four or-five differentLteams ' '
a 3 mixing systems and fragility people and adding NRC staffers, 4 and adding peer. review people.and we all walked through the
-5 control room and we.all walked through at different.' times of j
)
6 the day.
7 lI walked through- this area about .930. a. m. , and aust 8- looked at some security . lights, some work-sheet lights sitting i
. . I 9 on shelving. Just made's little note that it would be nice to l
..I 10 have those tied down.
11 Another group walked through that same area,'about 12 12:30 that same day and we had a post-day meeting, that evening ]1 13 and we got into an argumen' about whether these were tisd or
() 14 not.
15 The process of tieing-these down,,went on between J
16 9:30 and 12:30 and during our review, the. state:of'theLplant J 17- was changing.
.i 18 Again, this was an area, that in our general terms, 19 we call vulnerabilities. But when we say,.vulnerabil2ty, we 20 really mean areas that also have high uncertainty, in our j 21 terminology for this project. Your definition, I think, is a l 1
22 better one, we could not really.say that tied down sneurity, ;
23 lighting 1-s going to be a core-melt contributor.-
24 DR. SIESS: That was emergency. lighting?. )
1 25 M R. MURPHY ' Emergenny lighting, yes, in'the control O Heritage Reporting Corporation-
i (202) 628-4888 l '
'I i
82 1 room and then-throughout the plant. 1t turned that the tie-down' L 2. ~ bands had been fabricated prior-to our visit'but it just seemedf
- l 3 Ilke a good time to do it.
4 DR. SIESS: Let me-Just mention something that a i 5 couple of ACRS members had noticed in the walk-down of certain 6 plants. We saw several notebooks containing. emergency operating 7 procedures simply sitting.on a shelf.
8 And we fiDured that it would not take much of an 9 earthquake for one of those notebooks to end up on.the floor, !
10 and if they are like the notebooks that I get, by the time that 11 they gut through the U.S. postal: System, they have already been i
12 opened up.
13 Does anybody look at little things like that?' -
l
() 14 MR. MURPHY: I don' t know in this plant. I have an '
l 15 exact example on the Savannah River Reactor where all of the 16 emergency procedures were on a shelf that would spill and u l
17 possibly become fuel for a fire. '
l 18 MR, CUMMINGS: Garth Cummings and th1A is a much more 19 Deneral thing as you well know about. human factorniconcerning 20 seismic which there has been very little work done;on 2t.
21 I think that'is one issue, the whole business of how 22 the operator is going to respond followinD an earthquake that 23 has really shook him up. This particular program, we really 24 did not specifically address those issues. I mean, that'was 25 one of the things that we identified up' front in the program-O. Heritage Reporting Corporation (202) 620-4B88 l '.
6
, y ,
y ;r. >
- 5
+
J ,
,. 17 l l;
- 4
, j. % ., ' r .i Je 'r' ? t ,. < L3 -
y ', ,
'N i f i plan, O 2 But I think that it is stil10$.n outstanding issue.
.t 3
? . .
, i.
~'
3 DR. SIESS: I have serfn'plarts that did it quite f /
i 4 : differently. ;7 5 Mh KENNEDY: Bob Kegredyc on th h security lighting <
3 J ,, ,
6- tie-down, Hifact,' Bill Henr{:ls from well, . Maine YQkee and I' ' j 7 walked through the plant the day before the walk-down teams s 8 entered the plant, which was#a Sunday that de werA throu[ih it, -
i,
' . ,/ J 9 and we saw.the unsecured lighting fi xt u p%Mnd I ' acked ,him to 1 t- t
) 's
- f 10 secure them. , g. /; I h ! j 1.1 And they were being secured 'et the time that the 12 walk-down team was going through the very' rwt day. . Jt is not- .
13 a safety issue, but it is an example of/ ttdd ,i f .[yo'u c an get ,
F 1 14 these things-pointed out-to the ut ility, 'ify46 ! nave got a 15 cooperative utility that quickly fixes dned,/12 is not a s&fety 16 issue, but it is an issue that alh of these emerp5ccyM igh' ting
<j ,
17 are of f their bracktets and on the floor, khef dith not really t
a f ~j >
18
~
serve their purpose after an earthquake.
j A. I 19 And the probi pm tha,t ' I ses,in it,' is that almost 20 every plant that ! Adik tbrough, they kare not - t led down. There l'
has been never any requirement for ;;/ hem to be t ied c'oan,' Unless -
21
) ,u, they would hit some safety equipment $ I 22 [ 3,
'23 MR. MURPHY: They are undit11y sitting Just 'on a'Iflat 3
\
- /' y , ; '
24 /.
horizontal plate, up high on the weill with no positive
'f} ,
+
Lb 25 connection. j 1 /3 i
,/
O Herita29 Reportir.b,'r Co'rhoratica -
s j (202) 628 ,4888 .p j 5 L r ', ' ! \
A 4 ,
y -
, .i-i .$
P- ,1 ,* '
- _ _ - - _ _ . _ . _ _ _ _ _ _ - _ . - _ _ _ _ _ _ . _ _ . _ _ _ _ _.L..._ _ _ ..
'k <
Q-s
'" 84 1 MR, KENNEDY: Just'high up on.the wall'over doors.
1O I will.have to go;and look at the ones.
2 DR. SIESS:
3 out ipere.
4 h MR. KENNEDY: I suspect'that they,.are not;; tied down i 'S either.
1 6' MR. REITER: -You want them to.be fairly free-because 7 they take them out to charge the batteries occasionally, so l
(3 that tha device that was used at= Maine Yankee.was a strap that 1
F , ,
9 is easily openable to change and replace the light, but.once it 10 is in place, you have a positive connection to prevent it from
. < ./
- 1. t moving off the shelf. It is'a very good design. ,
1
- '1E MR. KENNEDY: I think that the key is to get the i l
- i. i 13 plant operation people used to' thinking seismic and then they
) 14' catch these things themselves, because I know that is what is )
i 15 happening at Maine Yankee, they are paying attention to these i
'416 things and catching them.
1
[ f17 DR. SIESS: You know when you are talking about 18 thinking seismic, a lot of people think that: seismic is shaky, r >
19 And I think that seismic is pushing..I think that it makes a 20 difference.
21 MR. REITER: Chet, the ones in the hallway here, are 22 very solid.
23 DR. SIESS: Well, they cannot be gotten to, to 24 replace the review, t 4
., [<25 MR. MURPHY: The issue of this ACE battery also came l .'
t Heritage Reporting Corporation (202)-628-4888
, 'f j * >
L3 < , V - - - --
85 1 up during our walk-down. The batteries at the plant were late 2 anti-monte. batteries (sic) that had been in service for a.
3 number of years and each time that they were tested 4 electrically they functioned fine. And our fragility team 5- identified these as a potential problem, the problem being that 6 in the aged led anti-monte batteries, .the internal plates 7 tended to become brittle with time,. and :an an earthquake 8 occurred, those_ brittle plates may then break.
9 Lots of effort was then spent'trying to find data on 10 age batteries here going back to the battery. manufacturers, as 11 well as tryinG to gather from experienced data-base.that 12 nothing was obtainable so the utility made the decision to' 13 first upgrade first the batteries one and three,.which were
() 14 deemed by the systems analysis to be the safety-related 15 batteries, and secondly, they decided durzng 1988, their next 16 refuelinD outage, to also replace the other two batteries 17 during the process.
18 Rgain, this is an area where --
19 DR. SIESS: Replacing them with what?
20 MR. MURPHY: They were goinD to use. lead-calcium 21 batteries? Direct-fill?
22 DR. SIESS: Why did they use the lead-antimony in the 23 first place?
24 MR. KENNEDY:- That was all that was available in 25 those days.
Heritage . Reporting Corporation (202) 628-4888
86 1 MR.. MURPHY: So these are-your-original batteries?
3O 2 The original' batteries had'been in service forJ10 - 12 plus 3- years.
4 The'last point I want to make on.this slide,'is'that 5 we probably spent of our overall resources maybe 20 percent of' 6 our effort reviewing large, outdoor tanks. And the? analysis 7 .are on the order of 70 to 80 paDes. The; reviews'are quite B extensive.
9 And this did' turn out to be what.overall' governed:the:
1 10 plant HCLpF. The plant HCLpF at the Maine Yankee plant,.
11 assuming the small LOCA, came out by our calculations'to be 12 0.21 Gi the dominant component was that'large, outdoor 13 refuelling water storage tank. The ability to site it during
() 14 the 1988 upgrade to replace and improve the anchorage to bring 15 the capacity of that plant up to .27, they later decided not to' 16 do in 1988 but to do it in their outage this year. I 17 So that anchorage has been upgraded. The point'I.
18 want to make here is, we all feel these are.very conservative 1
19 lower-bound est imates' on the act ual capacity of that plant, but 20 the analysis capabilities today for the large tanks do not
]
21 allow us to go any further than this. So it seems to me it is 22 an area that.we do want to point out that some more work needs .
i 23 to be done in better. understanding the large anchorages'and how 24 .to better get at their actual capacities. l 25 DR. SIESS: Are you referring to those as large, O Heritage Reporting. Corporation (202) 628-4888 i
1 1
87.
1 outdoor tanks?L What makes 'thern different. outdoors than
.(:) 2 indoors?
3- MR. MURPHY What I arn saying i s, I guess the major 4 . difference is'it would probably be on the. indoor tanks on the 5 order of ten . feet or lessLin diameter, and rnaybe a skirt or leg 6 mounted, where we are looking here at a tank that is 40' feet l
l 7 high and 35-feet or so in diameter, sitting on a concrete; ring 8 wall outside.
9 DR. SIESS: But it is not the ring. walli that rnakes 'it 10 vulnerable?
11 MR. MURPHY: The problem here is was that the chairs i
12 and the anchor bolt detailing, and possibly the fact that --
13 DR. SIESS: What I am getting at --
() 14 MR. MURPHY: was today we would have more anchor 15 bolts.
16 DR. SIESS: Simply building a building around that 17 t ank wouldn' t have helped?
18 MR. MURPHY: No. I think that tanks designed by 19 today's methods would calculate much higheri would be above our-1 20 review level, wouldn' t ' be a problera. I 21 Tanks at older, existing plants, where future margin i
.22 studies may occur, are going to, I would think, find-problems -l 1
23 in the tank ar eas.
24 The point.is, we feel these are possibly." staffer" !
?
25 problems, not real problems. The anchorage.was incurred to get.
1 O Heritage Reporting Corporation I
! (202) 628-4888 '!
l a
1 I
88 1 rid of the uncertainties in this area.
.2 The final point.here is.for our de novo assumption we q 3 were able to show that this plant had an overall capacity than 4 our review-level earthquake. So in our LOCA' situation, a small l 5 LOCA dominated the plant leveli HCLPF came up to be.2.1_by our 6 analysis team review.
7 Part of the overall process is.very.important in this 8 whole study, too. We learned a lot of things about.how to do l 9 it againi how our methodology could be improved by applying l I
10 this procedure tonan actual plant.
11 And I listed some of the items here, the definition 12 of a' review-level earthquake, should be firmed up up-front. We l 13 feel that would probably save a lot of time and discussion
() 14 during the process. I don' t really think that the .3'and NUREG 15 0098 spectra definition would changei I Just think'we would 16 spend less time discussing what it means.
17 DR. SIESS: Are you going to go into that further?
18 MR. MURPHY: I am going to let Robby go into that, 4 19 and then, I am sure Bob Kennedy will have some words to say on 20 that level.
21 The methodology report and our guideline, we see some l 22 areas that'can be strengthened. Some of these'have already i 23 been incorporated by conducting'a review of the EPRI 24 methodology and then modification or suggested modification to l 25 some of their fr80ilit'y screening guidelines.
l l Heritage Reporting- Corporation 3 (202) 628-4888 i
l
1
'89 1 So these are areas that, if we revised our guidelines.
2 criteria, we would make modifications. The use of 3 qualification data we felt was a very important ingredient in 4 this overall program, beinD able to Get our hands on some of' 5' the squad' data and use some of that information enabled us to i
6 screen out a number of components in this process. So it was a
]
7 very important ingredient that we used to assist us.
8 DR. SIESS: That was screening beyond the-basic-l 9 screening in the other cateDori7s?
10 MR. MURPHY: Right, i
11 MR. MURPHY: -Walk-down procedures, we were able to, 12 by interacting with the EPRI team, and by working closely with 13 the utility, we were able to streamline our walk-down and make
() 14 it very efficient: avoid lots of dead time while we were 15 waiting for health physics briefings and body countings, so we l
l 16 were able to make this a very efficient procedure. And the 1
17 secret there is good utility interaction good planning prior 18 to when our team arrives and planning on the utility's part to ,
19 inform their people that there is a team coming aroundi they' re 1
20 going to be taking picturesi they' re going to be asking '
21 questions -- please, plan according~1y. So that has been an 22 important aspect. l 1
1 23 We found that we would be able to perform this review :
24 on a future plant, a. future pressurized water reactor.With l 25 about a 2.5 man / year effort. This took about a half man / year !
1 O. Heritage Reporting Corporation (202) 628-4888 ,
1
' '90- l 1 ' effort'on the part'of the utility supplying!the drawings, the' ,
2 information, the stress reports, t he ' calculat. ions, and l
.I
- )
3 answering questions that we would need, and we. estimate about :
l 4 1/2 to one man / year ~ effort in the peer review process.
l 5 Prob' ably one man / year of effort was used here because we were l 6 learning, trying t'o. understand the approach.
l 7 So we feel that the peer review .is a'very linport ant '
l 8 ingredient for a future margins review, and somewhere in the 9 order of 1/2 to 3/4 man / year'is a reasonable estimate for. ,
I 10 ' conducting that peer review.
] 1 11 Finally, we feel that this was a successful review.- j
')
12 We feel that we successfu]1y completed the trial margin review 13 applying the NRC methodology guidelines to the Maine Yankee
() 14 plant. We were'able to perform this with multi-disciplinary l 15 walk-downs integrating the peer review people with our analysis l
16 teams and the utility teams, and learned a lot about.the plant.
17 This interdisciplinary concept is really vita.
18 Bob Mennedy's point' that this.is a very useful 19 benefit out of this margins program.
20 We identified. vulnerabilities ~and, again, we 21 stretched the use of that word to mean items that also have l i
22 large uncertainties -- not really called " vulnerabilities," but !
23 in our report, in our documentation, we use that word for both.
24 So we identified a number of artfas that the utility decided.to' l
\
25 upgrade, and we feel that was a positive. benefit. l l
( Heritage Reporting Corporation (202) 628-4888 .
I i
1 n '- '
__--___n_ _ _ _ -
g ...
l91 1 Also this review was used in the. licensing arena and !
(:) 2 a SER was prepared' based on our stud'y,- and was issued in the. ,]
l 3 March - early April t'ime frame to; resolve.any outstanding 1
4 seismic issues at the Maine Yankee plant. ;
5 DR. 51ESS: Now, the vulnerabilities that you ' found, 1 l
6 were those in the. systems you looked at -- in other words, you 'l l
7 screened out --
8 MR. MURPHY: Generally, yes, but we screened the-9 security lighting was an' example of sort of a broad, general l -
l 1 10~ vulnerability.
l 11 DR. SIESS: What about other systems, though?
12 MR. MURPHY: We extended our methodology a bit to 13 look at some of the not-Group A systems that would be needed
(} 14 for long-term cooling,that we identified some fans'that, based 15 on our methodology, when applied by the book, we would notLhave 16 examined. So we stretched a little bit.
17 DR. SIESS: Let me ask you specifically. Suppose ,
1 l
L 18 there is a seismically-induced failure in the fire system that !
l 19 blows out somethinD important. Would that have been gotten i 20 here?
I 21 MR. MURPHY: Yes. We have almost exactly that ;
l 22 example. In a diesel room. 4 i
i 23 DR. SIESS That would be an interaction? ,
24 MR. MURPHY: The systems interaction -- during our ;
~
25 walk-down, our systems people pretty much kept their head up- !
l Heritage Reporting Corporation (202) 628-4888 l
__-_ _a_____ --
?
92 1 looking for interact' ion-type items. Our fragility people 2~ tended to crawl around looking at anchorages ~and details of.
3 that . level. We then interacted and exchanged ideas.
'4 The example that did come out at. Maine Yankee was of.
5 the 15 to 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br /> we spent in the diesel generator room,.one 6 of the overhead piping,.'he t fire-water' piping was a; threaded 7 fire-water piping, essentially cantilevered through the wall, 8- free to swing at the far end. It. passed over.the' diesel 9 generator down underneeth for itssprinklering. 'When it first 10 came in the wall it passed.over the control panels for the 11 diesel generators. Our worry was that, if this pipe broke, it 12 would' flood and maybe prevent the control' panels from 13 operating.
() 14 After more hours of discussion with the utility, we 15 found that the line was dry. It needed another actuation tube:
15 to put water in the line prior to the sprinkler. )
. 17 So that is a systems interaction concern: threaded 18 piping has beer: shown not to perform very well, based on scrub 19 data. We were worried about water on the control panel.
20 Those types of interactions ware ~ identified. You-21 identify them by having this walk-down and by having different _
l 22 background people looking at different things. Systems people l
l 23 help a lot in identifying these areas. Their general effort 24 was to answer questions to our fragility people as to what 'are 25 important systems and components they neededi as well as-l v Heritage Reporting Corporation (202) 628-4888 l
93 I
1 identify some of these potential interaction issues.
O 2 That is a-very important benefit. At Maine Yankee it 3 turned out to be a non-problem, but at other plants with l
4 threaded piping, it may be a concern. ]
l 5 DR. WYLIE: Concerning intermetion, what, or did you !
i 6 have, a criteria for defining what you could look at?
7 MR. MURPHY: We had developed what we call our 8 " guidelines document." It was_an outline.of the overall' 9 approach. And then the basic criteriaLthat identified areas, 10 systems, regions, that we woulo concentrate on. Once we were 11 in those areas, the basic idea was keep your eyes open, see if 12 you could find something, use as'much of the experienced data j 13 base as you can. That is how the threaded piping came up.'
]
() 14 Threaded fire-water piping had shown in experience data to be 1
15 of concern, so that there was no specific step-one, step-two, J 16 step-three type criteria, was used heavily relying heavily on j l 17 the background experience of the review teams.
18 MR. WYLIE: I guess what I am driving at is, did you 19 look in areas that are not, let's- say " safety" areas, but where 20 failures of steam lines or water lines causing flood and the 21 like, go over any of the " safety" areas and affect those safety 2e -- 1 23 MR. MURPHY:
~
I am sure the answer is "yes." Maybe )
24 Garth?
25 MR. CUMMINGS: Yes, you know, the way this process Heritage Reporting Corporation (202) 628-4888
94 j l
1 . worked is we'did studies of which systems were important, and j O 2 it turned up in other PRAs, and those systems not only i
3 concerned front-line systems in the~ safety-grade and support 4 systems.that go along with it. .And.so you lock'at all sorts of 5 inter-dependencies, how these support. systems. feed in and the-
'6 systems analysis, you look at co-location ~ problems. Generally l 7 when you. walk through the plant'you look in those areas where l
8 both the safety systems and support systems-were and see if I l
l 9 things could fail to-Jeopardize that, like<you mentioned the 10 fire-water system. So you look at both the location arid the ,
i 11 functional independence.
12 But this was not a systems interaction project. That 13 was part of what we did, but I mean the wnole focus of the 14 margin review as 'n aeffort wasn' t ' j ust on ' syst erns interacticas.
15 MR. WYLIE: I know there isn' t. It's a part of 16 determining the margin --
17 MR. CUMMINGS: That is very true, and something we 18 paid a lot of attention to, but I'mean there are other programs i
19 that NRC has done where we tried to more specifically. address !
20 Just that issue and we tried to incorporate.that.
21 MR. WYLIE: What I.was gettingEat was, suppose for l
22 example, over in the feedwater area, did you look to see 1 23 whether feedwater lines were adequate? Or whether they.would 24 have failed. Or a steam line? Did.you look at those?.
25 MR. CUMMINGS: To the extent that they would affect O. Heritage Reporting Corporation (202) .628-4883 1
t I
--m J
95 1 the systems that we had already focused on.and jeopardized.
O 2 MR. WYLIE: No, I' m talking about if you burst a 3 feedwater line, you are going to have steam all ove your-4 plant.
5 .M R . CUMMINGS: Yes, that is true.
6 MR. WYLIE: And this would very. definitely affect 7 your safety assessment. I.mean,.what does that conserve?.
l 8 MR. MOORE:' Dave' Moore, with the Systems groupe And 9 we, looked at several things toLtry.to determine -- weLlooked'at 10 systems inactionsi floodings, steam, and things'like thatl 11 although we excluded seismic induced fires. We didn' t look at 12 that particular topic. But we did look for areas'for 13 separation primarily, so if we saw the. motor-driven' feed water 14 pumps were off in one area totally separated from all' the other
((}
15 feedwater lines, the turbine driven pump in another area, then 16 we felt much better about things like that. We.tried to look 17 to see if there was anything, piping was, looked at but semi-18 excluded in'the sense that we have a very high confidence that 19 the piping is going to survive.
20 But that is not to say we Just carte-blanche say not 21 to look at it at all. We did try.to look at.it. We. looked at 22 places in the diesel generator room to see that.they.had 23 curving, so if they were flooded outside in.the main pCS 24 systems, that water is going to go out of the great big panel 25 door they' ve got rather than going into the' diesel generator Heritage Reporting Corporation (202) 628-4888 l _ - . _ - _ _ _ _ _ _ - _ _ _ _ _ - _ -
l
'96 q I room, flooding.out those rooms. !
O.. 2 So we tried to look and use'these separations as'an i 1
'3 argument to see if we could knock out types of systems i
4 interactions that could occur. 'Now, that is not to'say that we 5 looked at every.little thing. But we certainly looked at everything in piping to see what the impact would'be if it 6 ,
7 broke? But we also had high confidence that most of-that )
J
.8 piping was not going to break,. and there was very high 9 Geparation of the equipment.
10 MR. MURRAY: The ending of this overall project was- i l
11 to put together a three-volume set of reports. We decided 12 about. half-way through the project to break this down into an 13 overall summary pulling everything together into one volume.
O 14 T"
- ee erv rener* 1 e 8 e toe eetimo imetee ef *"e neer 15 review group activity to be trackable during this crisis, as 16 well as the expert panel conference call meeting minutes. So 17 we were able to include as .an appendix all'the pertiner.t 18 supporting documentation, as well as input from the utility on 19 their impressions their overall estimate of the effort involved I 20 in their part and some of the recommendations they may have had 1
21 for improvement.
i 22 The volumes 2 and 3 that our" Systems and Fragility )
23 people put together completely describing their overall 24' approach. These are now out as a three-volume NUREG available 25 for discussion.
O Heritage Reporting Corporation (202) 628-4888
i 97 i
~1 i
1 I would like to -- I shot'my time budget here, but i
l 2 the next phase is systems. And how would you like to do that? l 3 ' D R. SIESS: Well, what I.think we wil.1 do is take a 0
4 short break nowl come' back .and hear the systems stuff and then d 5 break for 1unch. 'Let us put the lunch break some time between l l
6 12 30 and one, okay?
7 MR. MURRAY: Systems takes, what, on the order of 45 8 minutes, Dave, or --
9 MR. GUZY: Yes,.although a lot of work has already H 10 been done. So probably more like 30 plus. I 11 MR. MURRAY: 30 plus?
12 DR. SIESS: Why don' t you tell'them you' depend a~ lot I 13 on the questions.
() 14 Toff the Record.)
15 (Back on the Record.)
16 DR. SIESS: Mr. Moore, are you ready?
17 < MR. MOORE: Just a moment. As has already been said 18 several times, E. I . Services was responsible for the systems 19 task, when they conduct the margins study. .What I'would like 20 to just sort of do reasonably quickly was not' talk about the I
l 21 methodology itself as much as how we implemented methodology, 22 where Twe saw some of. the little bottlenecks werei what caused 23 us some'issuesi and then how some of these.were resolved. A j 24 little bit of methodology.of what we are doing, and'then, of 25 course, some of the results that we develop that' we then fed to O HeritaDe Reporting- Corporation .
(202) 628-4888 i
i
_ _ _ m. .._.____._______s__
1
'l I
l 98'
.i 1 the' fragility ~ analyst. task group.
2 Of course, as you see in the drawings, we. met and
]
3 talked and were integrated throughout the contract. I think :
4 that was a very important facet of this whole step. You.can 5 see from the dotted lines the portions that we primarily'were 6 involved with. We at systems.
7 The first thing we did, of course, is gather' 8 information. This' involved getting all.of the normal types of. .)
9 documents you would gather for.a systems analysis effort. Like i
the SSAR systems descript' ions, emergency procedures, pNIDs,.
10 i
11 electrical diagrams, that. sort of information that allows us to 12 know something about the system.
)
13 We also through the organization established contacts ,
(} 14 with Yankee Atomic -- Maine Yankee, so that we could ask 15 questions. As always, when you are going through an analysis, 16 to model a system or model a plant,.you have a lot of questions 17 about the way the system is operated at that plant, and any 18 special features about that, and that is what we try to find 19 out here. We gather information.
20 Notice we did this primarily for the Group A 21 functions. Dan mentioned that those from the methodology 22 document.were on reactor sub-criticality and emergency core 23 cooling early. See, emergency core coolant isn' t Just ACCS in 24 terms of injection, but it is also the oxidator in cores for.
25 decayed heat removal. And that is what 'we'are talking about, O Heritage Reporting Corporation (202) 628-4888
99 1 core coolant for decayed heat removal, or if we have to have 2 water in order to make up a.large reactor' adding water to the 3 reactor. We went a little farther than'that,, partly because we-4~ didn' t feel as. comfortable about ruling' other things out, 5 partly because.the methodology itself, Just to make sure that 6 the screening' material'upplied to this plant. And that in the 7 -- plant you need' features. . And in this case what we.did was r 8 we looked at that that stage right after early core cooling, 9 that is, injection phase of ECCS to the switchover phase of --
)
10 -- and a system -- it will be required the systems and a. l 1
11 support system -- that would be requiredLfor that phase. And 12- that led us to a few things that Bob. sort of alluded to and I 13 will get into as we' talk about it a little bit.later.'on.
() 14 But we did extend the search initially to look a 15 little bit beyond and do a screening process on the screens 16 that have already been done, Just'to make sure th'ose. screens -l 17 apply. We developed for the first plant; walk-down a list of 18 all the components, a list of systems, dependency diagrams that q 19 indicated which components and which systems required which 20 support systems in order to operate, and then we also looked at l
21 -- we developed some initial event trees, that allowed us to 22 see how those systems were geing to fit together to mitigate 23 certain initiated events. And one initiating event was, of-24 course, a seismic event. But then we separated that down into 25 seismic events that also had small LOCAs-associated with them, Heritage. Reporting Corporation (202) 628-4888 i
L _ _ _ _ _ _ _- _
103 1 either induced by the seismic event or in fact it.could have O 2 been induced by something like stuck-open'p0RV.
3 It turned out not to be important, but it was at.
4 least included.in the initial part of the analysis.
5 The other event was Just a seismic event that was a 6 transient and had no major leakage in the pritnary systems, so 7 that aux feedwater alone could handle the cooldown as.
B. necessary. j l.
9 We also had a lot of questions. .And I.think that 10 more than anything else, that first walkdown was to meet our 11 counterparts at Yankee Atomic, and Maine Yankee and to start 12 asking all these questions that start giving rise to the way we 13 develop a model that catches the subtleties involved in the
() 14 system operation and the systems at Maine Yankee,. notJJust the 15 normal run-of-the-milli here is a. aux feedwater system from 16 another plant. You would probably use most of it here because 17 they are probably pretty similar. We didn' t do that at all.
I 18 We developed our models from scratch. We had a whole list of l 19 questions that we used to go through and document what we knew i
20 and what we learned.
21 In the first walkdowns, Bob sort of says we walked in 22 and we looked around what we.might fall on something -- we-also :
'23 just served es a resource person to say whether or not a system.
24 was important. Whether a component was important to these 25 Group A functions or'they extended some of the. Group B I I
i
\
Heritage Reporting Corporation (202) 628-4888-
I
'l; 101 >
L v4' 1 functions:that'we had included in.our screening: process. ;
a L
(J . We also just looked around'.to see what we could find' 2
3 with plant-unique' features . that; you' don' t pick up off.the PNID 4 :and electrical diagrams. .There are a number of things l'ike 5 that. ,
J 6 After that, we developed fault trees and event trees )
1 7 for each of'the systems. We solved the' fault trees for system- )
I 8 level cut sets. By that I mean the failure that cause system.
- 9. level failures. Not necessarily core damage, because you might j l
10 have to fail in more than one system. But we release all the -l l '
4 J
11 fault trees to find out which components would either singly, j
.I 12 or in doubles or triples even, would cause a system failure,- !
~
13 and that gave us some feel for what was important when we went
() 14 down to the second walkdown.
o 15 We didn' t know entirely.what was important. There 16 were still things we hadn' t worked out. The failure' rate was 17 all in the components yet, or the operator action failure rates 18 entirely. Nor had we, did we have a feeling for,all the HCLPFs 19 as to exactly which components had high capacity and which 20 didn' t . Again, it is all part of.the screening process.
21 The second plant walkdown'then served more to look at 22 the leftover items that we still had on our listi we. looked at 23 test and maintenance outages to make sure that the Maine Yankee 24 plant did not have any out logs in terms of how often systems 25 and components were out for tests'and maintenance.
O-Heritage . Reporting Corporation (202) 628-4888 l
I 102 ,
. l 1 And finally we, I .think, got a lot of guidance from j O. 2 the pRG, especially from both meetings, but at the second 3 meeting, the review group that was' monitoring the" study helped 4 us look and resolve some of the issues and the problems'we had 5 as well. So that was very useful. :
6 Finally, after the second walkdown we had most of the.
7 information if not all of the information, we needed. And'we )
l 8 developed our PEP sets, which are essentially the combinations 9 of failures that can cause core damage. This-is for our -]
10 s'equences. 'And this included both for the small LOCA 11 sequences, and the non-LOCA sequences. l 12 We then developed Boole equations, which are just a 13 mathematical representation of these cut sets, after some
)
() 14 screening on some probabilities from failures that were not 15 seismic related failures. I will talk a little bit about how 16 we did the screening:
17 You will notice we turned over to the fragility 18 analyst to run-through and develop plant HCLpFs. And all along 19 the way, although it is a little hard to see there was'always j 20 talk and discussion when all of us here had gone through this.
21 From the systems interaction point of view, we did 22 try very hard as a walkthrough to look at course block walls -- -j 23 I forgot to mention when I made my response from the audience, 24 brick walls are one of the main concerns. What do the block i 25 walls fall on? Maine Yankee.had already done in their O Heritage Reporting Corporation (202) 628-4888 i l
103 1 . responses to sotue of the INE bulletins, had done a very good
.O. 2 analysis of what . block walls. were Doing to fall on, and what i 3 was safety-related?
-l 4 We looked,at all that again and.tried to determine ]
i l
5 what we could see'that with. As I mentioned, we. looked at 6 fire-water, not just in the diese1' generator rooms, but thero 7 was also fire-water piping over all the component cooling water 8 pipes, and that turns out.that the pumps themselves were made 9 so that they would shed the fire water. Their ventilation 10 heads were made such that the fire-water would not get inside 11 the pumps.
12 We looked at ways you could get floods if.you had 1
13 major piping failures in various places. So we did try to dota j i
14
~( ) fairly good look at what would happen if those types of things 15 occurred. We also did walkdowns on some of the piping, like 16 the component cooling water piping, to see if there were 17 components along it or places along it where you could get a !
18 failure of-a component cooling water system such that you would 19 degrade that system's. capability to cool anything.
20 And so we did do those types of items in trying to i 21 look at what our systems interaction was.
22 DR. WYLIE: You did do steam?- l 23 MR. MURRAY: Especially when we were -- the main 24 condenser and everything like that, and the; turbines, were out 25 in a very large hall, and the only other --
there may 'be some 1
1 1
( Heritage Reporting Corporati'on (R02) 628-4888 1
l'
_-m.-_m_m_-_ i____mL-_
'l
-104' 1 i
1 components out there were' things like component'coolirig water 2 pumps, that were out.
3 The motor driven aux feedwater pumps were way off.in 4 another' room totally separate from any steam piping whatsoever.
5 And then the turbine driven aux feedwater pump was in'the main-6 - - -I guess you call it a penetration area where the main i
7 feedwater and main steam piping went into and out of. ]
I 8 containment.
9 DR. WYLIE: What about the main feed water lines?
10 MR. MURRAY: They wert. .through that same area.
11 D R. WYLIE: But I mean the failure of one of those 12 would do what?
13 MR. MURRAY: THe failure of one of those could cause
() 14 severe problems to the turbine-driven auxiliary feedwater pumps 15 because it is right down there below them. So the Stmosphere' 16 in there is a little steamy already. It certainly was on the ;
17 days we were in there.
18 DR. WYLIE: What about the instrument room, the cable 19 room, the switch feeder rooms?
20 MR. MURRAY: On those essentially we had no --
21 correct me if I am' wrong, but those essentially had no main 22 steam piping and no fire-water piping.
23 D R. WYLIE: What 1'am saying, if.you had a'feedwater 24 line break, would.the steamLget in those rooms? j 25 MR. MURRAY: Very likely. 'They are closed off room Heritage Reporting Corporation (202) 628-4888 4
I
105 1 and considered - well, they werei in the generalr area of'the
.O 2 main feedwater and everything else, but'they~were closed-doors-
'3 on all those rooms.
4 Now, the other item there was a large-bore, piping,-
j 5 welded piping, was essentially screened out because of its high 6 capacity, hi gh HCLri~ capacity, . and we ' didn' t . try to go- around i 7 and postulating breaks in all of that large" bore welded piping, ;
8 We paid a lot more attention to the fire-water piping as a 9 threat.
10 Group A' systems we. looked at. These were reactor 11 sub-criticality early core coolant. And the reactor protection; 12 system had split CRDs in the internals. The auxiliary.
13 feedwater wystem for decayed heat removall. primary pressure
() 14 release system-for feedant lead. capabilities,-and the high l 15 pressure safety injections, in order to do feedant lead, and to 16 handle small LOCAs.
17 Ar;d of course the support systems'we looked at were 18 those that kept the front line systems going. .Your A C.-power 19 system, D. C. power and your on-site emergency power. They had 20 two component coolant water systems, a primary and SEC 21 secondary component cooling water, which reported like l ~
22 redundant trains at Maine Yankeel the service. water system and 23 then actuation systems for all of the above systems.
24 Now, of course you had the tendencies here between 25 electric power and service water that were also modeled I
() Heritage Reporting. Corporat'ioni (202) 628-4888
1
, i 106 '!
1 internal -- ~1
-(2) 2 So these are the" main systems.we looked at.: Wei 3 developed dependency matrices and related these,to thesejand 4 these to themselves. As I said, we sort of included some
~5 systems sort of beyond this system as well, so.th'at weLmodeled 6 high pressure safety injection,.we extended the high pressure- )
i 7 safety injection and the start of recirculation mode to make )
8 sure that we had a field here.that included the things that 1
1 9 would have to take place.- It is nice to say that the. plant is l l
10 going to go great for four hours, or six hours,'or something in j j
11 a small LOCA, but i f you are st uck then :-- -you don' t ' want to be 12 stuck then.
13 So we did include the next sort.of' partial phase. We
(} 14 didn' t do enything like the 72-hours oxygen cold shutdown and-15 things like that. Let's see. Using our systems, we then 16 developed what we call our system models developed fault-tree 17 models for each of the. systems. We did this in several 1
18 diffr;ent ways. The first time we did it we went through and 'l J
19 developed fault tree models for all the components that were in ~
20 the systems that were important to step system success. And 1
21 then the support system that supported those components. i 22 At first this 1s.a tree that is the HPI, the-HPSI.
'l '
23 tree. This particular flow from one area HpSI, and you can see i
24
~
that this area HPSI had a motor. drive pump and then some valves 25 we' re modeling here. Essentially what we are'saying is not Heritage Reporting Corporation (202) 628-4888 !
o 1
.1
_____________ ____ _.___________1_____.___ _ _ _ . _ _ _ _ _ _ . . _ _ _ _ _ ._____m_^.__._.._._____m.
. .__.____.___._n____.__.. ._._.1._._._ ____________
i i
107 L 1 only do you need the motor-driver pump to work,'but under here ]
2 we may have.X exits for our failures, essentially.
The first time through we H 3 We fill these in later on.
4 d idn' t know whether seismic coding would be important would 5 that pump, or a random independent failure of that'p' ump, or 3
-1 6 some sort of. operator failure that'miD ht'be-necessa.^y for it, j 7 So we left it sort of open at this point until we learned more 8 about the system. )
9 We modelled, however, all the support systems we i
10 needed, whether D.C. power, an actua. tion signal, component I 11 cooling water, your large A.C. power busts -- we modeled all 12 those in the tree as well. Now later on, the failure of the 13 HpSI pump, failed to start, failed to run, test and maintenance i
() 14 were all very small, random unavailabilities. And we were able 'l 15 to screen them out using our random unavailability screening 16 out rules. But if we hadn' t kept this in the model et f i r-st ,
17 that is the way we' develop the support system dependencies.
18 So although we can rule this out, we certainly may 19 not be able to rule out the A.C. power to it and especially the 20 transformers that went to some of those.A.C. power busts. Or_
21 down to the D.C. power, the batteries.
22 So the important point was, althottgh you may be able 23 to screen components out at a later stage in the' analysis for 24 various reasons, you have got to remember you have those 25 support systems that are feeding it too, and you have got to ;
Heritage Reporting Corporation (202) 628-4888 1
- . _ - --_____________.___U
jk 108 ,
1- model those;in there, even thouDh the component itself. falls.
O 2 out later on.
3 And so we use fault. trees to develop that force. And 4 that is sort of an important point. There was one case,.a.very-5 -minor type thing, but'with the PCC,ihere were isolated 6 ' procedures developed to isolate the piping -- non-essentiel 7 piping that_went in'and.out of contain;nent, because we couldn' t 8 Do inside containment, walk-down and go look at them.
9 Essentially the' operating procedures, PCC -- I can' t-10 remember, is that_ sort of like the A. train or the B train --
11 orie or the-other.' But some of the valves here are opposite:
12 trained power. That is, they have.got on the containment 13 penetrations, you have got an A train valve and a B train
() 14 valve. And it so happens that in the procedures the one valve 15 was picked to isolate a particular line, was on the redundant l 16 train causing the line that is not the one that is'normally PCC <
17 operative.
18 And so when we did our PEP sets ~we found out all of a 19 sudden that you are getting this' failure from one power supply 1
20 -- that is just one A.C. power supply failure caused failure of 21 the whole PCC and SCC. That is, all.the total component --- -;
i 22' water, because of its non-isolation failure.
23 So it is the type of thinD that you have to model it 24 down to enough detailinD that you get those subtleties. You 25 can' t j ust say PCC is B train and SCC is A train, because there I
Heritage Reporting Corporation (202) 628-4888
(
f 109 1 .are a lot of real subtleLdependencies in there thatfyou have i
<:)- 2 got to use something like this to'model.
3 We then developed screening . tables.: .This is an 4 interim process. We, of course, supplied the randomffailure.
I 5 probabilities. .And we used generic failure probabilities.for I 6 everything' generic from other nuclear industry daba -- for 7 everything except the diesel generators, and then:we had a B report that had some Maine Yankee specific number.- a 9 This is for the auxiliary feedwater system, which 1 10 . includes both aux feedwater1and. emergency feedwater. .They .
'11 called their turbine driven pump by.a different-name. But 12 essentially once we'know just how much HCLPFs have been 13 calculated, we are able to start excluding some' components.from
() 14 the analysis from our trees. These are all air-operated 15 valves. But they have a high capacity. <
16 By the way, we looked at almost every valve, I 17 believe, that we had here. There were feeders that1were'down 18 in places that my wife wouldn' t let me go. 'But.other than 19 that, we'had -- we took a look at almost.every valve'that was j J
20 en our list. And that was a little farther'than what the. I 21 guidelines methodology said. [ ]
22 So we were able to start screening. things out' based '
23 on high HCLPF capacity, and also~a-very' low probability-of
]
24 random failure. There were some' components which had a l 25 relatively high probability of random failure'. This was a~
.e r I
( . Heritage Reporting . Corporation (202) 628-4888: j i
i i
i 1. 1
<f l
110 !
=!
1 turbine-driven pump. An'd so because the _ t urbine driven pump 2 had -- although it had a high HCLpF, it had a.high seismic !
1 3 capacity, it's availability was less than what our screening. .i 4 rules were. And we had various probability screening rules 5 depending on whether it wiped out a component or one redundancyL ,
1 6 of a syntem, or a multiple trains like en A.C. powered.-- a-7 diesel generator,'like some of our failbre diesel generator. ]
l 8
placed on a lock.
9 So we nad-the'more_important component was more f) 10 things that went down the lower level of availability we allowed 11 it to have -- or unava'i l a b i l i t y. So the diesel generator. d i
12 unavailability would be screened out to the 20 minus 3 level l 13 ( 2- 3, whereas a single component or pump, if it had an
() 14 . unavailability of less than le - 2, it was screened out because 15 all it failed was one part of your system.
16 We also screened in of course things that had low 17 HCLPF values, like the water tanks for the aux feedwater i
l 18 system, the DWST and PWST, and also ttings th6t at the time of i l
19 the analysis we didn' t know what the HCLPF'was. And in this ]
l 20 case this was a large air receiver, an accumulator _for some of' i 21 the valves. So we left in those types of things. Common-cause 22 failures, some we left in when they were less than our ;
1 23 screening values, or greater than, I guess I should say, when 24 they were more unavailable than our screening values.
25 This allowed us to prune down a lot of the faults out l i
O' Heritage Reporting Corporation I (202)'628-4888 q 6
[ .+ f. s 4 .
9
^-
ylh
- m
i . , ,'.
- n. / I 111 h .e a i
1 of our trees. . So we had systems fault. trees. j that were much )
2 smaller, andlprobably we pruned oct "y.I #94uld say 80 percent, 75
,. .t. fd .p
-3 percent of all the fault's"'in our-faultJgtrees. We got down to a 4 lot smaller fault trees. t. .)
, <. TA l 5 We then combined those 'aystem' Yault' trees into.
sequences which say you m1Dht need to' hate more than one system
~
6 J
l-7 fail'in order to go to core damage. We did it for'both cases. l 8 We loci <ed at core' damage to distinguish between a=HCLPF for'a
> 0 '
t 1 n. , c.
9 LOCA and non-LOCA' cases primarily. ,: i rF
,s 1 J 10 To show you an. evente et v:e, real quickly to get somef.fGy J
- t. ,
idea.+ This is for the seismic loss of~off-site power. I a
' z' 11 ,
.y f \' < 4 l
12 should have said that first. WeidsuIned that the seinie event --
13 would cause .a' loss of of f-site poher tco the plant. That is u 4 14 part of,thh, ground rules and methodologies. Given that'y'ou
/
i /
15 have a seismic loss of off-site power, then you ask about
-) ~
hbu 16 reacter tub-criticality, did the rods go ihl does the reactor 17 shut dowa? This is an area that I will discuss a litV.le' bit
.ss 18 more bechbse the analyses were performed by EQE, we Were able' 19 to say that its HCLPF was hirgh but at high capacity, and its -
f .
e l 1
20 reliability from a random poipt of view or from a non-seismic - l
, , i 21 - point of view ( it was very availablet, so we didn' t have to d it a 22 evaluate teat sequence fur.her. It was not a. contributor to 23 the plant HCLPF.
y, 24 Then we would say, "Well, if you have got a v
25 transient, what do you need?" We need decayed heat removal, so >i i
? i Heritage Reporting ' Corporation (202) 628-4888 ' .s 1- i l
,,- ., i }u
'; t ym 1
.. y
]
W, e ll V.;
a.
i 4 ,' -l
'112 1 you need aux feedwater emergencyffeedwater. As'long as you-2 have that1then you asre'okay. You are not going to go to core
( 3 damage. Given t' hat you don' tihave aux. feedwater and -- develop
- t h4 ,. success criteria for these, of course. In this' case it was one e < t) .. .
y [ out of three pumps. had to be . pumping water to one of' the steam -
3 Ldperators. If you don' t have. aux feedwater then you can still
- s. l i, 7 go to Tcedant lead. The' operating procedures when you'go to
",, a i 8 feedEnt lead for a transient case, you had to open two PORVs 9 and the block valves, and then you had to have'-- so that would.
10 be.a success, and.then you had to-have high-pressure safety it injection in order to perform feedant lead.
12 You read this series and you can see that here you 13 have got successful sub-criticality for this sequence you 14 migh% say, "I don' t have feedwater, but I do take the operator 15 actions to the feedant lead and my PORVs work, but1I don' t have.
16 PhSI, so I go'to a core damage accident. And it is Just the-1 U gargon we use to confuse people.
-i 18 $/ We did a very similar thing for --'this is the small y
19 LO[jA case repostulates small LOCA because of'the seismic event
+
1 1 20 or because you have a reactor pump seal LOCA or because you
(
p 21 have a stuck-open PORV. Again you have reactor sub-(
l\j 22 cr i t P{> al i t y. If you have aux feedwatert even if you have it-
.x 23 you still have to supply HPSI. If you don' t have HPSI you are
^
24 going'to go to core damage again, even though you have aux i u I
25 feedwater. If you do have it then you are successful and' O Heritage Reporting Corporation (202) 628-4888 li l\
q s _ _ . _ _ _ - - - _ - _ _ - _ _ - _ _ - -
a 113 -
1 everything is okay. -'
1
..O 2 You'can follow the rest of the, logic through'to see
-3 what other things? This, essentially, is developing success i
4 and failure paths that remodel each'of these within our fault-5 trees. And we hook our fault trees together such that-you have 4 1
1 6 combined failures. You have an electric power system failure. j i
7 it' fails both diesel generators. You'will not-only fail the 1 8 motor driven aux feedwater pumps. You might still have a 9 successful turbine driven feedwater, but it will also fail HpSI 10 itself, and so you will fail here in any case for a small LOCA.
j 11 For a transient, that wasn' t the case. For a i 12 transient, as long as you had that' turbine driven aux feedwater 1
13 pump y>u were okay. So we developed these different success- l 1
() 14 criteria. We developed our system'models, including all of.the
)
15 support systems that were needed to provide for the~ systems -- 1 16 the motive power, et cetera, and we developed these models.
17 The results, totally un-understandable Boolean I
18 equations. These are the Boole' equations that we essentially i 1
19 gave to Rob at EQE, Dr. Lindbrook. These have actually been '
20 pruned quite a bit because we found that a lot of the failures 1
21 were not important, so that these are ones that became more 22 important to the total plant HCLpF when taken a lot of the l 1 23 things out that were not important, j 24 The next page contains an explanation of what each of 25 these are. But essentially, for the Nolo case,'you have. tor l
'i O. Heritage Reporting Corporation (202) 628-4888
- _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ - - - _ _ - _ . _ .- - - _ - - -. ._ -- -= . _ _ _ . _ _ _ _ . ._------.---a
. .)
I 11'4' ]
~
1 1 have two types of. failures. You can read this as being sort of'
'O 2 logic. You have to have some'of these and one of these in 3 order to- go to core damage sit uat ion. . Dr.this, and'one of 4 these, or these two together. That is what.it essentially told 5 .you. Now, we'have got a lot more but~the rest of them all fill-6 out either because they are of very low probability, or because 7 they had components that had high HCLpF, Which we determine 8 later on.
l 9 This -- let me get my cheat-sheet here " this was .)
.{
! 10 failure of the two transformers. I take the top from the-41-60 l l
l 11 volt emergency bust it down to four equal: --
12 So these two failures here essentially fail your
')
13 motor-driven pump. This failure is from poweri this is the l "
l 1
(} 14 seismic failure of the cire water pump house which has the
- 15 service water pumps in it. So if the service' water pumpscfail, .
l \
l 1
l 16 eventually you are going.to fail cooling to your motor-driven 1 l
17 pumps, 18 So both of these can take out your motor-driven pumps l 19 -- or either of these will take out your motor driven pumps.
20 These are essentially the failures that fail that turbine-l 21 driven aux feedwater pump. The first one is. failure of the.
22 demineralized water storage tank, which was the, primary suction 23 for the turbine driven pump. The secor,d one is, the rest of 24 these were all' non-seismic f ailures that didn' t come-into play ;
25 very much in the results. But we did include both seismic and O Heritage Reporting ' Corporation (202) 628-4888 i
.e - _ . _ _ - _ - _ _ _ - - - - - - -
l i
115' 1 non-seismic type failures. And one of these would be that O 2 turbine-driven pump is unavailable and fai.ls to start. That 3 sort of thing.
4 Again, another type was failure of the DWST which 5 enhales the turbine to the-pump, 14 and 16 are common-cause-6 failures of the diesel generator, where the operator error to 7 maintain fuel oil. Afterwhile the operator had to run down and i 8 take fuel logs used in long term storage tanks for the day 9 tanks and do a model of that -in there.
10 And again, similar types of failures here. This was 11 the RWST which fails feedant lead essentially. You can' t do 12 feedant lead without it. And this is a common-cause failure of 13 all auxiliary feedwater pumps. This is a seismic failure.
() 14 This is a non-seismic failure. And so we have those-15 combinations, i
16 The small-LOCA feeds that came out -- remember we 17 need not only. aux feedwater for decayed heat removal but we 18 need to have HPSI or in fact if you have HPSI with.feedant 19 lead, you will need aux lead. So primarily what'you had here I
20 are failures that cause failures in HPSI. The first one is the j 21 transformer failure again which loses all'A.C. power. The 22 second is the RWST, which of course fails your suction source
^
23 for your HPSI pump. And the third is the seismic failure of 24 the circulation water pump house area, which fails service 25 water which eventually f ails your HPSI pumps. And there are a 0 Heritage Reporting Corporation (202) 628-4888 i
i
_ _ _ _ _ _ _ _ - _ - - _ _ - _ _ _ _ - _ -. ._ n
-l
.116. -
1- number of other failures that could be added here.. In the - ;
.O 2 slides that.'emain, r I.have put a' lot more information in~more i
3 detail than other types of failures that could be there, These' .
i 4 are the ones that are important. l 5 So that is essentially what was d'elivered for better !
6 or for worse over at EDE, but we still had a few-insights from 7 a systems point of view, and Bob'has touched on some. We 1
1 8 touched on quite a few. I think we had both insights that were 9 important to Maine Yankee and then we had some other insights 10 that were more methodology oriented. And finally a few 11 insights about the execution of the project.
12 I think, in terms of the operational i n's i g ht s, Bob 13 mentioned that they made some changes in the plant, some
(} 14 upgrades in the plant, some of which are Just to remove the 15 uncertainty involved whether something is or is not below some 16 sort of cut off. It does power the system as such that you are 17 able to show that some things are more important'to .fix than' 18 others. And I don' t remember what the strengths are for each 1
19 of the tanks there. I would like to have the DWST, the pWST 20 and the RWST -- well, strengthening the pWST really did you 21 very little good, no matter what its HCLpF was and I don' t 22 remember what it was. Just because of the way which tanks were 23 most important.
24 The DWS HCLPF value was more important to you, and j l
l 25 the RWST value was even more important to you. .The,way it is ,
1 1
-I Heritage Reporting Corporation (202) 628-4888 1
l 1
1
I i
i 117 1 modeled allows you to tell' how much of an impact you are going O 2 ^to get quantitatively in terms of the HCLPF by changing things.
3 Yes,.I' feel that there are some very real insights you can' gain ]
4 from modeling, not that you can' t ' find some things Just. from 5 walkinD around and saying, "that is a lousy anchorage. That 6 needs to be fixed." You can also look at some of the failures ;
7 where you are not sure whether they'should be fixed'to i
8 determine, "well, that we don' t have to worry abouti this is !
9 important to our final results.. This is'important to core I l
10 damage." Now, the pWST was very important to the availability 11 of the motor-driven auxiliary feedwater pumps because it was 12 the lowest -- I remember it had a higher capacity than the DWST l
13 did. The DWST was important to the turbine-driven pumps.
() 14 So you were'able to make some of these decisions by is seeing some of these relationships brought out through the 16 Boolean equations and through the models.
17 I think~that we brought out_a few things where there I
18 are other systems that were -- that you wouldn' t - normally -look' i
- 19. at in terms of their importance, to say that they weren' t I
20 particularly important because of the function they served, .but 21 they were important because they were a pressure bouncary for a 22 system. Or they were chillers, or the computer room and some 23 other stuff -- the lab room, were on one of these -- a cooling 24 water line. And it' turned out in the long run that their HCLPF i
. . . i 25 was higher anyway than the .3 G 1evel -- I believe. l l
Heritage Reporting Corporation j (202) 628-4888 l
\
1 118 i But it was one 'of those things where you' didn' t need i LO 2, those particularffunctions, .but you wanted those chillers to 3 stay' intact, or you were. going to lose you'r cooling water. So- 1 1
4- we went through and we modeled those: types of-components whose l 5 function had to be maintained because'of pressure boundary and
'l 6 diversion purposes, not because.they needed to work in order l- 7 for a system to work. Maybe that.is not quite clear. I think 1
you understand what I am'saying.
9 In terms of methodology, I think'I have already 10 mentioned a little of how.I feel about the power and the 11 'usefulness of the fault tree methodology.and being able to 12 model the systems, model more than one success pattern'-- both 13 success of auxiliary feedwater and success of feedant: lead will 14 help you out in terms of a non-LOCA situation. Either one 15 could be used. This is a success path.
16 We modeled both of them to give the plant due credit ]
l 17 l
for having two different ways to be.able to keep cooling during 1 i L 18 a transient seismic event. '
19 We did have a lot of issues that came up that cost us-20 some grief at times. For reactor sub-criticality, you don' t 21 just need the rod. You can also have reactor sub-criticality 3 22 by using your boron systems to pump boric acid in your reactor.
23 Normally boric acid systems are not particularly seismic grade l 24 one. There is an awful lot of components. The model would 25 have been complex, and developing the HCLpF would have been y O Heritage Reporting Corporation (202) 628-4888 l
i
)
119 1 1 very complex. So it is very nice if you are able toute11 if-( 2 the rods are indeed going to go in and the internals are going I
l 3 to stay in enough of a line and'if'you can get the rods i n.
_ l 1
4 We did a component identification for the boric acid .i 1
5 transfer system. And when we did the first walkdown we' looked 1
6 at a lot of those components. We were fortunate that EQE was 1
7 able to say that the rods were going to go in with high j 1
1 8 confidence, and we didn' t have to model that system -- the l l
9 boric acid transfer system, any further. Essentially that was-10 taken out and the officer was then e.ble to say with high
]
11 confidence that you are not going to get an --- for one of .
12 these seismic events for this PWR hearing.
13 Emvrgency core cooling early and late -- I sort of
(} 14 talked about early -- was in the guidelines. Yet we included 15 the first phase of late in order to have some confidence that i 16 you would be able to make that switchover.
17 Battery depletion was an issue that we didn' t ' cover 18 to the extent that maybe one would cover it in a seismic PRA..
19 It was left just a little bit hanging. For some plants it 20 would be a much more important issue than it was at Maine 21 Yankee, because Maine Yankee does not have quite the same 22 problem with seal LOCAs in reactor coolant pumps that give 23 rise to a lot of problems in other plants, perhaps. So I don' t 24 think that was a real issue here.
25 But it would be something that I would include in a 1
Heritage Reporting Corporation (202) 628-4888 i
l l
'120 i
1 Snethodology ' guidance document ito c inc.lude, and to th' ink about
=
ii
.,[
2 more when'you are doing a seismic margin review. . We talked 3 about looking at feedwater and some of the problems 'that --
4 ' excuseL me, : fire-water, and fire-water pipe. interaction. We 5 also looked at . fire-water systems as an alternate system to-l 6 provide water to the steam generators. They have a Bendix R-7 system that is designed to provide water to the steam 8 generators.
1 l~
9 We ruled out looking at it after'the first walk-down 10 based on a couple items. One, there were some tanks that were
] 1 11 not anchored particularly seismically strong,.although those-12 anchorages could have been improved. Secondly, the aux 1
.i 13 feedwater and emerDency feedwater systems were fairly.
() 14 seismically capable already, adding fire-water system on. ops 15 level when you rec 11y want to affect the HCLPF.
16 For isolation, we talked about isolating piping goinD 17 into and outside containment in the PCC' system. Some of those 1 10 things were important to look at. They enabled us to not have 19 to go into sit e cont ainment. The secondary cooling water also 20 had automatic isolation on it, so we were able'to rule out 21 those areas of the component cooling water systems and you l 22 don' t have to walk down. That supports your systems model. It 23 is extremely complex. You have to walk down a lot of component 24 cooling wate- for essentially diversions or if a pipe breaks -- .l <
25 not pipe breaks, per se, but all sort of things where the i
O Heritage Reporting Corporation (202)'628-4888
q 121
'i '
1 connection toLother equipment are there, we.have to look at all.
10- 2 those other little heat exchanges-in all those-different' pieces 3 of equipment and it could be an intensive level of' effort. . It )
i 4 . wasn' t necessary because of isolation. J
'5 But a lot of recovery actions in that regard I didn' t; l
6 notice whether or not the binders to the emergency procedures ]
7 are on the shelf or if they could fall down, but we did review .
I 8 The guidance and the methodology document was to
.all those.
9 not use special factors for seismic events. That is, when you j
)
10 are doing human factors analysis you don' t consider'the fact -
11 that the operators might hit their heads. We did look at the l
12 sealant panel and make sure that those weren' t going.to. fall on l i
13 them, but the types of things we did do'is when we selected our
() 14 unavailability, that is our operator failures,<we used very
)
15 conservative numbers. I would say if you were doing a real PRA 1 l
l 16 and those became important factors to you, you could' whittle' j 17 those numbers down quite a bit, especially.for the levels of 18 earthquakes we are talking about here.
1 19 We talked a little bit about how we treated the 5 20 support systems and the fault trees, how it is important to 21 keep components in early on so you are.sure you have got all 22 the. support systems to their components. They actually screwed 23 up on it, so I will admit it. We left an air-operated valve on 24 in one place and we had,to go back in and insert it'and it 25 caused much grief later on to Robby, to have to put something O Heritage Reporting Corporation (202) 628-4888
122 j in at the last' moment.
1- 'So it is important to make sure your
'O 2 trees are correct and ycu have all those dependencies in.
3 'In the cut set evaluation process, we evaluated our. a 4 ' cut sets a couple more times than we thought we were' going to- ]
i 5- hcve to, because our models became better developed.. We._added.
6' so!ne things and we subtracted some things, and we.tried to.get 7 the guidance a little earlier in the process'to see what-was 8 going to be important than the methodology' document;specified, 9 and I think that it was probably important. The fault tree 10 evaluation was reserved unt'il after the second walk-down~. We
- 11. went ahead and formed a systems evaluation before the second 12 walkdown, and I think that was helpful'to us. .i 13 And then the overall types of things you would expect I
() 14 in an execution, close. interactions were extremely good, not J 15 just among the Livermore staff and the Fultons,. but'between the 36 Yankee staff as well. That really. helped. And between'the 17 review groups. That really helped as well.- That was a source 18 of real good help to us. j q
19 I think that is about all I'have. Are there ]
20 questions?
21 DR. SIESS: There has been some talk -- I hope for 22 future reactors dedicated fault. risk, highly qualified decayed 23 heat removal system. Like some European coantries have u r,ed .
24 If you are doinD this kind of review on a plant that had such a-25 system, would you just be able.to-look at that system? .]
i O' Heritage Reporting Corporation l
'(202L 628-4888~
I j
a l
123-1 MR. MOORE: In a sense that, if you were doing a 2 HCLpF fault review and you just wanted to do a stress pat'h of 3 that system, it would be possible to do that.- That would give 4 you high confidence that that one system worked. Remember 5 those systems are not very redundant. They are usually --
6 DR. SIESS: Well, some of them are. Some are 7 quadrupally redundant. Not failure proof additionally. That 8 would be a success path approach. That's the a f ter approach.
l 9 MR. MOORE: You could still do a fault trees on a 1
10 system to find out some system detaiIs. I am not sure how much 11 would be necessary if those are really separate from all your 12 other systems, if they are implemented to be highly separate 13 from all other systems. Very self contained.
() 14 DR. SIESS: Now, the systems you looked at here, are 15 considerably larger set than would be incorporated in what I 1
16 understand the dedicated heat removal system, am I right? )
l l 17 MR. MOORE: It depends.how you implement the j I
l 18 dedicated - yes, we definitely looked at more system hers, I 19 definitely right there.
20 DR. SIESS: These, the ones I have seen, have their 21 own diesels, their own fuel supply --
l 22 MR. MOORE: Their own water.
I 23 DR. SIESS: Their own water supply -- inside j j
24 containment they use all the existing piping for the aux steam 25 water, I see.
\ >
1 Heritage Reporting Corporation (202) 628-4888-
124 1 MR. MOORE - Again, I' d say that if you had a small l
.(:) 2 LOCA case here, which'was our domin&nt, failure in terms of the J
3 HCLPF, that dedicated decayed heat removal system would not 4 have helped you. That i s, it would only remove decayed heat,. ,
i, 5 and you have still got a loss of inventory inside the. .]
6 containment. .So.you have got to make sure you define what .you 7 are going to' protect against when you design a system like 8 that.
9 DR. SIESS: I guess what I am trying to feel' for is 10 there some basis we could design plants where certain~ systems 11 would be designed quite conservatively for seismic, and other ]
12 systems we could forget about -- not completely but --
13 ' MR. MOORE: .I guess my answer would be'that sometimes j) 14 we'need to look at a grander sesle of all the' things, all the-15 initiators and all the events that can occur, and why you might_ I 16 be able to design a system that is highly reliable for. seismic.
l 17 point of view, or maybe for a fire point of view, or a' flood.
18 What does it do to your core damage frequency for ird ernal 19 events as a whole, and are there other ways to go about doing.
20 it?
21 I think that there are certainly some real advantages 22 that are dedicated and highly reliabis and highly capable- 1 23 seismic lead system that you are trying to reduce scismic risk.
24 DR. SIESS: Well, I was getting away.from the 25 dedicated system now, Just say a graded seismic design, or Heritage Reporhing Corporation ,
(202)-628-4888 l
-i
125 i something of that scrt. Because some things we would really 2 beef about the seismic, and others we would say that, you know, 3 .are not too important. The' cable trays maybe not being as 4 important as' steam generator support.
5 MR. MOORE: With adequate redundancy in the system, 6 and the diversity, and taking care of both transient.and LOCA-7 type cases, then you could feel very comfortable about a 8 seismic event. But then you start thinking about, is'it also 9 separated such that fires -- it can take care of the fire type 10 events as well. -And what about main steam line break. case? Or 11 something like that or a LOCA outside containment case. There 12 are lots of other types of events which require you to have 13 different functions. That is the whole reason we have got 20
() 14 systems --
15 DR. SIESS: I am simply talking about seismic. design.
16 Can I simply concentrate my seismic design effort on' portions 17 of the plan without spreading it uniformly over all so-called 18 " category one" systems components and structures?
19 MR. MOORE: In a sense Don has a little bit here --
20 DR. SIESS: You mentioned'three tanks and you said l 21 two of them weren' t nearly as important as the other.
22 MR, MOORE: Yes, but do you want to put all of your 23 eggs in one basket, too?
24 DR. SIESS: Well, I~ guess if I was just looking at 25 the success path I would have them all in one basket, but'you v Heritage Reporting Corporation (202) 628-4888 L_ ___ - _ _ - _ . .. -
a 126 ,
1 1 are looking at more than one path here.
2 MR. MOORE: I think that redundancy and diversity is.
l 3 an important issue. j l
4 DR. SIESS: I really feel that diversity had nothing '
5 to do with seismic design. Does diversity raise the HCLpF7 If j 6 I have got two diesel generators made by two different 7 manufacturers am I going to get a higher HCLPF than if I have ,
8 diesels built by the -- -
9 MR. MOORE: No, I think I used to think of diversity 10 more in terms of having A.C. and D.C. powered systems and if 11 you are going to talk about.an A.C. power system, diversity 12 would be a diesel, a turbine -- a couple diesels and a gas 13 turbine generator. The types of things where you don' t have
() 14 the common cause failures that would get you otherwise.
15 DR. SIESS: If I had two diesels and one was oriented 1
16 north and south and the other oriented east and West?
i 17 MR. MOORE: .That doesn' t help you out. Robby is ]
i 18 probably the better person to ask. l 19 DR. SIESS: Will it help me out on common cause l l
20 failures?
21 MR. MOORE: We don' t assume that both diesels fail i 1
22 with in the same given ground motion. They don' t have the same ]
23 --
24 DR. SIESS: Suppose it is a northeast -- southwest 25 ground motion?
O Heritage- Reporting Corporation (202)'628-4888
J 127 1 MR.' MOORE: Well, Robby, maybe you can cover how we.
O. 2 calculated some HCLpFs for components oriented similarly?
- 3. DR. SIESS: I have often' wondered about that. I'have 4 often seen designs with'four diesels, one at each corner of the 5 plant, complete separat ion --- you know, and wondered what.I 6 gain ren seismic count on those failures by having four diesels, 7 and what I would gain if they were in orientation?. How could I B get diversity?
9 MR. MOORE: I think Robby can talk a little about 10 core dependency.
11 DR. RAVINDRA: That has beenJone of the things 12 pointed out in the rev'iew of'the seismic pRAs.that that kind of 13 seismic redundancy orienting the components in different; 14 directions and having them redundant components located in 15 di f f erent floor divisions and probably different types of 16 mountings would definitely help in reducing the seismic risk by-17 including the seismic monitor.
18 DR. SIESS: I am wondering to what extent'this is an 19 artifact of the modeling and calculation process versus 20 reality? Bob?-
21 MR. KENNEDY Bob Kennedy. As part of the Diablo 22 Canyon long-term seismic program, we have recently done a 23 correlation st udy of responses at different floors of different q 24 locations with the same building to try.and get a handle on 25 them. And the end resu'It of that study. was that equipment
-( Heritage Reporting Corporation j (202) 628-4888 '
1 P
_ _:= _a
128 1 mounted at different locations in the sarne building, 'if the -
2- iterns of equipment had essentially the same-natural' frequency 3' content, two identical diesel generators rnounted far _ apart from 4 eachother'in the same building, the cross-correlation was_very, 5- very high. The cross-correlation was much lower if they were
~
6 'different frequencies, but.it turned out that equipment with 7 dif ferent frequencies at the same floor had low cross-O correlation and almost as low as if they were at different 9 floors, and my basic opinion _is that you are~not buyinD much 10 increase in seismic capability or reduction in seismic risk-11 alone, by_having the same item of equipment mounted at 12 different locations or'in different orientations, that the
'13 conditional fragility of one of those, given that the other has 14 failed, is very tight, relative to all of our other sources of l
15 uncertainty, p. particularly uncertainty in theLfirst risk place, 16 and that I don' b see much reason to have much redundancy for 17 seismic. There is a lot of reason to..have it for the internal-18 events --
19 DR. SIEES: So we would be better off simply to go in 20 for robustness in the component than trying to --
21 MR. KENNEDY My opinion is that there is a lot of 22 benefits from having a single robust --- put your money in a 23 single robust system for seismie, but when you tall about a 24 system, you have to talk about all the support things anyway, 25 so there is a lot of support things that are common to all of O Heritage Reporting Corporation (202) 628-4888
c 129 1 t he sy'st erns.
O 2- DR. SIESS: Thank you very rnuch.' We will break for
'3 lunch. Just one hour. Be back at five ininutes after two.
4 Rernernber that.
5 (Off the Record.)
6 (Continued on the next page.)
7 8
9 10 11 12 I
13 0 24 Ii 15 l J
16 17 18 3
19 20 21 22 23 j 24 25 O Heritage Reporting Corporation >
i (202) 628-4888
1 130-1 DR. SIESS: ~ Dr. .Ravindra, have you got a handout?
~O 2 DR. R9VINDRA: It was attached here.
~3 Hello, I am Ravi Ravindra from'EGE, Inc. The purpose 4 of my presentation is to discuss the fragility analysis aspects 5 of the -- review of:the Maine Yankee atomic power station. 'The l .
\
6 objectives of this study were to test the seismic margins 1 1
7 review 1 methodology insofar as the fragility analysis aspects 1
B' were concerned and also to estima'te the seismic margin of the 9 Maine Yankee power station.
10 At the risk of repeating what we mean by theLseismic il margin, it is defined in terms of the high-confidence - low-12 probability of failure at its initial capacity, and it is 13 determined using the fraDility analysis' method in this
(} 14 particular study that we did where.the' fragility parameters of 15 the median ground activation capacities, and the beta-R and 16 beta-U signify the log standard deviation representing the 17 randomness in the capacity and the uncertainty of the median 18 capacity. The particular capacity itself is expressed as a 19 function of these median and beta-R and beta-U, are shown 20 there. s 21 The fragility parameters are estimated based on the 'j 22 margins and variabilities in the response analysis, and the 23 capacity calculations. We have utilized the earthquake 24 experience data, actual earthquake data, and the earthquake j 25 equipment qualification test data that are available to perform l
i Ov Heritage Reporting Corporation '
l (202) 628-4888
v, 131' 1 the fragility.analysisLand the margin calculations.
2 The rnajor struct ure . in a seismic margins st udy1 that ,
1 3 pertain to the' fragility analysis are.the ones that are shown 4 in bold' bullets here screen-the components. based on the generic 5 HCLpF capacities, and the of generic HCLPF capacities of the 6 cornponent types, exceed the review > earthquake level, we screen.
7 that component out but later on come back and make sure that.
8 there are no more weaknesses in those components.
~
9 That is done in the plant walkdown -- actually plant 10 walkdowns -- we-did two of them. And in the process of. walking 1
11 down the plant, we also gather' additional information on the 12 screened-in components in order to estimate the HCLPF.
13 capacities of those components. And using the system. equations, 1
() 14 the Booleans or the processes that are provided by'the-systems 15 analysts, we estimate the HCLPF capacity of the plant.
l 16 That is the overview of- the seistoic margins review l'
17 methodology from a fragility ar,alysis standpoint, and how we 18 applied it to the Maine Yankee seismic project review, I want 19 to discuss that and these topics: the review earthquake l'evel, l
l 20 how it was selected and how we interpredict the review 21 earthquake leveli the initial screening of, components where'the 22 systems analyst provided the set of group A components,.and q 23 that set consisted of about ten'differentistructures, about l
24 two-fifty components. '
25 These are the valves, heat'exchangers, the MCC, et i
i O%- Heritage Reporting Corporation l (202) 628-4888 ,
1 i
1 I. )
n.,
132 l1 cetera. And these also included;five outdoorLtanks,.and about l
(:) '2 50 block walls. Looking at those components, we screened them
)j
, . l 3 out. based on the expert panel's recommendations in _the NUREG: ,I 1
J 4 43-36, I believe, and-in order to confirm;that there are no- 1
'5 weaknesses that.would violate the screening, we' performed the 5 plant walkddwns and for the screened-in components, we l 7 evaluated the structural fragility as well.as for the, equipment I
B -- their seismic fragilities, and based on the-walkdotvn-9 findings and based on what was perceived to be seismic 10 vulnerabilities, the utility performed some modifications to 4 l
11 the plant, which were discussed by Dr.,Murray,-and using the i I
12 information that was generated in 4'and 5, along;with the l
.]
13 systems equations that there were talked about, we calculated j
(} 14 the -- level HCLpF capacity for the plant.
15 And finally we summarized the results and some of the ]
16 conclusions, conclusions more from the Maine Yankee seismic 17 margins. review standpoint, as well from the seismic margins 18 review methodology, this being a five-month review of the'
( 19 methodology.
20 The review earthquake level is needed to screen out 21 components because the expert panel report gives a different l
22 review earthquake levels, which components could be screene'd 23 out on a-generic basis, and which need to.be looked at a little I
24 more carefully. So we need a review earthquake level that also.
25 that earthquake level should be defined in terms of the peak Heritage Reporting Corporation (202) 628-4888' i
1
133 1 ground acceleration but also' accounts forIthe frequency content E' of the spectrum. And the numerie rise in the selection of'the 3 new review earthquake level.that was selected _as the NUREG 0098 4; medium ground responsa spectrum for the rock size anchored to 5 ,3 G, and the way it is interpreted in the fragility evaluation G was to say that this is a review earthquake and there is what 7 is called the seismic marginal earthquake level upgrade, which 8 is the review aspect of the component all could be screened out 9 on the basis of the expert panel's recommendat ion, then the 10 review earthquake level is in fact the margin earthquake-level 11 for the plant.
12 If in fact all the components could not txe screened-1,? out and later on fragility analysis has shown that the' seismic
() 14 capacity of the response is something less than the review 15 earthquake levels, make .3 PGs and you get .21 or .27, and that I i
1G .27 anchored to the unit 00 NUREG -- it is the margin of safety l 17 for the olant.
18 So, there is a difference between the review 19 earthquake level and the margin earthquake level. They could 20 be tha same if all the components had capacities larger than 21 the review earthquake level.
.j 22 Now the way we have interpreted that is the review 23 earthquake is a marginal earthquake now in this case if the 24 event exceeded so that that earthquake is the event when 16 25 percent of the spectrum margin in this exceed the margin Heritage Reporting Corporation j (202) 628-4888 l
i i
1
i 134 1 earthquake spectrum for the. frequency range of interest and the' ,
(:) 2 direction of interest.
.3 DR. SIESS: How Jo you get the 16 percent?
- 4. DR. RAVINDRA: It is actually 4 percentile.
5 DR. SIESS: I don' t understand. Suppose I have got a 6 component at, sey, 6 hertz, and I have an earthquake that 1
7 exceeds the seismic margin spectrum at 6 hertz by a factor 2, B and that component is going to fail, what does'that mean? I l
9 have to have 16 percent of the --
10 DR. RAVINDRA: Sixteen percent of the frequencies 11 within the the --
12 DR. SIESS: But what does it mean in terms of.the 13 seismic capacity of the plant? That says 16 percent of the
(} 14 spectral ordinance exceed the seismic margin earthquake. Now 15 the spectral ordinance would be the G values at this --
16 DR. RAVINDRA: These are frequencies, sixteen percent
.i 17 of the frequencies in the range between 3 to 12 Hertz.
18 DR. SIESS: Sixteen percent of the frequencies? Now, I l 19 what does that mean? The frequency could vary.from zero to --
1 l 20 i Dot one here that varies from .01 to 100. What is sixteen 21 percent of that ?
22 DR. RAVINDRA: I am looking at frequencies of.
I 23 interest to me, which is within frequencies of within 3 to 12 l
l l 24 Hertz.
1 25 DR. SIESS: Say that again? ]
( Heritage Reporting Corporation (202) 628-4888 <
I l
l
_- _ _ _ _ _ L
135 1 DR. ~ SIESS: Within 3 and 12 Hertz.
- O 2 DR. SIESS: What'is 16' percent of the frequencies 3 between 3 and 127' 4 DR. RAVINDRA: When you look at the spectrum, the new 5 upgrade that has exceeded the margin, supposedly exceeded thel 6 margin upgrads, When you look at how:many -- or What a delta--
7 frequency, okay, it has exceeded the margin upUrade. And you 8 calculate that delta to the ratio of delta d'ivided by 3 to 12 9 Hertz, these severities 6.16 or not. That is the definition.
10 DR. SIESS: I draw your spectrum between 3 and 12 11 Hertz, and then I draw a' spectrum for the actual earthquake, 18 and if ov*r that distance, the second curva lies about the 13 first one, over 16 percent of the horizontal ordinates -- and 14 the horizontal ordinate is logritmic.
1 l 15 DR. RAVINDRA: No, it is Hertz.
l 16 DR. SIESS: It is frequencies? And.I am looking at 17 one. And it won' t be logarithmic,'it will be arithmetic. So 3 18 to 12 is 9, and 16 percent of that would be over -- and it 19 would actually be -- that is all in the calculation. 'Again, I 20 don' t see how that kind of a 16 percent relates at all to any 21 kind of a standard deviation, a normal ~ distribution.
22 DR. RAVINDRA: This is the mean plus one standard 23 deviation state LOFA.
24 DR. SIESS: Sixteen percent of the horizontal - ,16 25 percent. of the number? Bob, can you explain it to me?
O Heritage Reporting Corporation (202) 628-4888 L
L 136 1 MR. KENNEDY: Yes. . Bob Mennedy. I can'try. I'can l 0 2 explain why it was'enosen, at least, and try to explain.-
3 Basically the' issue.is that real earthquake spectra.are spiked.
4 .They have peaks and they have. valleys. But you want-to do 5 equipment evaluation for a smooth' spectrum because no engineer 6 knows the frequency of equipment perfectly' and you don' t want:
7 that taken advantage.of the valleys or be too penalized by the 8 spike. And in'this program it'was felt that we.should take the 9 margins people out of the people of having to estimate the 10 variability of spectra, but that we had to cover.that 11 variability within a conservative manner.
12 For the.samel reason that Dr. Numeyer many years ago 13 decided that he would have his standerd spectrum at the 84-() 14 percent non-exceedance probability, it was felt that'the margin 15 spectrum should be at about'that same level, so that it was 16 biased toward the peaks. And if you did that, no' single peak 17 is going to exceed -- if you force it so that in the frequency.
18 range of interest no longer of 16 percent of the real spectra 19 will exceed your margin spectra, then no single peak of the 20 real spectra will exceed the marDin spectra very much. They' re 21 not that peaked and va111ed.
22 All equipment has some ductility,.so that if it wan 23 unlucky enough to lie right on the peak, a very small frequency.
24 shift would shift you off of it and below the margin spectrum, 25 And so it is set to cover the majority of a real spectra but O Heritage- Reporting Corporation (202) 628-4888
'l:
137 ,
i allow some peaks of a real spectra to spike above it. I 2 DR. SIESS: All right, there islstill something '
3 confusing me.. I' remember seeingLspectra drawn'which were I
4 means, mean plus one standard deviation, the mean plus two.
l 5 s'andard c deviations, and the mean plusjone standard deviation' j 6' was above the other;one. And that-might be schie percentage -]
7 'above there. (.;
8 Nowwetalkaboutvariat%[nsfor 16 percent of the 1
9 abscissa. '
10 MR. KENNEDY: Well, that is exact'ly equ2 valent to the il mean plus one standard deviation spectra.' At . 84 percent ~ of the 12 frequencies have to lie below that spectra, as the spectra .
13 accelerations at 84 percent of the frequency have to lay below 1
I
(} 14 15 that spectrum, the spectrum accelerations at 16' percent of the frequencies can Do above that, that'is then, a:mean plus one !
_\
16 sigma spectrum shape, t 17 But it is mean plus one sigma-all'the way out-through '
l 18 the CPA. It is not putting a,mer.n~plus'one sigma amplification 19 f actor onto a median CpA. i l
20 DR. SIESS: I think.I understand it. i, l El MR. KENNEDY: There was a lot of discussions over l l 6
i l 22 this and it has confused many people.
23 DR. SIESS: Now, the reason for h&ving a definition d I 24 like this after you get through arrivi' rig -it, is that if 25 somebody has an earthquake and you want to decide whether you HeritaDe Reporting Corporation "
(202) 628-4888 i l
i il l
iI L_____________________________________i________________________.______ _ _ . . _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ __ 'b
138 1 have exceeded the evaluation earthquake or not, instead of-2 going out and looking for thin ~gs that nave failed, you could 3 Just plot the curve and compare it?
F 4 MR. KENNEDY: Yes, I think that if you have a future 5 earthquake, you would like very much to know whether that l
6 earthquake had characteristics.that would have exceutted the i 7 marD i n review earthquake level.
a ,
8 DR. SIESS: Have we got a similar equally-useful -- I '
I 9 might even say, " equally valid," definition of the SSE7 10 MR. KENNEDY: No.
11 DR. SIESS: Dr the ope, for that matter? I mean 12 right now we' ve got some rumors that,say if you exceed the OLE 13 you have to do certain thirigs, and --
14 MR. MENNEDY: TFMee is an EpRI program working on it 15 right now, by John Reed, who heads that program. That i s, 16 trying to face up to some of these same issues for the OBE as l 17 to how do you decide if you did or did not exceed the OBE.
18 DR. SIESS: I am trying to Det any of this down to 19 one figure of merit. It is almost ridiculous, but I -- okay, 1
20 Do ahead.
21 DR. RAVINDRA: Also, we interpreted the NUREG 0098 i 22 spectrum as the height of the response spectra from the two.
I
- 23 ort hogona'l - ground -- directions. The principal 24 interpretat ion of the ma'rDin earthquake is the -- for the site
~
25 specific for Maine Yankee, it'~is defined by this in terms of I
Heritage Reporting Corporation (202) 62S-4838
r
D k
)st ,r, l- :)
, 6 x 139 on 1 the 16 percent spectral laPgeness, andalsvlhe[rnarginof,
~O 2 safety is de' fined as the higher of the two;qrthodonal ,
l 8 '6 f, r .
3 direbtions. '
y ) \, -g y ,
4 DR. SIESS: If I have got a spectrum /ji i n t ao ' q' ' ^ {
V t/,
5 orthogonal directions, and they are let us say equal,.whatever 6 " equal" means,1 what would the _ spectrura' look ~ 1ike at - a direct ion 7 of 45 degrees to those? Same thing?
8 DR. RAVINDRA: It depends upon the particular o
9 components you are analyzing. y 9
10 '
Lt D R. SIESS: Are they.tvndom or is it - it is not ,
, y
~
11 like a resultants of forces where you take the square root _and s
12 "the ' sura of the squares -
1 l 13 ,
MR. KENNEDY: Could be higher. Could definitely be l
l~ A 14 higher. The most probable result would be somewhat higher, but U i 15 -odly very slightly higher.
16 D F7. ' S I E S S : Let us say I have an equal in two 17 directions, &' completely uniform di_stribution, it probably 18 seems to me more likely from tht[ fact' that 3t is an ellipse, m 19 and I Just happen to get the two equal diNections. -
20 MR.. KENNEDY: I' you go through probability --
21 DR. SIESS: You can gety anything you want..
3 f 22 M R. KENNEDY: - you can have in these two directions 23 equal and you have a probablistic combination, i . e. , they are 24 randomly phased, if some cut angle, the probable spectrum would 25 be trivially higher, and I say " trivially higher," one or two-Heritage Reporting CorporQ n (202) 628-4888 i
_ _ _ .1.___._ _ _ __ ___ _.t2 .
s 3 ,, v
~
! ,s t .
}
q 140 j i percent.
2 Now, if you believe the probability theory'really ;
3 '
applies-to that problem, then you'd believe the end. result.. ]
l 4 But if you go through the probability theory of these being J i, 5 completely random, equal components, the highest vector sum
- ( 6< will --
7 DR. SIESS: Put them in.the ration 2:1. What is'the l
8 probability that the ratio is greater than the two?
9 MR. KENNEDY: If the ratio is-like 2:1, by n 10 probablistic approach is the'one that was the two~is the 11 highest. The chances of exceeding that'at some cocked angle j i
12 are very long. l 13 There is no assurance that these instruments have i
14 been placed out in.the direction of largest response and least i
15 response, j j
16 DR. SIESS: What is the rationale-then for H 1
17 recommending average? The probablistic basis for that? That.
! l 18 is the recommendation, isn' t it? i e
19 MR. KENNEDY: There are margins program that I am 1
20 aware of, both Maine Yankee and Catawba and in-the EpRI l 21 recomrdendation which are more deterministically oriented. The j l
l 22 recommendation is to use the largest' of the two horizontal l l
23 components, an the description of the earthquake.
24 DR. SIESS: I think somebody somewhere said, "use the i
25 average."
4 Heritage Reporting Corporation (202) 628-4888 l
e
141 l 1 ME. KENNEDY: S-WRAP keyed their earthquake O. 2 experience data recommendation to the average to the horizontal !
3 components because we felt that the average might be a better i
4 descriptor of the damage capability of the earthquake. If I q l
5 was trying to predict a median damage capability, I would use 1 6 the average. But if I was trying to predict the HCLPF damage l
7 level, I would use the larger of the two.
\
8 DR. SIESS: Okay, good point. j 9 MR. MURRAY: The essay report says " average," and 10 then since then, John has changed his mind. John Reed. The
{
11 EPRI report, the compara2cn of the EPRI record where we asked I
12 them to tie down, p' ease define what you mean by the " margins 13 earthquake" so we can stop this discussion over and over again, j
() 14 And that point, which was earlier this year, the average was 15 stated. Since that, John has changed his position on that, 15 participating in his HCLPF comparison study back to peak, l
17 DR. SIESS: Peak will establish strong motions i
18 seismographs when they orient the two directions of any l l
19 seismological basis, or do they just choose north-south-east- '
20 west 9 21 MR. MURPHY: Ninety-nine percent of the time it is j 22 north-south-east-west, so they can find it.
23 DR. SIESS: Maybe we ought to lay our plants north-24 south-east-west and use diesel.
25 MR. MURPHY: Orient'our diesels at that? You put in ;
O Heritage Reporting Corporation (202) 628-4888 i
142 1
You 1 digital recorders and it is not'that much of a problem.
-(:)- 2 can get the correlations while you are_doing it.
3 MR. KENNEDY: It will not affect the calculated HCLPF y
4 value. It will effect the interpretation of the value, the 5 fact i s, because they'took this factor and they applied it in.
6 the direction that was the direction that would cause the worst l 7 damage to the component. But it' depends on how you interpret.
8 the HCLpF results. I think the way Dr. Ravindra's viewgraph -
9 said it is the better way.to interpret the HCLPF.results'by the
- k 10 larger of two horizontal components. .)
11 DR. SIESS: You say the " actual HCPLF value" as if it 12 were an actual.
13 MR. KENNEDY: Yes, I am sorry. ,
1
() 14 DR. SIESS: But it involves a little interpretation, 15 too. i 16 DR. RAVINDRA: Okay, we go onto the second topic.. I 17 The initial screening of components, the screening is 18 based on generic capacity, the seismic capacity,-as derived-19 from the seismic PRAs, the experience data and the 20 qualification test data. Combined with the expert panel's j
21 testimony, the screening table was built u p, and that table has ;
1 22 been used in screening the components in Maine Yankee, and the 23 focus is en the group A system f unctions on the components that 24 support the group A system functions as described by Dave, and i
25 once we do that with that focus, we have identified the target Heritage Reporting Corporation (202) 628-4888 H l
i
143 1 areas and components and systems for the walkdown based.on what 2 was screened out and which components had to be looked at more 3- thoroughly.
4 Here is an example of the kind of screening that was 5 .done based on the generic seismic capacities. For containment, 6 the expert panel said that the containment has in .act, a much-7 higher capacity - mu'ch higher than . 3 G, and as such, one need 8 not look at the details of the containment that are in the 9 review of the plant or in the seismic margin review, or in the 10 walkdown.
11 We confirmed that because an independent study was 12 done on Maine Yankee on concrete containment and.in fact the 13 HCLPF capacity of the containment was shown to be in. excess of
() 14 1.0 G, and as such we agreed with the panel's . recommendation 15 and screened that particular component out, that particular 16 component was not looked at further.
17 DR. SIESS: Hashimoto's study was specific to Maine 18 Yankee?
19 DR. RAVINDRA: That is right. Again.I have said 20 there are duel roles that we are playing. One, total plant 21 review of the plant -- of the current plant review of the 22 margin methodology, and the Maine Yankee seismic margins 23 review. So the remarks will pertain to how we interpret the 24 expert penel's interpretation and hcw we confirm them.
25 Similarly, for the HCLpF and the supports, the second O Heritage Reporting Corporation (202) 628-4888
144 1 is done by Livermore aslpart of the DNC evaluation of the 2 combustion engineering reactors, we found that.the l
/
3 interpolative support being 5, have high seismic capacities, 4 and as such, we could screen out interpolative supports.-
5 The reactor internals, this is a combustion 6 engineering reactor that has undergone'either a seismic margins j 7 review or the seismic pRAs-did not' include the seed reactors, q 8 and as such, the expert panel's recommendation was you had to; 1
9 go back and look at the internals and-convince yourself that it ,j 10 had to be screened ou', t or do some fragility evaluation, 11 In the process of this study, we did talk to'the 12 reactor internal' vendor, Combustion Engineering, and got their 13 stress calculations, on a comparable plant and also specific to :
(} 14 this Maine Yankee, and we made some fragility calculations to 15 convince ourselves that the HCPLF capacity is indeed larger 16 than .3 G. That is all we could resolve that question of 17 whether to screen in or screen out the reactor internals, i 18 Similarly for the control rod drive mechanism,-
19 although the expert panel's recommendation was that it could be ;
1 20 screened out at .3 G during the peer review,.there was a 21 concern expressed. The control rod drive mechanism, what.they 22 call the " control element assembly," mechanism in-the C-23 reactors, they look different to-the ones we hav'e seen in the 24 other pWRs, so we did look at the drawings, we looked at the 1 25 stress analysis that are done by Combustion Engineering, and' O Heritage Reporting Corporation (202) 628-4888
'- # d' _.._-_. ____.________ __._ _ __ _ ___
145 also we talked to the. people at C-E, and asked their opinion
()
1 2, and.their judgments as to how similar it is to other plants 3 where they will qualify it with a peak level.like .67 G on a 4 LOFA and they are also data to show that they are qualified for 5 a higher earthquake level at Calvert Cliffs and Millstone, so 6 we use all that information to convince ourselves that the CRDM 7 at Maine Yankee is indeed, has a high capacity and it could be 8 screened out as recommended by the expert panel.
9 One other area here. The steel structures. There 10 wasn' t any recommendation in the expert panel report because 11 not many steel structures were reviewed in the seismic PRAs and 12 we did look at the steel structures. There are three of them, 13 and one, the circulating water pump house structure, we looked
() 14 at iti we evaluated the seismic fragility based on some of the 15 analytical models for the delta, and came to the conclusion 16 that indeed it has a high capacity, higher than .3 G, .and as 17 such, we could eliminate it at a second round of this ]
I 18 calibrations. l 19 The block walls is one thing that the panel 20 recommended that we need to look at the block walls, and it is 21 also an important system interaction question, so I am going to ;
22 switch to the detailed viewgraphs just to go over what we did 23 in the block wall evaluation.
24 For Maine Yankee,' we took -- undergone a seismic '
25 retrofit according to the I&E Bulletin 80-11, and there are O Heritage Reporting Corporation R (202) 628-4888 r
146 1 several block walls located at Maine Yankee, and many are O 2 -adjacent to circulator equipment, and typically these are 3 unreinforced, and of course, the soismic experience and.also as 4 recommended by-the expert panel, it is shown that the walls are 5 .in fact seismically vulnerable, and that was responded to by 6 the facility by the retrofit that'was doneJto I&E Bulletin 80-'
7 11 and in that there were identified and categorized all the ;
8- block walls by their safety significance, and'their line on the 9 retrofit, such as original dressing and the block' catchers and- i 10 -- way components, include not only the front leg components,.
11 but also the support systems, the light lines, such as cables, 1
12 piping, ducts, et cetera.
13 We will exclude our data on those components that
(} 14 will not cause damage even if they. collapse but because they 15 are shielded. Or if they collapse they will not hit the group 16 A components. Or they may collapse but even if they hit-the l 17 components, they will not damage the components, for example, 18 the large piping, the failure of the block wall -- will not 19 affect the functionality of the piping.
20 So with that kind of --' ceiling approach, we identify 21 certain components, certain block walls that needed to be 22 further evaluated as to their capacities -- particular 23 capacities. And our analysis'-- shows that all block walls 24 that were screened in at this point, do in fact have capacities-25 larger than .3 G, because of the confirmation.of the analysis.
( Heritage Reporting Corporation (202) 628-4888'
147 1' that was done by Maine Yankee in their response to 80-11, hiDh j O 2 -- event, are because the' consequence of the retrofits and that 3 the previous retrofits.are the ones that are' upcoming at.--
4 Some of the retrofits on the block wall.
5 Based on all that.we confirmed th,at the'HCLpF ] a 6 capacity is large -- it is larger than .3 G for.the block 7 walls. And one of hate retrofit, for example, is a one that is 8 added to the block wall around.the containment spill pan.
9 So this is one way that we have' looked at the systems 10 interaction effect and also to one of the things'that'was 11 strongly emphasized in the panel before, i
12 Going on to the plant walkdown itself, after initial I i 13 screening of the components based on the generic capacities, we 1
() 14 wanted to conform those generic capacities are in fact present
- 15 in the plant as it exists, and that is done because there are l
T6 seldom caveats that are included in the screening guidelines of 17 the expert panel, and you wanted to make sure those caveats are 18 addressed in the study, and also to confirm that there are no 19 substantial weaknesses in those components.
20 Going on to the walkdown,.the procedure was we l
21 located all the components for walkdown for group A components 1
22 that were identified by E&I on the 80 drawings and identified
]
23 building materials that require access to the walkdown, and j I
24 this time the initial screening, we made a list of things that ;
i 25 we want to look at when they walk down by the' side of the i i
Heritage Reporting Corporation (202) 628-4888 l
a_-__-_--_-_ . _ . _ _ _ _ _ _ _ _ - - _ _ . _ - _ _ _ _ _ _ _ _ - - - _ _ _ _ - _ _ _ _ _ - _ _ _ _ - _ -
i 148.
1 particular components we build up walkdown data. sheets: some of O 2 them are included in the volume III of the NUREG, that Bob 3 Moody mentioned, and the results confirmed'that indeed the 4 screening criteria of the expert panel is_ satisfied for each 5 component.
6 At the first_walkdown it was also intended to 7 identify areas that will require additional review. -The first 1 I
B walkdown findings -- this is~Just an example of three or four 9 pages of this symbol, Ewe-looked at the --- fine water pond and 1 10 the little dam there, but on the embankment, and we concluded-11 that if there is a. seismic failure of that there is.not enough 12 uater to flood the plant:and there is a good chance because of 13 the topography, the water will run out in the other direction.
() 14 So that is a finding of the walkdown.
15 For the -- trace, we -- the -- trace for aluminum ---
16 trace, -- -- we had very difficult for the industrial -- t race 17 so we can prove that it does compare with the data base systems 18 and it is not a problem.
19 And also we confirmed that Maine Yankee has down some 20 -- o f t h e insert to the ceiling, so the concern about supports 21 does not exist.
22 DR. WYLIE: So there is no' massive seismic-supports 23 on the trays -- they are just rod-hung?
24 DR. RAVINDRA: Yes.
25 DR. WYLIE: Great.
j O Heritage Reporting Corporation l (202) 628-4888
j l
-]
' 149 1 DR. SIESS: That is t'he way they did'it in those
.O 2 days.
3 DR. WYLIE: Great. That is the way to do it.
4 DR. RAVINDRA: The pumps, we found that most of the 5 pumps -- we found all the pumps were well-anchored and :J 6 consistent with the panel guidelines, and there were three 7 pumps or so that did not meet the init'ial steaming guidelines ;
8 of 20 feet unanchored cantilevered shaft, but we found~that one i
9 of t he -- the bottom of the shaft was inside, there was aLbi D.
10 connection so there was no problem with excessive. displacement 11 because of the cantilevered shaft.
12 So, as such, we could conclude that the pumps'are in.
13 fact well within the panel guidelines and we could steam them ,
I
() 14 from the pump size shown in the previous table.
15 We talked about these presentations. Morie and Bill 16 Moore there talked about --
batteries and the utility house 17 decided to replace them so that is not a. problem, and we looked 18 at the diesel generators and we found that they were well-19 anchored including the --- of the diesels.
20 DR. SIESS: Excuse me. I noticed you have a-21 hierarchy of anchoraged. You have " anchored," "well-22 anchored," and "more than sufficient anchorage." Do those have 23 engineering meanings or are purely rhetorical?
24 DR. RAVINDRA: During the retrofit for external- ,
1 1
25 seismic evaluations that were done at. Maine-Yankee, anchorage, j
.i Heritage Reporting Corporation (202) 628-4888 ;
1 j
_ ____ _ - - - - _-. --_--- -_-_.2__- . _ _ __ -
150 1- I mean, original anchorage have been put~ on'the components.
O. 2 Sometimes the-anchorage has been much larger in terms of tho 3 anchorage and the lateral. supports, that'have been added. ')
4 Typically:much larger than what if you are desi Dning the 5 equipment you.will not add such beefy' anchorage. It was done
'6 because of'the water level of that' time and also Just to avoid 7 .a problem they added additional anchorage.
8 DR. SIESS3 probably didn' t even calculate the
(
9 anchorage. .
10 DR. RAVINDRA: So there is a hierarchy.in those three
.i 11 terms.
~
12 So for the screened in components we' talked about the 13 three structures and block walls. We looked at the outload
() 14 banks, the RWT, pWT steam-in water to storage water tanks, so :
15 we evaluated the possible that it will not meet.the expert 16 panel's accommodation stage that you have to evaluate these l 17 tanks because typically they don' t meet the p.G. review I
i 18 particularly well. I 19 At first we targeted these for seismic criticality 20 evaluation, the vertical. storage tanks, and the details of that l
'l 21 argument in the report come back to the capacities for them a i 1
22 little later.' 1 i
23 DR. SIESS: I think you.can skip the details.because 24 those sie all in the reports.
25 DR. .RAVINDRA: Yes, they are all in the report'. s Heritage Reporting Corporation I (202) 628-4888 I
1
-151 1 DR. SIESS: If somebody has1 readL themland has.a. ,
2 question they want to raise about it?
3 DR. RAVINDRA: Th e -- - t h at i+e looked at were the 4 pumps, compressors, valves,. clean air changers, coolers,Etanks, 5 et cetera. They are all listed here. Let us go to -- the.
6 valves, for example. The ones that were,.when we, the expert 7 panel recommendation said that'we have to do a sample revie'wfof 8 the valves to convince yo'ursel f that there are --- that: they. jI 9 meet the criteria such as the eccentric . length'of.the? operator 1
10 and the weight of the operator and L the size of the five ----
11 such criteria have been established by Escraft & Scruggs and 12 one of the recommendations of the expert panel was to convince
'13 yourselves that indeed the wells would not -- indeed the whole
(} 14 criterion.
15 So we looked at the valves that were something like j 16 5%) valves or so and we looked at each one of them in the 17 walkdown, and compared that with the experienced mounting 18 spectrum --
19 DR. SIESS: And that was?
20 DR. RAVINDRA: Developed by Escraft & Scruggs -- and 21 convinced ourselves that indeed the valves' meet the criterion.
22 Similarly for the heat exchangers and coolers, we looked at the 23 -- heat exchanger and the diesel direct heat exchanger, and 24 soma other heat exchangersi We looked at the critica11 failure 25 modes as the silent supporters and the connection' group in the 1
( Heritage Reporting Corporation (202) 628-4888
-1 I
l
-152 1 saddle to the support, to the' foundation, and-that was down in
- 2. the already evaluated seismic capacity or the incapacity of the 'l 3 particular. component.
-4 And the whole purpose there was for each component, 5 we are identifying what'are the critical failure modes and what 6 each failure mode, what would be a realistic capacity of the 7 component and what is the realistic response that the' 8 earthquake level would impose on the corpponetst and using that 9 information in the process you would also evaluate-the ,
l 10 reliabilities and uncertainties.in those capacities'and
]
11 responses for each critical failure mode and y~ou would identify l
12 what is the median ground --- capccity that it would withstand -l i
13 and derive from that the high-confidence - low probability of ]
1
() 14 failure capacity for the component. That was.done for this l 15 entire list of equipment types of review, and in the report 16 these are explained in a little more detail for these three i 17 components, the -- , the diesel - - , the ---- tank in its 1 l i 18 modified condition and the containment spread pans, also in 19 this modified condition. ;
20 We looked at the next --- this time the. pump water 21 and the review of the plant, . certain items were identified as 22 needing modification and these have been discussed by Walt 23 Murray in --- 1-4801, system batteries, diesel spread pans ---
24 and the ones that are also system interaction related high ---
25 such as the welding cart, ---- and emergency lighting.
O Heritage Reporting Corporation (202) 628-4888
l
'le 153 1 So.far.we'have discussed the screening, and'the l
2 capacity evaluation, the acute capacity evaluation [for'the
'3 speed in' components. The next. topic would be'---- components L .
P 4- that have high seismic. capacities, but as a system, whether the L 5 block level capacities,-the' block level HCLpF,'is it. higher i-6 than the lowest component capacity?
7 Now, in order to evaluate this, th6re'are two 8 approaches that are' suggested by the expert. panel second 9 report, the guidelines report, where you could either do cn a 10 deterministic basis'where having. evaluating, having developed 11 the ---' inspection like this, because thesefare on the union of.
12 components for 7 and 20, the lowest capacity, the. field 13 capacity of the lowest component, is the plant level HCLPF'
(} 14 capacity. )
15 When the components like this with an ---- that'is, 16 either these components are to fail, one'of these or one of 17 those are to play together and this sequence.---- and in order 18 to evaluate that, the deterministic approach would include, 19 would say, that you take the minimal Duide, the minimal guide 20 and take the higher of those two terms in parentheses. So that.
21 is the deterministic approach.
22 The probablistic approach will consider.these i
23 components as described by their seismic fragilities, we not ,
24 only look at the HCLpF capacity but the entire fragility 25 course, in order to calculate what is the seismic fragility.of Heritage Reporting Corporation (202) 628-4888 L:
l
154 1 this small LOCA, LOCA sequence.- Similarly for the low LOCA O 2 sequence.
3 DR. SIESS: Slow down. I think it_would_be_a lot 4 clearer to me if you would' limit this1to'one simple case and 5 compare the two. Don' t put these in with the.small LOCA and 6 the no-LOCA. Take the small LOCA case and show.mo the two.
'7 .different procedures.
8 DR. RAVINDRA: Th'e small LOCA' case. Let me put up 9 'ths --
l 10 DR. SIESS: .The first one you are calling 11 deterministic versus probablistic?
12 MR. MOODY: Yes.
13 DR. RAVINDRA: I will go through.the deterministic 14 one first. The' deterministic- part 'of these compenents -- !
15 DR. SIESS: What would be your answer? I can see the f l
16 numbers up there. j 17 DR. RAVINDRA: For these components, 4, 7 and 20, the 18 HCLPF capacities are calculated to be .3 G --
19 DR. SIESS: We can read them. Just give us the j l
20 answer.
21 DR. RAVINDRA: point three-zero, .21 times something 22 else.
l 23 For the deterministic evaluation of the HL'LpF ' '
24 capacities for-a LOCA, a small LOCA, it would be the lowest of 25 the HCLPF capacities --
O- Heritage- Reporting Corporation (202) 628-4888
l i
155 1 DR. SIESS: All right,.so 177 :
H 2 DR. RAVINDRA: It would be .21.- )
i 3 DR. SIESS: I.am sorry,. 21. Okay, now what would it 4 be by the other method?:. -i 5 DR. RAVINDRA . By the other= method,'if this example, 6 because of the kind of dominance of one low-capacity component, 7 it would also be .21.
l 8 DR. SIESS: Then I don' t have the. slightest' idea what l 9 you are talking about.
10 D R. RAVINDRA: What we do in the other method -- can 11 b e *-
12 DR. SIESS: In the first place I don' t quite see how 13 this one is deterministic? 1.believe your language would
() 14 assume all of method A and method D --
15 DR. RAVINDRA: It is deterministic in the sense --
16 DR. SIESS: No, don't.try to educate me on the 17 difference between " deterministic" and "probablistic." Just 18 give me up and call it '
"A" and "B." And where is this 19 described? I read a lot of stuff'that I can' t . recall seeing 20 this.
21 DR. RAVINDRA: It is described in --
l 22 DR. SIESS: Volume II? ,
1 23 DR. RAVINDRA: Volume III.
24 DR. SIESS: That is all right. If it is in volume II~
25 I will ---
l Heritage Reporting ' Corporation (202) 628-4888 i
j
l 155 -1 1 DR. RAVINDRA: It is.in volume.III,' section 5.
O' 2 The other methods we tahe into account the seismic-3 fragility of component 4,. discussed lby the. set of fragility 4 curves -- as shown by this family of curves, okay?
5 DR. SIESS: Instead of Just taking the one point . of f-1 J
.6 you use the whole curve?
l 7 DR. RAVINDRA: Okay so what you will do is if this is~ q d
8 one component, then another component that is.similar.to this i 9 which has fragility growth this far, let us say there;are'three 10 curves. So what you will do is to combine this curve with the-11 similar curve for the other component and combining in the 12 sense a failure record, either this -- let us say this is Ll 13 component 4 and let me show you another component --- t h e !
() 14 failure record is either 4 or 7 at the phase. So1that means 1 15 that at any acceleration we'take this thousand year failure and j j
16 that similar positive failure, and typically I average the 17 probabilities of failure. I am not at there --- okay?
18 And later on I did another --- the code that is --
-j 19 that code represents either this component fails or th'e other j 20 component fails. So that I can build up a' family'of codes for j 21 the failure of. components -- 3 1
22 DR. SIESS: It is a new conditional probability? -)
23 D R. RAVINDRA: .And the advantage -- j i
24 DR. SIESS: Which gives the lowest value? J t 25 DR. RAVINDRA: If there are a lot of componenta in O Heritage Reporting' Corporation-(202) 628-4888 .i L1___ ._ ___ _ _ __ _
l V ,
j
'157 1 the system, okay, there.are in a union -- if any'ons of them-0 2 fails, the system would fail, then this method of adding the 1 3 specific codes, this would give you the lowest'value. But-4 aDain,'it is a function of how'much faith you have in the 5 entire fragility code. I am willing to attempt.to have more 6 faith, i
7 DR. SIESS: 'You tend to have more faith?- l 8 DR. RAVINDRA: That in done. In this particular 9 study we looked at two accident sequences, but both sequences-10 must be combined in'the fragility codes'as described. Now, the
- l 11 fragility codes pertaining to.the accident sequences and 12 derived from that what are the HCLPF capacities? You are going )
l 13 to have 95 percent confidence and five percent probability lof' l
() 14 failure, and that is because what is the peak zone accident 15 that corresponds to that point.
16 And one other thing that we could do,' using that 17 approach was we could influence seismic and non-seismic 18 failures. Other wide abilities. And the non-seismic ---
19 abilities, is the shown by Dave in his presentation. They were 20 given to us in this fashion where you had the median and other i
21 viability for demand and the other --- curve there represents ,
I 22 the uncertainty in that other wider probability. j 23 And in doing the so-called probablistic - load and 24 group capacity calculation, we.could also include any concern )
l 25 to automation within failures. That i s, if you could take that l
1 v Heritage Reporting Corporation ,
(202) 628-4888 l
- _ - _ _ _ = - - _ _ _ _ _ _ - _ - - _ _ _ _ _ _ _ - _ - _ _ _ _ - _ . - _ _ - _ _ - _ _
158' )
1 component B and 14 on highly correlated, you can. treat them in O 2 the probablistic method to some extent you can treet it also in' ;
j l 3 the deterministic method.
4 Now, the sum of the results are still on.this table,.
L 5 whereas we looked at'the small LOCA and calculated the ---
l-6 small LOCA level in this capacity ofL5-21, where you assume 7 that the seismic schedules are. totally dependent and if you
)
8 suspect that-there is some dependence,-because again there is. 1 l
i 9 one component that'had a 20 corresponding to thus , mponentsin 10 a --- capacity that is among the dominant, as ,ach,.you would 11 not discern much difference between these two cases.
12 From the -- nuclear --- case, the.non- --- capacity 13 is final,
() 14 (Continued on the next page.)
15 16 q
17 -)
i 18 19 20 1
4 21 !
22 23 ;
24 25 O Heritage Report i r.g Corporation !
(202) 628-4888
159 1 DR. RAVINDRA: The total sequence with non-seismic O 2 failures, we found a slight difference in HCLPF capacity. We 3 thouD h t whether to treat non-seismic failures or not cannot be 4 answered by looking at these results, because there may-be 5 other plants where there could be a lot of components with
- 6 non-seismic failures, then they.would show up'as important 7 contributions.
8 And'because we looked at two different, two-in 9 sequence apparently and the fact that'we had to concede that ;
1 10 there would be a small LOCA, because we.could.not go'inside a' ~!
11 rertainment to look at the in-post lines position, the total 12 fact of non-HCLpF capacity, we had to assume this .213 and 13 maybe it was larger than .21, it would be clo'ser to .32.
14 And by looking at -- by saying that'the small LOCA
( 15 break occurs in one-hundredth of 1 percent of the times as 16 opposed to the LOCA cases, and if you combined these two 17 sequences like that, you would find that the HCLpF capacity of 18 .32-G, 19 DR. SIESS: I don' t get the same number as you do, 20 using the equation that is in volume I. Why is it that when P 21 is 5/10, I don' t get simply the mean between the~two cases? 22 Twenty-one and 32 divided by two is not 23, 23 MR. MOORE: Some of your components are in both 24 equetions. And so, you cannot simply combine them. 25 DR. SIESS: Well, this'is a sigma'in front of all of O Heritage Reporting Corporation (202) 628-4888 _. _ ._ _ - _ - - _ _ _ _ _ _ - _ - . _ _-_-_. U
160 1 this, you mean' somewhere? O, 2 .It~goes item-by-item? 1 3- MR. MOORE: Go back to.the bullet and you'will see 4 it. 5 DR. SIESS: Well, all that I know is that it says 6 that the core damage is p-10, the small'LOCA core damage, plus 7 1-minus O times the r.o-LOCA core damage. That is what it says-8 in the report, and I'have got'a table up there.andione of it is 3 labeled no-LOCA and one of it is labeled small LOCA, because it 10 does not mean that. 11 MR. BUDNITZ: That thing is not correct, tha't;p and 12 1-minus p. 13 DR. SIESS: Thank you, that is good enough, I will () 14 buy the numbers, I just could not get it1out of that equation. 15 MR. BUDNITZ: Seven is also invoked although it does 16 not affect. 17 DR. SIESS: The equation that in in the report is 18 wrong, you don' t do it that way. One of them is wrong. Either 19 page 0-9 is wrong or page 3-10 is. wrong, and frankly I don' t 20 care which. 21 Sometimes you try to make things too simple. If you 22 make them simple enough I can understand them, because I 23 calculate it and don' t get your answer. 24 DR. RAVINDRA: Okay, then the conclusions -- 25 DR. SIESS: I think that those are pretty obvious, O Heritage Reporting Corporation (202) 628-4888
161 1 let us read them and we are running about an hour behind.
-(:) 2 D R. RAVINDRA: Okay,-in terms of the methodology l- 3 i t'sel f, we have identified several areas where the methodology 4 could be improved and there is aga.in, in.the last chapter'of 5 this report.
6 Thank you. 7 DR. SIESS: Thank you, very much. 8 Now, the next thing that we have is Mr. Budnitz who 9 is. going'to comment on the Maine Yenkee thing and then I am 10 just looking a little. bit ahead. 11 And then we are going to have comments by Maine 12 Yankee, right? 13 And then we will -- it seems to me that ' the fut ure () 14 plans go somewhere down the line. 15 Which hat have you got on for this now? 16 MR. BUDNITZ: I have got on my west coast Pacific 17 Daylight or whatever time it is. 18 We made a mistake, I mean all of us here, practically l 19 are from California and we should have held it out there. 20 I have fifteen minutes. l l 21 There is a little rowboat with about 25 people in it, 22 and for different tasks five of them are picked and they1 row 23 the oars and anyway, the peer review group for this had Mike 24 Bond, and John Reed, and Jim Conniff, and Loren Wylie and the 25 objective of the peer review was to assure that the th'ingLwas O' Heritage Reporting Corporation-(202)'628-4888 l
162 1 done properly and'there is more work there, which I copied out O 2 of the charter, but that is sort of the idea. 3 And just what peer review i s, Just what you think it 4 i s, but in fact, it did not work that way and I want to explain 5 that here, in 30 seconds before I go on. 6 This was not a peer review of the 'tradit ional kind 7 because it was a trial review, and as such, the peer review 1 ! 8 group was intimately involved in the conduct of the study, 9 rather than merely its review. 10 At almost every stage the peer review group was 11 consulted and various technical issues that arose, were bounced 1 12 off of the peer review group, and we told them some things, and 13 tney sometimes took our advice and sometimes we argued out, and () 14 decided to do something else, but because.of that, it was 15 almost as if the peer review group was a part of this study in 16 its execution. 17 And that was mostly because the guidance in the 18 guidance documents was not complete and one learned that as one 19 went through it and so more guidance was needed, or else some 20 of the guidance was there, but did not make sense. So, the 21 objective is what I said. 22 How the peer review was accomplished -- the peer 23 review group participated in all of the walk-downs, that is, 29 that there were two walk-downs and we walked around with it, 1 23 and St the same time, we were there for, sometimes two. days, O Heritage Reporting Corporation (202) 628-4888 L
163' f-1 and then they were there another two days and that sort of V 2 thing. 3 DR. SIESS: All five of you? 4 MR. BUDNITZ: Yes, and all of us were there for all 5 of it, except that one guy missed part of one, I guess. Is that 6 right? 7 Yes. And again, we were basically there all the 8 time. And we interacted'a lot and we had various meetings. It 9 was a much more intense review than your normal peer review, i 10 which comes along afterwards'and you put it on the: table, and 11 they review it and they say, gee, that looks great or does not. 12 So I will repeat the central findings which are all 13 in the letter that was written because I want to get on to the (} 14 substance quickly. 15 The central findings were, that.everything was done 1 i 16 fine, the project team did follow the guidance, although it was. .) 17 ambiguous and we had to amplify it. The study was executed in 18 a competent professional manner, and used state-of-the-art 19 methods, which, by the way, they were developed, in some cases, 20 as we were going through it. I 21 The state-of-the-art was a moving target which is 22 fine, that is what the notion of a trial is supposed to be. And 23 they took cognizance of ours, we could tell of all relevant 24 information, including-information, that in some. cases, was not 25 published, or that only some of us knew;about and others did Heritage Reporting Corporation (202) 628-4888 1 i
1 4 164 1 not'get their hands on. There was'a lot of just sort of O 2. feeling your way.
.3 There were a whole bunch of other comments'which I 4 have labeled A through H, and H is 8, and1there are eight-5 different comments whien was in the letter which was written --
-6 i f you' don' t have it, this is what I am.saying is a summary of J
7 this letter of February 27, and you can get the details from' B that, and if you don' t have it, I will make copies, but it.ls ) i 9 in the first volume. 10 And okay, eight comments, other comments besides - 1 11 these, which by the way,.these are sine.qua non, if any.of 12 those on the top had been, you know, gee, 'it was lousy or it 13 did not follow the state-of-the-art,.then the rest would nave () 14 been moot, but these are technical comments which are important 15 to the methodology. 1 16 First, selection of the review level earthquake. We 17 felt that more guidance was needed to clarify some of the ; 18 ambiguities, and in fact, that has now been accomplished 19 through interactions with the EPRI Program. If the guidance 20 documents were to be rewritten we would now understand better 21 how to specify the review level earthquake and I won' t go into 22 that any more here. 23' DR. SIESS: You don' t mean the E Val ue, you mean the - s 24 -- i 25 MR. BUDNITZ: I mean the specification of it,- the j
. .]
O Heritage Reporting Corporation l (202) 628 '+888
?
i _ _ _ = _ _ _ _ _ _ _ _ _ _ - - _ _ _ _ _ _
f., ' 1 i 165
. 1' way, what the'84 percent means and what the frequency' range of-2 interest and the interpretation of the, - and the horizontal 3
3 components and so on. And interpretation thery/.gof,.and-so on. x, 4 And I won' t go into that more, except just'to clarify J 7 5 one thing that was on Robbie's slide, I think was not.quite
.e 6 right. ,
7 At Maine Yankee, the averagefof the two components 8 was used, is that not right?. 9 DR. RAVINDRA: No, the larger, o 10 MR. BUDNITZ: The larger, okay, I had misunderstood 11 that. I stand 1here, to this day,'I did not understand that. 12 But in any event, we felt that better guidance was needed and 13 that guidance has since been developed, and it is in the EpRI-() 14 thing, and it is now understood., And surely.whenever we get 15 all of these things put together in one package, that is going 16 to be okay. 17 Combining seismic and non-seismic failures. We 18 commented that the methodology used at Maine Yankee, is not 19 rigorous and in fact, we Just had a little discussion here, 20 about some of that. 21 The reason that it is not rigo'rous is that if you are 22 going to use a deterministic method, there is no way to 23 incorporate a rigorous unavailability of-a diesel with as f 24 failure of the diesel day-tank or of's. pump or'something, that 25 can produce a rigorous answer, unless you do the full PRA i
' Heritage. Reporting Corporation !'
(202) 628-4888 o S ! - _ _ _ _ _ - - - _ _ - _ _ - _ . - _ _ _ _ _ - - _ - _ _ - . _ _ - _ _ - - - _ _ _ _ _ - _____ __ -_ _ _ _ i t' c __ _ .__ - _ _ _ _ - _ -_____J-_'
f
/
166 1- thing. O- 2 If you do, then there .is rigor,. but short of.that,.
/
3 there is no rigor,' and it ' has to be an approximation, but 4 approximation does not bother ine, because this whole. thing is 5 ' an approxitnat ion. I mean,.I don' t ' know what 5 percent of 95 6 percent means, and I don' t know what. arithmetic really tells 7 you in terms of uncertainties, you know, small uncertainties, 8 so that -I am not really nervous about that. 9 But the point'is that it is not rigc rous. l I 10 DR. SIESS: You don' t mind that appearing in the 11 record in quotes? ., 12 MR. BUDNIT7: I dc# t mind that, you can put it 1.n 13 quotes or whatever. l-14 But the peer review gecup did recommend that this be 1 1, i 15 developed more fully and is goir D to be. I am not.sure how it l 16 is going to come out, but 'it is a small pro,)ect that Dave Moore 17 and I are going to do in several months, which we hope will 18 give a bit more light on that, without saying that we are going Vi to reso ve it. But at losst we are going to give more thought
' / 20 to it, and there is a report already,'a few months' pid, that
, j 3 21 did a little more on that, h , ,
\' ^
L,. ( 22 r . And we also thought Ihat clarifica' tion was needed as i .> .i 23 to pow the neismic margin ward tp be' interpreted. You see, for
'l 24 example, Jim Thomas said,' w 11,let's suppose that' you have a 25 'ydiesel failure, a random dt;ksel failure that occurs at .~ 1 -G ? '
I
/ i ,-
/
O , Heritage Reporting Corporation i: j (202) 628-4888 'l
/' .;
( i i t y ,) a. l _2 m ,
~
}- .-.
b_.] l
if 14
")Y ,
,r _
' b 167. /
t
.o
'1' fYou know, it is~ going to occur the same at ~1.1-G-as
-a 2 it would ata.'g3 - G. , because it'is s- 3 not seismic. Or it-occurs'
~
4o 7 I l' 3- without an earthquake.
'4 So the question of how much margin there is, depends 5 on'what you are building on. Now, some of us did not like 6 that, and I did not agree with Jim Thomas about this,'but he 7 thougH that the margin ought'to Just be the seismic component.:
~
8 But I . don' t see it that way. I-think i. hat an. 9 earthquake comes along and you have to asi.Nhether it makes it 10 or r.ot? i " 11 DS. SIEPS: That is not'a margin at'all, p p 12 MR. BUDNITZ: Yes, seu , so tYere is in ambiguity lt k 13 there. , j
.,e c/'
14 DR. SIESS. My que tiEh is, why do you want to 15- combinosejsimicand non-seismic? 16 t MR. BUDNITZ: Well, suppo'se that I have an' earthquake 17 that cov.es along, at .2-G for Maine Yankee, a11'right,'andLit 18 fails component.A, but if component B worked, everything weuld s
. p
<f
! 19 be ghgt down. Then you woulu says q'ir e . terrific. I i 20 , All rfght, I am not saying that this is true of me, 21 but suppose that failed component.A,'and component,B w9s-22 unavailable with some very high. probability because it was a 1 23 lousy CoNponent and you haVe got a core-melt out of it, with. t 24 some rh sonable frequency. We* 1g 'I would want to worry *about 4 , 3 25 that. < t,
^ d' Heritage ReportinD Corporation ,
(202)'628-4888 y, 4
168 rs 1 You see, because, in fact,' it was the earthquake that f 2 caused it, the fact.that other stuff was not there is 3 participatory, but it is not - it seems.to me, if you are , 4 going to ask them, does the earthquake cause a problem at'some l 5 number, you have to stick that in there. Otherwise, you are l
'6 peing illogical.
7 . DR. SIESS: I Just don' t know where to stop. 6 MR. BUDNITZ: I agree. 9 DR. SIESS: .If both diesels are out of service, in 10 the plant. 11 MR. BUDNITZ: Right, two out of two. \ 12 DR. SIESS: They'have to shut.down within a certain 13 time, there is a'certain time available. () 14 MR. BUDNITZ: Right. 15 DR. SIESS: Now, if the earthquake occurs in that 16 interval, you have got a problem. 17 MR..BUDNITZ: .Yes, sir, so that'ycu have to ask t 18 yourself in a probablistic sense, whetherLthe probability of 19 these non-seismic failures or operator errors, whic'h is exactly 20 the same logic, whether that probability is high enough so that 21 added'on to the earthquakes, you should worry. 22 So that the cutoff that was established by the' extra-f 23 -- was sort of 1 percent, that i s, if the unavailability,was 24 sort of 1 percent, if it was 7 percent you kept.it in and if it j 25 was a 10th of 1 percent, you did not' keep it in and that was
'/' Heritage Reporting Corporation (202) 628-4888
169
- 1 based on the fact that the recurrence level, which you are.not 2 supposed to worry 1about in detail, but which we had in the back 3 of our minds, of these reviewer earthquakes, was sort of ;the
.4 10-minus-4 range, and so that sort of made it small-enough.
5 But I insist that if an earthquake _comes along.and 6 fails component A, and component B has failed for this other 7 reason, and it is a worry, I warit to worry about it. I . don' t 8 think that you can walk away from this by saying that it is 9 non-seismic. 10 DR. SIESS: And but i sn' t this something that gets 11 some insight from PRA's? 12 MR. BUDNIT2: Yes, and from PRA's we learned that the 13 ' combination of seismic and non-seismic failures turned out to () 14 be important as important as pure seismic in many of the 15 studies. Now, although and for some of them it is the other 16 way. l 17 But certainly there are enough of them wit'h those. 18 combinations appearing, to tell me that walking away from it, 19 would be erroneous, and therefore, you do have to do'some of l 20 that. 21 And the question about what the guidance ought'to be 22 and how you combine them, tnat is, you know, this 1 percent
.23 percent eutoff, or whatever you want to use, which is very, you 24 know, approximate, is something that has to be clarified in.the 25 guidance, O- Heritage Reporting Corporation-(202) 628-4888
170 l 1 And the peer review group'said that, as I said, there 2 were some, Jim Thomas, in particular, felt that it should not 3 be done, that this is a seismic program. I happen to disagree L 4 with that completely. This is not a seismic program, it is a l 1 i 5 safety program, but there is that difference of view, which is 6 fair to express. 7 DR. SIESS: And it seems to me 'that there is in some 8 respects that the challenge is within certain margins. 9 MR. BUDNITZ: Yes, sir, it does. 10 DR. SIESS: And he has to expand along with that. 11 MR. BUDNITZ: I agree. 12 DR. SIESS: And in fact, -shat is it good for? i 13 MR. BUDNITZ: Yes, nicely spoken'and when you come to () 14 the discussion of the EPRI methodology and the success path 15 approach, and whether or not that is done as well in this l 16 approach, that is an issue that has to be thought about. l l 17 DR. SIESS: The seismic qualities do not tell us 18 everything that we need to know about the actual situation. , 19 And it certainly does not tell us everything that we need to 1 20 know about seismic -- 21 MR. BUDNITZ: In any event they don' t, even for pure 22 seismic, but without including these' things, they tell you even l l 23 less. That i s, that leaving this out, I would' consider a ! L 24 serious or a major flaw in the methodology, not a minor flaw, 25 because so many of the PRA's have found seismic and l , Heritage Reporting Corporation (202) 628-4888
- -_ - - _ . . -___--._-----_-_-~_____a
1 I i 171 l 1 non-seismic failures combined as producing. sequences of 2 concern. 3 That.is a lesson which jumps right out of the page, 4 if you go and study all of these seismic -- a bunch of them. 5 have it. I 6 DR. SIESS: What is that? i 7 MR. DUDNITZ: PRA' s. l-8 DR. SIESS: I would not be surprised if the loss of 9 outside' power would not be one of the' major' contributors. 10 MR. BUDNITZ: No, no, loss of outside power is a 11 given here, with the earthquake. 12 DR. SIESS: I know, but it is likely to be.a given 13 with the earthquake in reality. () 14 MR. BUDNITZ: Yes, I mean, that is why. 15 DR. SIESS: But if you throw a loss of outside power 16 on top of certain random failures, they get a lot. worse. 17 MR. BUDNITZ: Oh, I agree with that, and that: is why ; 18 in PRA's without earthquakes, just PRA's, loss of outside power 19 is an important initiator for core melts, just generally. It 1 l 20 is not typically the dominant thing, but it seems to pop up 21 quite a bit, that is fair. 22 All right. 23 Just moving on, my fifteen. minutes is chewing up l 24 here. . Number C, the CDFM method and the fragility method, 25 there is a whole study' going on that Murray talked about, I: :j 1 l l Heritage Reporting Corporation ' l (202) 628-4888 l I l 1
4 172 i j
- 1 won' t talk about that any more.
2 But we urDe that that be done and it is being done, 3 but when we wrote it, it was not being done and that wan
- 4. basically because the budget that took time to get going.'
5 Level of' expertise needed, we felt that.the fragility i 6 team must have certain experience and that the systems analysis 7 team must have certain experience, and the fact is, that'there B is not enough experience out there, to go and do all 100 plants 9 this year, let's say. 10 If somebody said, do all 100 plants this year, God' 11 forbid that that would be' asked, there simply is not enough 12 experience now. The method that we'have, we did not . feel was 13 something that you could, at that time, turn over to AE's and l () 14 utilities and just let them go and do. 15 Now, much of what has happened with the EPRI approach 16 has been to try to find some method that could be more along 17 that line, but even that is not quite there. You know, there 18 are not, after all, hundreds of fragility analysts out there, 19 nor are there hundreds of systems analysts, but at least these 20 findings were in our peer review report and I think that it is 21 worth talking about, Just how much expertise,really there is 22 and when we come to the EPRI thing, ic.is' going to be raised 23 again. 24 There is more directed by intent to routine use than 25 this, but I am not sure that'there is any factor there, more l Heritage Reporting Corporation l (202) 628-4888 I
.I I
l 4
l 173 1 than about 15 percent. 'But it is his opinion. l O- 2 Number E, earthquake experience and test data base. 3 We felt that there was a need to' document the exp7rience test 4 -data base, that there were places, at.least in Maine Yankee, l 5 the experience and test data base that was used to underlay 6 some of the germinations, had not been documented and reviewed. 7 But a lot of that is now underway. Things are better 8 now than they were 8 months ago, SRAP and SQRT.are.doing some 1 9 -things and more GRR's were out than more'and more is coming 10 out, as I understand it, so-that much of that is coming along. 11 Yes, yes, I am talking about things like1 experienced. 12 data base, or in-test data base, which tell us that many' 13 components are real rugged even without seismic design. You () 14 know, from fossil plants and this and that. 15 And we also thought that the -- and this emerged in. 16 the study -- that the guidance on configuration details, in the 1 l l 17 guidance documents, was not as explicit as it should be and 18 more should be done. I won' t cover that too much, I have got 19 to move on. 20 Relay chatter, not covered in the Maine Yankee trial 21 review, it just was not covered at all. Today, if we were to 22 rewrite the guidance, it would be covered, because first, the 23 methodology Sas been developed and used. I did1some of that 24 with Howard Lambert, at Zion and Lasalle, and the EPRI people 25 and -- we now think that we know how to. do that, although it is. Heritage Reporting Corporation-(202) 628-4888
174 1 complicated, we now think that we know how to narrow the scope 2 enough so that.it is doable, and that recommendation would,. I .. 3- think, result in some guidance if we were to'do that over. 4 And that is an advance. 5 Correlation, t h er'e is a problem. Robbie.had'four, 6' plus sever., plus 20. That is on the viewgraph.- 'And now,- four 7 was the weaker component, but seven and'20'both had .3G for the j i ! 8 HCLpF, and if seven and 20 are correlated, you get a very 9 different answer, than if they.are uncorrelated, even though 10 they are in or. There are cases, where you know, you get 11 lower, some where you get higher, depending or and and it is! j l 12 not at all simple to figure out how to do that. First, because ] 13 we don' t know what the correlations are and it depends on which 14 method you use. 15 And if you are doing CDFM, you are really don' t even 16 have a method.to do it, even if you knew, whereas at least when 17 we are doing fragilities, if you knew, you have the method, but 18 you don' t know anyway. So I am not sure what use it is. It is 19 a very complicated question about how'to do these things, and 20 how to imbed the sort of conservatism that the HCLpF concept 21 has embedded in it, in the rest that we have done. ] 22 That is, having put together this conservative
'i 23 approach, how do we embed.the proper level of conservatism in 24 that, without being so. conservative, ,that everything just, you. I 25 know, that things look unsafe that are not?~
O Heritage Reporting Corporation (202) 628-4888 ;
175 1 Well, that is a question which awe have to wrestle O- 2 with, and luckily it did not pop up in Maine. Yankee and they 3 got their letter from the NRR. crowd, because it-did not pop up. 4- But' God forbid, if we had had to argue about that, we-5 would probably still be arguing, because no one would know. 6 D R. SIESS: If _ you don' t know, you Just be' confused. 7 MR. BUDNITZ: Yes, and that costs money.and sometimes B it is foolish, and that is why we would like to do better and i 9 there is some thinkinD that we have to do. 10 And I am Just pointing out that this is an area where 11 there is a question. 12 .Next, screening methods used in the Maine Y&nkee, we 13 endorse the interactive and iterative method. . We thought that () 14 the way that the screeriing was done, was proper i'n an actual 15 review, the crucial thing is that by having the systems an'd 16 fragilities people sit down together with a 12st before they 17 even get to the plant, a whole lot of stuff can be screened out 18 and makes, even that initial walk-down efficient, 19 And you end up with bullions that are simpler, and 20 Dave Moore drew faultries of entries more complicated than he-21 had to because it was a trial, than'you would do, if you were. 22 doing this in a routine application. 23 And Render and his associates did more fragility work 24 than maybe you would do, more interaction and. iteration'would 25 be even more beneficial than was-done. It was a trial, it is Heritage Reporting Corporation ( (202) 628-4888 l L__.m______. ___.__1__ _ . . _ _ _ _ _ _ _ _ _ . _ . _ _ _ _ . _ _
176 okay, but we' endorse.that heavily. We think that that -o and a 1 2 the EpRI p<rople that I know say the same thing, they think that } 2 is a crucial way to cut the. cost and the scope without 4 compromising the technical validity, and that is perhaps the-5 main benefit of this stuff, compared to doinD, you know,. all of 6 the. usual running around, doing piping, which we did not think, 7 you know, is the right way to uncerstand the seismic map of 8 your plans. 9 (Continued on the next page.) 10 11 12 13 I'T 14 V 15 16 q 17 18 19 R@ 21 22 23 24 , 25 Heritage Reporting Corporation (202) 628-4888
c. 1 177 :l
.j s 'l Finally, leve11of peer review efforts. The. level of =f 2 efforts spent at Maine Yankee.was a heck'cf a lot more than you 3 would do at a routine application because of~so much 4 methodological development. .But it wasn' t a factor of'two L 5 more, it was'probably, you know, 30 or 40 per cent or 6 something.
l 7 We thought that the level of effort we spent.was 8 eesential, but there was a disagreement in',the peer review 9 proup on how much would be needed in a routine application. To 10 be direct, there were two members of the peer review group that 11 sort of thought that, gee, you ' don' t need anything special l-12 here. Until he does the study, sends (t in to NRC. They do 13 that every day of the week. They sends them in'on FLDs, and d 14 they send them in on LOCAs, and they send them in on EO, and l 15 they send them in on this and that. . I 16 And when they send them in, as this -- this is ene 17 view -- they Just send them in, and they do their cun peer 1 18 review because they are a responsible, effective utility, and j 19 l then the NRC reviews it. There are others that felt that this l 1 20 required a sort of peer review like this separate from that, 21 and that division of opinion on our group is manifested on this 22 slide. 23 Well, Bob Budnitz has his view, but I' m not sure what 1 24 the right view might be amongst, for exatople, the NRC or what. l q l 25 That is -- Jim Thomas, for example, would say, they send:in 1 O Heritage Reporting- Corporation- - (202) 628-4888 i i l ___i___________.___i______.________._.___.__
l J' 178 1 large LOC 4s stuff all t he 't i tle, they don' t have . peer l review for .
.O ' 2 .that and their license depends On it. ~ Why do they.need a peer 1
3 review for this? 4 4 Well, there is a question. I can' t answer it. I' m 5 not'a regulator. 6 DR. SIESS: It all depends;6n which member of tne 7 review staff.is looking at them. 8- MR. BUDNITZ: . Well, it's ' more: than that. 7t. depends ; 9 on whether or not the methods have been. accepted, and obviously 10 in a trial,_in a trial you need it, but'the question is, how 11 acceptable will this -- and bow acreeptable in routine 12 application, 13 DR. SIESS: Routine,.yes. it's a long ways from 14 routins.. 15 MR, BUDNIT2: Yes, well, thatwhy'I do -- I' ta t h r'ee l 16 minutes late. 17 DR. SIESS: Guppose that the only event at Waine )
- 18. Yankee was due to'the walk-downs. To get to the point of l 19 knowing what to walk-down and 'rioinD the walk-downs with some 20 experts to point out thu kind of things that might need fixing, 21 and fixing them, without doing any of the other' quantitative 22 stuff, in terms of a bottom line of improved hafety, how much 23 difference would there be?
l i l 24 MR. BUDNITZ: Well, you Just answered the question. l-25 If you did it all, that wou.1d oe Just as, safe as if you did the. ' l 1 l - L Heritage Reporting Corporation (202) 628-4G88-1 = _ _ _ _ _ _ _ - _ _ = _ _ _ _ _ _ _ - - _ _ _ - _ _ _ _ _ _ _ _ . -- .- ._- - _ _ = _ _ _ _ _ - -- - - _ _ _ - _ _ .
i 179 1 l I 1 ' calculations:and also did itiali.. 2 DR. SIESS:- But how much of that would you have to do 1 3- in order to -- there are certain modifications 1made which would. i 4 improve safety. ! 5 MR. BUDNITZ: Yes. 1
~
l 6 DR. SIESS: Now 'that's .over and above what we know 7 about the safety, f I B MR. BUDNITZs Rfght. 9 DR. SIESS: How much of the total effort that was 10 done at Maine Yankee, by the outside group.or by the utility, 11 could have been done to reach that point? Or how much'of what i l A l 12 was done was not needed in order to reach that point? 13 MR. BUDNITZ: Gee, that's a very hard one. Probably () 14 several of us in the room have different opinions, j 15 DR. SIESS: I mean, you ought to have a quant'itative 16 number in mind thut you' re going for before you realize that 17 you don' t know enough and you ought to fix it. 18 MR. bUDNITZ: There is a' crucial thing that 'I, at U
~
l 19 least, have an opinion about that others may differ with. The l 20 quantitative analysis here doesn' t mean a damn down to the last 1 1 21 digit, but quantitative analysis is an essential element _ of i 22 this method because it disciplines the process, j i l 23 It enables you.to screen in and out, to say this-is -l l 24 more important than that, and to prioritize your activities. 25 So without the quantitative analysis, the process' wouldn' t be a j l l l \ [ A. Heritage Reporting Corporation j
-(202) 628-4888 l
1 l 1 I
3 f 180 j J
-1 disciplina and therefore would either. cost more'or miss things. {
- 2. However, in the end the numbers aren' t terribly J 3 . interesting to me,-down to the last digit.- .I mean, .1 is not 4 the same as 3, you know, that's a di f ferent earthquake. But 1
5 . 27, . 29, L don' t -- or even .21, . 27, that's approaching the i l 6 range where maybe there's some si0nificance. 7 So that's what I would think, and -- 8 DR. SIESS: Gee, a lot of fixes were made, not 9 because it was .2 or. 27, but.because without the fix, there ] j 10- was a fair chance it would be zero, or $300,000 later would be 11 .15. 12 M R. BUDNITZ1 Yes, or something. That 's ' f or. sure. 13 That i s, it is easier to fdx it than argue-about, or certainly () 14 cheaper. 15 DR. SIESS: I visualice somebody walking into a. 16 battery room and saying, gee, these batteries are not tied to 17 the wall, why don' t you fix them? 18 MR. BUDNITZ: In this particular esta, the truth was 19 that we didn' t have any experience or ar.y other' data base to l
)
20 help us understand the batteries, and no -- money eculd have
]
1 21 been spent developira it, and they said, they chose another 22 solution. 1 23 DR. SIESS: There's been a let of ernphasis on the 24 walk-down. Now .the walk-down itself, of course, is not the 4 1 25 answer. You have to know what to walk-down. ] Heritage Reporting Corporation (202) 628-4888 ; i I?
1
~
)
,181 ~ q
'1 MR.'BUDNITZ: Right, and that's where tha discipline
( 2 of.the process -- 3 DR. SIESS: 'Somebody did some screening, some systems 4 analysis. But the net rAsult, presumably, is the result not of - 5 any of the analyses, not of any of the. calculations, the rmet ; 5 result, any improved safety that's there, is' purely. mechanical, 7 physical changes in the plant. 8 MR. BUDNITZ: Or procedures. 9 DR. SIESS: Gr procedures. And my question is, is 10 there another way to get there? PRA is presumably-one way,. 11 this is one way, is there another way, or even a'way to get 12 three quarters of. the way there ? 13 MR. BUDNITZ: Look, there are other ways, but what () 14 you' re probably driving at i s, is there a simpler way-to get 15 there? Everi simpler than this, because -- 16 DR. SIESS: Simple, yes, i 17 MR. BUDNITZ: Yes, but we are not interested in more i 18 complicated ways. We all know how to do.that. 19 DR. SIESS: You see 3 you could probably, after doing i 20 two or three of these things, get three quarters or more of the 21 way down the line simply by compiling a list of things that i I 22 have been found at previous plants that had to be fixed, and ! l 23 say, go look at yours. 24 MR. BUDNITZ: Yes. By the way -- 1 i 25 DR. SIESS: Now that doesn' t mean you' re going to i l O. Heritage Reporting Corporation ,
.(202) 628-4888 4 j
l 1
l 182 ! 1
,, 1 catch al1 of t hern. i
\
%)
2 MR. BUDNITZ: By the way, I agree with that j 3 corn p l et e l y, and that's probably already in place in sorne cases. 4 For exarnple, there was bolt and -- several years ago -- all 5 those batteries, you know, had spacers put in or tied or ! 6 various thinDs, so that's true. 7 DR. SIESS: You don' t think the batteries are tied 8 down in all hundred -- 9 MR. BUDNITZ: I don' t know. 10 DR. SIESS: l'11 bet you I could find one that had -- 1 11 MR. BUDNITZ: But the point is that if you want to Do IC' tv a plant -- Maine Yankee, Catawba, Hatch -- and be able to 13 state at the end that you have high confidence like -- and they l 14 got this licensing letter out of it, so it's terrific, you 15 know, I was a triuroph for law and order because, I rnea n, it 16 doesn' t always happen, it doesn' t always happen on a quick 17 schedule like that, either. l l 18 The fact is that if you need that because your 19 licenue depends on it or there is sorne other thing going ont l 20 there is no other way than a discipline process. Now what 21 we' ve tried to develope here is a discipline process less 4 i E'R expensi"e than a pRA, but gets you that conclusion that people 1 23 -- engineers and concurrinD rnanagernent -- you get sorne other 24 engineering that cornes in and reads it and says, yeah, I agree 25 with that. And that's i rn port a nt . t k Heritage Reporting Corporation (202) 628-4888
1 i 183 1 What EPRI's approach is trying'to do is to make it. C:) E even simpler than the NRC's, although they recognize'that there j 4 3 are certain questions that you can' t answer as well, but ' that's 4 okay because they are answering good the one question,.you 5 know, perhaps less expensively. They'think. 6 So it's a very difficult -- q
.l
'7 DR. SIESS: I don' t yet understand how this was found i j
8 an a licensing issue, but I think I' d rather not hear it. But 9 I don' t know, since it's ' already set t led on -- 10 M R. BUDNITZ: But'what i said I think is the crucial 11 point, this is without a disciplined process, a reviewable-l -1 l 12 process, and a rnethodology, and a bunch of people that were J i
.J 13 part of it and think to review it, you wouldn' t be able to j
() 14 produce, you know, such a result. 15 You know,- I ree an, to not even debate what went around 1 ! j 16 in the last week before the reorganization, get everybody to ) I 17 initial to it. I L 18 DR. SIESS: Well that result was a licensee action.- 19 That was probably wort'n evEry minute they put into it, and then 1 20 they had to go back and reanalyze everything. As far as I' m l 21 looking at it, the result to me of some physical changes in the 22 plant, which produces vulnerability to earthquake - 23 MR. BUDNITZ: That's right. 24 DR. SIESG: And the question is, is there.a simpler ] 25 way to get that? 1 i Heritage Reporting Corporation. 1 (202) 628-4888 , l l
~
1
- i. ,
l 184 1 j 1 MR. BUDNITZ: 'I think I have an answer. O- 2 DR. SIESS: Well, so far, since this is not involved
-3 in the regulatory process, it has been simple. -l
'I 4 - MR. BUDNITZ: I think I have an answer.
5 DR. SIESS: I don' t think .what you' ve Egot is simple. j 6 if it's something that is required. 7 MR. BUDNITZ: If you have a walk-down without this 8 sort of guidance, you' re pretty sure to miss something 9 somewhere. Maybe not in the first plant or maybe not in the ~ 10 cecond, somewhere along the line some things will be missed 11 which you will catch. I mean, t h a t ' s my -- 12 DR. SIESS: Okay. Suppose there's 45 things out 13 there, and I miss five of them. () 14 MR. BUDNITZ: Right. I 15 DR. SIESS: I don' t know how much I' ve improved i l 16 safety, i 17 MR. BUDNITZ: I can't judge that. 18 DR. SIESS: Probably not a much as if I had found 25 l 19 i 20 MR. BUDNITZ: But more than if you didn' t do 21 anything. 22 DR. SIESS: More than if I d idn' t do anything. And 23 remember that if I found 25, there'might,really have'been 30. y 24 MR. BUDNITZ: Another thing is that you are very 25 likely to find the most cbvious stuff. I mean, that's the way R O-
' Heritage Reporting Corporation l (202) 628-4888 j i
l i
h 185 1 the world is. , j 0- 2 DR. SIESS: Only because everybody goes around'and 4 3 looks at building's batteries, everybody goes atpund and looks-1 4 st ceilings in the control. room, and things,like that now, :j l i 5 because if you find it in one plant, it's easy to look at-in
'6 another, 7 But when this gets into the licensing process, it's -
8 still a rather formalized quantitative procedure, and I don' t 9 know where it is going to end up. That's for later. Later l 10 today, maybe, and hopefully after next year. \ \ 11 MR. BUDNITZ: I' ve got to get out of here.. ] i 12 DR. SIESS: Thank you, Bob. ] l 13 MR. BUDNITZ: If I don' t get down here, I won' t get - 14 home, so -- 15 DR. SIESS: You mean you' re through? 16 MR. BUDNITZ: No, I' ve got to come back later, which i 17 means D ve got to get down here so I can do it. I 10 DR. SIESS: Okay. Bob. 19 MR. KENNEDY: Bob Kennedy. In answer to your earlier 20 question, at least at Maine Yankee and at other plants, in my. 21 Judgment in the component capacity area, the key ingredients 22 that you have to have out of this i s, the people doing the 23 c:omponent capacity evM uation need. the interaction with the' 24 systems engineers. 25 I' m not sure you need all.of this rigor in order to Heritage Reporting Corporation l (202) 688-4888 1
186 1' .get.that interaction, but you' ve . got to have:that interaction. , 'O 2 Beyond that, the things that were modified at Maine'. Yankee l 3 weren' t modified because of calculations. You' re cig'h t -~ l 4 except for the tank.. They were modified -- i 1 5 DR. SIESS: Because you could' Jusi flat'out tell 6 people go out and check those tanks. 'j 7 MR. KENNEDY: Yes, they were argumentative, and.it 8 was easier to modify. 9 MR. BUDNITZ: And some of it, Bob, was common sense. 10 I mean, wiien you see it,.you do it. Right? 11 MR. KENNEDY: And once they were modified -- at least: 12 my Judgment, this is may be controversial, this is purely my 13 Judgment -- in my Judgment, beyond those modifications and ; l 14 beyond the analysis on the tank, none of the other analytical 15 effort was necessary if you wanted to say that you had high 16 confidence that the HCLF level was going to be over above l 17 2. 5 2-18 I doubt if you could have come up and said it was .27 19 g or .29 g, but -- 20 DR. SIESS: You could probably have said that it was 21 better than .18, 22 MR. KENNEDY: Oh, certainly, we could'have said that. 23 I was there when the peor panels were doing the. walk-downs. To 24- the best of my knowledge, calculations didntt influence any of 25 their comments. Heritage Returting Corporation (202) 628-4888 _:-_=__
187 ~
.1 MR. BUDNITZ: That's true.
O 2 MR. MENNEDY: Except on the tank. 3 MR. BUDNITZ: I agree with that, more or less. 4 MR. KENNEDY: And so with the exception of the tank, 5 you could accomplish this if you had the systems engineers and 6 the capacity engineers working'together by some other~ method. 7 Walking that plant down, I think you could have accomplished 8 the benefits, i 9 Now for the tank, unfortunately, I have the opposite 10 opinion. Having looked at'a lot of tanks, I don' t know how you 11 make a quick assessment of the capacity of these flat bottom 12 water storage tanks. There I think the analytical approach is 13 necessary. () 14 DR. SIESS: After about 20 or 30 have been. analyzed 15 it gets easier. 16 MR. KENNEDY: Yes, they each look different at'these 17 plants. { 18 DR. SIESS: By then you' ve got a computer program. ; 19 MR. KENNEDY: Anyway, a tank analysis is, you know, 20 maybe a two dcy effort. The rest of the time on this project .
-)
-1 21 was spent on arguinD whether the HCLF was .2 g, I guess which l 22 was the low number, to above .24 g.
23 There was probably at least a couple hundred hours 24 spent on that issue. That was'more.because it.was a research 25 progre, I suspect, than anything else.
) Heritage Reporting Corporation g (202) 628-4888 l\
.I
188 i MR. BUDNITZ: Davo Moore had a -- ( 2 MR. CUMMINGS I',d to rnake a point. I:am more of'a- l
'3' systerns engineer, so I have a different viewpoint.than you~ guys )
')
4 do, and this.has come up in regular PRAs. I agree with what~ - 1 5 Bob says. You have to have a certain amount of discipline.. 6 You have to make a certain time commitment juststo' learn about 7 your plant, if nothing rocre than look for out wires, and make 8 sure you' re not just going through motions,Lyou1know, 'to meet 9 some prescription. 1
)
10 I think'it takes sorne intelligence. Obviously, if 11 you have top people doing this_that.areLreally' dedicated and , i 12 motivated, then it.is a much simpler. process than if you have a.
]
{ 13 lesser skilled Droup of people doing it. So I think .in.the { 14 implementation,.you know, you are asking, how simple is simple 15 enough? That's what everybody wants, a' minimum. l 16 And I think you have to look at who's. going to apply 17 it and how it's going to be applied. Otherwise, you' re liable 18 to end up with something that is sort' of;a sham.
. . I 19 DR. SIESS: You' ve got a good point, and look, we l I
20 don' t know what's going to be the prescription, what is the l 21 requirement, who' s going to audit it, what are we going to use 22 it for. 23 I would think that the. utility, with anywhere'from a-24 few hundred million to a few billion dollars investment on a 25 plant, would have some interest in seeing'that they didn' t melt Heritage Reporting- Corporation ' l (202) 628-4888 s l l a E___._ . _ _ _ . _ . _ _ _ . _ _ _ _ _ _ . . _ _ _ _ _ _ _ _ ._ __ _ _ _ _ _.________________________________.___U
.!'- j 189 1 their core. You know that,'s alwayn one possibility. . .(
2 MR. DUDNITZ: 'Chet, but I don' t think that that's a 3 fair thing to rely.on because if you thought that all^ utilities- i 4 all took that to heart, I wouldn' t . think we' d see near what j 5 we' re seeing out in some of these-plants, which tells me that 6 there are some - I don' t mean components, but there.are some_ .j 7 utilities that don' t take that to-heart, or we wouldn' t be j j 8 seeing what -- L 9 Now let's get back to , the other: thing, which is even 10 more importants who's going to use this? Well, there ~ is a. 11 rumor cut there that this or something like it miDht become the 12 seismic part of the.IP. isodify them, I' m not sure. I can' t 13 Judge that. There's - a whole . st af f work -- () 14 DR. SIESS: Nobody knows:what the other part of the i 15 Yankee -- -j i l 16 (Laughter.) 17 MR. BUDNITZ: Well, let's suppose 'that some day-18
~
somebody calls the commissioner's ' order and says, we' re going 19 to go look at all these plans for internal initiators.and'some 20 for seismic, it's not going to be very different, I don' t know 21 what it's going to be, but it surely can' t be entirely 22 disjointed from what we' ve done here. 23 Mow if you do that, you' re going to turn it over to a i 24 hundred plants, some of who are going' to use very competent 25 people and some of who aren' t. They' l l be competent but not,. O' Heritage Reporting Corporation- i (202) 628-4888 1 f f _________.___._________._________.____1 . _ _ _ _ . __ _ _ _ - . _
l 1
-190 1 you know, true international experts, et cetera, et-cetera,'et I
2 cetera, and the method 'has to be disciplined enough so that 3 those-people can carry it out following routine. instructions 4 and get something that the staff can judge as a reasonable 5 finding about'that plant. 6' And that means this thing has to be rnore disciplined 7 than was necessary with Maine Yankee, where not only do we have 1 8 'a good utility, it rnore of a research project. j l 9 DR. SIESS: Well, it's got to be doable and it's got I l 10 .to be reviewable. And somewhere in there.it ought to be right. 11 MR. DUDNITZ: It's Dot to be doable, and I have to l 12 say this in the -- to rne it is an American tragedy because'it 13 has to be doable, also, and get the right answer by those few I l () 14 utilities who, by my objective evidence, don' t take'you advice l 15 to heart, or if they did they wouldn t be running the plants l 16 the way they are running them. 17 For whom the difference between $106,000 and $108,000 R 18 is a big difference, and they try to get that extra 2,000 out. ! 19 of overy project, and I' ve worked for some of them. It's a - 20 tragedy. 21 DR. SIESS: I know that's- true. I know some of them 22 are out there spending all their engineering time and money 23 making other fixes that have been required by the NRC that may 24 not have as much on safety, I don' t know. Without a PRA I 25 couldn't tell, and I wouldn' t know when I got through, I guess. ' Heritage Reparting. Corporation (202) 628-4888
191 g g. 1 Okay. Stand up, and you had better use that V 2 microphone. 3 MR. MOORE: Dave Moore, Energy Incorporated, PI. 4 There is one case where we had a tank at Maine Yankee that had 1 5 a tip of .17, was the DLST, which feed the aux feed water 6 system. And you might have been tempted Just, if you had not j 1 7 quantitative process for evaluating how important that tank i s, 8 to say, well, that's the lowest capacity we' ve got, it oughtJto 9 be fixed. l 1 10 When in fact, that tank only showed up in double-11 failures. That i s, it always had to fail in combination with I l 12 something else before you had core. So I think that is one 13 piece of the puzzle that you can get from systematic analysis. l () 14 D R. SIESS: But still it probably wasn' t a bad idea 15 to fix it. 1 I j l 16 MR. MOORE: I' ll leave that to -- 1 l l 17 DR. SIESS: If Dr. Oakland was here, I' m sure he and f 1 18 Mr. Michaelson could think of some combination of circumstances 19 under which that tank should be fixed, but that's another i 20 story. , l i l 21 Well, we' ve had a number of comments .about Yankee l l 22 Electric and Maine Yankee and how cooperative they were, and ) l l 23 how good a plant they had to begin with and'how they ended up. 24 I' d like to hear from you now. I think it is important to 25 realize that you 11uys started a few years ago. J l l Heritage ReportinD Corporation H (202) 628-4888 l l
i-e
.192" 1 MR. WHITTIER: t.'y narne is Doug . Wh it t ier. I arn 2 Manager.of Nuclear Engineer'ing and' Licensing at Maine Yankee.
3' I think in light of the hour ar.d lightfof the' fact that many of 4 the ecmments'I had hoped to make have already been. tnade, I' ll 5 keep rny. comments brief. 6 When the NRC asked up to participate in the seismic
'7 margins research program, we saw it as an' opportunity.to do B three thinDs. One was to support the use of experience data in
') seismic evaluations. Another was to confirm our impressions 10 about.the inherent seismic ruggedness of Maine Yankee.
11 And' third, to resolve NRC questions about the 12 adequacy of Maine Yankee seismic design. We also thought that 13 the prograrn was a logical extension of the seismic work that we 14 had been doing. Overall,-we'believe that the program was 1 j 15 successful. We think it did what we hoped it;would do. 16 From our impression, we found that the prograrn was ' 17 well planned and.well implemented, or at least it'seemed to us. :. l 18 I think this is a tribute to the high caliber ot* the people, j
.1 19 the NRC, and their consultants assigned to the project.
20 For our part we placed a high priority on this 21 project. We assigned some of our best people to interface with 22 the NRC team, and of course, we contracted' with Bob Kennedy 23 also to assist us. We made a conscious effort to attempt to- I 24 resolve any questions, any concerns quickly and expeditiously. 25 And typically, if it was possible, to implement a O Heritage . Reporting Corporation j (202) 628-4888 I 3
o
.1 193 i
i hard ware fix that was relatively quick and-inexpensive, we o 2 would do so instead of spending a long time discussing a j 3 particular issue, ) 4 Overall, we feel that the use of walk-throughs, the j-5 use of seismic' experience is a cost effective way to asses a 6 plant's seiumic ruggedness or sei'smle margins, and'also it's a 7 good way to identify prudent upDrades in the plant that might l 8 be warranted. 9 That's all I have. I would be pleased to answer any 10 questions. .{ 11 DR. SIESS: You' ve had a, what I' d call, small l 12 seismic scare a few years ago. The New Brunswick earthquake 13 was a pretty fair size earthquake not too far away. Do you () 14 feel any more comfortable now about seismic safety. af ter you' ve 15 gone in and fixed some of.these things? 16 MR. WHITTIER: I think overall that we feel more i a 17 comfortable, I think that there was an impression by many l 1 18 people in the plant that there was a lot of inherent ruggedness
]
19 in the design, but that was ' simply based bn gut level 20 engineering Jud D ments and not the basis of.any sound program. ~ l l 21 I think th,at now through our efforts and the NRC's .ef forts that l i 22 we certainly-have a better basis on which to base that i 23 jud Dment. And I think also that we have hdentified some i 24 questionable areas in the plant design.and we have' upgraded' 25 them. h
% Heritage Reportd ng - Corporation (202) 628-4888' i
I l
.t
q
-I,i
- j) 1
-I ,
y
..{; f e> 394.
Q" ,
,1 1 DR. SIESS:. I know<tha't'Bobisaid earlier N mething,
- q ; -,_- 9-if we thoughf laboIb~it we wou3d have des)gned it. .
I'
- you know, If
// 't j
i ,
.g 1
) ')
U '3< we designed i t', ' i t is gqing to have sortlof.. inherent margins.
.i 4, And the thing that we ' real'ly worried about , was that we did not 1
('a, 4'
,' 5 think about ith ~And it.seems to me from the list of fixe /s that 6 there p a nos "any of them there that wer?e.really things that
' 3' j- 7 were !averlooked in the des,tgn,' em I ri Dht7 I ' mean t anks arn 1 1 8
tanks. 9 M R. WHITTIER: In ptarts,, you do not firp the. 10 emphasis in the design on things li'ke anchorage 'od battery 11 racks and putting in the spacers. Maybe in the newe" plarsts i 12 you do, now with the heavy test proDram. I thir k the'/ all' of I 13 these are finding th3t hose are the areas where if here is 4 14 goi r.g to be a potential weakness in those kind of details or ar[ 15 expansion bolts that were added'and field installed, not the [ 16 items that the design firms put their' big horsepower on in-id .doinD the seistnic design.
.)
18 DR. SIESS: When was Maine Yankee built' l 19 MR. WHITTIER: The commercial operations were in the 20 1972 to 1973 time frame. 3. 21 DR. SIESS: Again operationally. 22 MR. WHITTIER: Yes, the DS. Tne' construction period i
/
23 was.sbout five years. < l d 24 / DR. SIESS: So you got your construction ptirmit in i I 25 19.67?
.]
')
O- .c Heritage Reporting Corporation. (202) 628-4888 ," ' i ! -c iy 0 , ..t c I. w ,
~
.)
-my .
195 _1- MR. WHITTIER: 1967, arounef'thaOtime.
'O. 2 DR. SIESe:
There w.Ts somethingrin the report on.how
- 3. ~much'your manpower effort in doing-this involved. 'And as I 4' recall, it was ncd a great deal.
5- MR. WHITTIERt hfo, sir. It was about a. half a man 6 ybar. I should say primarily that it was . Yankee Atomic's I j 7 efforts. L 8= DR. SIESS: What do you expect the fixas that'you are l' 49 9 making to cust you? It et 10 e MR. WHITTIER: Do you have an overall costffor the U" 11 fixes that were put? 12 MR. HENRY: Well, everything but the tank'was l 13 probably very low key. My guess would be in the $60,000 or (} 14 $70,000 range. The tank probably ran $150,000 to $200,000.. 15 Ch, excuse me, I am Bill Henry, and I work fcr Yankee Atomic, 16 but I do all'of my work on Maine Yankee, and I interface with 17 everyone on '.he program. 16 DR. SIESS: Do batteries come in under those figt.es? 19 MR. HENRY: No,.they did not. We were planning to 20 replace battery unit 1 anyway because.of aDe. .What this 21 program did -is we replaced the two taajor batteries,'one and l 22 three. I think that the prices were $40,000 a; rack,-but'I 23 would have to check-on that. 24 DR. SIESS: 'That is. cheap. I know that the whole 25 plant. was cheap. O HeritaDe' Reporting Corporation (202) 628-4888 L. u -
~
- -c 1;
lW , 1 h I; . 9 196
.y -
1 Do you have any' idea what you paid per installed [
.O + V 2 kilowatt for that plant?
L n; 3 MR. WHITTIER: No. .The plant, cost is around $230 ( 4 millson. 'It is around'800 mev;. watts electric br 850. 5 DR. GIESS' Are there<any questions?- 6 (No response.') 7 DR. SIESS: Thank you very much. I think.that Maine 8' Yankee _has somethinD-to'be proud of in taking.part of this 9 thing, and I have heard some nice things,scid about'the l 10 cooperation. 11 ., MR. WHITTIER: Thank you very much. 12 DR. SIESS: I hope you are happy. 13 MR. WHITTIER: Yes, sir, we are. () , 14 DR. SIESS: That concludes the application of' seismic' 15 margins methodology to Maine Yankee. In the rernaining portions 16 of this, we are going hear sorething about the EpRI l 17 methodology, and then the applications of whatever to cevere 18 accident policy. That is going to come from-research. 19 So I think that I would like to take a'short-break 20 -'now, and come back and go ahead with the EpRI, and leave.your 21 future plans stuff to fili i n. ' ;l 22 (Whereupon, a recess was taken.)- 23 DR. SIESS: Okay, Bob. .Let's see, you nd Bob 24 Budnit: are opening speakers on this? 25 MR. KENNEDY: I believe so, yes. O' Heritage Reporting' Corporation I
]
(202) 6C8-4888 l l 1 i d l
- l:.1 ' - -
197 1 DR. SIESS: Okay.
't 2 MH. KENNEDY: My'name'io Bob Kennedy, and I am going 3 to talk.today and make a brief presentation on the methodology 4 that has been proposed in the EpRI-saismic margin program for 5 doing 'these seitsmic~ uargin studies. Baaica11~y, the objective 6 of the EpRI seismic margin. program is the development of an l 7 engir~ering methodology and to provide practical guidelines and =
8 'procederes for evaluating of the' power. plant-for earthquakes 9 larger than then the design SSE levels. 1 10 .The EpRI program has been built upon and is very 11 heavily influenced by the NRC expert panel on quantifications l l 12 of seismic margin reports. I was a member of that panel at 13 that. time. And when we took on the assignment to put together
]
() 14 the EPRI margin program, I dropped off of that panel. But the 15 two programs are very similar to each other. 16 But we have tried to expand upon the NRC expert panel-17 recommendations in several specific areas. The first of these ) 18 areas was we wanted to simp?ify the procedure in some areas. 19 In particular, we tried to use more deterministic procedures 20 rather than probabilistic procedures to the extent possible, 21 and we feel that they are possible. The deterministic l J 22 procedures need to be selected to be consistent with the j 23 probabilistic approachec. 24 But basically, our capacity evaluation procedures'are 25 totally deterministic. And the systems work is a little bit O- Heritage Reporting Corporation ! f202) 628-4888
- _ _ _ _ - _ - - - _ _ - - _ _ _ _ _ _ _ - _ _ _ _ _ _ - - _ _ _ - - = _ _ _ _ _ _ . - _ _ ._- - -_
198 1? the same kind of a counterpart, because-it-uses'a success path,
.O-
'2 approach rather than the f ault-free approach.
3 Now the key purpose of this we wanted the procedures 4 to be simplified and usable by e greater number of; qualified-5 engineers. I can speak in the capacity area. I cannot speak 6 to much in the systems area. .But in the capacity area, mostLof L7 the engineers out there who really understand equipment and who 8 really under the design of nuclear pswer plants and who at 9 least I have great confidence i n, they understand this l l . 10 equipment and they understand its capacity, because they have-l 11 spent years designing or looking at past earthouake performance 12 on equipment. And most o# these engineers do not,have a 13 probabilistic background. 14 And I have found the probabilistic aspects of the 15 seismic pRAs are greatly limited to the number of Lthe people 16 who can really do the work. And so we have a very strong 17 desire to try to make this deterministic in the capacity area. 18 The other feature that we tried to accomplish in the 19 EpRI program was to greatly expand the' details in the 20 requirement on how to conduct the seismic walkdown and the 21 capacity evaluation over those in the expert panel report. .The 22 expert panel report laid out a philosophy and laid it out;very 23 well. What we wanted to do is expand there and get[ down. to 24 getting more specific guidance, again with.the idea of trying 25 to open this methodology up to where it could-be done'by a l O Heritage Reporting Corporation j j (202) 628-4888. i l J t 1 a 1
- - _ _ _ _ _ - _ - - _ .-=
199 1 greater number of people. O- 2 So what we consider is.that' the EpRI seismic' margin' 3 program is a gradual' evolution and some.expancion what'the NRC
'4 expert panel required, but represents no substantial change.
5 We have tried to keep the proDrams'rather close to'each-other. .q l 6 Now in laying out the EPRI program, we. felt.that'the L 7- very first thing that we had to do was to lay ~out veryLclearly' l 8 exactly what the goal was of the seismic margin program.' I. l 9 believe thrt these two points on the goal are the exact same ac 10 in the.NRC margin orogram. 11 Tc determine a ground motion leve2 for which you do 12 have high confidence of a low probab' lit'y of seismic-induced 13 core damage for this plant, and to identify any weaker link , 1 () 14 coraponents in the plant which tend to reduce.the seismic margin 15 capability to a standard larger than the SSE' level at least up 16 to some pre-selected earthquake levels. Now you do not. achieve 17 these Doals beyond the pre-selected earthquake level'that you 18 choose, but at least up to that level. 19 We also wanted to lay out what we felt was beyond'the l 20 scope of the seismic margin nrogram as we described it. We
]
1 21 wanted to simplify this program down to be the most cost -] i 22 effective and efficient way to answer the first.two questions, .d 23 and we were willing to give up information'on the next four 24 questions on this page.
)
q 25 One, to determine the seismic risk of the plant or ! l 5 Heritage Reporting Corporation- .i (202) 628-4888 ) u j 1 q _ _.__________u_.__.__.__
n 200
'l the annual' probability of core damage or risk to the public is 2 not being answered by the'EpRI seismic-margin program, and'I do 3 not think that it is answered.by the NRC one either. But . it. is 4 definitely not answered. We. are - not attempting to answer 5 seismic risk questions.
G . Secondly, to determine the range of Dround motion 7 levels which dominate the seismic risk, we.do not really answer 8- that, although we give a partial answer there in that we have 9 found from the review of past.PRAs, seismic pRAs,.that'the im ranges of ground motion'which dominates the seismic risk are il from about the HCLPF level on up. 12 So I think that we can answer when you complete a 13 seismic margin review the.t the seismic. risk is not' going'to be-() 14 coming from earthquakes below this HCLPF earthquake level, the' i 15 seismic risk. 16 The third point. To determine.the plant components 17 which dominate the contribution to seisrl.icl' risk, we do not 18 really answer that questionLin the review.- We' answer the 19 question what are the weaker links, which components will be 20 damaged by lesser earthquakes or potentia 11yJdamaged by lesser-21 earthquakes. 22 In most cases, the components which will be 23 potentially damaged by' lesser earthquakes'are likely.to.be the 24 same as those that. dominate the contribution to seismic risk,- 25 but'that it not a: certainty.- So we cannot guarantee that when O Herit' age Reporting Corporation (202)'628-4888 1
l 201 1 we find the weaker link components that we have automatically 2 found all of the components which dominate. the seismic risk. 3 Lastly, we are not attempting to determine the median l
'4 seismic capability of the plant. We want a high' confidence and-l 5 low probability.of failure capability,'and.not a median.
l 6 The one thing that we have-observed from past seismic: 7~ PRAs and from efforts to develop fragility curves is that this 8 median capability in every case that I am aware of is at.least 9 a factor of two higher than the HCLPF level.. So I think that 10 you have.a partial answer on thel median capability. It is at. 11 least-a factor of two higher than what the HCLDF is reported. 12 at, but we do not get a very precise indication of'the median 13 level. It could.be anywhere' typically by-past estimates from> 14 two ta three times higher. 15 So we have stripped the program down-to'a level.that l 16 will enable us.to do a.very efficient answering of the first l l . 17 two questions and not an answering of ths.latter. In fact, my l ' 18 own personal judgment is.that the best.way to answer those 19 latter questions is by doing a seismic PRA rather.than ! 20 attempting to answer them with a margira type program. 21 The seismic margin assessmentLprocedure consists of a 22 number of steps, and each of these are described in detail in
.1 l 23 the EpRI' report, and I do'not intend to go through them.iriany.
l 24 detail.- The first'is the selection of the earthquake level 25 that you are going to do the' margins review, the' selection of. l
-l J
O Heritage- Reporting Corporation l (202) 628-4888 1 i
)
'l
-i
_ _ _ _ _ 1__.____.__ _ i
202 1 an. assessment team and criteria. contained in the EpRI margins I' 2 methodology report. 3 Now.there are a whole series of steps of' preparatory
'4 Work that it is beneficial to do prior to the first walkdown.
L 1 L 5 Now this is simply guidance. They do.not all have to be done ]
]
6 prior, but we find that it is beneficial to do a lot of 7' preparatory work p'rior to the first walkdown rather than trying 8 to do it'as part of the walkdown. 9 Next is the~ selection of the systems and.the elements 10 which will be walked down, or in a simplified statement the
]
l 11 selection of the success paths which will be walked down. -And~ j l 12 performance of the detailed seismic walkdown.or capacity. And 13 a section in regard to criteria for doing the seismic' margin () 14 capacity assessment spelling out this deterministic procedure, 15 and then a discussion on documentation. I am only going to be ;
)
16 touching on a couple'of these points here t'oday. 17 First, I would-like to touch on what the program l
)
18 philosophy is of the EpRI program, and I will do a little bit 19 of comparison with the NRC expert panel program to show where 20 there are similarities and where there are differences. 21 The EpRI program starts with the same basic statement i 22 .as the NRC expert panel program'in pre-selecting an earthquake ' 23 level for which you are going to do the review. The idea is to 24 select that level sufficiently high so that'you are likely to 25 find the HCLPF capacity. of the plant and you are likely toLfind 1
.h O Heritage Reporting Corporation (202) 628-4888
\ _ - _ _ _ _ _ _ . _-_-__ _ O
203 1 the weaker link components that govern that capacity, but to O 2 select it sufficiently low so that.you can minimize-your work 3 by maximizing your ability to screen out~using judgment and 4 experience, to screen out as many components as possible from 5 the-margins review. E Now it'has-been my strong advice and advice of 1 peers 7 .strongly as'well in.the EpRI report that for most eactern 8 plants of the existing plants, for. plants basically with SSEs 9 of 0.2g or less that at least in my judgment the' proper margin 3
'10 earthquake level, the. pre-selective level to do this review at,-
11 ouDht to be somewhere in the 0.3g region. J 12 The reason for that .judDment is below about 0.3g we 13 have a very extensive earthquake experience data base,.we have (} 14 a lot of qualification data and other things that will enable 15 us to aggressively screen out form review the' maximum number'of ! 16 component s. When we get above about 0. 3g, our data base 17 rather rapidly starts to disappear. 18 So to try to do margin reviews at higher levels is 19 goinD to require a much greater looking at the equipment in the-20 plant, because you are not going to be able to rely as heavily 21 on judgment and experience to screen out items. But'~we have'a 22 lot of data up to 0.3g, and I see no purpose in? setting the-23 level much below that level. 24 Secondly, in past margin reviews and seismic pRAs 25 when we are dealing particularly with the. older existing plants O Heritage Reporting Corporation (202) 628-4888 j
204 1 and some of the' newer ones as well, we are ' l i kel'y t ci find - our - 0- 2 weaker _ links below 0.3g. "Everything;is not likely'to': pass at 3 0.3g. That was certainly in the case in Millstone 3 and-4 Limerick, and it was the case irothe Maine' Yankee review. And' 5 as I said, most of the seismic pRAs would-tend to indicata the G some position. 7- It - is not . the case at alf plants..' Frankly, doing the 8 margin review at Catawba at 0. 3g af ter the fact 'was too easyL to
- 1 9 do. We could leave too much conservatism left in the. margin.
10 review. We did not really have to push the methodology to show 11 that the plant'had a margin earthquake capability. bigger'than-12 0. 3 g , 13 DR. SIESS: What was it designed for? () 14 MR. KENNEDY: Oh, it was 1.5g,.but'it.has held. That 15 plant is tremendously over-designed for 1.5g. It has got very 16 conservative design practives. And there was no real push'to 17 show a margin over .3 D. 18 And I am not 100 percent sure that for'all of the. 19 plants that we are actually. going to find the weaker links'or 20 the margins level if we select this pre-selected earth' quake 21 level at 0. 3 g. But if we go beyond that' level, we are going to 22 have to increase. So we are going to have to have an answer do 23 we really have to show margin beyondEsomething.about'that 24 level. Because if'we do, we are going to have an awful lot of-25 extra work. O. Heritage Reporting Corporation I (202) 628-4888
-_a _ . . - . . _ . . _ - , . _ . . _. _ .- _ - -. -- -___-___--__._L--- '
1 l 205~ l' But'in this aspect, the EpRI' prograrn and the NRC-O 2 prograrn are essentially identical. We give"rnaybe a 1.itt le rnore L 3 ' guidance on how to select the earthquake level and how.to get a ) 4 specter shape, but they_are very sirnilar. 5 The next itern which in our philost.phy is this idea 'of' 6 trying to raake as rnuch of - it Las ceterrainistic as possib)e, we 7 have used a success path procedure or rnsthodology in the EpRI-8 pro D r > rn. Frankly, this is the. biggest departure or difference 4 9 between the EpRI prograrn and the NRC expert panel ] 1 10 recornrnendat ions, and it is the source of continued discussion. 11 The success path approach used in.the EpRI prograrn . i s , 12 very siroilar to that that has been prepared for. resolution of' 13 the 846 issues which is basically a success path oriented 14 approach also. And that was one of our stt'ongest reasons for 15 selectinD it here, because it was our desire that if this 16 rnet hodology is adapted is to try to eventually rnarry those two 17 progratas closer together, so that these plants go.through one 18 walkdown for both purposes. And it is easy to do that with-the 19 EpRI rnet hodo l ogy, to rnarry the two prograrns together and go 20 through the walkdown, for both purposes a single walkdown. 21 In the success path approach, we do require"that you 22 look at at least two alternate success paths'to shutdown and 23 raaintenance for 72 hours. -Ycu cannot Just look at a single 24 success path for successful shutdown. You need to look at two. ; 25 And you need to look at the support equiprnent .that nupporty. I Heritage Reporting Corporation (202) 628-4888 i
206 1 those success paths, and you need to look at the systems 2 interaction issues that affect those success paths as part of 3 the success path approach. 4 It is nct clear at least to me yet, and I do not 5 think to other people either, that if you do the success path G approach in the level of detail and rigor that we are requiring 7 that it is not 100 percer:t clear how much you save or do not B nave versus doing the fault-free approach in the way that it 9 was done at Maine '/ankee which was a simplified approach. 10 (Cont i nued on the next page.) 11 r 12 13 O " 15 16 17 18 19 20 21 22 23 24 -- 25 i O Heritage ReportinD Corporation (202) 628-4888
--_-_-------_------,---------_a,- - - - - - - - -
1 i
- 207 fr1 1 MRe. KENNEDY: (Continued) --
2n the way it was done. 2 at-Maine Yankee which was a simplified fault path. approach, but
~
3 we.do think it. enables us to key it much closer into tne A46 < , 1 4 program'if.we follow this success path approach, and we hear 5 from the utilities that a number of'them would be more 6 comfortable following a succese path approach., 7 In the success path approach, key questions asked'are
.)
8 will the. components in the success path survive and operate. I 9 That's basica11y 's capacity quest ion. The next question 10 requires significant-interaction between the capacity people' 11 and the systems people, and that is if you are going to go with'
- 12 a success path approach, you have got to-look at su'ficientf 13 instrumentation and assure the survival of that instrumentation
() 14 15 so that you know that the operators will.know how to go down that success path. They have got.to have adequate 16 . instrumentation. ! 17 Third, the success path has got.to be pass or success. 18 pass all of them, at least -- they should be pass that:the-19 operators know how to'use and will use. . It . can' t be some ; 20 esoteric path that has maybe very high seismic capacity,'but no 21 operator would choose to use it. 22 They also, if you use the success path 1 approach, it's ! 1 23 not on the viewgraph, but in' response to the kinds of comments 24 that Dr. Budnitz was making earlier, in a success path approach .I 25 you cannot have in your success path unreliable components from l
.i Heritage ReportinD Corporation
'(202) 628-4888' i
.__.._._.._______L
_-m_
1 1
.I
)
i 208.
]
- 1. a nonseismic standpoint. They have gutD.to,'again,fhave a l
I 2 reliability.somewhere in the'98 - 99 percent reliability range, ) 3 or they shouldn' t be included.in the success path, because if 4 you don' t have that level of reliability,' I don' t know how you 5 say because your components had high confidence, low I 6 probability of seismic failure.that'the whole. plant has high 7 confidance of low . probability of seismic failure. You need. l 8 high reliability.on the components to do that, i 9 Lastly, we strengly recommend, we don' t require, - we
.10 do give another option, but we strongly recommend that at.least .j 11 one of the selected success paths should include the.
12 consequences of'a small pipe break or leak equivalent to a-13 small LOCA. () 14 Basically, that means you need a water support, a 15 water source. The idea is it's very unlikely .that you could 16 walk down a plant so well that you could convince yourself 17 there was no way of small, small LOCA-type leak someplace.in 18 the plant after a major earthquake. 19 So rather than attempting to do that walkdown to 20 justify no such leak, we' re makinD the recommendation that the 21 success path should include the consequences of a small pipe 22 break. 23 Now in comparison of the success path approach-isfin 24 the margins program with a' success _ path approach is in A461for 1 25 the Spug program. Basically a margins program looks at more l l Heritage Reporting Corporation (202) 628-4888- .) i i _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ J
7 209 1' components than the A46 program. -It looks at tanks. .The A46-2 program does not currently do that. It looks at systems 1 3 interaction' issues. I 4 So.if the two programs eventually were married into a 5 single walkdown, that single'walkdown would have to. include 6 more components. It would have to include'all those components 7 necessary for the margins program, and that would automatically
~
8 include all those components necessary for.A46. 9 So the merging would require looking at.more' components, and 10 looking at systems interaction issues. 11 Secondly, a margins program requires'looking at these 12 for a higher earthquake level than the SSE. Frankly, I don' t 13 see that as a major difficulty. The earthquakefexperience data () 14 base goes up to about .3g as I said, and I think that the 15 people walking down a plant should have an earthquake level 16 about that in mind when they are screenin,q out components and 17 components below that you would like to have a capacity 18 estimate made on. But it is at a h'igher level. There will be .] 19 somewhat less components screened out. 20 The third difference is a margins program, at least , l 21 the way the EpRI margins program envisions it, the NRC program j l 22 envisioned it, will look at these. items in less detail of i 23 inspection ~than is likely to be required in the A46 program. I i 24 And the A46 program is likely to require a rather;significant , 25 sampling of actually putting a wrench to bolts to be sure that O Heritage Reporting Corporation (202)'628-4888-1
'210 1 they don' t twist' freely in the hol'e. We don' t envision ~that in 0 2 a margins program.
3 So that by marrying the two together we probably;have 4 to'take.the moro detail of the A46 program.. But you T+ould 5 still go by these items one time.rather than two separate 6 walkdowns. 7' A margins program'does not looking at piping in any 8 great detail, f,46 does not either, because the piping i s . n o't . 9 part of FA6. A margins program will not look st cab'le trays in. 10 great detail, particularly'if we. find the oldes trays like in 11 ind ust ri al plants, because we know they have had excellent 12 seismic performance in past earthquakes. 13 Now possibly some of the roully beefed up rigid trays 14 that we have ir new nut. lear olants, < we may have - to .look at 15 those in more detail., That may be.a biased opinion, but it's 16 my opinion that might not be as safe as the rod-hung trays ~that 17 we have in the older' plants and the unbraced cantilever trays 18 that we have in the older plants. 19 But there is the capability of marrying these into a 20 single walkdown with having to use somewhat more detail 'of the l 21 A4C and the more components and higher earthquake level of.the 1 22 margins program. 23 The margins program is described in the EPRI 24 methodology concentrates and provides more. discussion on the- ; l 25 iwsue of seismic systems interaction than is' described'in the
]
i i O. , Heritage Reporting Corporation ; (202) 628-4888-l l E_ _ _ _ __
1 211 1 .NRC expert panel report. Things like valve' impact, cabinet- > 0: 2 bangin0, seismic-induced flooding e.re discussed.in more' detail q 1 J 3 and'more guidance is given.in the a- bec.a use it was written- .i
- 4. after the other program.and was intended to exoand upon it.
5 These are the locations, at least in our.. Judgment,
]
6 where if .there is weaker links, a fair amount,of'these cea!'.er 7 links.do tend to be in the seismic ~ systems interaction.. issues,
.I j
6 er in anchorage issues. And so we want the seismic system
.l 9 interaction issues looked as part o? the mar gins program.
10 A very heavy emphasis.in the EPRI margins program on > 11 se?.smic walkdowns by experienced engineers with the heaviest of : l I 12 engineering Judgment and seismic ruggedness' pre-screening, both I 13 pre-screening of components out of the review and pre-screening 14 of components into the review. 15 It's indicated it follows a conservative l 16 deterministic failure margin procedure for estimating these 17 high confidence, low probability-of failure seismic capacities, 18 and the purpose of that is that we feel that deterministic 19 approaches can be used by a greater number of experienced
.]
20 earthquake engineers, and that the calculations are simpler, l l 21 and that they can be expected to lead to more repeatable and 1 l 22 substantially more easily reviewable results than the fragility .l 23 calculations. -j 24 Having been heavily involved in fraDility 25 calculations for over 20 nuclear power plants, I can. assure you Heritage Reporting Corporation (202) 628 4888 1 I
I t 212 1 the review of those calculations' is very dif ficult,. nnd I' 2 susnect the NRC people who have had to review such' calculations 3 also understand the difficulties of really understanding how 4 sensitive the end results are to some of those uncertainty 5 estimates. And the uncertainty estimates are highly uncertain 6 in themselves. 7 Ycu are trying to estimate a. capacity, then you are L. 8 trying to estimato.your uncertainty in that capacity. And'the
'i 9 end results are sensitive to your estimate of your' uncertainty..
10 Now we have. stayed away'in tho'EPRI program'from 11 defining high confidence, low probability ~ofLfailure by any , ;q
'l 12 mathematical statement that it's 95 percent confidence, 5 i d 1 13 percent frequency'of failure, because, frankly, we are'not sure
]
() 14 we could defend any such specific numbers. Those are about the 15 numbers we are trying to achieve, but we don' t ever say that in j 16 the report itself. 17 I am heavily involved right now in an effort, andcI ) 18 have done a number of calculations.both in that'and others, l 19 comparing the results-from this CDFM conservative deterministic j N 20 failure margin methodology with the fragility methodology for ' 21 estimating these HCLPF levels, and I find the two agree j i 22- extremely accurately, extremely closaly in every case that I-23 have made such a comparison, agree within-20 percent in.every
.I
'j 24 case, and that in most cases the fragility results tend to be 25 slightly hi Dh er, so that there may.be a little excess i
O Heritage Reporting' Corporation R l (202) 628-4888 : l
213 1: conservatism' thrown in to. achieve this deterministic: approach
- a. 2 of,a. deterministic methodology.
3 Dr. Ravindra showed you the fragility curve, and I am 4 not going to-spend time on it. But in the" fragility 5 methodology, first have to try to estimate 'a niedlan. on l 6 certairity bands on the-median, random slope of the fragility _l 7 line. -And from that median and these two estimates of. G variability, you back calculate'your high confidence, low I 9 probability of ~ failure as a 95 percent confidence, 5 p?rcent
,10 frequently of failure.
11 In the conservative deterministic failure margin 12 approach, you are attempting to go directly: to this point .by 13 deterministic calculations where the degres of conservatism'in j 14 the calculations has'been pre-selected from comparative. studies 15 to try to hit this poin." cn past fragility curves. 16 So it is .nore repeatable. It gives you less 17 information though on the full.' shape of the fragility curve, 18 It doesn' t give you that information. 19 DR. SIESS: To get some feel.for what that 95/5. - 20 means, have you ever applied that kind of an approach to a 21 conventional structure? 22 We' re sitting in a' building with say a floor loading 23 of 150 pounds per square foot. What would the 95/5 give you- l 24 for floor loading. l 25 - M R. KENNEDY: I have reviewed'some work'in that area. Heritage Reporting Corporation (202) 628-4888
, '214; !
1 I have never specifically.tried to apply this criteria to floor 2 loadings. I suspect the floor. loadings that we design for 1n 3 buildings.have lower. probability of exceedances'than 5. percent. 4 In other words, the floor loadings we design for I 5 doubt are at the 95' percent confidence, 5' percent 1 frequency of . l 6 exceedance. They are probably lower than 5 percent ' frequency. H 7 Why this confidence level'and frequency'of exceedance 8 were judged acceptable for earthquakes was because the 9 earthquakes we are dealing.With have a very ' low probability of 10 occaterence in the'first niace. So it was Judged -- 11 DR. SIESS: What about seismic -- i 12 MR. KENNEDY: And because of the seismic'pRAs that -- l 13 the low about this level was not contributing to the seismic () 14 risk in these seismic PRAs. i 15 DR. SIESS: What about seismic design of buildings in 16 California? 17 MR. KENNEDY: Well, in seismic design of buildings in 18 California, it's so hard tc compare that with nuclear plant 39 criteria, because buildings in Cal ~ifornia are designed taking 20 tremendous credit, tremendous credit for inelastic energy 21 absorption capability in the building arid credit for that on 22 the input to the equipment-in the butiding as.well. 23 So.that, you know, if we take a stendard building. l 24 design procedure, we' re designing for snywhere from 10 percent ! 25 to 20 percent of the-forces that we design a nuclear power O Heritage Reporting Corporation (202) 628-4888 - ' l
- . - - _ _ ~
215 !
. 1 plant for. So we are far more conservative in this methodology l I
2 of evaluating a' nuclear power plant than current building code j
~3 design practice. I 4 The. building code design practice, because it takes 5 so much credit for inelastic energy. absorption, it's very, very 6 hard to compare them with nuclear power plant practice. People 7 have tried doing-that, and the numbers continuously come out at S the same location,1the same facility. Nuclear power plants are 9 being designed for five to ten times the ferees, the lateral 10 forces, and the anchorage is being designed for.five.to ten i 11 times the lateral forces that a building is being designed for.
12 This conservative deterministic failure margin , j 13 methodology philosophy is described on this'viewgraph.. We find ! (} 14 that if you fit this, that it leads you pretty close to this 15 fragility curve HCLpF level. 1 16 Basically it says that for a margins review we' re l 17 only going to combine the margin earthquake with other normal 18 operating loads, no other extreme loadsi just normal operating 19 loads that could be expected to be present when the earthquake I 20 occurs. 21 Now if certain extreme loads might be caused by the; 22 earthquake, then they have to.be in that combination. For 23 instance, the issue of whether.an earthquake causes safety 1 24 relief valve discharge in a boiling watcr reactor, safety I 25 relief discharge is reasonably likely in a major earthquake, O' Heritage Reporting Corporation (202) 628-4888 I
l l
'216 '
i and so they need to be considered. But in general, it's normal j O. 2 plus the earthquake. 3 :Then in the ground response spectrum area, feel very 4 strongly.that-that is an area we must add conservatism. If you 5 are going to.be able to say.you want high confidenceJof a low l 6
.)
probability 'of failure of this plant in.some earthquake,.you' 1 7 .want a conservative ground response spectrum, because.there,is i 8 significant variability there. 9 So in the EPRI program there is a strong 10 recommendation that the ground response specter should.be q 11 . defined at'the 84 percent nonexceedance: probability, site-i 12 specific spectrum ~ shape, if possible, i 13 And it goes into further discussion of this issue'
)
() 14 that came up when Dr. Ravindra made his presentation, and-the 15 basic reason for wanting it at that l'evel of conservatism is a 16 desire ~to remove'. variability of the ground motion out'of 17 consideration of a seismic. capacity engineers, keepLit with'the 18 seismic" hazard people. They have a better knowledge of what 19 the variability is than the seismic capacity. people do. And if 20 you set it about this level of conservat ism, it ' doesn' t too 21 much me.tter what the seismic capacity people' assume the 22 variability;is in the fragility method, and it doesn' t matter 23 at all in the deterministic' method. 24 Damping structural modeling,-soil' structure 25 interaction, the_ idea is to do a median-centered analysis. No l Heritage Reporting Corporation (202) 628-4888-
,q
- j.
a 1 217: 1 intentional conservatism in the' response analysis in a margins-2 review beyond the conserva.tism-that you put'in by a 3 conservatism specification of'the' ground response spectrum, but.
'A to treat your uncertainties'and to envelope:your uncertainties.
I ! 5 So it's a median centered, but you have to recognize" 1 6 you don' t exact ly. know everything, and you have got to' cover I 7 that with some treatment of uncertainties. Andsthat means for~ 8 damping selecting a conservative estimate of.the median- b 9 centered damping. If you don' t know what'the median-centered : 10 damping is, you should conservatively estimate the median i 11 center. But we shouldn' t be out ,here t trying to estimate- !)
-) .
12 arbitrarily low camping values to add conservatism. We should-i 13 be trying to estimate the median. q () 14 In the case of structural.modeling, median centered 15 but with some frequency variation to account for uncertainties 16 uncertainties of cracking of concrete because,. frankly, you 17 don't know the natural' frequency of an existing as-built 18 concrete structure, particularly sheer-wall structures.very 1 19 well. It's taking year to prove that as we got into earlier 20 today, but you don' t know it very well. I don' t' think ' that l , 21 it's necessary to prove it. I think that people who have [ 22 worked around concrete already know'that. 1 . 23 Soil structure interaction, an attempt , at median- l 24 centered evaluation. In other words, taking full credit for ! 25 vertical special variation of the. ground motion,.taking full l-s 1 O. Heritage Reporting Corporation j (202) 628-4888 1 e
- l 1
F t m 218 ~ 1 ' credit for base mat averaging effects whichlsometimes.have been'
.O 2 called towel effects,. butErecognizing - 'but when'you'take full 3 credit for those also include.the neDative. kinematic 4 interaction' effects to getta median-centered soil structure 5 interaction. analysis,.but recognize that you..have got 6 considerable uncertainty in-that analyses,-and do some. l i
7 -uncertainty sensitivity studies and envelope.the results of 8 those. 9 So you have added some conservat' ism i n, in the i 10 uncertainty part,;but that's the only place. l-i 11 Under capacity parts, to specify' material strengths 12 at about the 95 percent exceedance.of the' actual strength i f. j 1 13 you have the data. If you don' t have the data, then usenthe l () 14 code specified minimum strengths. - l 15 Now 95 percent exceedance on actual strengths would ; 1 16 be somewhat higher than code specified minimum strengths in 17 most cases if you have the data to justify that. 18 Static capacity equations, aDain if you have the data . 19 to justify liberalization of code equations in a margins
- 1 20 review, that's recommended to be, done, to; basically set these 21 equations at about the 84 percent exceedance level.
22 If you set the equations at the 84 percent.exceedance 23 level and you set the material strengths'at-the 95 percent 24 exceedance level, in both places you have added substantial-25 conservatism. O. Heritage Reporting Corporation (202). 628-4888 =__- _--___-:_--_-.. - _-_ _ ._ _ - ._ . _ _ _ _ _ - _ _ - - - -
219 1 1 Again, i f you don' t .have the data, go f all back on O 2- the code equations. Dut places like low rise sheer walls, 3 there is substantial data to show conservatism in the code 4 equations, or buckling of flat bottom, fluid storage tanks 5 there is considerable data. 6 In a margins pecgram recommendation to incorporate 7 some inelastic energy absorption capability. The fact that 8 under; dynamic loads you can take a greater,. calculate' linear 9 elastica 11y calculated responses than you could if those were' 10' applied statically. It's recommended in the EpRI margins 11 proDram that'if you have.got ductal components you can multiply j l 12 the seismic stresses by .8 before being combined in with other l 13 loadings to account 'for this without doing nonlinear analysis (} 14 es long as you know it's a ductal failure mode. 15 Alternately, you can do a nonlinear analysis.and take 1 16 further credit, but the recommendations are not to take further 17 credit unless you actually do a nonlinear analysis, l 1 f l 18 In the case of in-structure specter or floor specter 19 in the plant, the recommendations are to keep them unbroadened r 20 like they' d act ually be produced in an earthquake in a 21 reasonably linear otructure, keep them unbroadened and 22 frequency shifting'to account for your uncertainty, because in 23 the process of broadening them, which is simpler,_you create 24 'more energy in that floor specter than.ever existed in the 25 earthquake to start with, and it's a source of . substantial O Heritage Reporting Corporation (202) 628-4888 u_____________u_______ _ __ . _ . . __ _ ___ __
220 1 conservatism. O 2 So those are the kinds of- recommendations whicia are 3 now . being compared to frag ility . est imates. 4 DR. SIESS: It wasn' t clear as you went down the list 5 whether you were including an uncertainty.in each' case. E MR. KENNEDY: Well, basically, in the ground response
-7 specter,.no. uncertainty, We' re put t ing it-right'at the.85 8 percent --
9 DR. SIESS: Okay.- 10 MR. KENNEDY: -- exceedance. Structural response 11 analysis.should be median centered, but you should do-multiple i 12 analyses or frequency shifting to - .the structural model,' you 13 need to frequency shift'your median-centered analysis to () 14 account for the uncertainty of-the frequency of the structure. 15 Soil structure. interaction, we are recommending doing i 16 three analyses -- a median-centered analysis and a softer than 17 median and a stif fer than median analysis,c but taking full 18 credit for the full -- that these analyses;give you of vertical End13 19 spacial variation of tne groundwater. 20 (Continued on the next page.) 21 2R 23 24 25 O' Heritage Reporting Corporation (202) 628-4888 s.
, q:
7' ' l
'y y 221 1 hc T14_ '1 MR. KENNEDY: We' re trying to sts.y out of the 2 probablistic terminology.
3 DR. SIESS: I know'but -- [ . . . 4 MR. KENNEDY: .But~ that 's wha't t hey . are there . for,' t o . 5 cover the beta use. 6 DR. SIESS: . What happens to-the' beta sub-Rs? o ? 7 MR. KENNEDY: The bpta R moytly comesTin here.( This
;c 8 is the biggest ' single source of. beta R, and you cover eihat,by 9 specifying it at the 84 percent not exceeded probability."
)
A 10 Another place it comes in is down here and you do that by , 11 frequency shift. 12 DR. SIESS: What about its material strength?- !' 4 13 MR. KENNEDY: Basically-the way -- f .
,h
(} 14 DR. SIESS: You got it in after the 95 percent. ' 15 MR. KENNEDY: In the most of the more recent seismic i 16 pRAs, the way the decisions have been made,' the only things 17 that go into Beta-R are those things associated with ,, 18 variability of the ground motion. Material's strength is not., 19 so that all goes under uncertainty. And the,ideajis'From
./ ..
20 assistance engineers is you may not know what ' thp struct ure's \ : ': l 21 capacity is but.the structure knows it. 22 Now, if the structure L. nows it, that's not . random. 23 That's your . uncertainty about' it, and it gets very esoteric.. 24 DR. SIESS: That's not ithe way struct ural engineers 25 treat it. Heritage Reporting Corporation (202) 628-4888 * [ o'
/
.]
(,( l-1
( ' r
.i ?
;g 9 222
< ? l, i[./
1 MR.' HENNE!N: It's not the way :we' ve thought ..in the..
~e !
2 past, and in the tparliest fragilities, we put beta-R over-s 3 there. In the most. -scent PRAs, it's .all ' been moved under the 4 - uncertainty term; There isLrandomness term here, and there's s' 7 randomi 3 ss associated with the elantic energy S substantial
\'/ I >
..p .
.c i6 absorption. 'That'h. very dependert/on the3 details of the
,.i , "p7 earthquake. / l h 8 D R. SIESS: And yet you uay going throuun this i
~
9 process., you come out within 20 pcMcent?
)
m
.10 MR. KENNEDYQ I come out ' witbhn 20 percent.- in' '. i 11- approximately.25 casee thht I made the comparison. I can' t
. ... 11 12 give my specific ' numbers on these recent; ones because we' re 13 doing them indopar!dently of each other and we'll see how we 14 '#
each agree wllth each other. 15 When I do it by the deterministic. method and then I i 16 do it by the probablistic or fragility method, keeping the same 17 assumptions, keepi re; the same basicLknowledge of the component 18 as I had, I get atawert, within 20 percent of each other, av.d I i ~, . l' 19 guess so far they(ve varied-from the fragility. method.can be 11 .s > act ,i anywhere from midus 5 percent to plus 15 percer.t of the - ) e 1 21 deterministic result. .Now, there is a slight conservatism to 1 22 the deterministic result for those' cases. 23 That's- npt to say another engineer will got answers s 24 within 20 percent of my answers.. I think there'is bra a bigger j 25 di f ference between two ' engineers' estimate-than there willcbe.
]
l O' Heritage Reporting Corporation. I (202)~628-4888 i l 9
-)
- ]
1 d 2235 . 1 due to the methodology. ,- 2 DR. SIESS: Depends on who they work'for.
'3' MR, KENNEDY: But we' 11 find that out, guess, in 4 the near future. f l 5' Again, I. keep wanting to come back and emphasize 6 this,'but it is the essential ingredient of the EpR1 Seismic.
.7 Margins Program, as.well'as the NRC Seismic Margins Program, in 8 this feature, I don' t think there's any real difference. I 9 think maybe we spent more time expanding upon it, again, '
i. 10 'because that was the purpose of our effort was to expand on the g g i 11 details, but the essential agreement is a seismic walk down by 12 ' experienced seismic review team. 13 What e,e basically envision is a three- to five-man 14 team whose purpose is to go into this plant, once they.are 3 15 preselected and they' ve done their preparatory work, to go into 16 this plant and aggressively screen out from the margin review . 17 all those elements for which they are prepared'to state high 'l 18 confidence low probability of failure at that earthquake level, 19 based upon.their combined experience and Judgment, using' 20 earthquake experience data, these qualification test data, the 21 curves that EPRI has put toDether, all of this data' base to 22 aggressively screen out from the review, everything they can so , 1 23 they concentrate their effort on the remaining items and do the 'q 24 detailed review there to reduce the scope down to the-1mportant 25 items. 'And then to clearly define the failure modes that they l '. (J serit.ge negortin, co,poration
-(202) 628-4888 '
'224 :l
, 1 want reviewed for those itemsithat were not screened out. l O- 2 To define duringithe walk down,; failure modes that 1 3 they.are' concerned about and that need a detailed review, I 4 probobly calculational review. And the:second part of the walk 1 5 down, af ter they' ve 'agg 9ssively screened out, is to add items-C in and that's just' as important'that while they' re in .this walk I I7 down mode, they are to be looking for seinmic systems 8 interac -3,on issues which they judge to be suf ficiently serious ' l 9 to warrant review. 10 Dur real concern is that'if we Just open up the door -
~
11 to seismic systems interaction issues -- because I' ve been 12 involved in a couple of cases -- back comes a detailed. review 13 of 6,000 issues and then you have to. spend a year checking.them () 14 off one at a time. The idea here is to start the.other=way 15 that these people, when they' re doing the walk down, they are 16 not to have blinders on. They are to be aggressively looh.ing 17 for these issues, but ~they only put down those' issues that are la considered suf ficiently serious to warrant a-detailed review. 19 Such. things as hung ceilings in controller rooms,- 20 space heaters over batteries, vulnerable non-seismic water I 21 storaDe tanks that didn' t have any. seismic design and'could be l 22 expected to likely fail in an earthquake or.possibly fail in an 23 earthquake and the consequences of that. Serious impact 24 situations, things like that. 25 Now, because these are judgmental and. experience
' Heritage- Reportinn
~
Corporation (202) 628-4888
l
;i i
225' , 1 . based decisions, We' re strongly recommending that the . decisions !
;O; have to be made by a consensus of this review' team, that"none l R I
i ! 3 Of these decisions are made by a single person, no matter who j 4 that person is. They take at least two. 5 That if they cannot reach a. consensus, then.the 1 1 6 conservative decision is made, i . e. , .you do not screen out-
]
7 items, and you do screen in systems interactions if they cannot 8 reach a consensus. 9 Now, the items which are most'significantly i' 10 emphasized in the EpRI Margins program and pretty close to the 11 same list as the NRC Margins program, maybe slightly different 12 emphasis, but active electrical components,'in particular, the ~I i 13 diesel generators and their peripherals, by that, the: day () 14 tanks, the control panels, the air lines, if they' re needed, 15 any batteries that might be needed, systems interaction' issues 16 in the vicinity of the diesel generator for'their peripherals. 17 The batteries and their racks, and particular 18 emphasis is on the racks. Are there. adequate spacers.between 19 the batteries? Are the batteries supported by end and side 20 rails? Acreage of the racks and any systems interaction. issues 21 in the vicinity of these batteries. If you hung a Spalding 22 space heater can be extremely disastrous. 23 Invertor, chargers, transformers. ' Basically the 24 issues are the inverto.s and chargers, there are issues 25 associated with the way those.are supported'down to the civil Heritage Reporting Corporation (202) 628-4888'
1
-1 226 1 structure. Because they tend to be raised _ up .off the- civil
.O 2 structure for airflow, and then the support ing struct ure _ that 3 takes them down to the civil structure may be very different 4 than what was used on the shake deck. And we' ve seen that case 5 af ter ' case, and there needs to be a careful review of that l
6 detail, and that's discussed, i 7 'How the transformer is supported inside the invertor 8- is an important issue. You need to open up these cabinets.
'i 9 You' l l never be able to do a successful margins review if you ;
i 10 can' t open the cabinet s. 11 The transformers are a tougher issue. It was this 12 issue that was found at Maine Yankee. It's been an issue that 13 I have been a bug on and so has Loring Wyllie on the other side () 14 on as well. A lot of these transformers before we went into 15 seismic qualification programs, the dry transformers, if you' ll 16 open up the enclosure, you' ll find that the' transformer is 17 basically lose inside of that enclosure. And the reason is, if l 18 it's not loose, it creates vibrat ions. And those in some 19 applications are considered detrimental and so there's shipping l 20 bolts and the instructions are to remove the nuts from the 21 shipping bolts when you inst al l the transformer. And so plants I 22 that were built before seismic qualifications became very 23 prevalent. 24 That feature needs to be looked at. We' ve had a hard 25 time sometimes convincing people of that because we_ can' t point j I 1 Heritage Reporting Corporation (202) 628-4888 ;
)
\
_ _ - - - _ - - - - -- _ a
227-1 to a seismic failure. But'it looksivery suspicious. And we 0 2 have suspicioned that there's been - some seismic failures. 'We 3 know there have been seismic failures of. transformers and-the' 4' failures that have been repcrted sound like they could have-5 been due to this cause, basically a burning out of cones, which 6 could easily occur if this thing shifted a little bit'inside 7 the enclosure. So'this is going to be.a tough'one. It 8 requires opening of that, but we think it's important. 5 Motor control centers and' switch gear, the-real. issue 10 is adequate anchoraDe. If you can' t judge the adequacy.of the. 11 base anchorage because you can't.get in'there and-look, do'what-12 was done at Maine Yankee, install al lot of top anchorage.- iThe. 13 Maine Yankee cabinets are known to b'eihell for stout anchorage () 14 because you can see the anchorage. There's big angle ' irons, 15 and they go between cabinet to. cabinet, and over to the walls. 16' That's why they were not an ' issue at Maine Yankee. 17 DR. SIESS: Bob, getting-back to the transformers, o 18 why can' t you put those locking nuts back on? 19 MR. KENNEDY - That's what they did-at Maine Yankee. 20 They create a little vibration. 21 DR. SIESS: So? 22 MR. KENNEDY: I mean, t h ere' s lots of_ vibration in a 23 nuclear plant anyway or in any other heavy industrial facility. 24 Before putting them back on, we wanted to talk with 25- the vendor, in that case, G.E., to be sure there was no harm O Heritage Reporting Corporation (202).628-4888-
1 l' l
' 228l .
d 1 they could be.doing to it-electrically by'doing that. And'they J 2 got that assurance. I do think it'h'as to be' carefully.
)
3' approached. Youoneed to. talk to the_ vendor to be sure you , i 4- aren' t doing some harm some other place, j l l L 5 DR. SIESS: Oh, yeah. ] 1 l 6 MR. KENNEDY: The relay chatter and breaker trip 1 7 issues are quite a bit discussed in the EPRI-Seismic Margins 8 Program, again because there was a chance to expand in that'
- 9. area. Primarily theseLare. issues that are going to be solved 10 primarily in the domain of the systems engineers and the 11 electrical circuitry engineers, because fragility engineers 12 almost have to throw their hands up. .There is-so much 13 variability there from a capacity standpoint that it's almost i
() 14 unsolvable by the capacity engineers. 15 Water storage tanks, and'I should probably add to the j l 16 list, fuel oil storage tanks for the outside fuel. oil' storage 1 17 i f it's stored in above ground tanks, definitely need a I 18 concentration, and there's a very extensive appendix to the H 19 EPRI Margins Program on details of how-to do a water storaje l 1 i l 20 tank evaluation. 1 21 MR. WYLIE: Let me ask you a question. l 22 Of course, a lot of this equipment in newer plants 23 have been tested you know on shakedowns, and so forth. And 24 they' ve learned some things about them,'- like motor control l l 25 centers and switch gear, they found the doors flew open, things Heritage Reporting Corporation f (202) 628-4888 l 1
- l 1
1
'l
'229' /
.' 1 'like that, and the man'ufacturers went'back'and made I
2 modification's. I i 3 How'is that being factored into this program? 'l
'l" 4 MR. KENNEDY: The cases where doors have' flown open 5 that I' ra aware of have generally been qu'alification testing .,
.6 which was quite a bit higher than what you would' expect-in'a a
7 .3G earthquake. We' ve had 'some instances of that in the L 8 earthquake experience data. 9 It's not easy to factor in. The walk down. team needs 10 to be looking for indications of weak door' latching toechanisms. 11 I' m not aware of any door having flown open with any decent l l 12 latching mechanism on it. There's no emphasis in the Margins 1 13 Program on that. () 14 MR. WYLIE: I know in some of the early ' testing, the 15 manufacturers did that. 16 MR. KENNEDY: I' m aware that some 'of the early . 17 testing had that problem. I 18 MR. WYLIE: Then they went back and modified their 19 latches and things like that. ; 1 20 MR. KENNEDY: We are not seeing that problem to any- 'l 21 si Dnificant extent in the earthquake experience data. We 22 overtest in these gault tests to a large extent. I don' t think 23 that' it's a major issue below .3G. But it is an area that (. 24 maybe our data base is.still weak. , l 25 Well, other. items on here, the non' load bearing Heritage Reporting Corporation (202) 628-4888 _ _ _ _ _ _ _ _ _ _- _J
230 1 ' masonry' walls, emphasis,again.on the detailed plant walk downs 2 with particular emphasis'on'doing those. walk downs from 3 anchorage of equipment and seismic induced systems 4 interactions. 5 Lastly on the list-is soil' failure modes, liquid 6- factions, slope instability and earth dam. The EPRI. Margins 7 program has a complete' chapter on that. subject on criteria for 8 the soil failure mode evaluation as part of a Margins review. 9 iso so far, th'at chapter has not.been fully exercised 10 -because the. Margins programs so far have all been done on rock 11- site plants. It is our desire to exercise that chapter and 12 that's the strongest desire for the next plant ought to be:a 13 soil site plant. And it's my understanding that the desire is () 14 for the next the plant to be a BWR. plant,.an older BWR plant, 15 preferable Mark.One, on a soil s i t e.' 16 . DR. SIESS: It seems to me, _ Bob, unless' there's 17 something new done in the way they go at liquefaction, t hey' re i 18 going.to be some sites out there' awful hard.to prove three-19 tenths on. 20' MR.' KENNEDY: That appears to be the case. That l 3 21 appears to be the case. Well, first of all, the older 1 22 ' liquefaction evaluations are pretty conservative, j 23 DR. SIESS: Oh, yes, they.are. 24 MR. KENNEDY: For instance,.on the Vogel' plant, there 25 was concerns about potential forfliquefaction slightly above 1
. Heritage Reporting Corporation.
(202) 628-4888 i I ;
^ ,
i I l 231 8 i 1 the SSE. Professor Seed went down.there as part of a group with ' O' 2 Bechtel to do a detailed evaluation on liquefaction to see if. j l 3 it was going to be an issue for seismic ris'k in a mini-seismic t
.1 4 ERA effort. And concluded from his looking at the 1 boring-data !
i 5 that liquefaction as long as.we were talking.about magnitudes 7-1 6 and lower earthquakes, liquefaction was-not an issue.up'to .7,. ' l 7 . 8G, the highest credible. earthquake he could1think of, it was 8 'not an issue. 9 So I think there is margin.tnere.in a lot of.these-l 10 evaluations,.but it's. clearly something that we' ve got to feel 11 our way along on. Beca uese I' m not aware of a rigorous margins 12 review really looking into that, and I am_ aware that there's. a 13 lot of concern of liquefaction issues for~ plants in the () 14 southeast and there's are well 'below . 3G. 15 DR. SIESS: I can' t recall that'anybody ever did what i 16 I'd call a consequential review. 17 MR. KENNEDY: That is discussed in the EPRI Margin's 18 work. It's not just incipient liquefaction we' re worried 19 about, it's actual significant soil movement ~due to - 20 . liquefaction which will lead to other consequences. 21 D R. SIESS: Liquef act ion doesn' t bother me. We have q 22 one plant designed to float in the Atlantic Ocean, you irnow. 23 MR. KENNEDY: You' ve got to have significant 24 movements, either significant soil settlements or lateral 25 spreading, one or the other, so that you break piping. O 1 1 Heritage Reporting Corporation
]
(202) 628-4888 l
232 1 DR. SIESS: Relative movernent. O 2 MR, KENNEDY: Relative. movement.so that you; break 3 piping. Or.such that you have such severe distortion to the 4 sfoundation of a building like a diesel generator building., I 5 can' t conceive of liquefaction being an issue on an auxiliary-6 building or a containment. They' re so deeply founded.- .! 7 I'think the issues have got to be'either looking: at 8' the piping, looking at distortions of diesel generator 9 buildings,a nd things like that. But that's an area that .is 10 going .to be tested in the next Margins Review program.if the
.11 plant that we think is going to volunteer actually volunteers.
12 That completes my presentation. ; 13 MR. WYLIE: Back on the systems interaction review, I ()- 14 we discussed earlier the NRC progtam4 and some of the ways 15 they re going about that. Is the EpRI program similar to that? 16 Is it expanded to that extent, or? l 17 MR. KENNEDY: Well, there's much more substantial l 18 guidance of the things that need to be concerned about in the j
\
19 EpRI program from a walk down by capacity engineer' standpoint. l 20 I don' t think there's really anymore guidance from the systems 21 engineering aspects. The program certainly comments and: l I 22 discusses that f ou' ve got to watch out for seismic induced [ 23 flooding and points-out some of the areas where this is 24 possible, threaded pipe, cast iron piping, non-seismic category 25 one tankage in the plant. l l Heritage Reporting Corporation (202) 628-4888 i 3 - __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
i 233 II An issue'we had'in one seismic PRA was that an ~O 2 outside water storage tank was at a higher elevation than;the' l
- 3. plant, and the inlet.line into that storaie tank went up 4 through the top of the tank but because they didn' t' want 5 splashing, it?went all the way down. to nehr the bottom of the 6 t .a nk. . Thatuline then went down and through the turbine 7 buildinD which was not seismically designed,-and the.line was 8 'not seismically designed.
9 Now,.-in that. plant there was a systems 1 interaction- , ( 10 issue associated with siphoning, siphoning the water out of the .i 11 tank in an earthquake such that the water wouldn' t be there. 12 M R. WYLIE: We'll, for example, islthe analysis of'the 13 fire protection systems strictly one of-breaking or if its'a () 14 deluge system, are;they looking to.see whethernor not the 15 deluge valves misoperate during'an earthquake of that magnitude l 16 and cause it to sprinkle down the plant, or? 17 MR. KENNEDY: To the best ofimy recollection, the 18 only issues that are really raised on the firewster piping 19 system is breaking of the pipe. In other words, the issue you j 20 just raised I do not believe.is discussed in the EpRI report. 21 I cannot be 100 percent sure of that. I didn' t write the 22 systems part, but I don' t think it is. 23 DR. SIESS: It.seems to me there was a pipe break , 24 recently and it activated the alarm system. 25 MR. WYLIE:- That.was at Surrey. Heritage Reporting Corporation j (202) 628-4888 l _m-_ _____.&_.__.___._m____.___.-___________w
1 1 254- ) 1
.1 DR. SIESS: At-surrey? !
O- 2 . MR. WYLIE: Well, t' hat's a classical interaction , 3 where the feedwater line caused moisture to get into;the halon 4 system and initiated that and sprayed into'the control area. 5- DR. SIESS: How'far down.the line do you look for 6 . interactions, you.know? I mean that was sort of Rube Goldberg. - 7 This thing broke'and that'affected that, and that affe'eted-8 some' thing else. 9 M R. WYLIE: 'And itJended up. th' rowing CD: balon all 10 through the control area.. 11 MR. KENNEDY: You' re getting out of my. area of 12 expertise as to how far you should'look.for systems. , 13 interactions. 14 DR. SIESS: Those are the ones that bother us. 15 MR. KENNEDY: I know. Systems interactions,are 16 something that we shouldn' t ignore, and I' m. sympathet. ic with 1 17 your statements. I don' t know how, I mean, . I' m not a systems i 18 engineer. I don' t really know how far we should describe for a 19 seismic margins program to go out into those kinds of systems 20 interactions. I' d almost prefer, those types of systems 21 interactions could occur without the earthquake. 22 DR. SIESS: Oh, yeah. ] 23 (Continued on the'next.page.) 24 j 25 O Heritage Reporting Corporation
\
-(202) 628-4888 .]
i __.________________n_______ 1__'.___ _ _ _ . _ _ _ . _ _ _ _ . _ . _ _ . _ _ _ _ . _ _ _ . _ . . _ _ _ _a_______ _ _ _ _ _ _ _ _ _ _
1 I 235 I
.I i . . .
q l 1 MR. WYLIE: Controls those things as seismic - l 8 susceptible. Sitting there like a trigger. 3 , MR. KENNEDY:: In some plants, it might be. Now,'in a 4 lot of the. plants, it is my understanding they are dry lines H l i 5- until'they are turned on. And that that takes more than - - l 6 MR. WYLIE: Well, there is a mixed bag ~. lI think that. f 7 you go to one plant and you:have got gas systems in the control 8 and' switch gear areas. You go to another plant and they have-9 got sprinkling systems. And'one plant has got deluge system-l 10 and the other has got, dry pipes. Some have'got wet' pipes with . 1 11 heads on them. It.is a'whole mixed bag.
~
12 MR. CUMMINGS: There is a concern ~that I' don' t think l \ l 13 a whole lot of attention has been paid'to seismically induced 14 fires. And when you talk.about fire protection systems,.you! ] 15 have got to think that if there is correlation between a fire . 16 and a seismic event so that the~ fire system is activated or,. l 17 indeed, it could be just your alarms and.your sensors are 1 18 trigDered by the seismic' event, independent of having a fire or 19 not, you might have a chance of actually activating a fire in a 20 deluge system rather than gas or water that would .have, you l 21 know, some sort of correlation to seismic events. As far as.I . 22 know that has not been looked at in any detail. 23 MR. KENNEDY: I have to admit that being a seismic 24 capacity engineer, I am prejudiced, but I want to make th'ese 25 walkdowns simple enough and' straightforward enough-thatLevery: Heritage Reporting Corporation-(202) 628-4888 ) l
l l 1 236 I 1 plant goes through thern' because I think every walkdown that has !
'u
~.O- 2 been done, sornething has been found. And I think that that is 3 a way to irnprove these plans, but it will not; cover those 1
4 issues. -We are going to have to find another way to cover'- , -l l
-5 thone- ,
-6 MR. WYLIE: One of . the iterns that we raised-in review )
'l 7 of the A-46 resolution was interactions, again. And it had to .
8 do with things failing in one' area'of the plant, ' like a tankf 9 full of-water or whatever, and then that being' discharged into-10 another part of~the plant where you have safety equipraent. . 'I s 71
~
11 that being looked at?' ] 12 MR. KENNEDY: That part of systerns interact ion needs l 13 to be treated in this seistnic walkdown because it 'is the, 14 seistnic capacity engineers who can assess whether that. tank -is . 15 going to fail. And then the systerns engineers' ought to be able j 16 to assess what the consequences are. 17 DR. SIESS: But sorne of the systerns, interact ions that 18 people have talked about and I believe sorne that they have 19 looked at haven'.t worried about the rnechanisrn of things. They 20 sirnply say if that pipe fails or if that tank fails, we have a l- 21 syst erns interaction.. So, that's ' fixed by drains or whatever, j 22 then we don' t care what causes the fa' are, i 23 It is irnportant, though, ;that' you look at those' kinds of failures"that could be caused only by an earthqjtake and are
~
24 ) 25 not covered by sorne other systerns interaction.
- Heritage Reporting' Corpora' tion
-(202) 628-4888 1
l y b
._ __m .___E.._____i.____ ._O
_.y 237 1 MR.. KENNEDY: Well,-several offthe plants - i . ( ! 2 DR. SIESS: Some of them are non-mechanistic. 3 MR. KENNEDY: Some of the plants that I have-4 participated i n, it was that was the way that they actually got 5 around this seisniic value retained issue is they could'show i 6 that they had done. sufficient studies already as-.to where this ! 7 water wodid go for water coming out of any non-seismic. safety 8- tankage, t h a't we . d i d n' t have to do any assessment of the 9 seismic. capability of the tankage ~because:they could 10 demonstrate that the water. wouldn' t get.-from there over to the 11 safety equipment, f i 12 That is probably likely to be the most cost effective ) J 13 way to resolve that issue, but I don' t know.. () 14 DR. SIESS: Well, thah is what they did at.Surry, 15 they went back and added a dam to' keep it out. 16 MR. KENNEDY: Right. 17 DR. SIESS: If they have already done a. good systems l 18 interaction, your seismic system interaction gets a little 19 easier. 20 MR. KENNEDY: Yes. 21 DR. SIESS: Garth was talking about'the seismically 22 caused fire and then somebody can look at what.the effect the' l 23 earthquake is going to have on the ability to t'ight a fire. It 24 is the human factor part, again. 25 DR. CUMMINGS:' Unfortunately, if you look at real Heritage Reporting Corporation
'(202) 628-4888 t
j l \- = _ . - _ _ _ __ : . _ - . _ _ _ -
238
.. 1 World accidents, you get involved in things like that.'
2 D R. SIESS: And if you-areidepending on the fire 3' department down the road, they.may be awful busy with the fire ', 4 in a village hall. Depending on your own fire crew, they may 5 be out doing something else. Because'the. earthquake that wei re
- 6. talking about is going to be fairly traumatic, both on site and- q 7 off site.
8 Okay, let Bob Budnit: come tell us what's wrong with 9 all this. 10 MP. BUDNITZ Nothing wrong with it except that'it is i' 11 quar ter after two. 12 -DR. SIESS: You get no sympathy whatsoever. 13 MR. BUDNITZ: I am not asking for sympathy. () 14 DR. SIESS: Anybody who wants to live out in that 15 part of the country that far away from everything -- 16 MR. BUDNITZ: I like the life. l I 17 DR. SIESS: You didn' t always live in California. 18 MR. BUDNITZ: What? l 19 DR. SIESS: You didn' t always live in California. i l 20 MR. BUDNITZ: I' ve been there for twenty . years, 21 though, with a two-year errata. 22 There was.a panel.to review -- well,;the Latin. word, 23 " error" means to wander. And I sort of wandered east for..a 24 while. 25 The panel to review to EPRI methodology, the NRC, L Heritage Reporting Corporation (202) 628-4888
.l
+ , j i 239 1 throug'h Livermore, appointed a panel to review the EPRI j t 2 methodology last spring and'it is the same row boat, different; ! 3 guys pulling cars-this time. Amico,' Mike Bohn, Bill Mills,. 4 John. Reed, Bernie Safel, Lorin'Wyllie and the charter, which is 5 on the thing, here is to review and critique the EPRI' seismic 6 margin methodology. And next, toLmake recommendations. 7 concerning~the general applicability of the' methodology for !
~
'8 determining both the seismic' margin and identifying.
9 vulnerabilities. And those words are taken straight out of the 10 charter which I cite for you in the letter of -- excuse me -- 11 July 15, 1987, there is a' letter, which if you don' t have, you a 12 should have, which is the. basis for this. summary, 13 Now, the review panel looked at the EPRI methodology () 14 and did not specifically review'the Cat'awba report,7although we. 15 used the Catawba report because it'was.the-example of how to' .
'I 16 use the methodology. So, We didn' t find ourselves saying, '.
17 "Yes, the Catawba report was done properly.and therefore the 18 Catawba plant conclusions are accurate or whatever."- We didn' t 19 do.that. 20 What'we did was we used the Catawba report to ... 21 understand how the methodology was used in an' actual trial. I 22 With that as a background,;I.have one slide. This is- q 23 my 30 minute,.15 minute, 2. minute talk. There'is one slide l 24 here. The central findings are, and this is a critical point,. 25 because there is a distinction here between. objectives. That' Heritage Reporting Corporation (202) 628-4888-f
. _ _ . . - - . - - .__- ____-_L__.-.-_:_._--.__x
1 i 240 gs 1 the EPRI methodology is capable of meeting its~ objective, of 0 2 determining the HCLPF based seismic margin of a nuclear plant. 1 3 And, as Bob' Kennedy said, that is only one of several.possible 4 objecticns that one might identify for a seismic review. But-5 if your objective is to' identify the.HCLPF based seismic q 6 margin, the panel found that that methodology. works. It works l 7 fine. It can work. 8 It is implementable and it is reasonably accurate. 9 Reasonably accurate means it is as. accurate as, for example,. 10 the NRC supported methodology which means you don' t bel'ieve the 11 last little bit of that second figure on the HCLPF, but I won' t 12 go into that any more. 13 Despite that, the review panel has identified'-- and () 14 there is along 20 page thing here that tells you something 15 about some of the limitations. And the most important 16 limitations are involved, I think, with the application of this 17 to do something different than that other objective, which is 18 to identify vulnerabilities or to go beyond seismic margins 19 themselves, for which the EPRI methodology.was not designed. 20 And, as Bob Kennedy said eloquently an hour ago, for which it 21 is not as suited as other approaches. 22 That i s, if you really want to understand risk, you 23 have got to do a PRA. And if you want to understand 24 vulnerabilities, you have got to do a little more than that. 25 But ' for the narrow object ive, it works fine. O Heritage Reporting Corporation (202) 628-4888
241 1 Now, the key limitations-are that it only finds the. l 2 margin for the success paths analyzed. And we need to
~
3 understand that in plain English. That is, if there is another
.4 success path that is.not analyzed, there ain' t no way on earth 5 that this method can tell you anything about1it. j I
6 Now, that is not as different from the'NRC approach 7 as you think because, in fact, the NRC approach does some of 8 that screening, too. Just the most obvious example of Maine 9 Yankee was that because the rods -- the reactor internals were-10 judged to be sufficiently stout and the rods.were judged'to go 11 i n, one didn' t analyze the part of the' success path on that' . j 12 It does not work, you have got to borated system. So, that 13 screening was done and-only one of those two ways of doing that () 14 particular function was analyzed and found to=be stopped and we i 15 walked away and went on to.the next. But that is an important 16 point. 17 And it important to realize and Kennedy'said it,'that J 1 18 in this approach, you have got to use the success path the 9 19 operators are going to use. Even then, there are-some. issues 20 here which are going I' m going to come to in a minute. 1 21 Next. Non-seismic failures and operator errors.are l l 22 treated only in rudimentary way. And there is no hiding that 23 here. That is that if those things contributed in an important 24 way, you shouldn' t be analyzing that success path. The success 25 path requires you to do (a), (b), (c), .(d) and:(e) in order to O. Heritage Reporting Corporation (202)' 628-4888-
\ n i i
?42 1 get home. And (d) has an important non-seismic failure' i !
f 2 contribution or maybe it is out for test or maintenance. Or 1 3 maybeLit is an operator error that isn' t relat ed -- t hat the-4 human factor isn' t related to the earthquake or something. L 5 Then the EPRI methodology exhorts you,.it says, " Don' t use that ; 6' .one. Use (f), (g), (h), (i) and (J) or something."' 7 And the EPRI methodology is intrinsically incapable ) 8 of doing more than.that exhortation in terms of quantifying in , i 9 its impact. Now, the NRC approach on the'other hand has some I J ' 10 ability to do that, but it is not exactly' quantitative, it is 11 sort of semi-quantitative. .And this comparison, again, then 12 depends on your objective. If your. objective is one kind, you r 13 can do one thing. And if you have broader objectives, then 14 this maybe has some limitations. 15 The panel found and the letter goes into some detail 16 about why the success path approach is more prone to errors of 17 omission than the fault sptce approach .and - I won' t go-into that 18 in too much detail unless you ask me a question, which you j i 19 probably will. l 20 The fact is that in one sentence of detail, there has 21 been a history of success path analysis of various kinds in 22 which errors or omissions are -- that is omitting things, steps _ 23 that must be accomplished in which you have to be.very' careful. 24 Naw, in the hands of competent analysts who are. familiar with 25 the plant and, especially, in the hands of plant personnel ; l S . Heritage Reporting Corporation-(202) 628-4888 - - - -_-_ _ __-___ - _______ ___-__ ___ __-____-__ _ _ . - _ _ - _ _ .-. a
i i
,J 243 ,
1 working with-such analysts or maybe they are the same.that-2 likelihood is reduced. In the hanos of people that1 don' t know l l 3 muchLabout the plant, that dontt have much interaction, that i I
.a 4 : likelihood is increase'd and you have to koigh that. l
- 5. .I am trying to get through this fast on-account'of-6 it's' late.
7 Finally, and this.is the crucial point: The. success i 8 path approach, the EPRI approach does not pretend to identify l l 4 9 vulnerabilities, per se, in the same way as'the NRC approach j l 10 would because the NRC approach -- the fault space approach does 11 look at the whole plant,lat least the Group A function. 'And 12 for BWRs, where there aren' t any Group A, Group B distinctions l
-l i 13 quite the same, there is -- it does look at the whole plan. I l I
() 14 There isn' t any pretense that it can find' 15 vulnerabilities in success paths'not_ analyzed. And that is~a l 16 distinction which is important because if an objective is to l 17 identify vulnerabilities, then this approach, then, isn' t that - 18 suited. j i 19 On the other hand, I'want to repeat, if your j l 20 objective is that first thing I put on this other slide, that j i 21 is it is capable and it is implemented. 22 Now, Just to' reiterate that I have to make an 23 important distinction. You see if the EPRI approach. picks a j I 24 success path and starts marches down it: '(a),. (b), (c), ( d ') , l 25 and (e), and they find that - (d) is-in real trouble. You_know, I O.- Heritage Reporting Corporation l' (202) 628-4888 I i 3 ,
I: 244 1
'1 it's Dot a low capacity or something.
O. 2 There are two things that one can do and they'are.in L 3 the guidance. You can go and fix it, put a bunch of-bolto in.. 4 Or you can go and try, th), (i), (J), (k), _ (1), which is I l 5 another one. . And' either . isn' t acceptable. You;know, i ! l 6 eventually, you find.one'that. passes.~ Of. course,' it had better 1 l 7 be the.one the operators use. It does not do any goced to find l , p 8 (h), (1),- (J), ( k ) ,- end (1) i f they don' t use or they wouldn' t 9 use it or.they swear up and down they are never going'to use 10 it, which is, by the way, ari issue. 11 DR. SIESS: You' ve got two success paths with a hi Dh' 12 degree of capacity. A vulnerability in some other path isn' t 13 the vulnerability of the plant. () 14 MR. BUDNITZ: Th a t ' s fair. That's absolutely fair. 15 DR. SIESS: Bob, I' ve read your whole report. And I 16 think it raises a lot of interesting questions, but the most 17 important question it raises in my mind is what-is the 18 objective of doing seismic design margin -- is it a legitimate 19 objective either for the utility or for the NRC as a. protector 1 20 of health anu safety of the public? I don' t think anybody has , J 21 answered that. 22 MR. BUDNITZ: Of course, but that's because the crap 1 1 23 game is still floating around. l 24 DR. SIESS: I know. 25 MR. BUDNITZ: And just to say here, because it is
/
Heritage Reporting Corporation. (202) 628-4888
-245 1 important to say, there is a rumor floa' ting'around that.some O <
.2 day somebody may decide'to do IPEs. And, if'they do, they may 3 decide that the seismic part looks something like this.
4 : DR. SIESS: That may be it. 5 MR. BUDNITZ: I said this before. :And' if that'is so, 6 IPE -- the word, IPE, means examination of what?. Examination 7 of outliers, vulnerabilities. But'someone is go'ing to have to 1 8 define them because 1 am not interested in vulnerabilities-that i 9 occur 3(g), you know, I mean maybe some people, butsthat is 10 not me. 11 DR. SIESS: Well, the vulnerability or outlier has to. j i 12 be defined as one that effects the safety of the plant in the. 13 event of some -- () 14 MR. BUDNITZ: In the context of some overall -- 15 DR. SIESS: There is also something floating around 16 called a safety goal, which can only be established with PRAs,
. . I 17 which is defined in probabilistic terms.- And there is.no way I.
18 can take seismic margins, HCLPF and relate them to a safety 19 goal. 20 MR. BUDNITZ: Well, actually, ~ that is not'true. You 21 could do that fairly easily if you are willinD to make some 22 approximations. For example, if the margin is sufficiently 23 great, you can use a generic hazard curve and still show you 24 are okay. 25 DR. SIESS: .Okay. I can show that there is one. 5 Heritage Reporting Corporation (202)~628-4888
- i 24s' I
I l' . chance in 20 of fit failinD given this.certain weight. It k , I What I mean is -- suppose that. the 2 MR. BUDNITZ: ' 3- seismic review level that you' chose, point 3. Well, you ' don' t '
'I 4 knsw exactly what it is at that site, ' but you use Livermore's [ !
J 5' generic, you know, eastern stuff and then you try to co'what-6 you are Doing to do. AndLyou;say, well, . it's 10 minus 4.- 7 Okay? 8 The review level-is 10 minus 4 and then you.say,. 9 "I' ve got a high confidence and. low probability of failure 10 there." Well, that is some f' actor which you are going to have 11 to figure out. But in any event, it is probably 10 minus 5 or i 12 something and then maybe that's okay. The' staff is going to 13 have to figure that out. () 14 DR. SIESS: Well, I get a factor of 20 out of-it 15 without, you know, accepting the confidence for what it might 16 be, which is 5 times -- 17 MR. BUDNITZ: Yes, well, whatever you want' to call 18 it. It may be more. j 19 DR. SIESS: Okay. You can go at it that way. 20 MR. BUDNITZ: Yes, I am not going to argue the case. 21 It is not quantitative. . It 's ' sort of semi, you know. 22 DR. SIESS: Again, why are we doing it? I know why 23 it was done for Maine Yankee. I know.Why they.were interested. 24 I know how we got into the whole issue. 25 MR. BUDNITZt So, do I. Heritage Reporting Corporation
-(202) 628-4888
247
.i 1 D R'. SIESG:. But how we gct into it is so far back in 2 history that I am not sure what we are doing is~~-
3 MR. BUDNITZ: Now, you scee, this is worth discussing i 4 because I was sitting over there hhen Bob Kennedy was talking 1 5 about the recommendation that is in his head t. hat this sort of ,4 6 thing be combined with an A-46 type of review so you don' t have 7 to do everything twice, But I' m telling you that IPE1 are 8 coming and it seems to me -- all tight? And you don' t have to 9 agree with me that they are coming. But I thought the 10 Commissioners told people that they were coming. 11 DR. SIESS: Oh, the IPEs are cominD. I' m ,j ust not 12 worried about them. 13 MR. BUDNITZ: But it would be crazy to go walk down a ! () 14 plant -- for A-46 and come back nine months later and walking 15 down for margins and come back nine months later and walkir.g 16 down for IPEs. That's nuts. 17 DR. SIESS: I think we told them that a couple of' , 18 times. l 19 MR. BUDNITZ: Right? 20 DR. SIESS: But you know the origin of this thing in 21 a bunch of ACRS letters -- the first letters we wrote, we had a l 22 reason for looking at a higher G value. Some was .12 and now 23 it was .14. Then it got along the line to where it wasn' t a 24 reason. We Just thought this should be looked at for larger 25 earthquakes of lower probability. Heritage Reporting Corporation (202) 628-4888 i h_-m___. _ _ _ . _
.s t
,21 b 248'
' g-' 1 MR. DUDNITZ: Make sure there wasn' t . eclipsed, q 2 DR. SIESS: Well, an eclipse. !
3 But, again, ,all of that was still more or less-within. I 4 the context of design basic type events to the extent.that we 5 were limited to design basic type events. Obviously, we 6 weren' t worried unless it was going to be a core melt -.somewhere , 7 down the line. 8 And none of it was probabilistic. 9 MR. BUDNITZ: And to its detriment in that sense, 10 because without the probabilistic understanding, certain things 11 that are worth screening out can' t be screened out. 12 DR. .SIESS: Yes, but the only trouble with 13 probabilistic when it comes to earthquake is that the () 14 uncertainty of the hazard is so great that those people who 15 have lived and died by uncertainties go crazy. You know, 4 is 16 a magnitude uncertainty. It's silly. Ard all the research we i 17 have done has done nothing but increase the uncertainty. 18 M R. BUDNITZ: Yes, but at Maine Yankee probabilistics l 4 1 19 were important because they'took tFat tank that was 21-7 G, 20 told us t'mt it didn' t govern the capacity of the plant because 21 it was in and with a strong component. And that emerged from 22 probabilistic systems type sequence analysis. - l 23 DR. SIESS: Yes, on the plant, but I am talking j
.1 24 about --
25 MR. BUDNITZ: Yes. So, in other words you can' t
~
l i Heritage Reporting Corporation' (202) 628-4888 i
-q j
j i
249 i ignore those sorts of lessons.1 2 DR. SIESS: Andy wants to argue that the research"is 3 reduced to uncertainties, f 4' MR. MURPHY: No. I won' t- fight ' about that.- S The.po2nt that I wanted to make was thet, . unless I'am 6 mistaken, part of this program was set up to address the 7 Charleston issue gaestion. 8 DR. SIESS: That came'up lonD'after.it got--- that' 9 was one'of the good reasons we had, I think.. .
%', i 10 MR. MURPHY: Yes, but I mean that's one of. the thinys
.0 11 we are still developing a tool for. And it is next winter, J l 12 next spring, the folks at Livermore are going to come up withf 13 the ranking, if you'want, of the power plants in the eastern
(} 14 United States -- 15 DR. SIEES: l Uniform fazards assessment. type thing. , l 16 MR. MURPHY: Right. So that a tool like this woulci 17 be available to -- ,t, 18 CR. SIFSS: .-If t he MSG can make it work. I't doesn' t ] 19 fit'within the present regulations, the design basis. 20 MR. MURPHY: I' m not sure what doesn' t. 21 DR. SIESS: Well, by the tirae I take the ASME code 22 and Appendix A and define an SSE and 'so forth, HCLPF . isn' t 23 goinD to get me anywhere. A stafT rhviewer' says, "It doesn' t . 24 meet the code." , 25 M R. BUDNITZ: Yes, but what about' plants that are
-A]
( Heritage Reporting Corporation ( E'Le) 628-4888 e
.l (
, . f-
_ _ _ - - - - - - - - - . - - - - - - - -- . - - - - . - - - - 1
L , . 4, . M' A
', , :i i
t 250
; ~f 1 running? .Therefore,:they reest'the code and still'have
.O 2 vulnerability.
, . , 3 DR. -
SIESS: I know!. cu x N '
./. 4 MR. MURPHY:- Like was .manipghted with Maine Yankee, 5 DR. SIESS: Lipat are you;q6ing to back fit. That's a E, - Dood question. You k r.d.4, how do you back fit.. And the 4
7 Commission has never ,addres, sed these issues. SCP, let's' face 8
}l 1
S it. They tried to address it. SCP was one attempt, it all 9 hung on.a PRA. It is still up in the airs
. . i I fli .W hat do you do about older plants that do not meet' 11 current criteria? You whittle away at them here. Some you- -3 12 just don' t ' worry about. And all of a sudden, you start ~
13 worrying bout this aspect of all the plants that don' t meet - .j y 14 it. It cight be earthquakes this month,..it might. be' pipes in 15 the feodwater system at Surry, et cetera. You know, .those
]
16 weren' t even catpgory safety-related systems. 17 Wehbve got a problem now. Even in a severe accident!
\
18 thing, we are outside of the regulatory framework. There is j I 19 . nothing in 10 CFR -- j i 1 l 20 MR. BUDNITZ No, no. That's not accurate. You are j 21 outside of whet's the 10 CFR, but the Commissioners have stated )
. . l 22 something that as policy becomwc part of the regulatc,ry 23 framework.
24 DR. SIESS: That's ri ght. But outside of '10 CFR, ; i 25 which is one level of think'Ang about regulation, it is in ternes -I l : l l
;J Heritage iReporting Corporation l (202V 628-4SSO- !
+
b y 4
) lf l
, h.
______-__h & L'- rl (
- __ ____--- _w I
- q. -
. )
1 1 1
- ) l
.? 251~ \
\\ l 1
1 of a policy statement,'but so is a safety goal'. And nobooh{'. l O 2 understand it yet.
^
3 I mean we' ve got a policy statement' dad three j " 4 explanations of it which don' t agrees the six-safety goals. t f , 5 And severe accident-is, you know -- read it. Read wWat it 7 L
..A l 6 says, not what they.are 2oing to do. And we are still gapping !
.r 1 y ,. ,
7 for a way of looking at safety of plants with these things / .; , j
'L a
! I 8 What do you care if you worry about an outlier, ifi l
9 you are already at 10 to the minus 6 on severe' accident. .Is i 10 your mind on' core melt? ! 1 11 Suppose you got it down to where.it is all seismic, l j j 12 but it is 10 to the minus 6. Do you worry about it'p I don' t 13 know. Somebody says, yes, because.it is to uncertain. , 14 How many people can you get to talkL thdd ians wnoin they , 15 are looking at the seismic? They talk of uncertainties. It is J 16 a problem. 17 MR. DUDNITZ: Well, Chet, I have a suggest ion and; i 18 that is: the ACRS can help in this regard.
.h 19 DR. SIESS: Oh,- no, we would mess it-up.
20 MR. BUDNITZ: No, no. By commenting on this very 21 point that: a methodology has beenqdeveloped that has two 22 different but very similar manifestations,-the EpRI' developed ~ ;;H 23 approach and the NRC developed approach. They are similar. 'In 24 fact, they are closely related. One'fd11owing on the other. 25 Now, t, hey are good for a heck of a' lot, but sp411 not
,c!
i[ O Heritage Reporting Corporation ' (202) 628-4888 ,
.I b -:l
.g
~
(.
m c5 1
> 252 l 1
'l be implemented by anybody unless specific objectives are 1 L
]
2 delineated which haven' t been. l.- 3 DR. SIESS:. Yes, but you see, you are asking me.to go 4 .to the full committee and talk about the EPRI method and the 1 5- NRC method. And that is the least of my worries. tW MR. BUDNITZ: Oh, I don' t care which method it is. T DR. SIESS: I am tryinD to figure how I.can go to the L 8 full committee and explain what the seismic margin program is -
;/ 9- i'n relation to the PRAs 'that they have been looking at, in 10 relation to searches for vulnerabilities which nobody has every 11 defined and try to tell them what this means in terms of'the j l
> 12 things they have been thinking about.
i .13 MR. BUDNITZ: So, what do you think is the answer?- () 14 DR. SIESS: Do you see what I mean? l 15 MR. BUDNITZ: Yes. What do you think is the answer? 3. 16 DR. SIESS: I' m not sure. I wasn't sure when I came I 17 in here. And I don' t think I want to take this to the full i l l 18 committee right now unless it comes out as something that has
]
19 got to be used in a severe accident. Or,'if the Committee is 20 going to be faced with'a seismic upgrade,. Charleston-type 21 seismic upgrade for a plant, you know, it was our hope that if 22 you did this HCLPF sort of thing: Here's a plant designed for l 23 .15 and Livermcre gets through with the uniform hazards things, l 24 it comes out .16, somebody will say, "Oh, hell, forget about 25 ft." O Heritage Reporting Corporation (202) 628-4888 , i _ _ _ - _ - _ - __ . - _ _ _ _ _ - . _ - - . _ - ._.--______-_--_-___-_--__-a
253
, 1 I don' t know how they would do it, but this is one - 'd %> \
2 thing, you know. Now, that is common sense. And I hope it 3 could be used there. Now, if it is .1 to .18, like Millstone, 4 I would still be inclined to say forget about it. But, 5 obviously, once you have done the seismic margin study and come 6 out with . 27, anybody would say, it's a hell of a lot better 7 than we thouDht it was, you know. It's nothing to worry about. 8 I don' t know how that fits into the regulations. 9 But as far as how it is going to fit into severe
]
4 10 accidents, that is pretty much up in the air and I think the j i 11 full Committee would not really be receptive to this thing l 1 l 12 until it got into that stage and they could see it, i 13 MR. BUDNITZ: Well, Chet, there is an important point . 1 i 14 l l []- ( which I think is important to make, i l 15 DR. SIESS: What I am saying is I am not sure the 16 full ACRC is ready much more than I or Charlie are ready to get 17 into the methodology. ! 18 MR. BUDNITZ: Yes, I am not going to dispute that. 19 But there 1.s an important part which you need to understand. 20 And that is that depending on which -- I won' t use the word, 21 " objective," singular, but which set of objectives are going to 22 underlie this thing, different approaches are appropriate. 23 DR. SIESS: You will list there, you know, the fact 24 that the EpRI method does not find vulnerabilities. j 25 MR. BUDNITZ: Well, they do, but not in quite the l t Heritage Reporting Corporation (202) 628-4888
1 254
- 1 same systematic way. %.)
2 DR. SIESS: The fact that.it is more likely to be-3 incomplete because of being success space rather than failure i 4 space. It is easier to miss something. These are valid. points 5 for certain uses of it. They might be a more valid point for l 6 using it in severe accident than they would be for upgrading of. 7 plant due to a meandering Charleston. So, again, what are we j i 8 goinn to use it for? j 9 MR. BUDNITZ: And, again, if you are going to do it . 10 along with A-46, compatibility is a question, you know. 11 DR. SIESS: And, you know, some people, if you Just-12 do it and find some things and fix them, that is going to give 13 them a warm feeling, just because you found something and. fixed () 14 it. I mean I get a real nice feeling about Maine Vankee l 15 because they are going to beef up that tank. No redundancy in 16 that tank sitting over there. Just one of them. And it gives 17 me a warm feeling. 18 You know, I don' t know how to get a discussion going 1 19 on these objectives. It is not an easy question,.but I.would j I 20 like to see the ACRS do it, but it is going to.take more l I 21 background than any of them have right now. I have been 22 following this thing and I' ll have to' admit until I got into 1 23 these last few weeks and started. reading your reports and so i 24 forth and listening to possible objectives and beginning to 'l 25 think what we are going to use this for that these questions l O Heritage Reporting Corporation 'I (202) 628-4888 1 1
1 255 1 1 1 came to my mind: What are we going to use them for? There are 2 some legitimate uses. And for some ' of them, one method is Just 3 as good as the other, 1 4 - But ' if you want everybody to do this, if we' ve got 1 5 some good reason for wanting everybody to do it,.that has got ] j 6 to come somewhere out of regulations. IPE miDht be the thing. 7 But does it satisfy the people to get out of IPE what they i 8 think they-want out of it. I don' t think they know what they 9 want out of IPE. 10 I know what some people want. Some people won' t 11 settle for anything less than full PRA. Seismic PRA at that. 12 They' d like to see a seismic PRA on every plant. Personally, I 13 think that would be great. I think everybody who has done a () 14 PRA thinks they have learned a heck of a lot. 19 MR. GUZY: I would like to say something. The NRC 16 working group is reviewing the EPRI methodology now. And we 17 are looking, as the panel did -- we have listened to the panel 18 and I think we could state now, if we had to, to put it all 1 19 together was that we do believe the EPRI methodology is an ! 20 adequate way of addressing our first objective, which is the 21 Charleston / Maine Yankee type objective. 22 DR. SIESS: Now, have you got NRR's agreement on 23 that? 24 MR. GUZY: I think the people on~the working group ; 25 that represent NRR agree with that. Heritage Reporting Corporation (202) 628-4888 ! _-___________n
256 1 DR. SIESS: Okay, because you' ve got NRR people on O 2 the working group. You are speaking for the working group,' 3 now. 4 MR. GUZY: For the working group representatives. 5 DR. SIESS: Good point.
- 6. MR. GUZY: So, I don' t think there is any -- I think 7 as far as -- our first objective, to say something about the 8 plant for earthquake and the design basis, both' methodologies, 9 I.think the working group feels comfortable with that 10 objective. The second objective, we are still struggling'with, 11 And I think it is the criticism for -- !
12 DR. SIESS: Wait a minute. What was the second 13 objective? () 14 MR. GUZY: The second objective is define 15 vulnerabilities. 16 DR. SIESS: Oh, define vulnerabilities. 17 MR. GUZY: But I don' t that objective has been 18 clearly -- I mean it hasn' t been clearly answered for the NRC 19 program either. 20 DR. SIESS: Well, I don' t think it's been clearly l 21 questioned. I don' t think the question has ever been clearly ; 1 l 22 stated. Everybody knows a. vulnerability when they see one. l l 23 But which one of these methods would have found the buildings l 24 banging together at Indian Point. 25 MR. KENNEDY: They both-would have. O- Heritage Reporting Corporation (202) 628-4888 1
1 257 l 1- DR. SIESS: They would? They have looked.at that?- I-2 mean even before they found.it. ! 3 MR. KENNEDY: Oh, yes, I think they would have found 1 4 it.
.)
l 5 D R. SIESS: It's a new path somewhere. i 6 MR. GUZY: 'There is a popular feeling I guess amongst i 7 many of the' people in the working group -- many people feel 8 that a .3 Gs margin review for eastern U.S. plants is' adequate 9 to find whatever your unstated reasons for checking. 10 vulnerabilities are. .; 11 DR. SIESS: .3 HCLPF. - 12 MR. GUZY: .3 G. seismic margin review. 13 DR. SIESS: But it doesn' t have to come out .34 it ; () 14 might come out .27 or .25. 15 MR. GUZY: Well, if it came out at . 27, then you 16 would evaluate. But there are a lot of plants that would pass, 17 would have a HCLPF above .3 Gs. And there are a' lot of plants 18 that you may Justify later on'like something in ilorida that j I 19 you didn' t have to evaluate at .3Gs. But there is.a feeling 20 amongst many that you don' t have to do - you don' t have to 21 identify vulnerability. .,
'l '
22 Another point with a margin' review, you don' t . get a 23 ranking of all of them. A ranking in the sense of. 24 probabilist'ic range. You get a ranking in terms of.HCLPF and 25 that may not be a clear indicator of what a probabilistic PRA Heritage Reporting Corporation i (202) 628-4888 l
258. 1 ranking would be. And I th' ink it is difficult to try to turn-a O 2 margin review into the type of ranking that.you would-get out
- 3. of PRA.
I 4 DR. SIESS: Now, what do you mean by ranking? l 5 MR. GUZY: Ranking means.one HCLPF - if the HCLPF is l 6 .25 or .20, what is really the vulnerability, the'HCLPF'is. tied l 7 up with a confidence statement. And it doesn' t give youLthe B same -- you cannot make the correlation between median ! 9 probabilities of failure from a HCLPF. You cannot derive that 10 directly. 1 11 DR. SIESS: No, but why would you want medians? 12 MR. GUZY: Well, the median gives you a. convenient 13 way to rank things as far as what is going to fall first. I () 14 DR. SIESS: What are you comparing? You mean systems 15 in the plant? l i 16 MR. GUZY: Systems -- 17 DR. SIESS: Ranking one plant versus another? i l 18 MR. GUZY: No, I think systems in a plant and j 19 components of systems. 1 20 MR. BUDNITZ - Yes, i t doesn' t -- let me describe. If
]
-)
21 one component has a HCLPF .20 and the other'one has HCLPF .25, ) J 22 you haven' t learned what to fix. Without additional insight, 23 which you need other. things to know, like the systems that it 24 is involved with and you might call it the beta -- the 25 fragility curve and so on. There are other things. 1 Heritage Reporting. Corporation (202) 628-4888 1 w_-___.
1 1
. . .259
]
1 1 DR. SIESS ' I may not want to fix it. 1 2 MR. BUDNITZ No, no. You may -- exactly. Bob ; 1 3 Kennedy.wants to -- 4 MR. KENNEDY: In the seismic PRAs.that'hhve been done 5 for the eastern' plants, the part of that fragility' curve that' 6 dominates'the seismic risk in general.is from about the: 10 7 percent. point on the fragility curve up to maybe the 5. percent. 1 j 8 They may even go down as low:as the 5 percent on the fragility 9 curve, up to the.30 or 40 percent. 'I t is th'at space in the 10 fragility curve that is important for estimating seismic risk. 11 And the HCLPF value lies below or at the lower region of that 12 space. Actually, I.think it could go down as low as to about 13 where the HCLPF is. (} 14 So, for the eastern U.S. plants, if._you really wanted i 15 to know which components dominated ~ risk, the HCLPFfrank'ing g 16 would only give you that approximately. It would give it.to 1 17 you if the ratio between the'HCLPF to the median were 18 essentially constant. ) 1 19 DR. SIESS: But what would the median give me? j 1 20 MR. KENNEDY: It wouldn' t give you anything. ) 21 Frankly, it goes to high because,the' dominant'- the risk , 1 22 estimates are coming from the portion of the curve;betwenn'the .j 23 HCLPF and the median. And, actually, in most cases, clos er to 24 the.HCLPF. 25 Now, when you go to high seismic 1 regions: Diablo l
.O Heritage Reporting Corporation (202) 628-4888 '
;j i
C----_--____ _ _ _ - - _ - _
-p 260 1 Canyon, the two Taiwan PRAs that I am familiar with, the risk i
2 is dominated by the part of the fragility curve not much.above ; 3 the HCLPF. 4 In fact, at Diablo Canyon, it.is the fragility curve 5 portion.right in the vicinity of the HCLPF which is'the. 6 dominant contributors to the risk.because the seismic levels 7 are so much higher. And, so, I think if you look at-other B' PRAs, you can find that the HCLPF-is probably:a pretty good.
? ranker of contribution to risk, but it is not perfect.
10 DR. SIESS: But, Bob, does this mean that if'I.have 11 got a low value and I want to improve it, that raising the 12 median isn' t going to help me that much. I have just got to do 13 something to r2 duce the uncertainties? 14 MR. KENNEDY: I think that the uncertainties and
}
15 fragilities, the uncertainties and randomness and fragilities 16 are driving the seismic risk that are being reported in seismic 17 PRAs, not the medians. 18 DR. SIESS: And.if I just move the whole thing over, 19 I don' t gain that much. 20 MR. KENNEDY: Well, if you move the whole. thing over, 21 you are moving the HCLPF over proportional.to'the median. 22 D R. SIESS: That's right. 23 MR. KENNEDY: So, you do gain. 24 D R. SIESS: That's ri ght. Move the median and leave. 25 it down there. O Heritage Reporting Corporation (202) 628-4888
. _ _ _ . . -- . - _ . -- =_ - _- _ _-.___- _ - _ _ -
261 1 MR. KENNEDY: Yes, we have done. sensitivity studies 0 2 on that. If you hold the HCLPF constant -- if you hold the 3 HCLPF constant and adjust the median, it isn' t going to change 4 your risk estimates very much, at least in the sensitivity j 5 studies that I have had done on a couple of cases. I 6 But if you just move the whole curve over, that does 7 make a difference. 8 DR. SIESS: Well, you.said you think it is a good l 9 tool for handling the Charleston type things where we are l l 10 talking about increases and design basis earthquakes on the 11 order of what? Less than 2. 12 MR. KENNEDY: Right. 13 DR. SIESS: I think the ones that have been done give () 14 a good strong feeling that there is not a cliff out there Just 15 10 percent beyond the SSE. At least there is not a cliff out l 16 there after they have been done. What I am not sure is whether l l 17 there might have been a. cliff out there before they were done. 18 That a what ! was getting at earlier on the Maine Yankee 19 fixes. 20 I think for that particular bunch, the batteries 21 certainly were zero. They were Just highly uncertain. 1 22 (Continued on the next page.) 23 24 25 Heritage Reporting Corporation (202) 628-4888 1 1
262
.. . . a t16 hola' 1 DR. SIESS: At less than one in a million -J
() 2- probability or something. So.I think it's a : good device for u
;j 1
1 3 showing the robustness.of our plants for earthquake. Once we 4 catch these things that didn' t get looked at,. which is. fewer 1 J 5 and fewer. Maine Yankee would have.more than Catawba, as you-6 well pointed out. By the time you've designed 1your third 7 plant, you' re looking for things. 8 What it'll do for the IPE's. I don' t know. That's-l 9 something I think the staff has got to work out. We don' t know. ] l 10 what to do for the IPE's on .the other parts of the plant' yet. J l 11 Still being kicked around. -I 12 MR. KENNEDY: I question the desirability of doing of L 13 forcing a lot more seismic PRAs that lead to such-uncertain end- j 14 results. Uncertainties in farmout frequencies. If it's the
}
15 end results that you' re going to pay attention to. Now,'if 16 you' re doing the. seismic pRAs for the purpose of; learning what 17
~
you' re learning as you go along, then they are very valuable. 18 DR. SIESS: That's obvious to everybody. Everybody 19 will agree with you except some people up on the next floor-20 that have promulgated a safety goal which'is expressed in 21 probablistic terms and can only be tested against the bottom 22 line of a PRA. 23 Now, what do we do about that? j 7 24 MR. KENNEDY: I can change.the seismic risk' numbers? , I ! l 25 in terms of mean risk on a plant by_-making changes in my l Heritage Reporting Corporation (202) 628-4888 _-- . _ _ _ ____:-___-.-_. __ - - - - - . = - . _ _ . _ _ , _ = _ _ . _ _ _ . _ - ..-_-_-_..-___.__.- .__.._._._ -._ _= = .__._~_ - - _ --- - _ m _ _
i 263 l' calculations, changes in the system's calculations, changes in ] 2 the hazard curves, ' spread these changes out . among enough things
\
3 . . - l 4 - DR. SIESS: That nobody's going to know it. 5 MR. KENNEDY: And I can shi f t those numbers a fac' tor - ) 6 of at least five, and probably ten. So if I ' m t oo - '
'- DR. SIESS: You don' t think Brookhaven could find.it
- when:they reviewed it for the Staff or Los Alamos when they
'9 reviewed it after Brookhaven?
10 MR. KENNEDY: Well, t hey' l l raise. lots (f questions 11 but nobody will know who is right. 12 DR. SIESS: Oh, we l l , . t h at ' s -- a 13 MR. KENNEDY: And so, you know, if we get a1 factor of () 14 two below the safety goal somebody else could easily get'a~ i 15 factor of two above it. Or if I get two above it, somebody ) 1 16 else could get easily a factor of two below it. And we' ll be 17 in never ending arguments. ) 18 DR. SIESS: Well, of course, you know, a'PRA cannot 19 be used to prove that you meet a safety l goal, because you can 20 always be challenged on the basis of completeness. You know 21 that. 22 It can be used to show that you-h' ave not met 'i t . I 23 may give you all the wrong reasons and you may1 fix the wrong 24 things. If the bottom line is used. I don' t know anybody that E5 wants to use the bottom line of. a pRA. And everybody that's - i O Heritage Reporting' Corporation ! (202) 628-4880 ., I i
264 1 ever.done one has thought th'ey learned a heck of a ' lot about O 2 their plant from it, and it was worth it, you know. 3 If I owned a plant, that's what I' d dc. If I could 4 find somebody to do it. I' ve seriously considered asking ACRS' 5 to think about what would be the implications of just~trying to 6 say everybody should do a PRA. Is it possible, a hundred and 7 some odd plants, there won' t be quite that many because there 8 are some good duplicates out there to do a PRA? Are there 9 enough PRA people out there to do it? ! 10 How many PRAs have been done? 11 MR. KENNEDY: Seismic? i' 12 DR. SIESS: No, let's just start without seismic. 1 1 13 MR. KENNEDY: Oh, about 35. j 14 DR. SIESS: About 20 seismic? { 15 MR. MOORE: At least half of those PRAs are pretty 16 old. You couldn' t take them too seriously. 17 DR. SIESS: That ' s all right. They may be the best 18 ones. 19 M R. KENNEDY: There must be at least 30 seismic PRAs a 20 because I' ve worked on 23. 1 21 D. SIESS: Now, are the seismic PRAs less complete 22 than the others, or are they? 23 MR. KENNEDY: Scle are old. 24 DR. SIESS: When you say seismic PRA, you.mean total. 25 includinD seismie, or just seismic? () Heritage Reporting. Corporation (202) 628-4888 ) 8
s 265' 1 MR. KENNEDY: I believe every seismic PRA has been O 2 part of an internal PRA. I don' t think anybody's done 3 exclusively a seismic PRA. I' m not even sure I know how you
'/
4 could. 5 DR. SIESSL Given enough incentive. 6 MR. KENNEDY: But.the walk downs in the plant in the 7 seismic PRAs, the vast majority of them'have not been to the 8 same level of detail as the walk downs in the Margins Program, l I 9 DR. SIESS: And of course they' d have to be done for ' 10 more systems, too. 1 11 MR. KENNEDY: Because you' re havinD to do them for i 1 12 more systems, you' re having -- in the fragility, aspects, the 13 seismic PRAs, you end up doing a lot more quantity of work and
]
,( ) 14 the quality is not as good in the Margins Review aspects as in 15 the Margins Review Program, because you' ve narrowed the i 16 quantity down.
17 You could do the Margins Program effort in a seismic l 18 PRA, but then you' d greatly raise the cost of the seismic PRA. 19 Probably the Diablo Canyon seismic PRA is very.close, probably i 1 20 has got at least this level of quality, but Diablo Canyon - 21 seismic PRA is far more expensive than any previous one that 22 I' ve been on. 23 DR. SIESS: It's not even level 3 is it? 24 MR. KENNEDY: It's not even what? ) 25 DR. SIESS: It's level 1 i sn' t it? l i s- Heritage Reporting Corporation. (202) 628-4888 E________________________________________._______--- _ _ _ _ _ _ _ L _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ _ _ .
;l
'l) 266
~
1: MR. KENNEDY: It's only level one but it's a lot more- _O- 2 expensive than'any of the others because it's . got a lot of l 3 seismic fragility work in'it and a~1ot of hazard work in'it. 1
~ 1 4 M R. MOORE: A lot of cir'_uitry and electrical stuff il 5 too. A l
6 MR. KENNEDY: Yes, a11ot'of circuitry, too. 1 Well, look, gentlemen, it's gotten . l at e,- 7 DR. SIESS: 8 and Don, we didn' t let-you finish up and was somebody going to l. 9 talk about the -- Frank here? 1 10 MR. COFFMAN: You want me to do that now, or?. .
)
11 DR. SIESS: No, I' m Just' trying to think. I' ve read 12 through the rest of Don's presentation, and I don' t have any 13 questions I would ask on it if you presented it. () 14 MR. MOORE: Okay. 15 DR. SIESS: Because of this hour, although I think 16 it's something we need to hear about more eventually. H 17 But Frank, where do you fit into this thing? Are you l 18 on this work task group?- 19 MR. COFFMAN: One of our branch members is on one of 20 these panels. I got lost with all the panels. 21 DR. SIESS: Yeah, what branch are you on? l 22 MR. COFFMAN: Reliability and human ~ factors.which has 23 the people responsible for the IPEs were shifted over into that' 24 branch as it's carried on. 25 DR. SIESS: Okay, so your branch now irr NRR has got Heritage Reporting Corporation (202) 628-4888
\
._.._.m .__________._______,_____a
a-267 l
. 1- the Job of deciding what'to do on the IPEs?
, ) 2 MR. COFFMAN: Within'research,-yes. We' re in within 3 research. 4 DR. SIESS: You' re in research, y 5 MR. - COFFMAN: We have the lead. We' re tasked with i 6 that. 7 DR. SIESS: That's right, but that's in research, , l 8 now. 9 M R. COFFMAN: That's right. 10 DR. SIESS: I keep getting confused. 11 MR. CUMMINGS: You' ve got to remember, research 12 doesn' t only do research, hasn' t for years. 13 DR. SIESS: Well, doing less reports than you used () 14 15 t o. MR. CUMMINGS: That's what I said. 16 DR. SIESS: I' m sorry, I forget. That's under team 17 assistants. 18 MR. COFFMAN: We' re tasked with that, but before it 19 can go out, it will have to be concurred in by NRR, and there j l 20 is a package that NRR has right now. l 21 DR. SIESS: Where in NRR are they likely to look at l 22 that? 23 MR. COFFMAN: Well, first and foremost, Tom Burley, 24 the Director, has a package -- 25 DR. SIESS: He's got to ask.somebody. ( Heritage Reporting. Corporation (202) 628-4888 t
268 i M R .' COFFMAN: Yes, sir. But in addition, there are ! O 2 two other groups, one headed by Ship Fiddonne, and the other 3 Frank Consenserve and Rich Barrett, in the risk assessment 4 branch. 5 DR. SIESS: Frank wrote'the letter, he signed the i 6 letter that Leon wrote suggesting how to use.the seismic ! i' 7 margins in the IPE. i 8 MR. COFFMAN: Which, 'I got the people mixed up. 9 DR. SIESS: Seismic Issue The Severe Accident Policy 10 Implementation. This is a member from Conjoe to Zultan. 11 MR. COFFMAN: Yes, sir. 12 DR. SIESS: Which I assume Leon was on the -- 13 MR. COFFMAN: Yes. 14 DR. SIESS: And is that something .that's gone up now? [ 15 MR. COFFMAN: Well, it's something that we' ve 16 considered. It's reasonably old. It's something we' ve 17 considered. We sti.11 have questions that we' re addressing that 18 I don' t think that memo provides the answers to, i 19 DR. SIESS: On seismic? 20 MR. COFFMAN: Yes, sir, and it's at the front end- of i 21 determining in fact wh'at is the earthquake level that we should 1 22 select and how do you assure uniformity in two aspects: one is, 23 uniformity acrosr the plants because some plants may in fact 24 not need the margin that a .3G would provide them, and others 1 25 would need it. So across the plants, there's a uniformity -) i l o ( Heritage Reporting ~ Corporation (202) 628-4888' i
269
, 1 question. !
b, 3 2 DR. SIESS: Why would they need it? On what basis 3 would you say a plant would or would not need it? l 4 MR. COFFMAN: If .3G were adequate for Maine Yankee, 5 it may provide more than a necessary margin for Turkey Point. 6 DR. SIESS: In terms of a uniform hazard type thing? 7 MR. COFFMAN: Right, yes, sir, across the plants. 8 But then even across hazards, there's that consistency of l 9 uniformity consideration too. 10 DR. SIESS: I know but we can worry about being fair. i 11 and not get anything done. We might get the system so 12 complicated trying to be uniform that we can' t do it for 13 anybody. And is anybody looking at that? () 14 MR. COFFMAN: Well, we' re certainly not lookinD for 15 perfection and total fairness, but then again, the plants have 16 to make this more than just an empty exercise for the 17 regulators. 18 DR. SIESS: You go down to Turkey Point, what was its 19 design? It couldn' t have been lower than a tenth, I' ll 20 guarantee you. 21 MR. COFFMAN: I t 's .1G. 22 DR. SIESS: .1 G. You think you' ll have a tougher job 23 -- of course, Maine Yankee had already done some upgrading, e 24 But what could happen down at Turkey Point? You' 11 just have a 25 lot more hit lists to evaluate. You couldn' t screen out as Heritage Reporting Corporation (202) 628-4888 L___ _ _ _ _j
- ., 9 1
y 270 j i i much?
-( ) 2 MR. COFFMAN: Yes, ' that may. be true while the real:
i 1 1 3 hazard may well be high winds, because -- 4 DR. GIESS: Oh, well, I know the real hazard may be 5 high wind or it may be flooding or something else, but. We' re i 6 talking about the seismic part. Now, you' re also saying you j 7 don' t want to grade it higher. But you wouldn' t have to meet, l 8 it wouldn' t have to meet three-tenths. You' d just take three-9 tenths as screening because that's one level you' ve got. in your 1 10 guidelines. .I 11 And you' d st ill throw out a whole lot of components ' 12 at three-tenths. You' d probably have more left to evaluate 13 than you did at Maine Yankee because they' d already done. some 14 upgrading. But i f .you' re doing generic screening, Turkey Point 15 used the same kind of motors that somebody.else used. And then j 16 if they come up with a HCLPF of two-tenths, that might be quite 17 adequate. You don' t have to come out with three-tenths., I i 18 Three-tenths just says where you' re going to screen to. 19 MR. CHEN: That's exactly what I'was thinking, is how- i l 20 are we Doing to make a decision what level of earthquake will 21 be acceptable. 22 D R. SIESS: Thing ' is if you don' t make it three-I 23 tenths, you can' t use it, can you? How do.you do it for two-24 tenths? I i 25 MR. KENNEDY: Well, you' d use the . same screening .I l l
) Heritage Reporting Corporation (202) 628-4888 i
i u_____________._____.__________.__ _ .__. _ _ . _ . _ . _ __ _. . _ _ _ _ _ . . _ . . . _ _ . _ __ . _ _ _ _ _ _ _ _._...____..___J
271 1 tables at two-tenths but.I guess your walk 1down team would 2 throw a few more' things out by judgement, but you' d use the same 3 screening tables, because the' earthquake experience: data base 4 simply i sn' t that refined. We didn' t- have instrumentation out 5 where most of this data base came from. 6 DR. SIESS: It-seems to me, deciding in advance 7 what's acceptable 'is very dif ficult. Going ahead and doing the 8 thing on a practical basis an'd deciding whether the result is l 9 acceptable in comparison with the.other hazards would be a much 10 more logical way to do it. Now, it may not work in the 11 Regulatory system, I don' t know. 12 MR. COFFMAN: Yeah, I think thet's our problem is l 13 trying to motivate this because we' re not, we don't have force 14 of a regulation behind us Other than 20 CFR 5054F which says r( ) 15 that we know enough to be able to go to the plants and.ask them. I 16 to give us more information. So we know enough about seismic 17 risks that we can go to the plants and ask them. 18 DR. SIESS: The Severe Accident Policy Statement 19 doesn' t give you the authority to go out and do IPE's? I 20 MR. COFFMAN: That is our only authority, but it is J 21 less than a. rule, it has.less leverage than a rule. 22 DR. SIESS: I never knew that NRR had much trouble -]
~
23 with leverage. The next time they came.-in for..a reload, they 24 Just got held up on their reload core review until they did' 25 what they wanted. I don' t mean to put "words' in . somebody's ; i O'- Heritage Reporting Corporation (202) 628-4888
272 1 give them ideas. O 2 MR. WYLIE: Never happen. 3 Will.you guys becoming into ACRS with this as a part 4 of the IPE implementation any time in the'near future? 5 MR. COFFMAN: The schedule called for the IDE to 6 address external events about a year behind the internal 7 events. 8 DR. SIESS: Part of the rationale for that was they 1 9 didn' t know how to do it. 10 MR. COFFMAN: That's right. 11 DR. SIESS: And right now I' d say you know about as 12 much about how to do it for external events, if you' accept 13 seismic margins as an approach, as you do for internal events. 14 . M R. COFFMAN: Well, there are other external hazards 15 besides seismic, and I' m not sure that they' re as f ar along in 16 method development short of a full blown PRA, which is an 17 option. 18 DR. SIESS: I thought fire was.the only one that the 19 SECY said was at all important besides. The SECY to the 20 Commission on -- 21 MR. COFFMAN: On External-Events? 22 DR. SIESS: On External Events, 86- 162, : concl udes 23 that -- external events, is thiszthe one I read. 24 MR. CHEN: That SECY 86-162 did not conclude what' 25 external events are -- /~ (_h) Heritage Reporting Corporation (202) 628-4888
l i 273-1 DR. SIESS: Well, another one or something I read
) 2 somewhere. )
l 3 MR. CHEN: Yes. SECY 162 is studying -- 4 DR. SIESS: That only fire and earthquakes are. )
\
5 MR. CHEN: We are trying to address in a two-phase' 1 6- . program. In the phase one program, we are trying to identify. j l 7 what external events we want to look into. And Lawrence 8 Livermore has a contract to identify those external' initiators-and determine whether some of them are important enough to be I 9 i 10 looked at. 11 DR. SIESS: Okay. But right.now, are you thinking 12 that for earthquakes, this is the way to.go, or? Anybody that 13 wants to do a full scale'PRA can do it? I 14 MR. COFFMAN: That's right. That option is open to
}
15 them in the'IPE, open to the utilities, the IPEs, and we would-16 like to encourage abbreviated methods. And we' re looking, 17- monitoring this and the other seismic margins program and all ; 18 these related activitien that I think most of them were-touched 19 on today. 20 But we don' t think we' re at the point 1of being able I 21 to say that that method now exists in a form.that_can be used 22 to identify the vulnerabilities which are the objective of the 23 severe accident policy.
- 24. DR. SIESS: Have those vulnerabilities been defined 25 in any documents you could refer me to?
( Heritage Reporting Corporation (202) G28-4888-I o
)
274 1
'1 MR. COFFMAN: New Reg. CR 4920' lists the screening )
) 2 criteria we used to establish what we thought were -- we didn' t l
3 .use the word, vulnerabilities:-- we said these are areas of !
-4 potential--improvements at the plant.
5 DR. SIESS: To reduce core melt,- for core melt? 6 MR. COFFMAN: For nboth core' damage and unusually' poor ) l 7 containment performance, which is a bureaucratic way of.saying l
- 8. level two.
1 9 DR. SIESS: Do you have a report? 10 MR. COFFMAN: No, this is a Brookhaven. 11 DR. SIESS: Oh, that's the Brookhaven. 12 MR. COFFMAN: There are five volumes, and they list 13 them in there.
/ 14 DR. SIESS: That ' s the one that covers each of the 15 five 1150 plants?
16 MR. COFFMAN: Yes, sir. .But those screening criteria 17 that we use in there are much finer than what one would want to l i 18 say -- l 19 DR. SIESS: But those don' t have anything about 20 seismic do they? di MR. COFFMAN: No. Th at ' s a l .' -- i t doesn' t exclude' q 22 or include seismic. ;
.)
23 DR. SIESS: But it's based on the' New Reg 1150 24 studies and those five plants don' t - well, they don' t even- j 25 represent anything. They didn' t even represent the different' '
) Heritage Reporting Corporation (202) 628-4088 f
l_____________.-__. - - - - - _ - _ _ .__c_ _ _ - _ _ - . - _
I l 275 ; I containments that well. But they certainly don' t represent 2 seismic problems very well. I 3 MR. COFFMAN: Well, we not only looked at the 1150 4 plants, we looked at the available PRAs, the PRAs available to 5 the Staff, plus the in-core reference books. 6 DR. SIESS: But the five reports you mentioned, 7 weren' t they divided up the same way at the time? i 8 MR. COFFMAN: Yes, sir, they were. i l-9 DR. SIESS: PWR Mark One's? 'I l 10 MR. COFFMAN: Mark Ones, Mark Twos, Mark Threes, the 11 ice condensers and all others, the PWRs large drying i' 12 atmosphere. 13 D R. SIESS: So right now, you won' t be thinking () 14 seismic for another year or so? j 15 MR. COFFMAN: We are actively thinking seismic and if 16 in fact we could, the external event IPEs'have not been delayed 17 to the degree that the internal has been on its schedule. So 18 the gap between them is closing, and if in fact,.it totally 19 closes, we' d like to go out with one IPE letter that addresses 20 both internal and external. So we' re very act ively working on 21 external. 22 DR. SIESS: With the emphasis I' ve heard today on the 23 walk down, and with the understanding that even for internal 24 events, you have'to do walk downs, it seems to me there' d be a I 25 tremendous advantage in not having two separate walk downs. Heritage Reporting Corporation (202) 628-4888
276 1~ MR.'COFFMAN: Yes, sir. O 2 DR. SIESS: Two separate walk downs, and I think we 3 had that same problem in connection with A-46 and what was the' 4 other one, Charlie? 5 MR. WYLIE: We discussed' interactions. 6 DR. SIESS: Yeah. 7 MR. NYLIE: So they said, well, we' l l come back and 8 do it later.
.4 9 DR. SIESS: We just keep seeing walk downs for this I
]
10 and. walk downs for that, and I don' t know, maybe it's good to 11 have a lot of Walk' downs, a lot of'different people, but I 12 think a couple of real good systematic ones would be better. 1 13 MR. COFFMAN: But even if you do.Just one walk down, I {} 14 you would have to look at the different hazards even on that 15 one walk down. So a walk down is probably a gross 16 understatement of the work involved in analyzing the plant as l I 1' 17 is. There's a lot of work that goes into a walk down. 18 DR. SIESS: Now, wnat you heard today'about the 19 . relative merits'of the two systems, the two procedures that: I 20 EpRI tried, do you see any preference in terms of using them l 1
'l 21 for IpE7 l
l 22 MR. COFFMAN: I' d like to qualify my response here. 1 23 I' d like to give you more of an-individual response than what I ; 24 think represents an organizational response. 25 The problem that I would see as an individual on.a-( . Heritage Reporting Corporation (202) 628-4888 _____m .__m._ ____m.m-____._.__________-w
l
'S
.1 277 1 success' path approach is_you touched on it, the fact that as o
1 2 long as that success path.is there, and you can assure it with I 3 a high degree of confidence you' re all right toward that first 1 1 4 objective, but for the objective of finding vulnerabilities, 5 you would have to be assured that.that success path represented 1 6 over something like 95 percent cf the way that the plant could ) 7 get into trouble. y 8 My English didn' t come out very clearly but - , , e, ,
, j N6W, theoreview group. recommended 9 DR. SIESS:
e ~ 10 strongly that there be two success paths. t 3 rg 11 MR. COFFMAN: However many success paths you have, in t{ 12 aggregation, they would still have to cover at l eas.s 95 percent 13 of the ways that the plant could get into' trouble in order'for 14 them to meet our screening criteria. l 15 Let me put it in terms of the way the, screening 16 criteria read now. They read in falseface, and they say'that I 17 you can any sequence any functional sequence that contributes 18 less than five percent to the total core damage frequency can )
, i 19 be ignored on that merit itself,'although the screening !
20 criteria are banded together, -they' re aggregate. That was for 21 the purpose of making sure that if we' ve got 16 sequences at 6 22 percent, that they' re covered. Or if we' ve got -- 23 DR. SIESS: I don' t .know how that fits in with the 24 .HCLPF idea at all. Because all you' d get' 'out of this there are 25 two paths that given an earthquakW, yol could take, given an i Heritage Reporting Corporation (202) 620-4888 i I l' i
\
t' - - - -- a
p; , y - - - - - n
; /l 278 n
~1 earthquake up' tw this level which. now has sorne' probability.
' O. 2
- q. f .
There are two paphy you can take with D'$ percent confidence andl j 3 probability thM they' ll work. f ' There's two paths.. Now,.the
.]
4 failure probability .is . 05 t irnes . 05, I' guess, which is .025, i
! 5 arn I right. This gives rne 99.75 success with two paths, ;
I.
,i-y .J4
.q .
\- 6 assurning. that you can use both paths. ]
( 'l 7 Then I don' t know how you-' fit that"into the other 1 8 contributors without getting into the probabilities. > i
~
i 9 MR. COFFMANr You don' t. That's our problern. 'We l y 1 10 think that the coupling that provideo the consistene;/ across .] 1 11' the; hazards is.throu1h-the probabilities. , i ,> i , 12 'c D R. SIESS: Then<fou have to do a'pRA. g
) , .
/ , )
We - 13 3 ,'1 - . MR. COFFMAN: Phat's ~ one . way of doing it. a' ./ .
. 14 t DR. SIESS: Ip there another way.-
l 15 , [ MR. COFFMAN s We' re looking at ; a. uacond - - J I ifi DR. SIESS: I know that's one way.
.i 17' MR. COFFMANt' There's. no e><ist ing other , way.
18 MR. CHEN:t There's a possible othdr ut i l i z at i on, - ; 19 7 ! because seistnic has a catali::atoin prograiu because that prograrn e i l 20 essentially would : define also the seistnic hazards for the .j c21 plants. / 22 DR.. SIESS: What, . 'you going to cornbine the hazards ' .l :
~- 23 with the fragilities?
24 MR. CHEN: Cornbine the hazards with the fragilities. . -l 25 DR. SIESS: I don' t think that' ll work,, will it? Heritage Reporting Corporation. 1 (202) 628-4888 / , J
,. R-l l
- - - _ _ - . . . . _ . ~ . - - . - -
2-1 279 1- MR. KENNEDY:- Well, there's an approximation that + <O 2 seems to bn. coming out of the existing seismic PRAs. I' m not' ,
-r 3 100~ percent comfortable with it, yet, but'it's werked on all of 1
4 the existing seismic PRAs. That is that the if mean annual , 5 probability of core damage is in every case if rny memory surves 6 me ri Dht, I have those numbers written down and some ,j r V ., 7 recommendations I wrote for IDCOR, bat my memory serves tre, t hje g-8 mean annual probability of seismic induced core damage in each. 9 case is a factor of 20 to 40 less than the mean annual 10 probability that the seismic hazard exceeds the HCLPF. .; f il So if:.you had a HCLPF for a plant you enter into a 12 seismic gazard curve, read offna"mean annual probability of 13 exceeding that value from t*pi hakard curve, and divided that by
, o
(} 14 a factor of 20 to 40, it's an approximation of the mean annual 15 risk. But all you' re going to learn from doing.that is an 1 16 approximation'of the mean annual risk. You' re not : going to .get 17 any of the benefits out of having done a seismic PRA.
\
) !
i 18 And I guess my feeling is if you want. risk numbers, 19 you should go to the PRA. 20 DR. SIESS: And if you define vulnerability as things 21 that contribute to' risk, you' ve got to got to a PRA. ! 22 , 3 MR. KENNEDY: Unleus you can. decide that the things 23 that contribute to the,hisk are the things witt7 the low HCLPFs. 24 Now, dhat tends to be the. case. But'itsagain/not automatic 25 that it should be. t' e 1 x Heritsge . Reporting Corp 2 ration: . (202) G28-48C8 U o l i
? ]'
g Ay
1 1 o 280-
-) t
^
1 DR. SIESS: It's going to be .very dif ficult~ to form a 2 lot of this. 3 I'd like to suggest.that we close this' meeting with f 4 .an expression of thanks to all the people that have.made ; L presentations. You' ve certainly done a fine Job'of educating 1 l- G me. I know a lot more about this than I did to begin with. I l 7 Unfortunately, I have.a lot more questions about it, than I had l F 8 to begin with. 9 And I . don' t believe I'm ready to take this to the 10 full ACRS. I just think it would get them more confused than l 'l 11 me, until we' ve' had further discussions, either.in this 1 L . I s 12 ~ subcommittee or another subcommittee. I'm going to suggest 1 13 that the Severe Accident. Subcommittee ought to look at it, 1 14 whenever you have something to talk about. It fits'into the
} ]
15 whole picture there, and we won' t be looking at an isolated 16 part. 17 I' don' t think the staff wants any advice from the 18 ACRS particularly on the Charleston issue until it becomes an 19 issue. And it's a potential, but you didn' t ask us about Maine 20 Yankee, We didn' t see the safety evaluation report -- we were 21 never asked to comment on it, were.we? We didn' t review the 22 Maine Yankee. That was a licensing action. 23 And I' m. sat isfied. - I don' t know 'about anybody else, 24 but --
- 25. So I think we' ll Just let this stew for awhile unless Heritage Reporting Corporation (202) 628-4888
281 1 the staff.has'something they specifically want advice,on.- But' O 2 I don' t think the ACRS is the person to ask now which method do 3 you use. And if the staff gets to the point of wanting to how-4 do we get everybody to do this, this is a good idea. You know, 5 I think'that would be a good subject to come in and talk about 6 some time. What should everybody be doing, -if anything, or not' l H 7 everybody, should there be'some' people.more concerned-than i 8 others. l
- l. 9 Well, Charleston makes some people more concerned 10 than others, but I' m not so sure that some of the other plants 11 were that much different either.
12 And again, from the licensing point of view, 'an i 13 Appendix 8 point of view, the Charleston problem I' d say is (} 14 potentially a real one. So far, nobody's , challenged water at. , 15 Turkey point and some.of these plants and. places where they 16 never heard of earthquakes. 17 DR. GUZY: There? s one point I' d like to make about - 18 - I think it's been talked about but not really addressed 19 thoroughly, is the 846 is a lot realer than the Severe Accident 20 and there people who are going to start evaluating plants for 21 that. In the Hatch One evaluation,'essentia11y'their 22
~
incentives if- they go ahead and do this would be to be able to
]
22 double up their'L26 review with a margin review. 24 I think the staff in general encourages that. I 25 mean, that's something that's going to happen. This is a way O(/ Heritage Reporting Corporation. I (202) 628-4888 L.__-____._____ . _ _ _ _
282 1 to= conserve their resources and our resources in a review. O. 2 DR. SIESS: That'makes sense. 3 DR. GUZY: And I think there's a motivation to do 4 something in conjunction with 846. jl think-I. heard you say 5 that you encouraged that. 6 DR. SIESS: I think if you want to ask basic RS, 7 you' d get some questions, but I think in the long run people 8 would agree. I don' t think anybody's happy about doing these 9 things over and over and over'again. 10 MR. WYLIE: Well, that was part of our letter we 11 wrote on, I think we said it in our letter that we wrote on. 12 DR. SIESS: Not in relation to HCLPF. 13 MR. WYLIE: No, it was in relation to the () 14 interactions review. That it d idn' t make sense to go through 15 the 846 review and not.look at the other things while you were 16 going through doing the review. We said that'in our letter. 17 DR. SIESS: Well, you see, you' re working on a 18 premise that is logical. I' m not sure I' ve bought it 19 completely, that there's some good things to come out of the 20 seismic margin review and you should get as many people to do 21 it as possible. 22 MR. KENNEDY: personally, I think we should encourage 23 anybody that wants to do it to do it. 24 DR. SIESS: Okay. That - I' d certainly agree. Anybody 25 wants to do one, I wouldn' t discourage them. But I got t h'e O- Heritage Reporting- Corporation (202) 628-4888
283
. 1 impression you were trying to encourage people to do them. l 2 D R. GUZY: personally, I.look at.our first objective 1 3 saying it's whatever comes out of Charleston and any plant 1
4 specific case where they have to raise their review level. 5 DR. SIESS: And you think its.a good tool for that?f i 6 DR. GUZY: I think it's a good tool. I think it's 7 available. I think it's been used. 8 DR. SIESS: I think it's a better tool'than.any 9 alternative I can think of that, including go back and i 10 reanalyze it. i
'l 11 Now, that's something I think I' ll report to the. full l 12 committee. This is one of the expected uses of it, much the 13 same way that it was used for Maine Yankee. And'if anybody has !
'.i
() 14 any problems with it, they can speak now or forever. hold their l 15 peace. 16 And if they want to hear more' about it, I' ll try to: 17' arrange it again. Usually we use a subcommittee meeting as a 18 basis for deciding what we could present'to the full committee i 19 that they could understand and get involved i n. And frankly, j 20 coming out of this one, I' m afraid it isn' t easy. There's some 21 .very basic philosophic approaches to doing this, and it's going j 22 to get so tied up with'other approaches people have in the back 23 of-their mind that to get them thinking this way in two or-24 three hours isn' t going to be easy. 25 MR. BUDNICK: I have a suggestion about something ) i ( J O-. Heritage Reporting Corporation i (202) 628-4888 i 4 ___..-_____.m_ _ _ _ _ _ . _ _ . _ _ - . _ _ ._ b
284 I that the ACRS couldodo. It's not very different' from what.I O 2 said a half hour ago, but I' ve formulated it a little better. 3 I think that the ACRS could do a service to everybody
- 4. by convening a meeting in which the ' subject was what are all-5 the different things we ought to do at.these plan'ts for seismic 6 review. And how can they be best accomplished and coordinated.
7 DR. SIESS: And not liinited to seismic. 8 MR. BUDNICK: No, no, seismic, seismic. ' 9 DR. SIESS: Just seismic. 10 MR. BUDNICK: And the reason for that is that without 11 some thinking on the part of somebody that's doing some 12 integration, we' re liable to have,846 ahead of the others,. and 13 I see that as not necessarily a good use of all the seismic (} 14 resources in the country. And they' ll come back a little later 15 and do IPEs, and maybe margins. And it might be a way for the 16 ACRS to focus the staff's attention more broadly than it has 17 been focused. 18 There are just other things going on in other 19 branches in both research and in NRR. I 20 DR. SIESS: I thi nk that's an excellent suggestion, 21 and I' ll try to follow up on it. 22 MR. BUDNICK: Now, of course, that may provide some ! i 23 focus for what's going on about these other things, but'-- 24 D R. SIESS: It mi Dht. 1 25 MR. BUDNICK: -- I' m Just thinking about seismic now. Heritage Reporting.. Corporation (202).628-4888'
,l i
_______.__m______i__E_____
i
.l i
265 l i 1 DR. SIESS: You' re ri ght. ' I' m not sure everybody on l
) l 2 the staff knows what everybody else is doing. 'We sit up.here l 3 frequently and see that.
4 MR. 'CHEN: What we are trying,' in Severe Accident 5 Policy, is we are trying to see how we can integrate all those 6 ' seismic systems together and try'to minimize the duplication, j 7 including 846, 817, 845, and the characterization program, all l 8 those things altogether. i 9 MR. COFFMAN: And 845, right? ] f 10 MR. CHEN: By the way, the.846 has been delayed. 11 DR. SIESS: Now, . you say you' re trying to look at 1 12 this. Do you have the authority to do it? 13 MR. CHEN: Well, that's the reason we cannot tell'you () 14 all the things going on, because at this moment, we'did not 15 work how to integrate these together, and there's so many 5 16 different branches. 1 17 MR. COFFMAN: I : don' t know that it's a question of i 18 authority because the office is aware of all these programs and , 19 is on us to make sure that they are integrated. l 1 20 DR. SIESS: Our offices are looking to you on l 21 integration? 1 22 MR. COFFMAN: Yes, both NRR and ARCS. s 23' DR. SIESS: These things get implemented by NRR, 24 usually. 25 MR. COFFMAN: Well, yeah, but we' re preparing the O- Heritage Reporting Corporation-(202) 628-4888
286
,_s 1 program which we' d like to have as complete as possible. All ;
() 2 they have to do is say, yes. But this isn' t the only area, you l j l 3 know. There are USIs that are a factor too, and it may well ) 4 be, I would be cautious about proceeding, I mean there are the 5 other USIs. I would be cautious about proceeding on one USI, 6 like 846, at the risk of having to go back and redo it. l 7 DR. SIESS: What we really need then is an integrated j l 8 safety assessment program with a PRA for each plant, that they 9 could look and see what needs to be done to bring their plant 1G to an acceptable level of safety and how these various 11 requirements can be integrated as one repleces the other. s . j 12 Wouldn' t you say that? I 13 MR. COFFMAN: Yes, sir, and in fact, we are working l l 14 on that, j
)
15 DR. SIESS: You working on ISAP. 1 hat's what I just 16 said. 17 MR. COFFMAN: Wall, ISP is limited to current 18 requirements. We are beyond current requirements. 19 DR. SIESS: But ISP working on current requirements 20 is just a little bitty step to apply it to the next 21 requirement. 22 MR. COFFMAN: The real benefit in ISP is the PRA 23 methodology, not the ISP program itself. , 24 DR. SIESS: You' ll benefit if ISP is the integration l 25 if PRA is the only means of getting a perspective in Heritage Reporting Corporation (202) 628-4688 l 1 .
287 L 1 integration, .then, yes, PRA. I think there's 'some aspects 1of l . g\_/ ^ 2 integration that wouldn' t require a PRA, but there happens-to-3 be a key to that. l 4 MR. WYLIE: Are you talking about. integrating ~845,, 5 846, 817, all of those? 6 MR. COFFMAN: 817. In addition to steam generator-7 tube rupture concerns. But for the objective integrating them, 8 not for their own obJ ectives, but for the objective as a mere 9 action of policy. 10 D R. SIESS: We ought to have it done tomorrow. 11 MR. COFFMAN: It's what ? 12 DR. SIESS: We ought to have it'done tomorrow if we 13 talk about generic issues, because the complaint has been one 14 hand has been compartmental::ation of generic issues to'make 15 them small enough to handle and solve. 16 (Continued on the next page.) I I 17 l 18 19 l 20 t 21 { 22 l 23 24
.I 25 O Heritage Reporting Corporation.
(202) 628-4888 m ___--__ _ _ __-_____2_ _ _ _ _ . _ _ _ _
288 T1 1 DR. SIESS: On.the~other hand, .if you integrate them ) U 2 all,-they all reach various issues and anti-core melt, a 3 nuclear power plant safety. If you put them all under one 1 4 heading, because the objective of all the USIs is to prove l l 5 safety, and we take it -- they came up in different time scales 1 6 that haven been broken up to make them manageable, and then we 7 end up with this question you'aust brought up about how do you l 8 now apply some integration to this. 1 9 Maybe if we did 'four of them, the fif th one wouldn' t l l 10 add anything. ' 11 MR.'WYLIE: This is A-447 12 MR. CHEN: Yes, A-44 is in there. 13 MR. WYLIE: So you are going to look at all them and () 14 see about the integration of all of them. 15 MR. CHEN: Yes. And there are'real problems in each-16 one of these. 17 DR. SIESS: I know, but.if you had done that to begin 18 with, you see. 19 MR. BUDNICK: Yeah, I know. I could come up with a 20 good thing. I don' t know. 21 DR. SIESS: Bob, I think you have got a point, but 22 what Frank says they tried to do it,-and maybe we will get them 23 in and.let them tell us what'-- 1 24 MR. BUDNICK: I knew what Frank was doing, ,I was 25 saying that the committee might help itself and them by l Heritage Reporting Corporation l-(202) 628-4388 4
289 1 providing a sounding board. O 2 MR. COFFMAN: It's a good suggestion if the seismic 3 initiatives could be integrated themselves. i 4 MR. BUDNICK: I was restricting my -- 5 MR. COFFMAN: Then we would-only have to relate to ! l 6 one. ! 7 MR. BUDNICK: Yes, I was restricting my remarks to i 8 the seismic piece. You can go off on A-44, you know.. 9 MR. KENNEDY: I' m not too enthused about, you know, 10 slowing down the A-46 program. It's slowing down for certain 11 other reasons. But the plants have been walkdown things have q i 12 been found and you would rather have them found now than have- ! 13 the earthquake find them. 2 () 14 DR. SIESSt. You know, that should be written in 15 stone, because it's .been several times. When you walk'them ] i 16 down, you find something. 17 MR. KENNEDY: But you don' t want to' postpone it too 18 long in order to get it integrated. 19 MR. COFFMAN: I think there might be a helpful 20 distinction here. 21 DR. SIESS: But the second time you walk them down 22 you find some more, Bob. l 23 MR. KENNEDY: Probably. 24 DR. SIESS: The third time. 25 MR. KENNEDY: Probably. Heritage Reporting Corporation (202) 628-4888 l
I 290 1
'l MR. CHEN: I think the slowdown of A-46'is not 2 because of the --
3 MR. KENNEDY: Right. 4 MR. CHEN: There is some problems trying to work it 5 out. They have another trial panel. They cannot get enough' 6 training for those people to perform that, so they put off for 7- training for the next year, and then so their walkdown for the 8 A-46 trial on this will be next year. 9 .MR. KENNEDY: I certainly applaud the effort. 10 MR. COFFMAN: There might be a couple distinctions in 11 .that the plans having the primary responsibility for safety. 12 could go ahead and do these things, and'in fact take advantage ; j 13 of all these fixes that have been found. The staff?is a little 14 more sluggish because it's not directly involved'with safety, 15 but with safety assurance. So we have geS to' work.through our j 16 paperwork. 17 DR. SIESS: Yes, but when we.tell ;he' plants, you 18 know, you are responsible for safety,.you ougnt to do it, and 19
~
then they get real smart and take the init'iative and go out and 20 do it. Two years later the staff says, no, you did it wrong. .) 3 i 21 You have got to do it over my way, and ~ that's happened to too 22 many plants as you well now. Now at least they are catching 4 l 23 them early. They told pulman not to go ahead with that . i 24 . director spanning before they spent much money on it. But you 25 know what happened to people that went;out on' human factors, 1 Of ' Heritage Reporting Corporation 'I l (202) 628-4888 1 4 4 ^_ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _
291
- 1. rebuilt _ things and then they were told it was wrong _and fire 2 protection. So it's easy to tell the industry, you know, do 3 nice things and we will pat you on the back, but it ' doesn' t 4 work out that way.' They are going to wait until the staff 5 tells them what to do because otherwise they are-likely'to get 6 in trouble, not everybody. q uite that - -they.cre It's not .
7 always making improvements'to the plants. 8 Well, gentlemen, thank you all. It's gotten very 9 late. 10 (Whereupon, at 6:28 p.m., the meeting.was recessed, I 11 to reconvene at 8:30 a.m., Wednesday, September 30, 1987.) 12 13 1 15 1 1 16 17 18 19 x 20 21 22 23 24 25 O , Heritage' Reporting Corporation (202) 628-4888.
' ' ' ' ' ' ' ' ' ' ' ________---_____m_m._. __mm_ __.-__-_. -
/
j 242 1 CERTIFICATE ( 2 3 This is to certify that the attached proceedings before the 4 United States Nuclear Regulatory Commission in the matter of: 5 Name: ACRS SUBCOMMITTEE ON EXTREME EXTERNAL PHENOMENA 6 7 Doc!tet Number: l 8 Place Washington, D.C. 9 Date: september 29, 1987 10 were held as herein appears, and that this is the original 11 transcript thereof for the file of the United States Nuclear 12 Regulatory Commission taken stenographically by me and, 1 13 thereafter reduced to typewriting by me or under the direction
, 14 of the court reporting company, and that the transcript is a 15 true and accurate record of the foregoing proceedings. I 16 /S/ tur rs . e ,e 17 (Signature typed): IRNIN COFFENBER i
18 Official Reporter j 19 Heritage Reporting Corporation j 20 21 22 l l 23 i
)
24 j 25 1 1 1
- /~5 G
Heritage Reporting Corporation , (202) 628-4888 i
'l l
i f O SEISMIC MARGIN REVIEW OF NUCLEAR POWER PLANTS MAINE YANKEE TRIAL REVIEW PEER REVIEW GROUP REPORT (Slides Prepared for ACRS Subcommittee Meetine. 9/29/87) ROBERT J. BUDNITZ O Future Resources Associates, Inc. 2000 Center Street, Suite 418 Berkeley, California 94704 (supported by NRC, working through ! Lawrence Livermore National Laboratory) i O
? 2 PEER REVIEW GROUP MEMBERS O
Robert J. Budnitz (chair) Future Resources Associates, Inc. j Michael P. Bohn Sandia National Laboratories ! 1 John W. Reed Jack R. Benjamin & Associates, Inc. i James Thomas Duke Power Company Loring A. Wyllie, Jr. H.J. Degenkolb Associates, Inc. 1 PEER REVIEW GROUP's OBJECTIVE "The objective of the Peer Review Group is to assure that the trial O seismic margins review, following the guidance established in NUREG/CR-4334 and NUREG/CR-4482, is executed in a fully ' competent and professional manner, uses methods that are at the , state-of-the-art, and takes cognizance of all relevant information." i I HOW THE PEER REVIEW WAS ACCOMPLISHED ] o PRG participated in d walkdowns I o PRG had continuine interactions with the study team: o the study team asked for PRG advice frequently , o they got continuous feedback from the PRG I
s 3-PEER REVIEW GROUP's CENTRAL FINDINGS
- 1) The project team did follow the guidance in the NUREGs
- 2) The study was executed in a fully competent and professional manner. !
- 3) The study used state-of-the-art methods. ]
- 4) The study took cognizance of all relevant information. -)
1 ( l. 1 \ l O OTHER COMMENTS by the PEER REVIEW GROUP l l Al Selection of the Review Level Earthauake i o more explicit guidance is needed to clarify ambiguities ; l . ! i l 1 Ill Combining Seismic and Non-seismic Failures ) L o methodology used at MY is not rigorous. o PRG recommended that this aspect of the methodology be developed fully 1 O o cia,incation needed as to how seismic ma, gin. is to se /'l L interpreted-l q o
O Q CDFM Method vs. Fragilities Method for HCLPF Analysis o PRG recommended further methodological development in this area Il Q1 Level of Expertise Needed o Fragilities team must have experience in o doing detailed walkdowns o focussing on critical components o realistically analyzing capacities i O o systems team shouid have experience .ith o PRA methods o external-initiators analysis i El Earthauake Experience and Test Data Base o need to document experience data base o guidance on configuration details must be O - st,engthened
t ' O 31 Reiay Chetter o Not covered in the MY trial review o PRG believed guidance shouid be strengthened , in this area j l l G1 Correlations Amone Earthauake-Induced Faijpres O o PRG identiried this as a methodoiogicai i probiem area j 1 o HCLPF-type approach isn't weil suited to i handling correlations well ! l i H) Screening Methods Used in the MY Trial Revievy o PRG endorses the iterative / interactive methods l used
[ ] 6
'O L3 VEL OF PEER REVIEW EFFORT o The level of effort spent at Maine Yankee was essential to the project's success i l
1 I o On the issue of how much routine neer review is needed: o Panel was divided o Some thought peer review is essential o Some thought the need is no different ; than for any other utility-type analysis O O
L 1 8 I 9 B T1 , A R S BO U H O F G C R U R P TA A P M TE , L C 4 A E N 3 2 N MO 3 8 C 4 4 E E - 4 - M R F R - O O P C R N C L/ C M E C G/ A H N H E G R P O R E G EU R O L D H N U R A E T( N P N S ( R A D E S E BN ASS HC N T I X " A G E N HE O R I C VR A E G A 0 P M M S RO BP E A R A A N R R T O MP ' ' D G P S S I X L T 6 CA E I E S E E C K H S N A 8 I S E A R 9 MG A T D 1 S N U U P T N I I O N C O Z TO N D O EN H O N I S ETE R D S I NI I EN U G RC EB P U XS A T" A 5 MF C ES R A E A 6 I O G N T D S Ai E 8 R OS O B 8 N N D 9 O 9 N E I DN 1 F N A T S 1 A S N A N O L T N I I D EE O[U S U B S A P N U L L I YI T E C N C I P N I N L O TEA L R P L E BD S U T O L N G MI M N E A S R A O SI EF VI ESI D E A R T R L V S C GY L O YW A BP O EDES DE U G N G E D S G _ D O G R O I N F 4 N TM H TN L E E 'E W N L V A O 8 I RA TN I I O E 9 KO R E E K N H I D R W G 1 R P G G O _ E N O P O MR RRO AP TV E N H E I I N WU R WU W G R I T E V T I S P TN F EK R E C EC E MN E E D RL D I DRI D A V M E N A E VD E E S E EE P P S P I R D Y O T C P BE 7 A DI O R WOX E O RE _ 8 I EN L EL E S L S E S EN
/ TMH E D I E R VI
_ 9 I R C V E V VE A V E O A 2 N O EE P EE H C.
'. C M
/ I F TD O R D T D 9 L E T P
V S _ M E A D - L - - S - - - D - - - O-o o o
)
G A
) N R 3 ) I F
( 8 T _ 8 9 E 3 E H
) 2 0 M C
_ 1 A A 6 , 7 O 4 1 N 8 R 0 I 7 9 P 7 _ A , F8 1 P 8 S - A 9 N L R 1 O S L Y P _ I H N F E V MH F T L B _ 7 N L N M D C L , 8 E S R o O ( I _ N 6 9 V A _ L 2 1 M _ O FM 8 S D 0 _ L W E E N _ ( 4 H E H Z - N
- C F C I F 8 EI
_ E RR A L O8 S _ E C A D - 9 U D 8 K / M E T OA E 8 T X R N D 1 _ N G U E P I N DTY _ A EN ) N P F7 EH N E F _ Y R I 8 B A 8 C EL _ 6U 9 I P9YR TP R E8 N D( 3 R N XMO N 9 E 0 T I F U 1 BA EO F
- M I 1 U A S li P O C A 7 S I T L R YS D - D N D I C G L N N N YE MF 8 S W H U O I O 9 I I YJ I A D D YD F 1 F S D U R E LE D A D
_ OG L TT E F UNUE E T A N H E N 6 N EU TI LUC V S W WI C K S S S O A E R R N L 8 EL S V GNS R G I P A A L Y C P S D N OI P NY VS MY ( F EMI MN I C MI D E RO E I P U YN C E F A T R NN E B D ETB O T _ S N DD A I I N I U D I
-T TD A O F T O
C EE P N MA O C T O C S TS 1 S ER N T F O _ I EEM T D OC F I T I 1 G TT - S F FT C RA EN S A WT N O I R E EE _ A I L O MI L E IR O OKK N O J S S MTP ETA S R C GP O A A I MR T I N VP O I I E R H H CNO E S NI EE I R WKS R P P P _ I I C S YI F R R A M S P R S M I I R O B E W - C - - - - S - - S - - - B - - - - O o O O o
E C L R B I N S 7 S R 8 O O 9 P F 1 S T S D YN S E L E U T UM C U J M S T B O I E I N C D Y R I T E7 O D S WT 8 T S E I EA 9 P N D I R 1 R I VO C R D F Y EP R N / E D R R E S D U O B D E T RC O W N R N S I N T E E EI C I A H O V Y E A , B B B TO E W TN I R I R D T A D P E O T ET S G N A TF D N E 8 A C E A E I
, 8 N S L R T G S A A P D E R E R 9 O Y G
O L D N A M O TP C R M L A R E M D O W 1 Y MH Y E O O CN P L D D O Y I A R U / D G YP C M A A TS O O G E E S E H L OR E S I I R T O L B I R G G O Y E P R N E D S E O I S N Y MO D G O E T I B H 7 O O T D R N G T N P S T8 H L WE 1 R E O D N E 9 TOWE BE N A D C E I M1 E D E UI WETT I S R G MO S L TEI S C O R I F H VS A EWN C S S A R O I RER T EI I BTU W E A TN M P E R E N O EG P M P B TT BH I C V EE P C N E N R VS I F I DU E E MO R PWEO O C E G TE V E S E S RT R MN A R EL B I S S I I G N E I H R S A ET VV T / E T N O S RO IN REEG R I R E F W N R T R N IO P R G E O B I E P P A M TH I R E WO S KE E WE P XC P R E TA R R G 7 E L MER E I H O EN 8 I B D A TVL WG VI / VI K0 ES S L E F F EE I E I N S O AR RN ARP CR TR A O / 3 R S O F L E R U P O L R WD P E N N RW6 E E I P E O L V O O E - - C - - - R - - R - - - c o o o
S S E E I K T A I U L Q I H B T A R R A E E N L D U N V A T C S I H M T S I I W E S
'0 Y T
F S I T T N N A E P L D M P I D S R O A T F E
- o. O S
O T
- L C
U S N A E - N E V Y M I G F T O O T C L N E O S E J D E I B O I C O H T I T I F E L F M I E B W A L D E P E N I A V A V C E E L E R T V N N I C E G T I R I C M ES E S H E F I N D F E I E S E EH N F H T A 0 T E E T S V D N S O I E E B V M S A O P S - R O A P L E O V E D
)
1 1 2 F L ,
^ 1 i
_ .E ,N O S P I U T S O _ R A R _ O B G _ P P C W U I E S F I I V L C S S E A E N R R C P I E I S G B R N - R M Y E H K A E D E o M _ C S M 5 P E A LN T T C R LN E I E AE E D A M M HK K N S R N - A N I O . . A O E F WR Y S S P P D U E U E F O DN O T O R EI R C P G N A G E N U I M . L O O W M W E I R E R O E S T G I ET I A V T V S C E ER E P I W R DA R I F E L I H I V ) R EI P S T ) N E N E BMI M R N R A E D E A A P OS B U M M S Q P R R T _ M E I W N E I E A E PO M N A P H I I I O H C C V HT D A E ( E S I N L ( E R R S
~A E K EO N Z L U K Z R BP L Z N T S E A T OE 0 A I A I W MM S P I Y NNDML B EO R N C N DN D HEOL MC E T DI R EI E U0 EHY L T R UMOEH N BBR TW A R E P
EA C R S I I A . . A . . . . . R . H X . . . M RMJJL T - - C E RPA J M O - o o o o .
Y G O L O D O H T E O M N I G R A M ) 1 C 6 I 4 M 0 S A I
)
E ,R S 8 E I 9 G 3 A R 0N P ) A A E) N M N A S EA M N T HM P R I C S T TR I U I F E TR O A ( J N A WA R H A O A EH G C R H R T , I C ( D N C P S L I V( Y N O R ( O U E E S A N S R Z L DY ~. I N EYI N , T - L E U DYV E S A S O A OI O S E I T EA A E R S C R TN DCI NH LL DEFFL L S NR R DV E E R R A F N R U R R L U M0 I E A Y F EU . E T E , E BA BMR S W P KMKR S L MP H E N - A . . . L C . . . .- . N L L R P l T0 O E P R P MWJ BL - H RP MJ J - O o
~
o 0 o
- ' M -
- s -
M g C
.. T E
S x T - N E D
) I N C A C M A R
I E A R
- H E S C V R N I - E r O
m. E G P S a P B
- S U M N
- DN OI C ,
O E OTRA S YN 0 ~ R M SHA TE I A R 0 G K GH A MRNDE A M 0 R C I C RI A .HR S G E A NI E RE / N M R J KRBGLKTO F 6 - I
. . . . . . " 4 K O . .
( - R F RJJ LCGP A A O W , P S M N D a I S 'L G o R A M N B G S B EEBREPSE
/SI /////
BBBBB F E A XG HMRRS Y G O L O F O P
~
S E I D G TSESTATT S//PSA D I H U s. EEERERSS O S T - I S DD DDDDDD EE H N S , E T O
////////
D E P RSSSRSRR M I T M u REEERERR C A D I NR R RN RNN S L S ._. - M N E - S I R E
- I G R WY _
E ) R D U ED _ S N A N T I U r A ) M A U VT - C MN F ES R RA C S R N I M I E D F - A R M L N TP , u T HI S O A NL _ N C A I R AC a _ E - H S E G LH .. R 0C P S N N P , _ P C -
- O. I A Y -
U ( ON I L I I OYRFD U w C C O S NHL R T GDW I ( R ES I P U N UBST H E P OT VS CYE HKHR N SL GZD C O PE F I F A MR - A UNN H T OR I FI BGA EGCEI 0 C E E WRDB MLNC H N I HN R W D WBTCS w YS E E I E A EO I V I TI UN WH TN - ERN LUO V O V I E R E OA G DNEOEI J P NL AE L R P R - - - - O o o o a s
! ll, t i [' ! ' :'
x C T E S N E O C I N A S T A T I C T N A S A S O D Y R I I L E S T T A T E Y N N R G A S O U E A I L L T T A S S L O E D V & M M O E E E M H E T T E T M C S S E N Y Y R E O M T E S S C S I Y R 8 , OWV E I S 8 E P X Y T S E G N C I M E E I A O S R T L R I N I N E I O T O E S E K S G H A I S N N A N A C U T I O U W R N L A T G P O O F A A U U R M H D . V L O A S O T K N R E A B M E C R L E O V A V A A E F Y E C R E E W W T T I U Z T S I S N M C I T T 5 K L M E S M N N B C I E M I D I E A E G T E E R N C L N H A S T S A I E P O C R Y A Z M R I F S T C A T E T S S R H P O S V A E R R N E T E I R D O O T C C T S E H F F S N G N P t U I M O N A. P E O C 3 3 B S C I O T O N / / S O I N P X C I T 2 1 T R E F E R E R L S P E O O U L R C O F S O C C N W F - E N H S I T - - E - - R - o o o o o O o O
~
n. lV ~d Seisn0t0GiCAt ASeECTS OF sEisn'IC nARGINS APPROACH l i 0- SEISMIC HAZARD CURVES ARE NOT DEVELOPED j
-l 0 FIRST STEP IN A MARGINS REVIEW IS T0 SPECIFY A REFERENCE EARTHOUAKE O USE AVERAGE OF TWO HORIZONTAL PEAK GROUND ACCELERA-TION (PGA) VALUES-AS PRIMARY DESCRIPT0R OF GROUND MOTION AGAINST.WHICH MARGIN IS DEFINED 0 IN DEVELOPING THEIR APPROACH, THE EXPERT PANEL FOCUSED ON NUCLEAR PLANTS IN EASTERN U.S.
DESIGNED FOR 0.1G TO 0.25G SSE SUBJECTED TO GROUND MOTION WITH FOLLOWING CHARACTERISTICS: EARTHOUAKE MAGNITUDE RANGE: M5.5 TO M6.5 3 TO 5 STRONG MOTION CYCLES l j 10 To 15 SECONDS TOTAL DURATION AND LESSER STRONG DURATION { BROAD BAND RESPONSE SPECTRA j l l 4 l 1 I i l O ; i
E D U N L A A V
) Y F N T P O I L I C C ) T A H F C P P N A E L U C H C F T H C
( R I E E O M C U E ( S I L R I WA U M E TV L E S I T TF A S F S P F Y O A L W S EC F E E L H O I , R V T U TE Y E N S A H T R E A T I N E E L S O M B M I N P O B I M D O R A G O N TF O O B R P R A M C E H O U B D Y E L TT C T R C C W I E E E O M T W P R L S A O X I I L E D E E D O S E , E D N T B E E R I S T C O M U D A N F R B L L E E O U U D T T R O C I I E WL F R D A A N E YC O M E S T C R I I E F A S C B H O L E A G S E U P T I E E V L A O H R U E A C N U L L V N G A N A I VT F A C F N P I F A L D T P L C E U L P H M_ B C o H A A o o o O
N I G R
) A E M N S O A E I B R T U D A A L N OL T I A TU A A C D F D D L - E E A E C V S C H I I U T TE F S C A P S I N TL T N O A C I I C D H F M -
R E YO T E R TN I TE I . - EW L H F N D. S P G I C I O N G I I E F M ) I A H I' T VP D u N R WR A I L S l ( E F O L TC E R I U A H C , S R A O R C V - R R N C R F P L R TN O S P A E N O I T S E E 6 SA C , F D L N N H M A P N E " F O O T A ML L A VS P C P ( A U C E E L MR P P C H ,L S C D O E MS G L T S H N C D D I O A T S E A A N S S R C N E K P T G U Y P E T A " N ) N C L W N N U E A I L R A E I G O , Q T N FTE N N I R I P E H N O ( A N A V P S M C TE P - L A E E E E O N R N M S U P H YR DW C E A O O I C TT I N E I EP C S L T E I OI R M Y A R N L E - TV E D O L C E I I K D E P N C A P G N E S R X A N R X D A A P E E . E A O E E R Y O I A SH F P F L TB ETT Y E O E EI W KN TS H L A N VR A A E , I ET E R I E A T U N . L H VP A D L A Q O E I C Y E - M I C H P . G A B D S HM TM I A O - E N C I N R O ( R R D H R I A S I A C F P E C I O E D .P D A U D R ED YE EA N O Q E P R E C F D H E R E S P O S I I E T) M P R U A MU R TE MM P E N N H F O A M N E TD C F E D S O C E A D G I I B ( R F - - C - - O o o o o
' , L : i
4 l
.O FA HCLPF APPROACH 1.0 ' 3 i 95% confidence curve ,
e X
$ 0.8 - -
3 Median fragility
.,_ curve O
N O.6 - -
.n j 0.5 - - - - - - - - - - - - - - - - - - - -
0.90g : Am' e i
" 0.4 l c I I
O =c 5 o I l I c 0.2 - ' I x - o o HCLPF I (5% confidence . I curve - 0.05 - - - - l 0 l- ' I I I I O 0.17 0.30.4 0.8 1.2 1.6 2.0 Peak ground acceleration (g) HCLPF - mA exp'[-1.64 (SR
- 00 )) ,
i-O , i
- - - - - - - - _ _ _ .--_i_.-
\
l Conservative' Deterministic Failure Margin (CDFM) Method Load Combination: Normal.+ Earthquake Review Level Ground-Response Spectrum: 844'Non-Exceedence Probability Site-Specific
. Spectr um Damping: Depending on the earthquake review level, tihe following are the conservative estimates of the median values Structure: 74
. Piping: 56 Cable trays: 15%
Structural Model: Best-estimate - median j Soil-Structure Interaction: Envelope expected parameter variation l
-Material Strength: 954 exceadance actual strength .J
~
Static Capacity Equations:- 844 exceedance by test data or code equation System Ductility: Conservatively selected to be between 1.0 and 1.5. For shear wall structures,.should O (Inelastic Energy Absorption) not be less than 1.3. Floor Spectra Generation: Median damping value for equipment Frequency shif ting of floor spectra rather than peak broadening. 1 i l l
-l O
h - -_
1! O _ E H T E O C _ T) N D E _ U - TE L N S . A S F F TE N P RR I L O D C P D T H MA O I N T T N S O O A N N D L O P I E T _ TS A E . C A H N N E T I U L F E R S E M R E T R T _ N E U E T D A V _ G. L I C N P T D U S O H I C E D R I T C - N EA S E T S A E S 0 S S D S Y S O OR R E OI E E U L I H D I R L D A S P _ C TA D D C N N Y P U D N A A L A S R Y( T A I A _ S L S S S N H M N T E S T M A T E O L A B M N E A T E R Y E E T E P S R M P T D T N S E Y E I E S O Y R S S D E M L E Y P S T S I R O A N S M E _ S O R C G O T C C C N C F I N S T L . O TI N R U U C R D I L O O R A S O E R O I P E F S T I C O T P H S E C F - C C U T S E H A I B N S O E S C E T U E U S U _ A R N S R F D L U _ O E O N L H R G D R C L A A H C P N I O A C A PI TY R E T A O A T S : C L : E N U O R A N S A R S V I O R P TG O R E A R E L E P P N I I W R E W S N V P A ET T P B T E0 A _ R I C N E B I RM N R R O R A C R R P U U O O R C C F F - - F - F S N E o 0 o o o o o o - O -
e
*O Start Time axis l
' \
i I 1 Select an earthquake review level )
.i i
Gather inf ormation ,. .. , ; f ,. . Gather information on on systems and sort the plant. Determine which !
, ,. . . l.l l ,-* //
Group A functions.
,/ broad classes or groups of f
. *] g , ,
, // components have HCLPF
. . . .. .* . +----- 3 J
******* values greater than the review {
b.*** ,*
* * *,* ~<
- level. Possibly identif y plant-unique features.
l 4 1
- . First plant walkdown ,
. . Concentrate on identification of problems. {
Emphasize systems interaction. Confirm {
. .4. ,
applicability of screening tools. Complete l identification of plant-unique features. i ( l I v KEY: Task is performed by: . ; , . ,., ; i,*
. , Revision of systems C ***t'****. ***
. . . relationships established '
l
\ '
,1. , *, . Systems Analyst '.*.~.,'5*...-
,,,,,. in Step 2. Develop iault
,1.*,*.l,.*.*y trees and event trees.
) , Fragility Analyst y Both Second plant walkdown: l I Primarily fragility analyst for checks. ' 1 6 Collect specific data (size and other i physical characteristics) of components l requiring detailed analysis, o
.;.*. . ,.. .:t . .:.,..,;*......-
/ Finalize HCLPF value f or Determine minimal ***
'/, .' ; *. 7 ., / ,. ',.* components in final cut sets cut sets for end- , g
,t* . . , . . . . . , - (components not screened out).
point core melt. . ,
* *. .= .. =. ,. .. .. .. .. ....
( o Margin ' assessment complete SEISMIC MARGINS REVIEW
O N I T M R O N A R O I L F I T) M T A E I H R R V S C O E I A P D T S O P I C E R U S E R P S N F U P O F D A D C E E N C S A D E ON ER R I P E O P G R N M R O WA C S S EM N I I Y O VC G I G E I O TN S RM L C I S S O S N N E Y D I N ' U E R D N E H A F E G U A S C R R O T E P TC R S G C T N S P N R C A S I N K I L M N M N S M P E I E E E I S T T E H R I D S S S R T E N Y I Y C D S A S S S N ES W I I U R S R E R S L N W RW I W MS ' P B O P V B E P D T( E T R A N T A R SW A N I W Y P , E TE , S O , C I I D R C A E N V E R O I R R I E U W F MF R S B A X L S N , I A S I EE L L S R N D V U N E I Y I I A L D V G L V G P L E E R T O T S A N R E A M E P S : G T C 0 S I A C E Ol i R T H I R TT O S T M T E S Y A V S I L
- G N E E N E I I D S O S T U O S Y L R T P E D CW F R V U N B A U N T O R P I S C D o o o o o i
o Y L N A R O P E C S S _ I FN U _ M E O O S S G T I E , I A " P N T C . E ) E S ME0 A S F O C I A N E E. _ Z I D A F T U I Y S E F E A Y Q ( L R ( E RR E A O P TL N G O N EW EE S " T C A R E D V E A C N L I I O FS A I S S C OA R F D - P EV E L G E N E E E ) R R O R GL T N E L N E A R U KA C T F S L L R I R E M B : I A EO R B U S RYT VT EF H . H D O O R T T L O OA _ I H YR T R R R _ T F P E OY O ) 8 W GF A I I ET MP - T) W I H Y S T N I O KI S E E G E NL N _ I L E N C D N N , U A U I A ER D D A I O _ V O S O T U G R LN R BS E R A O U A A I R E H P Y R P S EQ A U S MC O EBR - P G A TA C) N OA N H ( . N EOD F R I R P _ T J A _ I G S I LN O YO MPM R E S P E E GR R" O D , R ,R % OR E C oM A N A R U L E R O N I G O C S R OC 1 RO0 L E O TR EOE C I S E I A O M R A M F O O F F N L A YN D N OA HM D OYT V I -_ M T F YER H TU TCA _ S A . EH NT _ I T C D 7 D S G EA T _ E S I 8 E N O R E M DEI _ S M , 9 S ) O L C " S G EU T E S ) 1 A K I OS S S N S QN A G I A BO S D ( , E N I U EA I D RU G A M S E R F H C F R
- O U L
O HYA YL GU RL EFQ B N A - ( R P B C TFS T EE I D D N O Z A M L A C E N O EI MT , R O A C MN I TGK _. N T N T HS C N OF OA A _ E N D S ETF O( N MM T A D I N E MD O N N D I I G E T S DDO A __ X L N D T SE L - T E P A U E R A G N A _ B L F RYI R N R UE - P K A L DT OU OR S _
- . M S C MEA X I L HO' .
R O TI VEE TI SC A - T C HR EI L A A " R _ C D C F W P _ E : GE S T J S Y I M UA S E I EO O R D S I VET FC I TE U NL OR TA I I V P R O DM EEC P T T I ( TEA M O S R VC A S S TI G DS ENS T MI I S R L O S E ST - I I G EE N A L B T N R S BC E TS A Y I S N I L L S S I S EO G EU E P PL R R F V O MA A D N C ON E O A N M M I S C A F - - O o o o o o o
( w m a E Y ) m Z G T 3 Y S O N L R L A H m A O O L G N T D P U A C A T; O H R' O R h R F T W H _ E C E B T T N A _ MY _ T A R GL ) E M T ) OA 1 B U N I L I H O - OR M . O C O T F ) B ED T R O I U S O - A HC F W I S HTR E E P EN D ' _ I L E L ( M/ E V O S A N E R R A H N O DDR EEP I R A m - E P I VVE E N I H T ( T A OO RRG L Y - - G S F R oRM A D N A E R U R E T P PN MMI I I M O S N O S E V O L N GGC S C C S I I NNP ) E E I E A l I I U EL R M T F E SS L S A UUOUN R I T C TD EO O E S I H T S M T YY A E S I DDYHD - R W U UGC E A G I A E , TTOS SSL SS AEEERP G N F A S
- S ON BTUOA L -
I D N E EA D O T E BO - YTI D D N T F O A K NA TNO WHTGSA N .F A I E N C A TEELE T X A L D N O YAMDOG C N DA CRI O E P A S DEOMMR S E A N I I EPHA ES R RVETDI A PODE E _T S C E MER(P MTS NN - A C E EEMWEA N M R T J O U C TTI A A ETL I AOP UNH R O E P S J U UYVU E O L L LEL R R R A AP RA V P VVP E - - I B E EEA - S R 8 E U 8 P T O ) ) ) i U Y C ) 4 F F S 1 23 O o O o
L L I W S D D O N H A. T E
, M G S N D D I O N T D H A E E T E E E D M R M E T 7
. H A 8 J T U 9 O L 1 B A A V R R G E E D D N B E N E I E MW S S I D S B E E Y D U VE U V I D O A L O V U H E R M S L N E T T R O F I R S E A . R P WD R M K F L E E R F S C S N TW O P L
C N O I E R U MR
.T O
H T N R O A I E F B A N 8 8 H T D ,P N TL D I 9 A E YP E C E 1 M L C A U N I S N D F U O R S O D N R E H D C R R P E O A T C C L P U D M R S E A R A MN O S P S L R A S C M A C T E O M V F . N N L E P F D R , F E I BR R A P C N O S N E O O F L ,O N O S T DC F C I YS S O P E A L N H R D N N I M C R U I D P E E O TO N O O E T E N V I A C E P H D L N N E TZ R R S L U E D ET C I E E O U D N N KS I N L F C S O E O A D A P F N TC H P . . E G M I I H C M S R J R R A D G S S O E P O S O I T C R TT S N T U : D TL N N E1 R F D S N N A A D I O O E R I E I C E 9 P C O L R T I I TV E N O T BE I FF N O R O R A F N I I A R S P G E F I N D C P R I I L I G O I M C P A T P D 5 I M F I / A F S I S I G P E TY E N E M C V L B Y B G R O R N U N I U C N I M - A S - N O o o o o o o o
Y G O L O 2 D 8 O 4 H 4 T R E M D C oSN I E T N E
/
G E R G I U R R N A 0 F M D - N P W E Y A L E C T R H C I I E T T H V M L S E S P W U R M R I M E D " E F E O I N Y D D S S C V I L A C M E D O E I T R S N R F T R R I C E B O S R P O P I I E Y E E F U Y T R S R T S T F O G S F I U N T F E R O I
~
L E E A O G L N Y E I M K H L O I T C A P I C W E G D M I N F O L E N N O R L A L - Y I A I H E I D - l E G A V P K T T T I R I I V U L E R E E U U W S 0 E Q E R W O M D " G B S S D S R E W I I I V C R R E R P P R N E - - E - - - - - - o o o o O .
T U B N , O S I N T O C I N T U C J A W N E O G I C N V N I E S R I N E E S D D C N A E I I M M L. - G R - R N O G A E F N - M E R I B E O C - I P G E R - I N - M V P E O S A E B I H 8 Y E DW 8 N - S S N E 9 A T A I 1 T N V O N E ,E T C R Y A M G L - L E R R Y P E N S A L R N E T R G ,I C W A R G S E - B E R W I R L W A E Y I - L A O MI A N D E - A N P V W O C - I I C ES R I D N R F A I R I E T EE T I M H D A I R O GS 6 T4 N I TE C N R I 1 T T F O E - D N U - R E I EE - N , ES A E M S N BR
/ Y G G A GA N N I
- I A O TE W N N R I R A I W O R E TI U P R T E N U E E E A MO T D .T N E A I WH O C D M V L U E N TTC NP O I E L F U E E C R T V O O R I - I BDU F G WS T E E S I R E R A A R G S EE K E E N O W - R A N O H R R TP F EE - E I P T A I I I V P S TO D R E VR O N ER HN MS A P T EE R O O EP T A , L P E A R S - C I M I R D T C WR 1 W O '0 I D S D. A 7 P R N L 8 G N B S N N T I
/ H A E I U
- T C G G O - O 0 C R RW G 3 T P A A E / A M N 6 - - H - - - G T - O o o o o o-o l! t
P I H S R E B M E M C D P N G A D R S N E E A T R R I A Y R H B _ P C _ E D A E Y R BH O
-. T S
'Y I N
_ L W O _ I P R P S _ A U MO E I R B R G _ P Y D W A E D E M O _ S E I _ U G V N A E O EN R I BA TS L L D E MR E A N U O L I _ I N P A T WA S VU _ A ' EB _ H D E I N C E EE " P _ O A D G K S T _ I O N A N MN TR U N A EO _ A P F A YTC R P M - S _ EA E EYS P B D N S E O S - N I R O N L A A C C I L . MR E G I D N L I R WN O " B R A U T - I _ P MWF ES E E - R TS
/ I I L A A O C R VL L R P
_ R P E I I A _ N E R WMP R I E E D E E S S S H EH H E T S TTR A O O P O R P - - - - - O o
1 , \ l
, ag, ,
/[g f&l0 1
RESEARCH PROGRAM 1 l 0 EARTH SCIENCES t ! o SEISMIC MARGINS O o FRAGILITIES AND RESPONSE 0 VALIDATION l l O 09/28/87 .VUGRAPHS j
]
i i
.' REGULATORY ISSUES
($)
. 4 SEISMIC HAZARD o QUANTIFY AND REDUCE UNCERTAINTIES IN SEISMIC HAZARD ASSESSMENT AND DEVELOP. TECHNIQUES TO DEAL'WITH UNCERTAINTIES IN A REGULATORY ENVIRONMENT, SEISMIC MARGINS ..
o ASSESS IF CHANGES IN THE SEISMIC DESIGN BASIS CAN BE (]) ACCOMMODATED WITHIN THE INHERENT CAPACITY OF.THE ORIGINAL-DESIGN OR WHETHER PLANT MODIFICATIONS ARE NECESSARY. 1 I ADEQUACY OF CURRENT CRITERIA [ o CONFIRM THE JUDGMENTS AND ASSUMPTIONS USED IN THE DEVELOPMENT OF THE SEISMIC DESIGN CRITERIA AND ASSESS THE. VALIDITY AND/0R CONSEQUENCES OF NEW DATA, l l i a
.O 4 l
l
'09/28/87- :VUGRAPHS
)
L RESEARCH NEED' 9 DEVELOP THE METHODS.AND DATA BASE THAT WILL SUPPORT SEISMIC CRITERIA DEVELOPMENT AND PROVIDE EVALUATION TOOL'S o IMPROVE ESTIMATES OF EARTHOUAKE' HAZARDS BY IDENTIFYING POTENTIAL EARTHOUAKE SOURCES AND DETERMINING THE PROPAGATION OF SEISMIC ENERGY WITH DISTANCE. o ASSESS THE EFFECTS OF EARTHQUAKES ON NUCLEAR POWER j l PLANTS BY DETERMINING SOIL RESPONSE AND THE CAPABILITY 0F STRUCTURES, SYSTEMS AND COMPONENTS TO WITHSTAND EARTHQUAKES LARGER THAN THEIR DESIGN ~ BASIS. : o VALIDATi CURRENT SEISMIC CRITERIA AND RISK ANALYSIS I METHODS.
-O SHORT-TERM RESEARCH o OUANTIFY AND REDUCE UNCERTAINTIES ASSOCIATED WITH i CURRENT REGULATORY REQUIREMENTS.
o REMOVE CONSERVATISM WHERE THEY.ARE UNNECESSARY AND ADD CONSERVATISM WHERE WEAKNESSES EXIST. ; 4 LONG-TERM RESEARCH o RESOLVE CONCERNS RELATED TO SEISMIC DESIGN MARGINS AND THE EFFECTS OF DEGRADATION NOT CONSIDERED IN SEISMIC DESIGN, MARGIN AND PRA EVALUATIONS. o DETERMINE HOW IMPORTANT EARTHQUAKES ARE IN THE REGULATORY PROCESS. , 1 0 ' L 1 09/28/87 VUGRAPHS L__ ___ _ _ - _ _ _ _ _ _ - _ _ - _ _ _ __ _
I l]l1llli e o lh 1 s O 0 8 t e r n u w d 1 e n a w e is m t c u
)
1 8 ise vn _ G. vn e e r eo _ YS _
. k R A eo t a s t
S 1 Rit _ RN ic ( Rit t n t r . g da n n _ a ix g da S
)
d o d s g w r d n s y op e 61 1 n p a G. inn ig n a e is OIO s n t c ep n e in) ipeg e u ic s s u a g u p r t c R ( i pe s e n w :sa De t TT u e pp Pmi n s l o e s . n e g Pm D o ie AA r Bse S r t s P R I G. i Se p i 9 .m ic v l c
.mlp e icmsims I
ir it* c A S o c LC s wu es e p 9ou E ,c tn n d c R ip ip e p 1 1c' m ea R - (g P- P i 1 e 0. 1 co m e s pi UI c es n S 0 ec m s eoi n im s s n a eR* ie s re A ic L n Nis op ic 1t eRe u i e ia dt a ie i n e tn o1 D u se* S nt n PR S i S e - G E P e .e nk e ic a seE s f r sed % S cn ct s S pmd o de p2 s2 la se 0, Ii EC r o e e 0 . RP I t i c o e Ae e e e1 r B t i e e 4a - i a C n e" I t sequa nR e pPR t A S t iciS tS k e s em i t e s D E R1 u t ci irm8 4ai Or Are tt- a t a sn ns Ar . r m- i p r . c e e g S n r e l n nm" d e e . h l rh is e C nmG
- F e E e vp el i g e e p SM d e oG u o.R r eo l t Si r ss oe il c l t Si r a ic at r c e m hi ic i t
r ha e it 0 eoO a iP A n l FI t S GC UC PR L L UC - a CE D S 1 GC( L Si F r L
- e * .
I
* * * *
- V . o e -
E * * * *
- a e8 e:l ,6 8 l s elliesl,n asgaa a sl3e'
.: ,s l e,8
.Bel' .:' la lse' _
, s.
, r i
o t
, a
. M i l c A p p R _ G A y O g
, r o
R n t a f P , to l u t A c te H m a r i o , o r m g e C g r
, o er P R
$ f
, C,
- d n R r
' s P nd ), i
* **re a e s ,, hc d A m s
er u atu y r n E a r ma u eds t c Hc au is it l
, a
, e a g t c e ar ,, s S o r
r g o r u oMt "** j orm ms t er c, n e s p _ E P c r P S t en*Pg a S te F t
, ie ih R u es e I ri taor in rg o sc se ir n
e '" ,,, w ni s Y wn wpo s n o o y r g Fs lap
" 'a Pr vn
*o o*
p' ,
, s
- e n T . s p e td m
- e ng Ar'a c , o na c"am "
f e s c to e o i e eo o ,q t E n e e d mR j R C mL "" uD ta te in av,o , e
, g a F r D cd , l iinc e 6c
,e o d ic H bi6o, i* , a c nt A n g
_ t S s s a m am *e m hj e r ht e s s ns
- u t
c in Hh i" s
, co r e
S r * + r e ie oe
- r ip *e s" i
t e" ,
, EP xr a % it S cS S P Rs t C E t
c. H 49 ,
- n I e
*
- e * * *
- a i st c F' t M r o g.
o c e S f. p r iI j o I m m r E a a o 1.o r P _ S * .d P r .c r _ 1 C s dw e R _ e d eo r o h p u N ,
, te wm eo i g
,, M sc F y
a s8e9 a el!8 88'I e ase g88l
- l8 !ea0' .sa ,e l 8 8 ea' leeal,a
.I.I!.e a
i p cr C I u or a e s TS l h G me s s _ SM E s ia io S c r a e n o es n o a R r y m a t en c i m e MA r d dr ic s s it it i t g it r g m s _ i u a ) e k n a i l a o r o ire Su ie S az BS mies b R t R is o ic it ic P r g r bam U R p f f e P p s OG i s ora s i x Ct o rg on eH N t - - s m la n l h DO Po
. S n ic t e a u o ua m a ar c I n
R D E Ed Po it kmA sr ma s ig is R r Q it Q' t H SS UT I _ DR s rP yr o a a urM eg er n r a l i g o n t a c t r g e s en Nt P R mn o e t i g M b d n K NP no i e sd M. qgE a n Pr e f e r f n /l a _ c da ur d hoF ir r b li K a AD n n ra n u arPF t r - gP a ic o a g mp a u m g g ia - iwTs t nt r n n n n I Men .. E e I H or En oN S En M m s P s s .m iu e .ep qm Q c qm ip u i t o y a NT om n T i c dnc f e ig l c im e yer i ip lt a - m S ai G ai , om t imEa A P Ca a nir GA I L h I mS iouC t ncS se pl t i Sa m IRorg ula tsys i l i g P Ea r R IR g 8 si Cg C C Cg r r d mI r R ae pp . E E Rs G tdS rt s Pr dn P Pr o e Rr o R R Ro r li e P ax . t i Pe U S aeU a a G E JE Nt( oe r i e EP iA n E EP S NP N N NP = R a ES i CS r V* OR
- a * *
- E * * *
- S *
- F * * * * -
F lillj!
l' l RESEARCH PROGRAM O EARTH SCIENCES o RESEARCH DESIGNED TO ADDRESS ISSUES THAT CONTRIBUTE TO I THE UNCERTAINTY IN SEISMIC HAZARD ASSESSMENT, L o UNCERTAINTIES' SEISMIC SOURCE ZONES PROPAGATION OF SEISMIC ENERGY SITE-SPECIFIC GROUND' MOTION RESPONSE
~
g o- PROJECTS l ; - REGIONAL PROJECTS ADDRESSING UNCERTAINTIES ASSOCIATED WITH SEISMIC-SOURCE ZONES AND PROPAGATION OF SEISMIC ENERGY, Q TOPICAL PROJECTS THAT DEAL WITH DEVELOPING SITE-SPECIFIC SPECTRA AND STRONG GROUND MOTION MODELS. PROBABILISTIC PROJECTS THAT DEAL WITH TECHNIQUES TO HANDLE UNCERTAINTIES SEISMIC MARGINS o RESEARCH TO ASSESS THE INHERENT SEISMIC MARGINS IN NUCLEAR POWER PLANTS o MARGINS IS DEFINED AS THE ABILITY TO SUSTAIN AN EARTHOUAKE LARGER THAN THE SAFE SHUTDOWN EARTHOUAKE. o DEVELOP REVIEW PROCEDURES AND SCREENING GUIDELINES TO O ASSESS SEISMIC VULNERABILITIES AND' EVALUATE SEISMIC 1 MARGINS OF OPERATING PLANT, 09/28/87 VUGRAPHS
a RESEARCH PROGRAM-(CONTINUED) l 4 FRAGILITIES AND RESPONSE 1 0- RESEARCH TO PREDICT THE BEHAVIOR OF STRUCTURES,. SYSTEMS- 1
~
AND COMPONENTS SUBJECTED TO SEISMIC LOADS CAUSING ELASTIC AND INELASTIC OR NONLINEAR RESPONSE. 1 ! o ADDRESSING LARGE UNCERTAINTY IN THE FRAGILITY OR l FAILURE DATA BASE USED IN SEISMIC VULNERABILITY AND j SEISMIC MARGIN STUDIES. < t o RESPONSE / FRAGILITY DATA'0BTAINED ON STRUCTURES, COMPONENTS, PIPING, AND DAMS AND EMBANKMENTS. O ! VALIDATION o RESEARCH TO VALIDATE EXPERIMENTALLY THE CALCULATIONAL METHODS USED IN PROBABILISTIC RISK ASSESSMENTS, SEISMIC' VULNERABILITY ASSESSMENTS AND SEISMIC DESIGN MARGIN , STUDIES. o- COOPERATIVE RESEARCH PROGRAMS USING LARGE FOREIGN TEST FACILITIES. FRG - PIPING' RESPONSE AT THE HDR FACILITY L JAPAN - COMPONENT DAMAGE AT THE TADOTSU FACILITY l TAIWAN - S0IL-STRUCTURE INTERACTION.'AT THE LOTUNG SITE - 0 09/28/87 VUGRAPHS
--- i 1 1- - _ _
k, R A oo t a s 1 c ( Rit S RN ix t n t n OI O i w s n a r t c e d n e nepeg i gda n) s e u S y ic o p e s su
- H 6
1 dn a r op p G. R . 1le e e d n a TT u p e p Pmi n s lo e s u
- I g D n AA r Bse S es e c t
r A p
. mlp st e n P Ri 4
. WS e P
E n c io t LC UI se wu p 0
. 9ou 1
1 co i g icm eo t nn d ic R (g P ip ip l ia i m c e pne cNseso n S 10 P m e GP L ne .n e i p ic t eRe u ec sed MSe cn r ai e s dt ia i n ai e m s n p t fD a r e e ir S t t s e nt Iir np E RA P ic o k a R e P s e i ac f r IseE t S Ae ct s S e ee a C imdna. " d e l a r C 0. 4e BOr EC o S su t eq n p R iciS r m-t i c es em it ns D E 'p '* t u g n Ari
- e t it a
h lrh at r ig s e f,C nmE S - F e B mn s e r e el vp i lig e eP SM e d'
' . cu r
* ,i
" ip I t Si e t t ss oe d
i t c r l r ha e a CE D S it eoO GC( ieLic em SI ai r L A In" g S t P ( UC PR a E e
- e f
- a . S * = F +
- e *
- e
- e
- V egell .l ,a s :esl e8l,l s s e j e Il il 8 i, a !la !aI'88 m
a r g _ e P M p _ A e e _ R m _ G _ ~ O R e n is ic
' P e e m
- H m6 se m c de s e ,. C . r e n p e R s ee e r ' d n A P. a O m e
, ity
_ E e,
, ru t c
Ms s o , l _ S r 7 t c ru
" j eo is o
E P , " r m F r R e ,, s t c e "
- P e c e
e
'e tn
n fe _ I _ o r
- n o ep e'
Y r ni i" t e mP o* c e w
- p
T at o
- e n E c n m ,
o te *
' D ie s 7., C m*
o* D "* .
, c s s e e F ie ie e. "
c A sc s . me n ic " ro e ne 's*s t
*e i**
m S n e s' e i h" e e . m, _ c e lo w ep 6e '"
- s. s s" s P C
I r - c. i r e
- e e
- e
.M st c e r ~
_ c re S I ic fot eP o r E mm m ee ism S ses m8 _ t e oe rC C eno* P re d, t _ R i* ca* ee sp N _ d" n n ne'* le e= ie ae v" sc d i n e e _ e v _ ..la8eleja!el lj el8 ajj ea8 e
,t8llas asa8 !e8l 8s8 p h C c u r
_ I or e TS e le i G e _ l t me c o SM. E it is s ie S c r a e n o s n iei o o R m t MA d dr c e s it t r g y a r ld s a t k n i R t u ar S I m e s i a l o it r g ems e 8 t R i o ic it ic r g o is R U P r _ OG b a or n H N e s c p f li a n f l h t e P DO r g . S se ma le n _ O DR P s trP Poyr io t Mo aur e kmA sr aM er eg m a s gr in a in it i l s e R rg Q u t iot a c Q u c r e e n en I t R E D H SS I WRe m t N qgE gM b d or n K e sd c ur d hoF ino r r b a nP e i f i ne f AD nda n z n t r . rPF gP n ai c o a g m l a u m g ee n K a NT E ieI Ha ou cd nc G r EnlaioN a S En f e t Ms c im m e P r yes se ne p e ip u qm Q c p iu, q ,, MAom n t n n - o y i n T GA om i S r it imEe a a - Cg C A P Ca it ai mS ncS. t se iSa t sy i P Ea r r _ I E EL h I Rs G ouG pl t mI g ola r l igR I s r Ro g s C Cg d mI r sRo on i t tdS rt e e R o, R R Ro il e P ESiU SNRI aeU CS oe R r Pe a i e EP Pr iAn aP r E PrEP S N, N N NP r aG E R V e O E *
- e '
- S
- e F e + e e e *
- F .
Oa Y G L - O Y - L A _ O N R E @ D M T OE O A HE N E R TK H @ _ EN P R _ MA YN L A A L S N i _ N EO N R L E I I NT @ Y T A GIAA :A :X M _ R 0 7 YR OE @ A MS T T 8 BR T E 9 0 T ME C DU DM 1 H R O EM _ EE 9 U TR M - C T EV I D T NC NT 2 r e MFV? -ET SR EX SE b E R - SO O P
@ EE E N m e @
I M RB RO t p EW
]
PO P e S C T E W - EI EIV M N R E T S R E H EO O T I V O TF RT A M M L O OE R _ NT OO H C B U S C N E R P_ I TT S R J V
- l_
S 5_ A C C A A L _ f_ I L aE P
- ~
P A
- L
SLE y d a r N a e n I r r A o i r v M u o a M M R F O S T L S w o w f UR e s e t nm i SE C v r ae yt v si e R A u e l tr t t v s i isR l F O O oy p g a e se OT d n ie m S F rRne ik N W a f e f gn r OI E n ok n n osn aa T V I i t o oa n oi oMY e A c i sY it sain c TR d u su en sl u umia NE o c s csM ci EE t r sa iae l i SK I n i DM DCSa t EN - RA PY 5 0 5 1 0: 4: 1 2 2 _ 1 5 1 5 1 0 _ A D 4: 1 0: 0 2 i* 1 N 1 1 1 - E G h A E
~
G S y N I g dr y o c G l a z e a R ? o ar u AE d o Fiqu ME h cqd e C K t e ie I N M mr a 8 g MA s il sdn euig 1 SY I n Sosce 0 EE i g e 7 d
- SN r a h I
t cc AA M fhmo i i D M i c n ws EF i m o ee EO i s ies t ak g NT e e f no 1 EN S cat 0 e rYn e WE M h pe e E S gi nm t DS I f o na s S DS iMe s a l g Y E S i r nr s S T aas he a _HWA A CNer r_ s s D E o O h E B N ~ 9
jfl5SP NUCLEAA SYST(MS SAF(TV PROGRAM 60 TRIAL PLANT REVIEW PROJECT TEAM WAS ORGANIZED O PLANT OWNER NRC/NRR PROGRAM NRC/RES PROGRAM NRC SEISMIC DESIGN ) REPRESENTATIVE - MANAGER - MANAGER - MARGINS WORKING GROUP (UTILITY) Co Chairmun: D. WHITTIER P. SEARS D. GUrf J. Richardson (RES) N. Anderson
! I Membiars:
PLANT SUPPORT TRIAL PLANT REVIEW G. Bacchi (UTILITY) TEAM LEADER L. Beratan J.Chen P. Anderson R.C. Murray N. Chokshi D. Guzy C. Grimes P. Niyogi L Reiter I I I I LLNL LIAISON FOR LLNL LIAISON FOR PEER REVIEW EXPERT PANEL SYSTEMS ANALYSIS FRAGILITY ANALYSIS GROUP TEAM R. Budnitz R. Budnitz P. G. Prassince R. C. Murray M.Sohn P. Amico J. Reed A. Cornell J. Thomas J. Reed L. Wyllie M. Shinozuka ! ENERGY. INC. EOE FRAGILITY ANALYST TEAM SYSTEMS ANALYSIS M. X. Ravindre, Leader TEAM D. Moore, Leader M. Quillci Structural Equipment Earthquake J. Young Fracinty Fracility FraciRv
^
P. Heshimoto G. Hardy S. Swan M. Griffin O
aL O I t e t s T E a n PE u q e MN K e n U A d o p S Y a m S a r o AE t c c DI N e a c e NA s c AM e t n ) s a r SO r u e n g p NT t c m o u OC % 0 u i n st c i a / d I TI F 5 t r a ou l e AI s - t n hc i f TC 8 n o tl ea i d 9 c o mc i I E 0 d d MP 0 n e F m e r e I LS d e R a i s vP n s n iL t o d C l e i aC d i s
/
n v rH e n G d o w e s o w E o sg a c E I R m d nn b t o V U g k oi c ne s n E n a l R N i l t u g g t s wdeer s n _ T N 3 i x o s c e i g LA s s L P L A I 0 E N U (s R A R h T S
L V iei ue t e n nn nn u uu u eeeeet t t aaaaaa t E doaoooooOOOOOO l t R hecomDD D D D D 777787 888888 SC 999999 N I 1 1 1 1 1 1 G R g A 1 M 2 0 C I T g M ,AS , 3 S I E CWRg O 0 8 0 6 L & S F l l t 7 aen2
> ed e
7 0 5 s s O mn0o- FP de
- d a ar r e
o S s pd a r g gp pu u e n B o&& m34 r a me f w T L r egn gaw o &s1 n L noc eca C ga , 'a r o oh od r d1e - 4 r seers Aa U 1 t i 2 4 vt r H it e r ee S in U h ci c nt nmE s ot gnfr a t g g a t r n aewaVasbb E a n e a r lt t ol f all r os sr R Fn u l a e ak eod gn ob edir wf caoh l aats s Piy P k aa t t ost cagocako sr nl a c Lmi/nA l t a tEy Gam /ignb lo or nh o b chlate y Coi t ra t oldoagr l t n nf CY f W H DU n e uRcl i gyoh t neneeen ae s t E I t O e n il t of o l er onicgali r t t ccldi a V n L is nl ciegt ppf r lal t d un n _ E R l a- - o a i r e o e e ar ps eee ci sDCWSMSUInRRR e t LN s s T A L P P N M D - - - - - - - - - ER L A I T
SL D s E e N W E D E nl i n
) I I N idoe A V R cu A
t i
.G E E SeG E R L eo f
R N k ead E O a E E u stac e S qU/ Do o W M O K N S hy L t E rg aoin nP o S R A D Eloa t
. T Y N e wdcoF F i
. H E Aihlvt kofd i
al G N T e eu I S I H RMQW A G N I S I M N Y E I N H A T
- M L
s s l E
i o n n g s i s W r a n o p n E w e I V m o m c E c d o i l y i k c oit R m l a t n t c E s a daep i e w pc a t E i r s d o l K f e s a p N o im ai c l A l a t aa m em ee / i Y i r d t d bs ne i E t n N a 2ry a a s ds da s ct n e en I A en i e Pa MD i E ml r e i mlp i l t Mlp O L T od ri oc i b Sg UE f r i fri s a en F L e u g ed pr r e hn i p/ n g t r S P y y e we l yt S E M l g l n un o c uo l l u i vi d n CO fl fh h o CC so st da erg dc sw i sd US eo ch e ct i f i t p e we s SA ct c n nu ou AW ue ua er hs Sm do f l Sp Sis i P_ Y R S A S__ M M U
~
s-
2NLI G RD AE MT N CE L E Q I MM N L l E E SU L / S
/
E C I / T S I I S R S Y SO D O Y L L E P A A E N EN R N A A KE Y Y NE R S M T AB I A E L I YS M T G
@ M S A EA 2 U Y R NH I @
S , S F AW 4 1 E 2 E 3 E ME I b M M M EV U U U HE C L O L O L O TR / V V V N R e *
- O I
M
- T A
T U N N E 2- M U 2 C FO ~ D
i
.h ?
o V RES' STAFF PRESENTATION TO THE ACRS , i
SUBJECT:
THE SEVERE ACCIDENT POLICY IMPLEMENTATION AND l THE ISSUE OF EXTERNAL EVENTS l DATE: SEPTEMBER 29, 1987 q PRESENTER: FRANKLIN D. C0FFMAN, JR, 1 L) PRESENTER'S TITLE / BRANCH /DIV: ACTING BRANCH CHIEF RELIABILITY AND HUMAN FACTORS BRANCH i
, DIVISION OF REACTOR AND PLANT SYSTEMS PRESENTER'S NRC TEL. NO.: 301-492-4609 SUBCOMMITTEE: EXTREME EXTERNAL PHENOMENA 1
O I
4 O. F THE NRC ISSUED THE SEVERE ACCIDENT POLICY-STATEMENT.IN AUGUST 8,'1985 5 0FR32138) AND THE STAFF DESCRIBED THE IMPLEMENTATION IN SECY 86-76 (FEBRUARY 1986) NEll APPLICATIONS EXISTING PLANTS o GUIDANCE ON.THE ROLE o SYSTEMATIC APPROACH OF PRAs FOR THE EXAMINATION (NUREG/CP-L1812) -0F INDIVIDUAL PLANTS I. O (GENERIC-LETTER) o . PERFORMANCE CRITERIA o IMPLEMENT MODIFICATION FOR CONTAINMENT THROUGH BACKFIT POLICY i SYSTEMS o CHANGES IN RULES AND REGULATORY PRACTICES, AS NEEDED (SRP 6.5.2 REV. 2 & 6.5.5) k O 1 1.
O
~'
. INDIVIDUAL PLANT EXAMINATIONS ('IPEs) .,
i OBJECTIVFS AND SCOPE o A SYSTEMATIC EXAMINATION OF INDIVIDUAL PLANT DESIGN'AND OPEPATIONS FDP. SEVERE ACCIDENT HAZARDS o IDENT!FY PLANT-SPECIFIC VULNERABILITIES
- TO SEVERE ACCIDENTS (FOR BOTH CORE DAMAGE AND CONTAINMENT PERFORMANCE),
o UNDERSTAND WHAT COULD PROBABLY G0 WRONG IN A PLANT. l i o IDENTIFY.AND EVALUATE MEANS FOR IMPROVING PLANT / CONTAINMENT PERFORMANCE (VIA HARDWARE ADDITIONS / MODIFICATIONS, ADDITIONS-TO PROCEDURES, TRAINING); INCORPORATE INT 0' SEVERE ACCIDENT < MANAGEMENT PROGRAM, j 0 LEADING SEVERE ACCIDENT. CONTRIBUTORS FOR THE PLANT q o 1 i 2 j
O QUESTIONS BEING WORKED G l t-DOES THE IPE NEED TO TREAT EXTERNAL EVENTS? IF NOT, WHY? l IF YES, WHY? l WHICH EXTERNAL EVENTS NEED TO BE TREATED? WHAT IS THE EXTENT 7 GENERICALLY AND PLANT SPECIFICALLY? !
. O H0w? l WHAT IS ACCEPTABLE FOR A PLANT?
o 1 l f 1 l 1 l 3
S R O T U B I R d T N O CE G B R s a W L AA P N N M O RA ED E G A M A f S E ID U T TE D S E E XR EO SC U O S R O C
$ P O
C S L L U F' H RT S T E N E V V E L L A N A ! R E N ! T R IN E T G I N ! : D N E G E L
Q g t-0 PAST IDCOR AND NRC WORK WAS FOCUSED ON INTERNAL HAZARDS. o RECENT WORK ON EXTERNAL HAZARDS IS SUBSTANTIAL. o CURRENTLY WE BELIEVE THAT EXTERNAL HAZARDS NEED TO BE CONSIDERED. 1 o GENERIC LETTER PERMITS JOINT CONSIDERATION WITH INTERNAL IPEs. O o REMAINING WORK IS: SELECTION OF EXTERNAL HAZARDS 1THAT SHOULD BE EXAMINED.- l l DEVELOP GUIDANCE OR ABBREVIATED EXAMINATION METHODS FOR THE SELECTED EXTERNAL HAZARDS. 1 L o INDUSTRY CAN: PROVIDE DETAILED JUSTIFICATION FOR SELECTING EXTERNAL HAZARDS l 0 - PROVIDE ABBREVIATED EXAMINATION METHODS
.; 5
l A 8 5 EVALUATI0tl 0F EXTERNAL HAZARDS O OBJECTIVE: OBTAIN EXPERT CAPABILITY AND EXPERIENCE TO ASSESS THE SAFETY MARGIN THAT PAST AND ONGOING PROGRAMS PROVIDE AND TO IDENTIFY PLANT-SPECIFIC VULNERABILITIES RELATED TO EXTERNAL HAZARDS l TASK 1: EVALUATION OF EXTERNAL HAZARDS FOR FIRES, EXTERNAL FLOODS, HIGH WINDS, TRANSPORTATION RELATED ACCIDEtlTS, I (A0448, 9-87) TASK 2: EVALUATION OF SEISMIC HAZARD (12-87) O l TASK 3: CONDUCT EXTERNAL EVENTS WORKSHOP (8-87) ! I 6
c I n _ a k . i c . r _ r a _ G
- d n
e e _ w o 7 L P , S 5\\\N\N\\NN\\\ d i s T r///////// k i c N P E ' V 5 E - x P
= _
L ' 7/////// _ A N R 4 - E E G T\\ P % _ T A M . X A D E E T _
= R N _
O A F C L - 3 P O L A T T\\ P _ N O T f _ O I O T . T 2 U - 5\ P E S T N B I 7// K A U E V R O H E R T T S RE E H N R
" A T
' E O -
O i 1 I P / V M C /',/j////////////////
- - ' - D 0 0 o 0 0 0 0 0 N 7 6 s 4 3 2 1 - E
. G E -
6 wS!$xE L 0 l
+ RELATIONSHIP WITH OTHER PROGRAMS: 1 t
-NRC-PROGRAMS i
NUREG-1150 REACTOR RISK REFERENCE DOCUMENT
.l SEISMIC DESIGN MARGIN' PROGRAM USI A SEISMIC.00ALIFICATION OF EQUIPMENT IN OPERATING
-PLANTS l
USI A-45 SHUTDOWN DECAY HEAT REMOVAL REQUIREMENT USI A-37 SYSTEMS INTERACTIONS IN NUCLEAR POWER PLANTS USI A-44 Q STATION BLACK 0UT USI A-40 SEISMIC DESIGN BASIS 3 SEISMIC. HAZARD CHARACTERIZATION OF THE EASTERN UNITED STATES: PROJECT J l SAFETY G0AL POLICY '! ROBUST TECHNIQUES FOR ESTIMATING THE PROBABILITIES 0F EXTREME FLOODS' i L o , ( 8 ? l L.. .
~6 n FIRE RISK SCOPING STUDY G
t-INDUSTRY PROGRAMS. IDCOR PROGRAM S0G EASTERN SEISMICITY PROGRAM EPRI SEISMIC t1ARGIN PROGRAM O i l l i 9 w
'(
G N U O Y A N N A E H M T O A N N W E O E H J I P V D N E L N O R A A T N G S K R , N N S E 7 I I I A T 8 C H G T X 9 I S R : E 1 L A A S O I W T E M I , U S M 9 Q C Y D E 2 T I L E R N O Y C M A T T R B R E S N N X E K I A E E A E S B M , S S E N M S M R O E , E E E P T S C E T E P E I K S E E N V N Y T S O R A S T J E Y I S M E E M N I N O A E I C I A B D M U , S E S R R O C O A M D I V A D
- )
(
?
S T L U S S E T R H G D I N S A N I _ T N Y _ E G M O S S L _ I S O S E D Y S O 2L S H - A A T EN E MA E M _ O U LS OM S I S C N E D N VE Y U A T N L Q 9 O Y A E G - S I G N S N S _ Y T O A I E - S C L T R C U O s N E N D D M E E E s _ O O E D N R E R H T I I E C - T T S C G F I N E Y C N E D _ I M S A E R N E - P
. . . . . . P 1 2 3 1 1
5 6 A - t O o
p
~ Start Time axis'~
O -1 Sele = an earthquake review level:
' Gather information . , .. , . ,'. : . , I Gather information on -
en systems and sort , ,. . . .*. l l , ' the plant. Determine which-Group A functions. Use information
.... 2
. . . , . . .* . . .. /. Interaction
/
[. broad classes or -orcups of
- - . - - - - - - 3 componenu have HCLPF -
en Table 2.3 and ';.;.. values greater than the review
/
R ef.1. "*/*.*.*. / level. Possibly identify plant- -
/ 'enique features.
. o.
[ , 'g First plant walkdown:I*'y Concentrate on identific
'./ t l Emphasize systems interaction.' Confirm
//h .y- applicability of screening tools. Complete identification of plant-unique features. , ,
KEY: Task is'pe-formed by: . .. f T.** * .l.
. : ,' Revision of systems.
. . . . . * - relationships es ablished .
*. .!* .*. : 5 ; *. , - -
;: .l.*. Systems Analyst
, .", ,/. * ., ; , , , 7 in Step 2. Develop fault
*. . .* /. *.,. ,,.* .*f trees and event trees.
! Fragility Analyst ,
Y I
**/.*g.:
Both Second piant walkdown: Primarily fragility analyst for checks.
- 6. . Colle= specific data (size and other I physical characteristics) of componenu recuiring detailed analysis. <
v ' l
.. ..* . : . . . * :* . . /.
/
Determine minimal ... * . * . * . * / . *. Finali:e HCLPF value for cut seu for end- */,,**.***** g componenu in final cut sen point core melt. *
.. t,,*,.. !,...;7, .,. ..;, ,' l
* -l (componenu not screened out). ;
j { v Margin asse= ment 1 complete l Figure 2-1. - . Graphic. Representation.of the Screening Operations. (Figure 2-6 from.NUREG/CR-4482) 2 f 1
)
) D S E L E A L N B R
E D T N N A I ) DI D E S N E P ) A Fl l C ( ( C S S S D MN ) M R) E O P , E CW T I EC T ( F S T OC S E Y C P Y l l / S E ,( S EW O A TF F N SA EI Y ( I J P C R D E T R S L Y , A U R E T PW) O N ER E C S R O T F A GW G I A E A ( NS T W R S I ( C D U R L - E E S E E O R T E S R W O E _ OF E U OC T R R S P A _ P YP S TW N _ R E C N O _ R A Y R I E E I O I RP R N C T T L A T O I A C I M l i C P V U E A X I G E M R T N E U R I L O E C RAPl ECSA l I L T
- R T_
N ee e e O. e e e e P O P R U F S O l l
1 S 2 O C T Y
- F T I 0 G E f
P 2 l F A I H 1 L S P 1 1 0 c t E E E 0Y0 E G G CR C R A R A 1 CS T P P C S E P E T l 1 C S C O E Y C NHS P R l I i _ S R P T 1 t H P 0 - R F R X X0 _ F O 4 _ 1 - _ W E - P _ O R P D _ L 0 8 U M _ F 2 1 L P - _ - I I M I O S P R U P _ N P H F P H _ l 5 M _ O E _ R R E c _ F H R R R U P NA T L _ O - s I I N A R T I R A UR F T c CT F A T T C I S
~
C A 1 t M 1 RS - S R T _ R S I S DI I E F H _ H S S S O R { i _ 5 1 t _ 2 o 1 1 R R _ S F ! t M U R O B HR T R F P
- sD I N R cR 1 L T I F R - A -
E S C _ U R T WU D T T C O B B < C A _ P - A S - C P R S OD C 0 I I _ HU l D H S S l 0 8 1 B _ S 4 8 7 _ 5 S 7 C M U M C O S B > O & R U sF R F B P 8 P cE _ L 7 L RC R M@ _ E R C C C WU A WC R O B O 1 t B _ P 0 - P - _ 3 P C P G 1 O O R C H4 P HU R l l l l W O _ L 0 L 5 R U M M O O - R X F W 9O F T E5 0 XO V O R -1 O L 3 4 U 1 i t F 2 1 L - W -
- I I I O I OS P RS L P HP l
l F HF I 1
i s y 999 9 - - 99 92 9 - 77) - - 4 l
? 333 3 - - 3 - - 33 - - 21 9 - - - a e . . 5. E 5 . 3 E n u 000 0 0 00 000 00 0 a l >>> > > >1 > > 0 a l f > 1 e V ( h t
Al l n i d e n i a m e r 3 l f 3 A1 C l ! TT5 1 1 1 l f 90 7 66 i5! 1 SS2 11 - t 000 3 34 1 1 1 R 2! l i K l F n 1 23 333 A PT T P T i PDT A e
- - - - - - - - - P -
P - - - - n l XXX X X X X XX - F - QQQ C o t XXX X X X X X F p
- - X - - XX L - EEE - -
. n - - - - - - - - - - - - X -
X
- - - - - - m e e VVV VVV -
- - V - VV -
P -
- KKK - - F o l v 000 0D0 0 CC P D P l l l C c b L AAA AAA A PP lD T lD il TIT i
C a - - - l l l l l
- - l
- - - f f
- - - I
- e t iii ii i l l l - l -
l lli i1 - l h FFF FFF i ii l ilL F t F F - i F FF FFF A w AAA AAA AA F AAA e A A A A d i n v a r e , v g o n s i g V a n D c pj t A d n e r c s r a c f r s ) p
)
p rr s p oe ii l i m m aa m ea F u u u ml A D D P 4 f f p i
) 5 oo t e a
m mm aa 2 ss r o l
- eb
. e ee P ss f h a 3 t S
t t eo tl 3 c S S. ) P D l l e r t a i i A1 C 1 . .. F T nn u av l e g 555 222 t e r mx t u P L h ou l i e o. b 1 - - - A AA C t nn a l a PPP ( (( i i ee f b a aw l T g w F pp n ooo A AA h P oo e l i FFF FFF t t t - F - FF F gF d L s i t n PP e u a PP P. Ll P. tPt tPP pip e L t L r FFl i lCl l l a Cf looLL LLI l l l l loL I vh e CCC CCC aaa f l f CCI C(C e PP aa C at r I II I I !
! I rrr II I I I d LL nff mI I
c I I I I ggg p ppI I i CC w n tF S hhh hhh eee uh uuhh h oh s I I I I oss o oP ggg ggg t t t ogoogg gd g n nVV m nL il i iii nnn rirriiii i ni o wwk0D m il l I i l liii l l I I I GiGGl l l i al l i C oonAA o sl
- - - - - - - - - - - - - - - R - - LLU - - C a ttt t t t t t t t ttt t t t - t t - - - t t - we uius uau uuu uuuuuu unu u nnnuu n ODO OCO DOO OOOOOO DI O O I I I OO O Ft P
L s Ci l l ai t e ah
)) ht
) . . t n mm muu s n ucc ea ccc t p t 1 1 1 890 cAA a n 000 334 231 583 A((
( cn e 1 23 333 676566 i i n - - - - - - 1 1 1 21 1 l 3 m d o AAA AAA Al C t - - - - - - A1 C 1 61 5l 2 e ne p - - - - - - 666 AA1 1 1 PT 1 555 1 22I 1 t i m l i l i il l i i i l l 888 - - - - - - 222 1 - - - - - s o F FF fF F - - - SS55SS - - - - KK KK y A a C AAA AAA EEE I l I I 1 1 1 1 l I i I PPP T T T U: 1 I S II
T N E M L S E E S S V E A S E S E C E L S S C L O A E C AA CC AL O N E U Q N S O E E E V R E TL N C E I E O L S T T T U L L A N N Q S A E U E E E OM G Q V L S- E NS A E EP R M T A N
- - D A - -
T E N E L E R O V O O E C B e e-e O I
l r e e c h n t r e u u F q g e D P S L I d g g e T T t a lu a v E t K K D D o s O O C C l f u t 1 2 3 4 5 t a S R
/
I S D P l l _ d e e O l B s 1
& nP doi et ec FA W
f E '
/L W:
M l a rit c oi . t r _ ccC ab ; eu s RS s e e r c c l u . i u a S F s c .
,q imp .
i SOT eO
- SL O.
,i 5 -
( l r
e c r n e e h u q t r e 2 u F S p D P
) J d e
S S S t o lu o v s E u t D K D D o K _ t _ t a O C O C C i t s 1 2 3 4 5 6 R _ / 0 I S _ P I I
) 2 p
_ n _ i o _ c _ A W F _ E
/ L W
F A _ l _ a ri t c oi _ t r _ ctC ab _ eu _ RS s _ s e e _ r c u P c il O u a _ S F _ O L - '
~
c, 2 S imA sC
)
ieO
% SL 'g i' F,
s Table 4-7. Boolean equations for no LOCA and small LOCA accident sequences. a
-i f
No LOCA Case Core Damage = (4+20) + (8+15 + 17+22) + 3 * (14+16) + 7
- 15 i
i l l Small LOCA Case i Core Damage = 4 + 7 + 20 ~ I
.s 1
I 1 i l l l l 4 { W L I 5 M -e
, 4 13 i %
s
i
. I i . 4 Table 4-1. Basic event descriptions.
Nonseismic Number Event Description- Probability 1 CCW-HTX-EQ-4B5A Seismic failure of PCC/ SCC coolers SF j 2 AF W-TNK-EQ-TK 25 Seismic failure of air tank for AFW SF -l AOVs j 3 AFW-XHE-FO-EFWXX Failure to align PWST to AFW MDPs 2.8 E-01 l 4 ACP-TFM-EQ-57X68 Seismic failure of station transformers SF l 5 HPI-CCF-FC-HPSI Nonseismic common cause HPSI failure 2.1E-03 ! 6 P PS-XHE-FO-FDBLD Failure to open PORVs for feed and 5.0E-02 bleed ! 7 HPI-TNK-EQ-RWST Seismic failure of the RWST SF l 8 AFW-TNK-EQ-DWST Seismic failure of the DWST SF ! 9 PPS-CCF-FC-PORVS Nonseismic common cause PORV failure 8.0E-04 10 PCC-XHE-FO-ISOL Failure to close PCC isolation valves 1.0E-01 11 OEP-PSF-FC-DG1B Random failure of diesel generator 1B 6.7E-02 O 12 OEP-PSF-FC-DG1 A Random fallure of diesel generator 1A '6.7E-02 13 AFW-XHE-FO-TRBLO Failure to start AFW TDP locally 1.8E-01 l 14 OEP-CCF-FC-DGN Nonseismic common cause DG failure 2.6E-0 3 15 AFW-CCF-FC-AF W Nonseismic common cause AFW failure 2.0E-04 16 OEP-XHE-FO-F UEL Failure to refill DG fuel tanks 1.0E-02 17 AF W-XHE-FO-TRBMC Failure to start AFW TDP from MCR 5.0E-02 18 OEP-TNK-EQ-TK 62X Seismic failure of DG day tanks SF l 19 CCW-ACX-EQ-C HILL Seismic failure of air cond. chillers SF j 20 SWS-BKW EQ-CIRC Seismic failure of cire. water pump SF l house
]
21 AFW-TNK-EQ-PWST Seismic failure of the PWST SF i
)
22 AFW-TD P-LF-P25B Random failure of the AFW TDP 5.0E-02 _. j SF indicates that the event is a seismic related failure, and a HCLPF will be determined 1 by the fragility analysis team. i O i M-d 4-7 'i i ____ A
{ll ,
- ~
O "
?'
Q D E T A R O
) S P
S E M R T T E O l i A T C G_ L S I Y N S S S I N V Y. I E v S T Y T G - Y. L A R , G H R N S g O G A R M S E L I S ( E E E S N D O N T T E R L S C E 0 I Y G E A Y O l i T N T S R L A D P O I T A I E N L L W N T M A O E A R N O C O N R OO N D I T I C O I S P I O N T I F N P T I S A A I E T O U A T M R R R E D I S U ~C S E G E C O L N T L A G T N P B C PE A C D A R N S I O U A N V E I Y R D S Y D N A E T T S E C - O Y N E E N R N Y I R Y T I E T N A O E R T E R E H R I G Y T G E A V A S. E O G N G C R T L O M A E T O C S M P N I T O A P E I O E M A S IE R U N L E U R L R E B R P C O _ E O I C T S E D T N O - - - - - - - U - - - I H C G T E N E X E N E 1 2 3 O (L l
~
O 4 iC C w W E I V _ E R S K N S S I A I I G T P R A A S E S R M I R E G S O C W C Y O I E V O I V I M S I L A N A M D I R E S P E V U S S A I K M D E C E E A E T B K S N Y A S Y E N I A M D C E
! lt L
1
?
b, Table 2-1. Definition of plant safety functions. l (Table 2-2 from NUREG/CR-4482) ) ICI'CIFICAT:CN CF SATI"Y PONCT CNS ' l
- 1. Reacter Sub:riticalit? - shutting dcwn the nuclear reae icn such that the only hea: being generated is decay heat. I
- 2. Nermal C:cid:wn - p cviding c cling to the reacter c:re through the use f of tne n=:=al power c nversien system, nor= ally defined as the main steam, turbine bypass, c:ndenser, cendensate, and main feedwater subsystems. l
- 3. Emer:ener C:re C:eline (rariv) - p cviding c eling to the react =: c::e I in ne es:17 (transient) phase of an even sequence by the use of one l or more e=e:gency systems designed for this purpose. The enac: timing i of "early" is sc=ewhat plant specific and sequence dependen . Ecwever, l for our purposes it can be deemed to be the time peried during which !
l
-these syste s see initially called upon -t0 eperate.
- 4. Emer:en:r C:re C eline (Late) - p;cviding ecoling to the reactor core in the late (stacill::ed) phase of an e' rent sequence by the use of cne !
or mere emergency systems designed for this purpose. In centent with the above definition of "early", for our purposes " lace" can be deemed ; fg to begin with the switchever to re irculation (fer LOCAs) c: with the Q achievement of residual heat re== val conditions (for transients) .
- 5. Centainmen: Hea: Receval_ - re=cving heat f:cm the contai=ent to the i
I ultima:e heat sink during the late (stabili:ed) phase of an even sequence by the use of one c: mere safety systems designed fer this . purynse.
- 6. Cen si=ent overpressure Protectien - (2r.1v) - controlling the buildup of pressure in the cenesi=ent caused by the evolution cf steam by j L
condensing this steam during the early phase of an event sequence by l using ene c mere safety syste=s cesigned for this purpcise. "Early" in i the centen of contai=ent functions is not the same as "early" f : core c cling. In this case "early" is cee=ed to be the time peried ce=encing when this function is required after the beginning of cere melt when these systems are operating in the injection mode.
- 7. Containment Overpressure Protectien (Late) - cont:cIling the buildup of 1 pressure in the centainn.en: caused by the evolution of steam by ,
condensing thi:: steam during the late phase of an event sequence using ! cne or =cre safety systa=s designef. for this purpose. In the centent of the previous definition, " late" in this case is deemed to sesr: when these systems are operating in the recirculation mode. l IDE!CIFICATICN CF 'iEE FUNC""CNAL GROUPS FOR PWRS AND BEPS. PWR Group A: Functions 1,2,3 j G cup NCT-A: Functienc 4,5,6,7 + All plant functions net related ; y to Safety BWR Group A: Functions 1,2,3,4,5,6, 7 G cup NCT-A: All plant functions no: related to Safety 2-10 _____ _ .)
)
R l Di _ S A ( _ L _ ) A I V _ S O _ P M _ D L E ( R _ E D _ U N T - L O ) A _ C I T E _ N T I L H _ I
) C S O T E ( R Y o A
_ 'T J T A O B N s N C ( I K O E N N C D S R Y A E T T E M N F W E R E S F N T U O S N A O S D N < A S S _ Y I T W S. R T E U O T E C _ R H D - R E P S L A D U J O - I S N Y E O T C S I R T C O A E A A N R Y D N L C P T N R A p I E O E M u R w F C T R o O o A E L O R B L S S A N G e e e e e e e O
EM D oo D o
, m I
l Sal rl - T N ,I l N - fiC
- Ul i ! , I ! I iii,!' ,I a l 8 l ii' 'l, g a , ,g i l l i! it 2 - 1 2 2
1
- 1 1
2 2 3
- - - 4 4J 4 1
4H U0 b t
- u O lp OL 1O k
- O!kt - I S
S I
- its I
5 1 l 5 I g 1 H i , n i2 f, f H f, R [" , P
>- "r
'fo I 7 .
s [*i O c_ V I l 0 4
- 4 2
4 3 4 _ - 1 - - - N N 1
- u u u
s 1 p 5 i t i s is s _ 1 P m u h '! !8 I 3 1
- i l
- u gl 4 l l 3 4 d 4 o 1 - O!k 0
_ t t 0 k b 4 _ ra n 7 1Ar - gI 4- - tl _ n 1 '~ l t oo F l C WC E - l o l.o tc n"~ m , i SL 1 ! c t x x t t v^ I Q' - - 1 } p l x x
- I l t l
X3 s r N B i X D j e
- g 4 -
e n 1 M a S t h - 4 s c P 1 lax 8 ) mid o uE sa t j P 3 r ea l l FRl 6 4 9 5 5 5
- - - I U
E < I S 1 1 1
- [( ;
58 H - ( i g! 4 gl a 1 3 g s k b 1 x u i x t t yi G lsO i(- l E ) y-g' i G EAl K l l E J Rt A - - fTA O T R^l l T O - E S /~ 3 R 0 3
- 1 8
1 4 5 OIk 1 1 [s t t x S nT K T s s o$
\
Table 3 5. AFT system cut sets. Nonseismic Ranked Sincle Faults Probability AFW-CC F-FC-AFW 2.0E-04 Nonseismic Ranked Double Faults Probability AFT-TNK-EQ-PWST AFW-TNK-EQ-DWST SF ACP-TFM-EQ-57X63 AFW-TNK-EQ-DWST SF CCW-HTX-EQ 4BSA AFW-TNK-EQ-DWST SF O CC W-ACX-EQ-C HILL SWS-BK W-EQ-CIRC AFW-TNK-EQ-DWST AFW-TNK-EQ-DWST SF SF OEP-TNK-EQ-TK 62X AFW-TNK-EQ-DWST SF AFT-XHE-FC-EFTXX AFW-TNK-EQ-DWST 2.SE-01 AFT-XHE-FC-TRBMC ACP-TFM-EQ-57X 68 5.0E-02 ACP-TFM-EQ-57X 63 AFT-TDP-LF-P25B 5.0E-02 CCW-HTX-EQ-4BSA AFW-TDP-LF-P25B 5.0E-02 C C W-ACX-EQ-CHILL AFW-TDP-LF-P25B 5.0E 02 S WS-BK T-EQ-CIR C AFW-TDP-LF-P25B 5.0E-02 OEP-TNK-EQ-TK 62X AFW-TDP-LF-P25B 5.0E-02 CC T-HTX-EQ-4B SA AFW-XHE-FO-TRBM C 5.0E-0 2 CCW-ACX-EQ-CHILL AFW-XHE-FO-TRBMC 5.0E-02 SWS-BKW-EQ-CIRC AFW-XHE-FO-TRBMC 5.0E-02 OEP-TNK-EQ-TK 62X AFW-XHE-FO-TRBMC 5.0E-02 OEP-XHE-FO-FUEL AFW-TNK-EQ-DWST 1.0E-02 OEP-CCF-FC-DGN AFW-TNK-EQ-DWST 2.6E-03 CEP-XHE-FC-FUEL AFW-TDP-LF-P25B 5.0E-0 4 AFT-XHE-FO-TRBMC OEP-XHE-FO-FUEL 5.0E-04 OEP-C CF-FC-D GN AFW-TDP-LF-P25B 1.3E-04 C'EP-CCF-FC-DGN AFT-XHE-FO-TRBMC 1.3E-04 SF indicates that -he cut se: events are seismic f ailures, and their HCLPFs will be provided by the fragility analysis team. O M-c/ ' 3-23
i. , . . :1 1 l ..) l \ l . Table 3-7. HPSI system cut sets. 1 i
~
,1 i
Nonseismic Ranked Sincie Faults Probability ACP-TFM-EQ-57X63 SF HPI-TNK-EQ-R TST SF CC T-HTX-EQ-4BSA SF CC W-ACX-EQ-CHILL SF 1 STS-BK T-EQ-CIR C . SF j OEP-TNK-EQ-TK 62X SF j CEP-XHE-FO-FUEL l CE-02 CEP-CCF-FC-DGN 2.6E-03 CEP-CCF-FC-HPSI 2.1E-03 l 1 l 0 . Nonseismic Ranked Double Faults Probability j a
= l PCC-XHE-FO-ISCL OEP-PSF-FC-DG1B 6.7E-03 OEP-PSF-FC-DGIB CEP-PSF-FC-DG I A 4.5E-03 l SF indicates that the cut set events are seismic failures, and their HCLPFs will be ,
l provided by the fraSility analysis tesm. ' l l M-,:/: 3-26
l
% l (d Table 3-9. PORY (no LOCA) cut sets. I i
l I l l Ranked Sint!e Faults Probability CCW-HTX-EQ-4BSA SF j CCW-ACX-EQ-CHILL SF SWS-BKW-EQ-CIRC SF OEP-TNK-EQ-TK 6:X SF i ACP-TF.\1-EQ-57X 63 SF PCC-XHE-FC-ISOL 1.0E-01 OEP-PSF-FC-DG I A 6.7E-02 OEP-PSF-FC-DG 1B 6.7E-02 PPS-XHE-FC-FDBLD 5.0E-02 OEP-TdE-FC-FUEL 1.E-32 OEP-CCF-FC-DGN 2.6E-03 i PPS-CCF-FC-PORVS 8.0E-04 l No Double Faults r SF indicates that tne cut set even:s are seismic failure!., and their HCLPFs will be provided by the fragility analysis team, i l
]
b a hl-c/h 3-2S
I-Table 3-10. PORY (small LOCA) cut sets. 1
' Nonseismic Ranked Single Faults Probabili v -.
CCT-HTX-E Q-4B5A SF
'CCW-ACX-EQ-CHILL SF STS-BKW-EQ-CIRC SF OEP-TNK-EQ-TK 62X SF ACP-TFM-EQ-57X63 SF PPS-XHE-FO-FDBLD 5.0E-02 CEP-XHE-FO-FUEL 1.0E-02
( OEP-CCF-FC-DGN 2.6E-03 PPS-C CF-FC-PORV5 8.DE Nonseismic . Ranked Double Faults Probability PCC-XHE-FO-ISOL OEP-PSF-FC-DC IB 6.7E-03 OEP-PSF-FC-DG I A OEP-PSF-FC-DGIB 4.5E-03 SF Indicates that the cut set events are seismic f ailures, and their HCLPFs will be provided by the fragility analysis team. l-l l l l M-c/3 - 3.gg L-___ -- - _ - ___-____-__-__ - _ _ _ _ _ _ ._ _ - . - _ - _ _ _ _ _ _ _ - _ _ _ .
] , j 1
Table 4-2. Sequence Tg LPi cut sets. 1 No Single Faults Nonseismic Ranked Double Faults Probability ACP-TFM-EQ-57X63 AFW-TNK-EQ-DWST SF CCW-HTX-EQ-4B5A AFW-TNK-EQ-DWST .SF CCW-ACX-EQ-CHILL AFW-TNK-EQ-DWST SF SWS-BKW .EQ-CIRC AFW-TNK-EQ-DWST SF - OEP-TNK-EQ-TK 62X AFW-TNK-EQ-DWST SF ACP-TFM-EQ-57X63 AFW-TDP-LF-P25B 5.0E-02 AFW-XHE-FO-TRB MC ACP-TFM-2Q-37X63 10E-02 CCW-HTX-EQ-4BSA AFW-TDP-LF-P25B 5.0E-02 h CCW-ACX-EQ-CHILL AFW-TDP-LF-P 25B '5.0E-02 SWS-BKW-EQ-CIRC AFW-TDP-LF-P 25B 5.0E-02 OEP-TNK-EQ-TK 62X AFW-TDP-LF-P25B 5.0E-02 i CCW-HTX-EQ-4B5A AFW-XHE-FO-TRBMC 5.0E-02 J O CCW-^Cx-Ea-CHitL SWS-BKW-EQ-CIRC
^FW-xHE-FO-TaBuC AFT-XHE-FO-TRBMC 5.0E-02 5.0E-02 -"
OEP-TNK-EQ-TK 62X AFT-XHE-FO-TRBMC 5.0E-02 OEP-XHE-FO-FUEL AFW-TNK-EQ-DWST 1.0E-02 OEP-CCF-FC-DGN AFW-TNK-EQ-DWST 2.6E-03 OEP-XHE-FO-FUEL AFW-TDP-LF-P25B 5.0E-04 AFW-XHE-FO-TRSMC OEP-XHE-FO-FUEL 5.0E-04 CCW-HTX-EQ-4B5A AFW-CCF-FC-AFW 2.0E-04 CCW-ACX-EQ-CHILL AFW-CCF-FC-AFW 2.0E-04 SWS-BKW-EQ-CIRC AFW-CCF-FC-AFW 2.07.-04 i OEP-TNK-EQ-TK 62X AFW-CCF-FC-AFW 2.05-04 ACP-TFM-EQ-57X 63 AFW-CCF-FC-AFW 2.0E-04 OEP-CCF-FC-DGN AFW-TDP-LF-P25B 1.3E-04 AFW-XHE-FO-TRBMC OEP-CCF-FC-DGN 1.3E-04 1 PCC-XHE-FO-ISOL AFT-CCF-FC-AFW 2.0E-05 l OEP-PSF-FC-DG IB AFW-CCF-FC-AFW l.3E-05 i ' OEP-PSF-FC-DG I A AFT-CCF-FC-AFW l.3E-05 PPS-XHE-FO-FDBLD AFW-CCF-FC-AFW l.0E-05 OEP-XHE-FO-FUEL AFW-CCF-FC-AFW 2.0E-06 OEP-C CF-FC-DG N AFW-CCF-FC-AFW 5.2E-07 PPS-CCF-FC-PORVS AFW-CCF-FC-AFW l.6E-Oi. j 1 SF incicates that the cut set consists entirely of seismic relatec f ailures f or wnica HCLPFs will be determined. O M-c/5 4-8
4 ,. l l s Table 4 2 Sequence S2 D cut sets. d Nonseismic I Ranked Single Faults Probability ) l l ACP-TFM-EQ-57X63 SF l l HPI-TNK-EQ-RTST SF l CC W-HTX-EQ-455A SF - CCW-ACX-EQ-C HILL SF l 1 SWS-BKT-EQ-CIRC SF .l OEP-TNK-EQ-TK 62X SF ! OEP-XHE-FO-FUEL 1.0E-02 l OEP-CCF-FC-DCN 2.6E-03 O set-cce-FC-seSt 2.1E-OS Nonseismic Ranked Double Faults Probabili v PCC-XHE-FO-ISOL OEP-PSF-FC-DGIB 6.75-03 OEP-PSF-FC-DGlB OEP-PSF-FC-DG I A 4.5E-03 SF indicates that the cut set consis:s enurely of seismic relatec f ailures for wnicn HCLPFs will be determined. O M-:/7 4-10
p z O ; 1 SEISMIC MARGIN REVIEW OF NUCLEAR POWER PLANTS i PANEL TO REVIEW THE EPRI METHODOLOGY (Slides Prepared for ACRS Subcommittee Meetine. 9/29/87)- I l I ROBERT J. BUDNITZ l l Future Resources Associates, Inc. ; 0 2000 Center Street, Suite 418 1 l Berkeley, California 94704 (supported by NRC, working through Lawrence Livermore National Laboratory) O
[ .)
-e 2 REVIEW PANEL MEMBERS Robert J. Budnitz (chair)- Future Resources Associates, Inc. 1 Paul J. Amico, Jr. Applied Risk Technology Corporation Michael P. Bohn Sandia National Laboratories William R. Mills Nuclear Engineers & Consultants Inc.
John W. Reed Jack R. Benjamin & Associates, Inc. Bernard F. Saffell, Jr. Battelle Columbus Laboratory l J Loring A. Wyllie, Jr. H.J. Degenkolb Associates, Inc. h PANEL's OBJECTIVES
- 1) "...to review and critique the EPRI seismic margin methodology"
- 2) "...to make recommendations concerning the general applicability of the EPRI methodology for determining both . seismic margin and vulnerabilities" PANEL's APPROACH o Review the EPRI methodology document O o ' Review the trial application at Catawba
t
-3 REVIEW PANEL's CENTRAL FINDINGS- _ l l
o EPRI methodology is capable of meeting its objective of determining the HCLPF-based seismic margin of a nuclear plant 1 o EPRI methodology is implementable o EPRI methodology is reasonably accurate 1 I i l l l 0 KEY LIMITATIONS o only finds margin for the success path (s) analyzed o non-seismic-failures and operator errors are treated only rudimentarily o success-path approach is more prone to errors of omission than is the fault-space approach l o success-path apprunch gives less insight into " vulnerabilities" O than aoes the rauit-space approach l C -}}