ML19345D112
| ML19345D112 | |
| Person / Time | |
|---|---|
| Issue date: | 12/03/1980 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-0803, NUDOCS 8012090094 | |
| Download: ML19345D112 (42) | |
Text
.
1 1
UNITED STATES OF AMERICA 2
NUCLEAR REGULATORY COMMISSION 3
O 4
PUBLIC MEETING 5
SUBCOMMITTEE ON RELIABILITY AND PROBABILISTIC ASSESSMENT 6
7 8
Nuclear Regulatory Commission Room 1167 9
1717 H Street, N.W.
Washington, D.C.
10 Wednesday, December 3, 1980 11 The Subcommittee met, pursuant to notice, at 2:00 p.m.
12 COMMITTEE MEMBERS PRESENT:
13 DAVID OKRENT, Chairman O.
14 WILLIAM KERR 15 JESSE EBERSOLE 16 NRC STAFF PRESENT:
17 FRANK ROWSOME 18 GARY R.
QUITTSCHREIBER i
l 19 l
20 21 22 23 O
24 25 www T
ALDERSON REPORTING COMPANY,INC, l
400 VIRGINIA AVE., S.W., WASNINGTON, D.C. 20024 (202) 554-2345
2 1
E E 0,,C E E D I E q E 2
CHAIRMAN OKBENT:
Okay, we are now in open session.
3 Th e meeting will now come to order.
This is a
'(/)
4 meeting of the Advisory Committee on Reactor Safeguards
~
5 Subcommittee on Reliability and Probabilistic Assessment.
6 I am David Okrent, Subcommittee Chairman.
7 The other ACRS members present today are Jesse 8 Ebersole and William Kerr.
g The purpose of this portion of the meeting is to 10 discuss the use of probabilistic assessment in upcoming 11 rule-making procedures, how to modify the single failure 12 criteria, and a discussion of the Interim Reliability 13 Evaluation Program.
()
This meeting is being conducted in accordance with 34 15 the provisions of the Federal Advisory Comm'ittee Act and the 16 Government in the Sunshine Act.
17 Gary Quittschreiber is the Designated Federal 18 Employee for the meeting.
19 The rules for participation in today 's meeting 20 have been announced as part of the notice of this meeting 21 previously-published in the Federal Register on November 18, l
1980.
22 23 A transcript of the meeting is being kept and will
! ()
he made available as stated in the Federal Register notice.
24 It is requested that each speaker first identify himself or 25
)
1 l
ALDERSoN REPORTING COMPANY,INC, i
I j
400 VIRGINIA AVE S.W. 'VASHINGToN, D.C. 20024 (202) 554 2345 l
3 1 herself and speak with sufficient clarity and volume so that
{;
2 he or she can be readily heard.
3 We have received no written statements or requests 4 to make oral statements from members of the public.
5 Okay, I guess, one item relates to the use of 6 probabilistic assessment in upcoming rule-making 7 procedures.
Do you want to comment on that?
8 MR. ROWS 3ME:
I have little to add to what I said 9 in the closed session on upcoming rule-makings, with the 10 possible exception of the work we're doing that micht be of 11 relevance to the minimum engineering safety features 12 rule-making.
And I did not do a thorough description of 13 that program at that time.
14 A package of programs we intend to pursue, if I
- 15. can market it to user offices, is to explore the middle 16 ground between the Band-Aid or add-on approach to the 17 present regulatory requirements and the opposite polar 18 extreme in which a criterion of acceptable risk renders all 19 the other regulations moot, one in which it is necessary and 20 sufficient to comply with a probabilistic risk measure of 21 acce ptability.
22 In outline, the program entails, first, some 23 methodology development in evaluating hypothetical f amilies
()
24 of regulation with respect not only to risk limitation but l
25 also whether or not it is an efficient use of NRC resources,
()
i f
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
4 1 whether it is an ef ficient use of industry resources,
)
2 whether it provides ef f ective avenues f or the consideration 3 of public interest and public input into regulatory O-4 decision-making, whether or not it can accommodate policy 5 guidance from either Congress or the White House or from the 6 Commissioners themselves to adapt to perceived needs and 7 senses of urgency about the use of nuclear power in power 8 generation, energy, energy policy issues, whether it is g compatible with the present regulatory structure, whether 10 you can get there from here or not, whether you would make 11 orphans out of the class that have been licensed under the 12 existing requirements, practicality and so forth.
13 Another facet of this family of research programs
()
is to examine the current regulatory practice, particuladly 34 15 the current regulatory documents, the general design 16 criteria, th e Eag Guides relevant to accident prevention, 17 and the standard review plans, analyze what they seem to 18 miss in the way of risk-significant coverage and where they 19 ma y be deficient in terms of staff registry resource 20 utiliza tion.
21 A third facets to look at what reliability 22 engineering of fers as it has evolved in aerospace and 23 weapons and electronics industries, chemical process 24 industries, elsewhere where reliability engineerino
()
25 techniques have been used, with a particular eye toward O
ALDERSoN REPORTING COMPANY,INC,
{
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345 l
5 1 reliability assurance to complement quality assurance for
{)
2 active components and human inputs, experience feedback, 3 management practices, problem identification and resolution 4 procedures, qualitative reliability techniques --
5 reliability techniques that would be applicable not merely 6 to the safety analysis exercise we go through but to the 7 processes designers and builders go through, the conceptual 8 design stage of a plan, the detailed design stage, the g design freeze, the procurement decisions that are made, 10 construction sequencing, start-up testing, operations and 11 maintenance.
12 And finally, a fourth element of this program 13 would be to develop and evaluate a wide range of options for
()
14 combining deterministic regulatory requiremen'ts with 15 risk-based or reliability-based criteria or requirements, 16 which could fill in some of the deficiencies we identif y in 17 the present regulatory process, among them those noted by 18 the ACBS single-failure criterion and the like.
19 I jotted down on a separate piece of paper but do i
20 not have available on slides a check list of some of the 21 options one might consider exploring in item D.
One might 22 be levels of assurance or formally codified defense in depth 23 with engineering margins and deterministic requirements for
()
24 each level, a probabilistic determination of how many levels 25 are require 1, and common-cause f ailure analysis to assess O
ALDERSoN REPORTING COMPANY. INC, j
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345 i
6 1 the independence of the levels; that might be one option.
2 Another, a spectrum of design-basis accident.s more 3 numerous and more varied than the present design-basis 10CA, 4 with reliability and/or redundancy and diversity 5 requirements for the engineered safety features for each, 6 according to the frequency and severity of the sequence 7 branch in this family of design-basis accidents, together 8 with conventional deterministic requirements.
9 A third option: hypothetical standardized 10 engineered safety features designs prescriptively required 11 with reliability and deterministic design bases.
12 Fourth: regulation based upon the emerging 13 requirements for siting emergency planning saf eguard s and (n) 34 containment systems, with everything else, that is, the 15 prevention end, lef t to a formalized procedure with which to 16 arrive at a finding that the licensee is doing a proper job 17 of protecting his investment in the facility.
This concept 18 is predicated on the idea that a utility 's self-interest in 19 making core damage rare might prove to be sufficient to meet 20 public health and safety objectives,- provided the siting 21 containment systems and the like are adequate and, further, that the licensee pursues his self-interest effectively.
22 23 The NRC might be prescriptive about the management
(~V) 24 structure, the style and thoroughness of the safety analysis 25 the licensee is to use, but not be prescriptive about O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
7 1 designs, technical specifications, or procedures in dealing 2 with accident prevention.
3 Fifth dual probabilistic risk criteria, a maximum 4 risk which must not be exceeded even when deterministic 5 requirements are satisfied and a minimum risk for sequences 6 below which the sequences are exempt from deterministic 7 requirements, something like the ten-to-the-minus-se ven a t 8 one end and the minimum that says thou shalt not have core 9 damage more often than ten-to-the-minus-four even if you've to complied with all the deterministic criteria, something of 31 that kind.
12 And sixths deterministic criteria for standardized 13 engineered safety features designs and procedures, with
()
14 procedures based in part on probabilistic risk assessment, 15 together with limiting conditions of operation optimized by 16 PR A, either generically by plant-specific analysis during 17 licensing or based upon feedback from start-up and 18 in-service surveillance tests.
19 Clearly, we could come up with other options and 20 other permutations and combinations, but those are some of 21 the ones I thought up offhand as the kinds of things we 22 would like to look at in phase D of this program.
CHAIRMAN OKRENT:
Now, if I were Commissioner, I'm 23
()
24 no t sure I would think I had heard how probabilistic 25 assessment was going to contribute to the upcoming I
l ALDERSoN REPORTING COMPANY. INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
b 1 rule-making procedures from what you have described, in the 2 sense that I might think that what I've heard was somewhat 3 general, not clear whether it was focused on the 4 rule-making, not clear what the time scale was by which one 5 could espect any kind of definitive or, at least, 6 significant impact on the rule-making, and so forth.
7 So can you help me?
8 MR. ROWSOME:
I can try.
But where we are is 9 about as far as the planning has gone; what I've told you is 10 almost as f ar as the planning has gone.
This is a program 11 that we conceived of in struggling with, among other things, 12 the concept of the minimus engineered safety fea tures 13 rule-making, with which we are not comfortable and we wanted
()
34 to look at other options.
15 So I could free-associate on your question, but we 16 do not have a planning base that I can report to you on.
I 17 think we have to look further afield; we have to, certainly, 18 consider the idea of a standard minimum set of engineered 19 saf ety features and a set of criteria to increment the 20 present regulatory requirements as an absolute minimum.
21 That looks to be the one that's closest to hand in time and 22 in feasibility and minimum disruption of the existing 23 process.
On the other hand, I think, a strong case can be
()
24 25 made that a rather more substantive overhaul of reculatory O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
9 1 requirements is necessary, in part for technical safety n
1 2 reasons -- the deficiencies such as the single-failurc 3 criterion and the numbers of others that we could list o' l
4 systems interactions and others -- and in part fo r 5 administrative reasons.
This agency is choking itself with 6 the work its present body of requirements and its present 7 evolutionary processes has ger.erated, and is at the same a time commandeering a very la rge percentage of the resources 9 that the owners can bring tio bea r ; and it's not clear that 10 these resources are being used in a safety-effective, not to 11 mention cost-ef f ective, f ashion.
And I think just the 12 efficient management of the agency dictates that we look a 13 little bit more broadly at the question of what are we now
()
34 doing that is saf ety-ef fective and what is not and how might 15 we overhaul it.
16 In that larger arena, I'm hard pressed to delineate a schedule except to say that we intend to pursue 17 18 it vigorously in fiscal '82 and will be spawning preliminary 19 insights about the direction in which the program will be 20 evolving and what looks to be promising and wha t doesn ' t 21 look to be promising, and be evolving future plans and 1
22 schedules and milestones as we go along.
CHAIRMAN OKRENT.
But, see, the Commission is 23
! ()
24 under some kind of congressional mandate to develop siting 25 criteria, if I understand it --
n\\J ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W, WASHINGTON, D.C. 20024 (202) 554 2345
10 1
MR. R0W53ME:
That's right.
2 CHAIRMAN OKRENT
-- correctly.
Now, are you 3 going to tell it, " Hold of f final action" --
4 MR. ROWSOMEs No, I would really 5
CHAIRMAN GKRENT:
-- or what?
6 MR. ROWS 3MEs
-- propose that this be aimed 7 principally at the prevention end of reactor safety, and 8 unless a compelling case develops from the internal logic of g the effort, to leave it at the prevention end of reactor 10 safesty.
11 CHAIRMAN OKRENT:
So the --
12 MR. ROWSOMEs That is, within the general scope 13 and envelope meant to be addressed by the minimum engineered
()
14 safety features rule-making.
15 MR. EBER50LE:
May I ask a question at this 16 point?
Frank, you used a phrase that always causes my blood 17 pressure to rise about 40 points: " engineered safety 18 features."
If you looked at a typical FSAR, you'll find 19 that these are defined as systems which are on stand-by for l
20 rare challenges, to cope with accidents.
They are by far the least important safety features at the plant.
The most 23 l
22 important ones are those which are in constant demand, l
therefore suffer the probability of going into disarray at 23 24 any point in time no ma tter what the mode of operation is 25 and then you suffer the consequences of that.
O I
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON D.C. 20024 (202) 554 2345
11 1
When you said " engineered safety features" were 2 you talking about those things that sit static and wait for 3 something to happen once in a millennium?
Or you were 4 talking about to include those which are running and 5 protecting the plant in active configurations at all times?
6 MR. ROWSOME:
Yes, I would certainly mean to 7 include those.
8 MR. EBER501E Okay.
Well, see, that's not the' 9 standard context.
10 MR. ROWSOME:
Yes.
11 CHAIRMAN OKRENT:
let's see, in the preventive 12 mode, now, we would have no containment requirements, is 13 that what you're saying?
()
34 MR. ROWSOME:
I'm saying tha't.to first order one 15 can start this project with the understanding that it will 16 be parallel to and have some cross-linkage with but be 17 largely independent of the degraded core rule-making which 18 will target containment systems, although if it looks 19 promising, if some alternative to the classical 20 deterministic, with perhaps an IREP overlay, begins to look 21 promising here that would suggest alternative ways we might 22 mandate containment systems, mitigating systems, consequence l
23 mitiga ting systems, then I would not by any means shut the 24 door to it.
25 The principal focus as a first approximation would l
j ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
12
(])
1 be to address the reactor safety responsibility upstream of 2 the degraded core rule-making, upstream in the sense of 3 precursors, prevention.
O 4
MR. KERR:
In a conversation which we had with the 5 Commission at one of the open meetings, if I remember and 6 can paraphrase Coasissioner Hendrie's comments correctly, he 7 said that he would guess that dealing with Class 9 accidents would involve a certain amount of mitigation down to some a
9 level and that one would then take care of the residual risk 10 by energency planning, or whatever.
I think that was the 11 substance of -- and that was, of course, a very preliminary 12 ODiDiOR
- 13 Suppose that one did take this route: wouldn 't 14 your input be fairly important in determining where this 15 cutoff point would occur and how one would deal with it in a 1
16 licensinc process?
17 MR. ROWSOME:
Yes, it can't be completely 18 decoupled from the degraded core rule-making.
On the other 19 hand, I don't think I either would be able to or desire to defer work on that while we tackle a rather broader scale 20 21 look at regulatory reform.
MR. KERRs But if one asks what you should be 22 23 doing in preparation for a rule-making, it seems to me that
()
24 that sort of consideration, if it is appropriate to any of 25 the rules, must go into your planning in some fashion.
ALDERSoN REPORTING COMPANY.INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
13 1
MB. R0WSOMEs Yes.
Well, as a first-order
)
2 approximation, I guess, I would target these programs 3 perhaps at two different levels.
One is at the scope of the
.Os 4 so-called standard engineered safety features rule-making 5 that name bothers me as much or more than it bothers Dr.
6 Ebersole -- and that another, perhaps less hurried pace at 7 the broader issue of more abstract and profound regulatory 8 ref orm.
g At this point the Commission, so f ar as I know, 10 has no efforts under way at all to systematically look at that universe.
A lot of people think about it and a lot of 11 12 people talk in the halls about it, but there's no organized 13 prog ram, there's no block of staff man-hours that is
( ).
34 thinking, gee, should we do something different than the 15 Band-Aid approach or the acceptable risk approach, is there 16 a middle ground to that.
37 CHAIRMAN OKRENT:
You know, I must say, I think 18 th e term " Band-Aid a pproach" is overused and it 's being 19 overused in this discussion, because it's used very much in 20 a derogatory sense and I could equally well look at what's 21 happened in the past in our political process, where every 22 four years a man says he's going to come to Washington and 23 make things better,
.a to me this is a man telling me, via (j'i this Vu-graph, he's going to make things better, and I would 24 s
25 say I have no better, no more assurance that what you 'll do ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
~_. __
14 s
1 here will make things better than that Mr. Carter did or the 2 people before him in the political world.
So I think we 3 need to be cautious about what we call the " Band-Aid" and O
4 what is going to be -- you define the adjective.
5 Again, I don't see in here something that is 6 focused on trying to help the Commission in what, to me, is 7 a very real problem, and that is how to come out with 8 something sensible, in the right direction, at about the 9 right level, defensible, whatever, however you want to put 10 it, in this combination of rule-makings.
That doesn't say, 11 in my comment, that exploring the kinds of things you've got 12 here isn't a useful thing to do; quite the oppesite -- it 13 just says it 's different, to me, than directly tackling the
()
34 question of is there a way of bringing probabilistic 15 assessment --
16 MR. ROWS 3ME:
In the short --
17 CHAIRMAN OKRENT:
-- into a position to~ bear 18 directly on these rule-makings, which -- which are really on 19 them.
20 MR. ROWSOME:
The rule-makings are going forward 21 whether this program exists or not.
And, you know, this l
22 program is not intended to overthrow, at least, in the short 23 term, the siting or emergency planning or degraded core
()
24 rule-makings.
It is intended, as one of its objectives, to t
25 explore options f or the "sta ndard" or minimum engineered O
I ALDERSoN REPORTING COMPANY. INC, 400 VIRGINIA AVE., S.W WASHINGTON, D.C. 20024 s202) 554-2345
15 3 safety features rule-making, which has, as I understand it, p)
(-
2 not got a time line and not got a short fuse as the others 3 do.
O--
4 CHAIRMAd OKRENT:
How can they do the siting 5 without having some reactor -- and by that I mean the entire 6 plant -- in mind, either via criteria, in other words, it 7 will be some probability of an accident involving fission 8 product release --
g MR. ROWSOME:
Well, I can a ttempt --
10 CHAIRMAN OKRENT:
-- f rom the primary system or 11 from the containment or something --
12 MR. ROWSOMEs
-- to give you the logic, although 13 I'm not by nature in apologist for the way that rule-making
()
g,4 has evolved.
But it has evolved based on a premise that one 15 would decouple design f rom siting and that the siting 16 considerations relating to risk would be based upon a family 17 of hypothetical core damage, core melt, and design basis 3 18 th rough 8 kinds of accidents, enveloped in some fashion in a 19 wa y that Roger Blond is working and with others --
CHAIRMAN OKRENTs But there are so m e si te s -- I'm 20 21 sorry, there are some reactor designs for which there is no 22 acceptable site in the U.S.
I mean, I could design such a 23 reactor that I would be unwilling to site anywhere in the U.S. because I considered it so likely to --
()
24 MR. ROWSOMEs The rule-making is aimed at light 25 O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W., WASHINGTON, D.C. 20024 (202) 554-2345
16 1 water reactors of essentially conventional design.
And I 2 believe ther have declared out of scope HTGRs and breeders 3 and the other -- the other options.
4 MR. EBERSOLE:
Dave, will you explain what you 5 mean by this?
The way a utility will buy a reactor is to 6 get a bunch of b'ookkeepers and auditors together, and with 7 no, virtually no engineering participation, and they'll sit 8 down and analyze the lowest bids.
There's no assessment of 9 adequacy or conservatism or whatever.
The general thought 10 is that they will all come up to the minimum standards which 11 will be required by regulatory and they can be planted 12 anywhere you want to put them.
There's no comparative 13 assessment of the safety of one over the other or a system
()
14 a. gainst a system in this context.
15 I think if -- I think it's an illusion to think 16 that a utility will sit down and assess a reactor on a 17 rela tive saf e ty basis.
At least, I don't know of'any that 18 do this, and I.certainly know TVA didn't.
19 They buy them -- they buy them from auditors' and 20 bookkeeping reports.
That's all you need, just go get the 21 bookkeepers and they'll buy the reactor.
22 CHAIRMAN OKRENT4 My understanding of the 23 situation is that what you say must be true at least most of
()
24 the time if not all the time.
But that's not, I think, th e 25 question that the NRC faces in its rule-making on sitina and O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
17
/N 1 on m1nimum engineered safety features, because if it's going
(
2 to adopt some kind of a regulation saying certain kinds of 3 sites are acceptable, it's going to have to have in mind 4 that these plants have certain features; and that's the 5 reason for the minimum engineered safety features, if I 6 understand correctly.
7 MR. EBERSOLE:
That gets right back to my earlier 8 comment about that phrase and its meaning.
That used to 9 mean -- it was sharply defined for a LOCA -- that you had 10 progressively better systems as the site got worse, to 11 diminish the releases that you would get from defined 12 release rates from containment.
And so you could build a 13 structure of mitigating systems, called engineered
()
14 safegua' ras, until you could make the worst reactor look like 15 th e best reactor, when, in fact, there was nothing real 16 about it, because all you were building was complexity 17 against a basic bad design.
You could just tack on one 18 thing and then another.
19 The worst accident at Browns Ferry was to drop a 20 stick of fuel.
It wasn't a LOCA.
That was measured in the 21 context of how much radioactivity got out into the 22 en virone.ent assuming you dropped a stick of fuel into the 23 secondary containment.
That was the -- that was what one 1
(])
24 would call the maximum credible accident in the context of 25 relea se.
l (2)
ALDERSON REPORTING COMPANY,INC, 400 '"~lGINIA AVE, S.W., WASHINGTON, D.C. 20024 (202) $$4-2345 6
18 1
Now, I think we all know that that by no means was
{)
2 the maximum credible accident.
It was a large pipe break 3 that was the maximum. credible accident.
But they had so 4 many crutches put on the design, called engineered 5 safeguards, so that the end result of those was that you 8 could calculate there was virtually no release from them.
7 Not so when you dropped a stick of fuel.
8 CHAIRMAN OKRENT:
All right.
Go ahead.
I mean 9
MR. EBERSOLE:
You. follow me?
10 CHAIRMAN OKRENT:
I know what you're talking 11 about.
I don 't think it --
12 MR. EBERSOLE:
Now, there's nothing real about 13 that process, and you should have had more diesels there,
()
14 for instance, instead of enhancing the so-called safety 15 features for large LOCAs.
I mean, one should have attacked 16 the real risks, not the artificial ones.
It made them look 37 good to say this on paper, but in reality it didn 't mean 18 an ything.
19 CHAIRMAN OKREsT:
Well, let's see, if'I can get back to the title of this agenda item, th ere is no other 20 21 focus program within the Division of Safety and Reliability Research -- l y that System of Reliability and Research?
22 MR. ROWSOME:
On single-failure criterion, Bill's, 23
(])
24 as he said this morning, that he is looking at several 25 strategies f or supplementino the single-f ailure criterion.
O ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
19 1 This too will look at the ways of dealing with the
[}
2 limitations of the design-basis accidents, the 3 single-failure criteria, the failure to include human G
4 reliability review or to consider human error in the context 5 of test against procedure maintaining correct stand-by 6 alignment and the like in the present body of regulations, 7 to consider option for introducing such consideration into a
- 8. body of regulations and to evaluate both the risk reduction g effectiveness and several other measures of acceptability of 10 such a regulatory structure.
11 The planning of this effort is far too preliminary 12 for me to talk about milestones, but it may well evolve into 13 something closer to your concept of the degraded core
()
14 rule-making plus a few others, o,r it might evolve in other 15 directions.
At this point, what I have here is merely my 16 own thoughts; I haven't even gotten input from the rest of 17 the division, much less the rest of the office.
I have 18 circulated it for comment, just as I am presenting it to you i
gg f or comment today.
Nevertheless, I think something with 20 many of these features needs to be done, and I have 21 allocated s corner of the '82 budget, amounting to roughly a 22 million dollars, to tackle something in this sphere.
MR. KERR4 But you have not, I take it from what 23 i
()
24 you say, had user groups come to you and say, "We're 25 preparing f o r rule-making and this is the sort of
(%
(_)
ALDERSoN REPORTING COMPANY,!NC, 400 VIRGINIA AVE., S.W WASH:NGToN, D.C. 20024 (202) 554-2345
20 1 information we need from" --
)
2 MR. ROWS 3ME:
Not on minimum engineered safety 3 features.
Siting, yes.
Emergency planning, yes.
Degraded O
4 core, sort of.
5 We've gotten no coherent direction from anybody on 6 what is neeled for the degraded core rule-making.
But 7 everybody has certainly been thrashing around discussing the 8 problem.
We made up our own program to our perception of 9 what the needs were.
10 Now, if you like, I can go on to the -- the other 11 package of research programs under the applications program.
12 CHAIRMAN OKRENT:
Has the --
13 MR. ROWS 3MEs I get the feeling you're not much
(')
14 sstisfied with this one.
15 CHAIRMAN OKRENT:
Has the Office of Research, by 16 the way, ever looked at the differences in design 17 requirements between the U.S. NRC and the Germans for PWRs, l
18 for example, to try to see which they though t made sense or l
l 19 not or so forth?
20 MR. ROWSOME:
Not comprehensively.
European 21 design practice with respect to decay heat removal systems 22 has been surveyed and will be in the report which is now in 23 draft on the alternate decay heat removal concepts.
24 CHAIRMAN OKRENT:
But that's --
()
25 MB. ROWSOME:
That's the closest to a survey of'it ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
1 l
l l
21 i
I that I am aware has been done by the Office of Research.
()
2 CHAIRMAN OKRENTs Let me ask a dif ferent 3 question.
Does your division look at existing practices for O
4 things like pressure vessels, for example, in the U.S.,
to 5 try to evaluate whethet the current situation provides the 6 necessary degree of reliability with the necessary degree of 7 confidence?
8 MR. ROWSOME:
We certainly have not done so from g the time tha t the WASH-1400 study team elected to borrow 10 from that prior vessel rupture study up until the time, a 11 few months sgo, when RSR came over to us in a flap and asked 12 us.to help resolve the pressurized vessel thermal shock --
13 some f acets of the pressurized vessel thermal shock problem
()
14 for them.
15 MR. EBERSOLE:
And did you resolve those?
16 MR. ROWSOMEt We did what they asked us to do, at 17 least, for the early phases of the procram, and we're 18 tracking the program since.
And that early phase was to 19 sk etch for them some reference cool-down accidents and 20 assess a recurrence interval for B&W plants for a gradel 21 spectrum of severity of overcooling transients.
MR. EBERSOLE:
Did you recall whether the 22 23 accident, which is fairly frequent, of simply stuck-open
()
24 by pass, not compounded by a run-on of main feed water, 25 whether that, the frequency of that presented a hazard to O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
22 t the vessel with incipient crack producing faults in it?
( j}
s 2
MR. R$WSOME:
I don't remember the d e tails.
Matt
/
3 Taylor, as I, indicated, sent over to RSR a family of O
4 scenarios, v ith frequencies established from operating 5 expacience where operating experience was available and j
6 questima' es, engineering judgment guestimates of frequency 7 of occurence of scenarios involving combinations of failure 8 that have yet to be seen in operating experience.
9 MR. EBERSOLE:
Did you have -- do you recall 10 whether you have had any such failures on secondary it blow-down that would have presented this kind of thermal 12 shock problem to the vessels?
13 MR. ROWS 3ME:
I don't know.
I don't -- I've
()
14 skimmed --
15 MR. EBERSOLE:
There have been stuck --
16 MR. ROWSONEs
-- Matt's report, but I didn't --
17 MR. EBERSOLE:
There have been stuck by-passes.
18 MR. ROWSOME:
I wouldn't be at all surprised.
gg That's being fed into the RSR work, and th e y will 20 assess the vessel margins, the likelihood of vessel rupture l
21 given those -- those scenarios.
And between the two of us, 22 we're expected to come back and collaborate in the program 23 at various phases, and I would inagine that the ultimate
()
24 result will be a vesstT rupture risk assessment for 25 thermal-shock-induced rupture.
O l
l ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WAShlNGToN, D.C. 20024 (202) 554-2345 l
23 1
If, as you.suggest, the numbers are a lot worse 2 than the numbers that were used in WASH-1400, it may prove 3 to be a really substantial problem.
And tha t'll be quite 4 messy, because perhaps the most grievous abuse of 5 probabilistic risk assessment, if you will, took place 6 before there was a W ASH-1400 when it decided that it was 7 acceptable to build reactors with essentially no prevention 8 or no mitigation of vessel rupture failure.
So all the 9 levels of defense in depth can be breached by vessel 10 rupture, except possibly those assoc ated with siting and 11 maybe emer'gency planning, although it would progress so 12 rapidly, or might progress so rapidly, that even emergency 13 response is in doubt.
()
14 That kind of issue is something I would target 15 hiire: can you get away with that: should you put enough 16 f aith -- even if you really believe when you've done all the 17 thermal shock analyses and the crack growth models and the 18 tran sient-induced stresses on the vessel, understand 19 thoroughly corrosion mechanisms and so forth, and you really 20 believe you have a ten-to-the-minus-seven per-year vessel 21 rupture probability, should you still have all your eggs in 22 that basket even then.
l CHAIRMAN OKRENT:
How would you arrive at the 23
(])
24 answer to that question via.what you have on the slide?
3R. ROWSOME4 One of the tasks in part one and in 25 O
I l
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554-2345
2n 1 part A and part D is to evaluate the pros and cons of the
[}
2 regulatory options.
And one of them would -- one of the 3 principal thrusts, although I haven 't verbalized it here C) 4 very well, is to -- is to assess the role of risk or 5 raliability in the regulatory process, how much faith you 6 should put it in it.
So that's kind of implicit, although, 7 you're right, it's not explicit on that.
8 CHAIRMAN OKRENT:
I don't care whether it's 9 explicit or implicit.
I wonder how you would answer that 10 question if I gave you the whole million dollars just to 11 focus on that one question.
And you certainly weren't 12 intending to spend it on thst one question.
13 MR. ROWSOME:
I think we would approach it by
()
- bringing in diverse perspectives on it and prs
.nting 14 15 in the results of this, which will, in effect, be a kind of 16 rule-making option policy paper, one of the outputs would be 17 a kind of policy option paper for the Commission; and one of 18 the policy option issues would then be, with appropriate 19 delineation of pros and cons, ought you to put all your eggs 20 in a basket such as this.
And the case for and the case 21 against would be summarized as best we can.
22 Ultimately, clearly, that is a policy decision that we would not solve by analysis, but would rather lay 23
()
g4 upon the Commission.
And my personal instinct.is that one 25 ought not ever to put all one's eggs in one basket like O
ALDERSoN REPORTING COMPANY,INC, I
l 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345 L
25 1 that, again.
2 CHAIRMAN OKRENT:
Well, it would seem to me, if 3 you were a Commissioner having to make the decision, you 4 would need more than qualitative or even semi-quantitative 5 discussions or just a range of opinions -- the range will 6 depend on who you get.
7 MR. BOWSOME:
Well, there are implications of the 8 positions.
If you were to adopt my -- my instinctive call, g my intuitive call, as policy and say henceforth it will be a 10 no-no to design plants in such a way that you cannot contain 11 or sitigate a vessel rupture, you have just bought yourself 12 a back fit issue for the operating plants because many of 13 them have it and then there are options of how to deal with
(-m) 14 that and that would necd to be explored and unfolded as one 15 of the concomitant facters associated with a change in 16 P0liCY '
17 CHAIRMAN OKRENT:
If you were the Commissioner, 18 would you want to know how one would design so all of his 39 eggs were not in that one basket?
20 MR. ROWSOME:
Yes, that would certainly -- we 21 would want to look at what could be done on the operating 22 plan ts or wha t mlgh t be done --
CHAIRMAN OKBENT:
No, let's even talk about future 23
[]}
24 plan ts.
Do you think -- and do you think you could give them the answer for a million dollars, as to what could be l
25 O
l l
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W, WASHINGTON, D.C. 20024 (202) 554-2346 r
26 I done?
2 MR. ROWSOME:
Well, we could give them a very 3 superficial answer for a fraction of a hundred thousand and 4 a somewhat better answer for several hundred thousand and a 5 somewhat better answer for a million, and a better answer 6 still for tan million.
It's a. question of where you --
7 where you wsnt a cutoff on the diminishing returns.
8 NR. EBERSOLE:
In a way, Dave, this relates to g this guarantee that you can depressurize and take the 10 mechanical load off these thermally stressed vessels.
11 CHAIRMAN OKRENT:
Oh, but there are other loads --
12 MR. EBERSOLE:
Yeah, the one I'm talking about in 13 particularly is the
()
14 CHAIRMAN OKRENTs No --
15 MR. EBERSOLE:
-- mechanical and thermal stresses 16 combined.
17 CHAIRMAN OKRENT:
No, I know, but there are other 18 concerns concerning vessels.
That just happens to be the 19 one that's current this year here.
But there's a different 20 one that 's current this year in the United Kingdom.
So --
21 and it doesn't involve necessarily the same scenario.
22 MR. EBERSCLE:
It doesn't involve the combined 23 phenomena ?
CHAIRMAN OKRENT:
No.
No, there's a concern about
()
24 25 one's ability to find flaws with a suf ficiently high O
\\s/
ALDERSoN REFoRTING COMPANY,INC, 400 VIRGINtA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2346
27 1 reliability that you don't know that there isn't something
)
2 larger than you like.
3 MR. EBERSOLE You mean flaws which themselves 4 without this will grow into trouble spots, right?
5 MR. KERR.
Well, either with or without th a t.
6 MR. EBERSOLE:
Either with or without.
7 CHAIRMAN OKRENT:
Yeah.
8 MR. 20WSOMEa I at -- first of all, the way the 9 Office is currently set up, this issue would not be our 10 responsibility, except perhaps in the context of risk.
11 CHAIRMAN OKRENT Whose responsibility would it be 12 to look at vessels to see whether, in fact, the existing 13 practice is giving the necessary
()
MR. 20W'SOMEs The metallurgy and stress --
14 15 CHAIRMAN OKRENT:
-- reliability no matter what it 16 is?
17 MR. ROUS3ME:
-- and so forth is in RSR.
And I 18 believe that this -- I c3uld be wrong that it's 19 Kugenbrachi's branch ; I'm not positive.
20 MR. EBERSOLE:
But the report that came through 21 the mail was addressed back to C.
Z.
Serpanik.
l 22 CHAIRMAN OKRENT:
Yeah, but they're not risk MR. ROWSOME No, not a t all.
23
('s 24 CHAIRMAN OKRENT:
-- type of people.
s/
25 MR. ROWS 3ME:
Thay would come to us as -- with, in O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
1 28 1
i 1 effect, a user request to supply the risk framework for it.
2 Or, conversely, we might come to them.
3 CH AIRM AN OKRENT s But you don't have any a priori, 4 oh, I guess you might call it a charter, that says you 5 should look at the existing safety approach and see if there 6 are potentially important weak spots in it; you don't have 7 such a charter --
8 MR. ROWSOME No.
g CHAIRMAN OKRENT:
-- from your own prospectus?
10 MR. ROWSOMEs Well, we have --
11 CHAIRMAN OKRENTa I don't expect you to --
12 MR. ROWSOME:
-- no written charter at all at the 13 soment, and have not had since Tony Buhl's day.
We have a
()
~
14 kind of common law or de f acto charter which is a mix of 15 tradition, what we have been doing, which has been 16 methodology, methodology development, and applied systems 17 analysis, risk assessment, systems reliability analysis.
18 And we have another chartar that is evolving from the way 19 various user groups have come to us for advice and 20 guidance.
For example, the Commissioners are beginning to 21 come to Bob Bernero inore and more to fill the role that used 22 to be filled by Steve Hanover when he was special technical 23 specialist to the EDO.
When the Commissioners have a 24 technical question and want to have somebody second-guess 25 NRR or second guess the industry, they'll tend to say, call O
ALDERSON REPORTING COMPANY,INC.
40G VIRGINIA AVE., S.W, WASHINGTON, D.C. 20024 (202) 554 2345
29 1 up Bob and say, " Hey, what can you tell us about this?
2 Write us a letter und give us a schedule in which you can 3 tell us about this."
4 NRR is using us, in part, to sprinkle holy water 5 on things like the Tedesco Task Force Reports and to cope 6 with some of the issues like single f ailure and station y blackout and what not that they don 't f eel comfortable 8 dealing with.
g RSR has come and SAFER is coming to us with 10 guidance on priorities and event sequence delineation, 11 scenario delineation for some of the work they're doing both 12 in reactors and in fuel cycle.
13 I think my prime sense of our internal mission has
()
14 a -- has two components in it.
One is a short-term 15 practical component; and that is to do what we can to try to 16 see the next TMI coming before it happens, to catch the i
17 outlines.
And second of all, to try to build the 18 foundations by which -- through which the safety insights from risk assessment and system reliability engineering can 39 l
l 20 be folded into routine safety analyses and safety i
21 eval ua tio ns both within industry and within the agency, 12 within NRR and IEE, the line offices.
l 23 MR. KERRs Well, in that box where you try to
()
24 preclude the next TMI, you could put any of a number of l
25 items that you considered crucial to saf ety.
So in some ALDERSON REPORTING COMPANY,INC,
(
400 VIRGINIA AVE., S.W., WA$ilINGToN D.C. 20024 (202) 5,54-2345
,...m
,,,,, - - +,. --,~
e-
30 i sense you do look on yourself as having some responsibility 2 to pick out things that may, in your view, be of current 3 importance.
C6 4
MR. ROWSOME:
Yes, that's true.
On the other 1
5 hand, everybody, wi th the exception of yourself, tells us 6 that we should be a service organization following user 7 request and not leading the pack.
We think we're doing 8 quite a bit of lesfing the pack.
On the other hand 9
MR. KERR But there is a conventional wisdom 10 MR. ROWS 3ME:
-- 36 people are stretched a little 11 thin with the scope we've delineated here today.
12 MR. KERRt There is a conventional wisdom within 13 RES, I gather, that it would be a more productive
()
34 organization if.it did have more freedom to decide what to 15 do.
1 believe I have heard that expressed on a number of 16 occasions.
17 MR. EBERSOLE:
Frank, may I ask this question?
I 18 have a little piece of paper here that was sent from the lab gg -- it came in my mail -- sent from the lab to a Dr. L. C.
20 Schau, Schau, S-
-h-a-u.
21 CHAIRMAN OKRENT:
Schau.
22 MR. EBERSOLE:
Schau.
System Director for General 23 Reactor Saf ety Research.
And it has to do with the HSST 24 thermal shock program,
()
l 25 My first reaction to reading this wast well, this j
(
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
31 1 is probably related to the LOCA related thermal shock O.
f 2 problems.
And I started smoking around into it to see if it 3 included these other kinds of transients that result in 4 post-thermal-shock repressurization.
5 What do you do and how do you interface with this 6 man when you see these things?
He's working on heavy 7 section -- HSST -- thermal shock pressures.
Do you feed him 8 the information on probability of occurrence of such aa g event?
10 MR. ROWSOME:
Generally speaking, no.
In this 11 instance, yes.
After, I think you're right, it was 12 Serpanik's branch, their people did this -- a conservative 13 analysis of a hypothetical worst-case overcooling transient O
24 1a e scw re ctor, watca areeictee thet the ve e1 woote 15 break, probability nearly one, they came to us and said, 16 "We've got this unpleasant surprise.
We want to know 17 whether any scenario approaching this bad can really 18 happen.
Will you do the systems analysis for us, 19 delineating scenarios with rough likelihood assessments 20 associated with them?
Give us a spectrum of the kind you 21 can expect to see once a year, once every ten years, once 22 every hundred years, worst in a hundred years, worst in a 23 thousand years: what are the systems doing in these 24 scenarios, what kind of failures are involved, what kind of 25 characteristic times, and so forth.
Do the scenario, system O
ALDERSoM REPORTING COMPANY. It C.
400 VIRGINIA AVE., S.W., WASHINGTON, D C. 20024 (202) 554 2345
32 1 scenario work for us to do our stress analysis, our 2 structures, materials, and so forth."
3 MR. EBERSOLE:
Yeah.
You gave him the 4 frequency-severity thing.
I guess.you started at by-pass 5 failure or close to it, secondary by-pass --
6 MR. ROWSOMEa That took historical records and 7 actually just providing them a strip chart recorder, Xeroxes 8 of strip cha rter recorder pa per --
9 MR. EBERSOLE:
Oh, you did it from records?
10 MR. ROWSOME:
-- from records of actual 11 occurrences for those that had a recurrence interval within 12 the histori:al ran7e of the order of 50 reactor years.
13 Beyond that he did it by postulating system failures that
- ()
34 have not been seen.
15 He searched the LER records quite thoroughly as a 16 starter.
And when he'd exhausted the historical experience, the worst transient was the Rancho Seco lightbulb transient, 37 18 in terms of the rapidity with which a pressurized vessel was 39 chilled.
They --
MR. EBERSOLE Was that a stuck by-pass?
20 MR. ROWS 3ME:
Let's see.
Feed water was shut l
21 l
22 off.
Aux '.f eed was shut off for some minutes.
Aux' feed 23 started ten minutes later.
No, I don't believe there was 24 any by-pass there.
MR. EBERSOLEs-No stuck by-pass?
25 l
l ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 564-2345
33 1
MR. ROWS 3ME:
No.
Aux' feed was started and the
)
2 operators manually initiated main feed water on top of it.
3 They were blind to primary -- both to the primary 4 parameters, temperatures, pressures.
And I think they had 5 two instruments left, pressurizer level -- that they trusted e -- pressurizer level and -- I've forgotten now.
But they 7 just socked the water to it.
Oh, ECCS came on, because of 8 the overcooling.
So they had ECCS, full feed water flow 9 filling up the steam generators with cold water.
Of course to the feed water heaters had tripped off.
So they -- they, 11 after some period, which I think was of the order of seven 12 to ten minutes without any feed water at all and with some 13 blow-down with the pressurizer relief valve, they had an
()
14 overcooling transient on the secondary side coupled with 15 ECCS injection of cold water.
And because of the failure of 16 instruments, they don 't have strip chart paper that tells 17 you exactly what the thermal transient was.
They know 70 18 minutas later they got their instruments back.
They were gg already cold.
They were at 200-and-some-odd degrees.
l 20 Analysis suggests that they may have come down most of that l
21 wa y in the five or ten minutes f ollowing the reflood of this l
22 -- both primary and secondary systems.
So that they may in f act have been, at J east, for a period of minutes, chilling i
23 I
{~s) 24 at the rate of many hundreds of degrees F per hour.
~
CHAIRMAN OKRENT:
Well, let's see, we probably 25
()
j ALDERSoN REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON. D.C. 20024 (202) 554 2345
34 3 should finish up in five or ten minutes.
I'm not sure what 2 topics or questions the subcommittee would like to raise 3 with Mr. Rowsome while we still have him.
O 4
Is there something on IREP that you'd like him to 4
5 cover?
6 MR.' KERR:
I guess I'd be interested in knowing 7 what he would put in a first priority category if he had 8 more resources and manpower.
9 CHAIRMAN OK3ENTs All right.
That's sort of a 10 general kind of thing.
You don't have to talk about numbers
.11 in a specific way.
12 MR. ROWSOME The first increment in both people 13 and dollars would be in human reliability research.
I think
()
14 that.is a number-one candidate.for undernourished program.
15 Second, I would give more attention to 16 rule-makings and accelerate that work, degraded core work 37 and the add-on decay heat removal systems, that family of 18 work.
19 Third, I guess, I would accelerate the training 20 program, and in particular the preparation of what I like to 21 call Shown's Outline Series kinds of training aids, written 22 documentation workbooks, solved problems, to bring people up 23 to speed in system reliability analysis and risk assessment f ()
24 techniques.
I think we need some front-end capital investment in this kind of material for our training courses 25 O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASNINGToN. D.C. 20024 (202) 554-2345
35 1 and to get these tools into use in the regulatory process, 2 not just system reliability analysis but event-tree event 3 sequence delineation and this kind of thing.
4 Another high-priority item beyond those -- there's 5 a whole bunch that come in perhaps at the third level, and I 6 can't prioritize among them -- but perhaps more work on the 7 acceptable risk program, broaden that a little bit and start 8 it a little sooner, flesh it out some more, put back in some 9 of the consequence analysis work that's been cut, put back 10 in some of the data analysis work that has been cut to meet 11 the budget constraints.
12 There's really hardly any program in our whole 13 budget that I don't think would benefit from a little bit O
24 more ao=rioa e=t-15 MR. EBERSOLEs You don't mention extending the 16 program to include more than merely diagrammatic 17 relationships in the IREP program, that is, to go on into 18 ph ysical reality of --
39 MR. ROWSOME:
That's --
20 MR. EBERSOLE You know, the -- to remove some of 21 the academic flavor from them.
MR. ROWSOME Yeah.
Tha t's th e -- the intent here 22 is -- there are techniques a vailable for tha t, of course.
23
'T 24 You can -- you can with codes like CONCAN and Back fire, (J
25 given that you have a f ault-tree analysis of systems, you O
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE, S.W, WASHINGTON, D.C. 20024 (202) 554-2345
36 1 can tag the fault events in the fault tree with codes that
)
2 delineate physical location, what room it's in, wha t it's 3 close to, what environmental effects it's sensitive to, 4 temperature, humidity, so forth, you can then sort the cut 5 sets for common elements using one of these computer-aided 6 techniques.
7 I -- we -- we are doing work with those.
We a re 8 -- in the procedures development for IREP and NREP it is g in tended to gear up for a second pass, a refinement of the 10 first IREP studies, that will pick up fires and floods and 11 that kind of common mode analysis, the environmental 12 ef f ects, the proximity effects, the environmental 13 degradation phenomena.
()
14 Yes, I would certainly put in my second-priority 15 catalogue of things that could use more nourishment in terms 16 of people and dollars to accelerate that work.
17 On the other hand, as I mentioned earlier, I think 18 it would be better to get all of the owners to be working on 19 an abbreviated program today in parallel with the resources 20 that are available in the industry, rather than to delineate a procedure for a massive study that could only be done four 21 22 plants or five plants or six plants at a time with the 23 resources available in the industry.
I think we can build
)
24 resources in the industry faster and tease out the outlier 25 sequences, if you will, more rapidly if we get the whole O
ALDERSoN REPORTING COMPANY,INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
37 1 industry working on an abbreviated study that can be pared O
2 down to about a five-man-year-or ten-ma n-year-pe r-plant 3 ef fort.
r 4
MR. EBERSOLEa Yeah, but the industry is going to 5 take your study clean and without this consideration and use 6 it as a vehicle to show how good they are, unless you call 7 out its shortcomings and even modify the codes and whatever 8 delineations you use to show that there are voids in the 9 logic in this context.
10 MR. ROWSOME Well, that's true.
And I've 11 attempted in all the IREP documents to make it very clear 12 wh a t this -- what the studies did not cover.
And those 13 features have been in those lists.
()
34 On the other hand, if -- if we were to suggest or 15 recommend or draft procedure guides for NRR implementation 16 that took 20 or 30 man-years to implement, with a very 17 substantial f raction of those man years put in by people who 18 are experienced risk-assessment analysts or 19 system-reliability analysts, we're going to find that many 20 of the current generation of plants are going to be retired 21 af ter their 40-year lifetime without ever having seen such a 22 stud y.
And therefore the iterative approach, I think, 23 though not without its shortcomings, is the preferred
(
24 strategy.
l 25 CHAIRMAN OKRENTs In what you do in IREP, how do l
[h
\\>
l
[
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., f.W., WASHINGTON, D.C. 20024 (202) 554 2345
38 1 you handle these failure modes of components which are not
{)
2 excluded by your exclusion, like massive failure of a steam 3 generator or of a flywheel inside the containment or the O
4 pressure vessel?
Do you assume that these probabilities are 5 low enough and just let it go at that?
Or do you look at 6 them?
Or do you say you haven't looked but state this?
Or 7 just how do you handle it?
8 MR. ROWS 3ME:
I assume you've read the procedure 9 guide.
It's a little vague on that.
It --
10 CHAIRMAN OKREr Yes.
33 MR. ROWSOME:
-- says that you should look at the 12 phenomenology associated with each of the sequences which 13 pass the screening criterion.
Some of the ones that do not 14 pass the screening criterion could be made substantially 15 more serious by such ef f ects, but we do not know ' how to, in 16 a resource kind of format, how to package that procedure to 17 do that in any other ray.
18 CHAIEMAN OKRENT:
Well, as a minimum, I would 19 suggest that --
MR. ROWSOMEs Perhaps we need an FMEA task in l
20 21 there too?
CHAIRMAN OKRENT:
No, what I'd say, as a minimum, 22 I
it would be well to identify those failures which are not 23
()
24 precluded by the original exclusion -- you know, seismic, l
25 flood, or so forth.
l
[J t
ALDERSoN REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASNINGToN, D.C. 20024 (202) 554-2345 l
39
(])
1 MR. ROWSOMEa Oh, absolutely.
2 CHAIRMAN OKRENT:
But -- but, nevertheless, they 3 are those f ailures which are not excluded thereby, but whose O
4 probability has been assumed to be so low, or has been 5 otherwise written off, so that it was nowhere included in 6 the review, because otherwise implicitly you're offering 7 your seal of approval on sort of a result that implicitly 8 has said this.
And in -- in effect -- I earlier asked you 9 whether, you know, you looked at things like was the 10 probability of pressure vessel failure sufficiently low --
11 it seems to me, when you do an IREP under your own ground 12 rules you would either you look at it or you decided you 13 looked at it last year and you don't have to relook at it
()
because whate'ver you did last year was valid, or whenever 34 15 yo u did.
But it's failure is not excluded from your ground 16 rules, and so even if there is another of fice tha t 's --
37 MR. R0WSOME:
The way that I would anticipate --'
CHAIRMAN OKRENT:
-- looking at the reactor 18 19 vessels, it doesn't exclude you --
MR. ROWSOME
-- vessel rupture will be treated in 20 21 these studies is that it will appear in the event trees and 22 be described, that no original attempt will be made to 23 quantify itn likelihood.
The quantification is mostly a
()
24 systems analysis f ault propa ga tion through the network of 25 systems kind of exercise, so that one would not expect to O
ALDERSoN REPORTING COMPANY. INC, 400 VIRGIN!A AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345
40
(~T 1 find in an IREP -- one of the IREP studies going on today an V
2 analysis of the frequency versus severity of overcooling 3 transients and its effect on vessel rupture probability.
s 4 However, the event tree part of it would be there, would not 5 need to be redone to accommodate that, and tha t the systems 6 -- most of the systems analysis to accommodate that would be 7 done, although we are not asking f or f ault trees of the feed 8 water system, feed and steam system leading to overcooling g transients, f or example.
We are asking for fault trees of to the feed water and steam system leading to interruptions of 11 main feed delivery.
12 MR. EBERSOLE:
You don't do any work en the matter 13 of predicting the probability of detecting or not detecting 14 a fault of whatever size --
15 MR. ROWS 3MEs No, none.
No.
16 MR. EBERS01Es which is important or not 17 important, depending on the transient that's being 18 experienced ?
19 HR. ROWS 3ME:
Bill has started a little bit of g work on modeling piping failures with enough mechanistic 21 d e ta il, probabilistic modeling of pipe f ailures with enough 22 mechanistic detail so that one can fold in data on the 23 ef fectiveness of various in-service inspection practices and
()
24 see what that does to the pipe rupture likelihood or pipe 25 rupture mode likelihood.
He's just started it in this O
ALDERSON REPORTING COMPANY,INC, 400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
41 1 fiscal year; it has fairly limited funding I think it's
[
2 125K, about one-and-a-third man-years, romething of that 3 order.
And this is being done in part for an NRR user 4 request and in part in coordination with the materials 5 people in RSI.
6 So that might be a tiny foot in the door.
But 7 basically the answer to your question is, no, we have not 8 done any.
9 CHAIRMAN OKRENT:
Well, all I was suggesting is 10 that there are, in fact, other things that really are 11 excluded from your look, although --
12 MR. ROWSOME:
Yes.
13 CHAIRMAN OKRENTs
-- by looking at the criginal
(/
14 listing one wouldn't think so, and it might be we.i.1 (or you 15 to identify them, at least, as a minimum.
16 MR. ROWSOM2s Yes.
17 CHAIRMAN OKRENT:
Because they can't tell.
18 Well, let me leave it at that.
19 Well, any last question or comment for this time?
20 Anything additional you --
21 MB. KERR:
No, I think CHAIRMAN OKRENTs Well, in that case, I'll thank 22 Mr. Rowsome and those who came with him this morning in 23 j ()
24 absentia.
And we'll adjourn the meeting.
MR. EBERSOLE:
Considering the time he has spent 25 l
ALDERSoN REPORTING COMPANY,INC,
' AVE., S.W., WASHINGTON, D.C. 20024 (202) 554-2345 400 VIRG t
__ ~ -.. _ - _ _. __
r 42 Q
1 on his feet and also sitting down talking, he has given a 2 lot of lectures today.
And I expect he's tired.
4 l
3 CHAIRMAN OKRENTs It's like at least two courses.
i
- O
~
4 (Thereupon, a t 3 :10 p. m., the meeting was 1
i 5 adjourned.)
6 7
8 9
10 j
11 12 4
13 O
15 16 f
17 1
j 18 19 20 j
21 i
22 i
23 l O 24 25 O
I i
ALDERSON REPORTING COMPANY. INC.
400 VIRGINIA AVE., S.W., WASHINGTON, D.C. 20024 (202) 554 2345
.. _., ~ _ _. -
41t'%
? A.
1:=b2 6
- n. o Qv.
e*
9 m**
9 m ae..
b...tys A.*.c G. m.(
e*
s 5Y v..
.t.,.
,,,.. s ~.y
_3
......g ACRS/Subcor.'nittee on Reliability and Probabilistic Assesstaent i ". Ohe 2300er 0f:
PUBLIC MCETIUG 220e Of.:rccesci.;;
December 3, 1980 Decke Nu=ber:
. lace *f ?receedi,g:
Washington, D.
C.
....s.-,,
..... _ _,... -..=
.. 5 c...
(.7.,.
,. e..c,.., :
O
.A.A f^< <J -
a n
/
.4, :
,s.
l 0
.c 5
0**]O *D T
Y h' l
.Jl..tIdl.
e o Ju e a
l i
I p
1