ML20134Q061
| ML20134Q061 | |
| Person / Time | |
|---|---|
| Issue date: | 02/20/1997 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-2093, NUDOCS 9702260233 | |
| Download: ML20134Q061 (428) | |
Text
. . - - . . .. . _. . _ . _ . . - .- - - - . - _ _ - .
OfficicI Trcnscript cf Preco dings -
[3 NUCLEAR REGULATORY COMMISSION l
! ACRST 2093 f
Title:
Advisory Committee on Reactor Safeguards !
Probabilisitic Risk Assessment Subcommittee t
TRO4 (ACRS)
Docket Number: (not applicable) RETURN ORIGINAL TO BJWHITE a M/S T-2E26 415-7130 2
THANKS1
+ 1 Location: Rockville, Maryland l O Date: Thursday, February 20,1997
~
L,,RS c-.ce a,opy- Re,a.1 :
i ide ZeoMe om:::ee
<..,r.
Work Order No.: NRC-1018 Pages 1-355 J
NEAL R. GROSS AND CO., INC. [ t Court Reporters and Transcribers 260047 1323 Rhode Island Avenue, N.W. 'r q i Washington, D.C. 20005 O <202) 2u33 9702260233 970220 09 PDR
DIBCLAIMER !
l
~x PUBLIC NOTICE s
BY THE i UNITED STATES NUCLEAR REGULATORY COMMISSION'S l ADVISORY COMMITTEE ON REACTOR SAFEGUARDS l SUBCOMMITTEE ON PROBABILISTIC RISK ASSESSMENT FEBRUARY 20, 1997 ,
1 l
l l
The contents of this transcript of the proceedings of the United States Nuclear Regulatory Commission's Subcommittee on Probabilistic Risk Assessment, of the Advisory Committee on Reactor Safeguards on February 20, 1997, as reported herein, is a record of the discussions recorded at the meeting held on the above date.
This transcript has not been reviewed, corrected and edited and it may contain inaccuracies.
l l
l
,c-k.s NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVENUE, N W (202) 2344433 WASHINGTON, O C. 20005 (202) 2344433
1 1 UNITED STATES OF AMERICA r~N 2 NUCLEAR REGULATORY COMMISSION U 3 +++++
i 4 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS l
5 PROBABILISTIC RISK ASSESSMENT SUBCOMMITTEE l
6 +++ ++
7 THURSDAY l 8 FEBRUARY 20, 1997 9 ++ +++
10 ROCKVILLE, MARYLAND 11 +++++
12 The Subcommittee met at the Nuclear Regulatory
,_ 13 Commission, Two White Flint North, Room T2B3, 11545 i'
14 Rockville Pike, at 8:30 a.m., George Apostolakis, 1
15 Chairman, presiding.
t 16 COMMITTEE MEMBERS:
17 GEORGE APOSTOLAKIS Chairman 18 JOHN J. BARM Member 19 IVAN CATTON Member 1 20 MARIO FONTANA Member l
21 THOMAS S. KRESS Member 22 DANA A. POWERS Member l
1 l 23 ROBERT L. SEALE Member 24 WILLIAM J. SHACK Member
(~h
(.f/ 25 DON MILLER Member l NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
f (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
[ J
! 2 1 ACRS STAFF PRESENT:
p 2 JOHN T. LARKINS Exec. Director i
l V 3 MICHAEL T. MARKLEY 4 SAM DURAISWAMY l
j 5 AMARJIT SINGH 6
7 ACRS FELLOW PRESENT:
8 RICK SHERRY 9
10 ACRS CONSULTANTS PRESENT:
11 EDWARD ABBOTT 12 JAMES CARROLL 13 ,
.A 1 1
D}
A 14 ALSO PRESENT:
15 MARK CUNNINGHAM 16 GARY HOLAHAN 17 TREVOR PRATT 18 BOB JONES 19 TOM KING 20 RICK GRANTON 21 MIKE CHEOK 22 GARETH PARRY 23 LAWRENCE MARTIN l
l 24 ROY WOODS l !p) y 25 NANETTE GILLES 1
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
3 1 ALSO PRESENT: (CONTINUED)
,3 2 MILLARD WOHL l l \
l
,s l
3 JOHN FLACK DAVID C. FISCHER I 4
1 I
5 W. BRAD HARDIN 6 JOE COLACCINO 7 TOM SCARBROUGH I l
8 STEPHEN DINSMORE 9 .
1 10 1
11 )
l i
12 13 l'h ,
V) 14 i
1 15 1
16 17 18 19 20 21 22 l
23 24 (9
s_/ 25 NEAL R. GROSS
- COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
l WASHINGTON, D C. 20005-3701 l (202) 234-4433 (202) 234-4433 l
a
I l
4 l
l 1 A-G-E-N-D-A l
2 Acenda Item Pace i
,f^x)
- C/
3 ACRS Introduction - Chairman Apostolakis 5 l
4 NRC Staff Presentation l
5 Overview 8 i
6 RG & SRP Appendices 51 7 RG, SRP & Pilot for TS 8 N. Gilles 150 9 RG, SRP & Pilot for IST 10 D. Fischer 251 I 11 General discussion of issues 12 13 r"s
() 14 ;
15 16 l
17 18 19 20 l 21 1 22 23 24
\.) 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON D.C. 20005-3701 (202) 234-4433 l
L a
5 1 P-R-0-C-E-E-D-I-N-G-S 2 (8:35 a.m.)
f- S N) 3 CHAIRMAN APOSTOLAKIS: The meeting will now 4 come to order. This is first day of the meeting of the I
5 ACRS Subcommittee on Probabilistic Risk Assessment. I am l
6 George Apostolakis, Chairman of the Subcommittee.
l 7 ACRS Members in attendance are: John Barton, 8 Ivan Catton, Mario Fontana, Tom Kress, Dana Powers, Robert 9 Seale, and Bill Shack.
10 We also have in attendance Jay Carroll and Ed 11 Abbott who are here as consultants to the Subcommittee.
12 ACRS Senior Fellow in attendance is Richard 13 Sherry.
1 f~s 14 The purpose of this meeting is to continue our l 15 discussion of the NRC Staff's approach to codify risk-16 informed and performance-based regulation through 17 development of Standard Review Plan sections and )
18 associated Regulatory Guides. The Subcommittee previously 19 met to discuss these matters on October 31, November 1, 20 21, and 22, 1996, and January 28, 1997.
21 The Subcommittee will gather information, 22 analyze relevant issues and facts, and formulate proposed 23 positions and actions as appropriate for deliberation by 24 the full committee. Michael T. Markley is the Cognizant (3
(_) 25 ACRS Staff Engineer for this meeting.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
a w, 6 .
1 1 The rules for participation in today's meeting !
l 2 have been announced as part of the notice of this meeting p!
%J 3 previously published in the Federal Register on February 4 5, 1997.
5 A transcript of the meeting is being kept and l 6 will be made available as stated in the Federal Register 7 Notice. It is requested that speakers first identify 8 themselves and speak with sufficient clarity and volume so 9 that they can be readily heard.
10 We have received no written comments or 11 requests for time to make oral statements from members of 12 the public, i
13 I would like to inform you that we have one
( )
14 change in the agenda. Messrs. Biff Bradley of the Nuclear 15 Energy I3stitute and Doug True of ERIN Engineering will 16 not be making a presentation tomorrow regarding large 17 early release frequency, as was indicated in the published 18 agenda.
19 Before we start with the meeting, our 20 Chairman, Dr. Seale, has a few comments.
21 MEMBER SEALE: I've been asked by the Staff to 22 remir.d the members that you have received a package that 23 contains resumes of individuals who have been identified 24 by the screening committee as possible candidates for s
25 appointment to this committee. We have scheduled NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
7 1 interviews for groups of three of -- three groups of g3 2 three, during the luncheon hour, so as to minimize the
() 3 interruption of the regular meeting agenda.
4 Arrangements have been made for the members to 5 meet with these people in various rooms as indics.ted on 6 your packet. Yours is right there at the top, George.
7 CHAIRMAN APOSTOLAKIS: Yes.
8 MEMBER SEALE: But if you wish you can just 9 find a table down in the dining room and eat down there 10 and then review your candidate; that's up to you. But we 11 do hope that you can talk to the individual, gather enough 12 information to assist us in making a recommendation which 13 we will want to be forwarding to the Commission fairly 14 soon.
15 I think that pretty well covers it, so thank 16 you.
17 CHAIRMAN APOSTOLAKIS: Okay. Thanks, Bob. We l 18 will now proceed with the meeting and I call upon Mr. Gary 19 Holahan of NRR, and Tom King, Rescarch, to begin. And Bob 20 Jones of NRR.
21 MEMBER SEALE: And Mark Cunningham.
22 CHAIRMAN APOSTOLAKIS: And Mark Cunningham of 23 Research. We only have four seats so that must be it.
l l
l 24 MEMBER SEALE: I think I know who's the bass rO
(,) 25 and who's the tenor in this group, but could you identify NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
8 1 your other roles? Baritone and whatever.
l 'f~y 2 DR. KING: Same people, same presentation;
)
-/ f' 3 just different day. ;
4 DR. HOLAHAN: Same story. ,
5 DR. KING: Same story. The past couple of 6 meetings we've had -- first, I'm Tom King with the office 7 of Research, and Gary Holahan and Bob Jones from NRR and 8 Mark Cunningham from Research, just for the record.
9 The past couple of meetings we've had with 10 you, both the subcommittee and full committee, we've 11 concentrated on the General Reg Guide and SRP for risk 12 informed regulation, the approach, the proposed guideline, 13 the process and so forth.
- l i 14 The next couple of days we want to concentrate 15 on the application-specific Reg Guides that supplement 16 this general approach and are consistent with it. And we 17 also want to go through this morning, the appendices to 18 the General Reg Guide and SRP which you may have seen in 19 various draft stages but are now -- have been finalized 20 and tailored to conform with the latest version of the 21 General Reg Guide and SRP.
22 And leading into that I thought I'd take a few 23 minutes and just refresh everyone's memory on the general 24 process and approach and the background leading up to --
s ,
) 25 before we get into the appendices and the application-NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4 433
9 1 specific activities.
l 2 As you recall, this activity stemmed from a I
- .<w\
Q) 3 Commission Policy Statement on PRA that encourages the use 4 of risk information in the regulatory process, and the l
5 Commission expected that to lead to better decisions, more 6 efficient use of resources, and potential reduction of 7 unnecessary burdens.
8 We've developed Reg Guides and SRPs for 9 reactor licensecc to use, and they deal with a process and 10 an approach for making plant-specific char.ges to the 11 current licensing basis that utilize risk information.
12 And the --
13 DR. CARROLL: Tom?
n O'
I 14 DR. KING: Yes?
15 DR. CARROLL: I'm coming into this sort of in 16 the middle. One thing that I wondered about as I went 17 through this material was that you had these various pilot 1 18 programs ongoing now. Do the licensees that are working 19 on those programs have the latest integrated versions of 20 the guides and SRPs to guide them?
21 DR. KING: No, they do not have copies of the 22 latest -- or they don't have any copies of the Reg Guide 23 at all. They have an earlier version of the SRP which is 24 very out of date.
/'
()s 25 We have gone to the Commission asked, could we NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 10 1 release these in advance of the formal public comment i
i r~s 2 period, not only for that purpose but for other purposes iv) 3 as well. And the answer we got back was no, the 4 Commiselen preferred we keep these pre-decisional until 5 they're actually ready for the formal public comment 6 process.
7 DR. CARROLL: That sort of inhib,its the 8 interaction between the Staff and the people that are 9 working on the pilot program?
10 DR. KING: Well, it inhibits them from 11 actually having a piece of paper in front of them, but it 12 doesn't inhibit the Staff and them from talking. They 13 certainly have access to all the viewgraphs and
[.
s' ,/
14 presentations, which really present the meat of the 15 approach and the guidelines that are being developed. 1 1
16 So I think you can argue that yes, not having 17 it is some hinderance, but I don't think it's a major i
18 hinderance. Maybe NRR wants to speak to that.
19 DR. CARROLL: So you believe that they are 20 getting at least the flavor of where you guys are through 21 the mechanism of looking at viewgraphs that have been 22 presented to ACRS and others?
23 DR. KING: And a number of meetings on the '
l i
l 24 pilots themselves. That's my opinion, yes.
/3
(,,) I think what I would say, Jay is, i
25 DR. JONES:
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
11 1 the reality is in the pilots because some of these rT 2 activities are in fact, several years old initiated by the
! l
\J 3 industry. And we --
4 DR. CARROLL: Long before this activity.
5 DR. JONES: Long before this activity. And in 6 fact, I would also say that in my mind, a lot of the 7 framework that you see embodied in the General Guide as it 8 exists now was basically constructed about six months ago, 9 and I think it evolved from the interactions with the 10 committee and got to that state.
11 Very few -- the pilots basically, are not 12 totally consistent with that framework. A lot of the
, ,,s 13 focus of the pilots has been on looking at what they were
( )
' 14 doing, how they were dealing with specific technical 15 issues: IST -- both the IST issues themselves and testing 16 issues, along with how are they ranking components.
i 17 And the interactions with them is how they 18 were ranking components, what kind of treatment they were l 19 going to give to low-ranked components, or factored into ,
1 20 the development of the specific Reg Guide as we went 21 along. Not always are we in agreement with their 22 position, and I think you'll hear that later on today.
23 Probably the one where I'd say we're closest 24 is in tech specs. I think in the tech specs area is one
(
(_,)/ 25 that we've worked. We've had kind of an ongoing review NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS; 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
12 1 and the thing evolved very closely together, so that one's
,m 2 probably closest to where we are. In fact, we're arguably (x s) 3 looking at issuing SERs in the next few months. So that 4 one is moving along reasonably well.
5 The other pilots are lagging behind and in 6 fact, were put on the back burner to some extent, in order 7 to finally finish the guidance documents and then we'll f I
8 finish those up after these go out for comment over the l 9 next few months.
10 CHAIRMAN APOSTOLAKIS: How long will you take 11 to finish the pilots?
12 DR. JONES: I believe most of them are 13 targeted for roughly the summer. I think IST is targeted I\ >) 14 around the summer; I'm not sure on graded QA. You can ask 15 them when they go this afternoon. I 16 CRAIRMAN APOSTOLAKIS: So will there be time 4
17 then, to include lessons learned from the pilots into 18 these documents? That would be after the --
19 DR. JONES: We have at this point, what we've 20 identified issues are clearly -- were brought forward into 21 these and what we think the staff position is on the way 22 to approach these issues. Obviously we're going to need 23 ongoing dialogue on many of those yet, to finish up.
24 DR. HOLAHAN: I think there's only one o
k ,) 25 potential inconsistency with that. Because I think the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C 20005-3701 (202) 234-4433
13 1 draft documents hopefully will be up for comment in the fs 2 April timeframe. So I think they will be out there for a I i
\~/ j 3 fair period of time while the pilot activities are let's 4 say, in their later stages.
5 So the only potential inconsistency is if 6 later in the year we were to make any major changes in the 7 Reg Guides or SRPs, then I think that would probably make 8 them inconsistent with the versions that we're looking at 9 the pilots on.
10 But at this point, hopefully we're not 11 envisioning any major changes, although there's probably 12 lot of opportunity for the details to be refined.
13 CHAIRMAN APOSTOLAKIS: Now, the public comment Ng (J
4 i k- 14 period is 90 days, right?
15 DR. HOLAHAN: That's our current --
16 DR. KING: That's our -- yes, proposal.
I 17 CHAIRMAN APOSTOLAKIS: And after that, how I
18 much time do you have to revise if necessary, the 19 documents before you release them?
20 DR. HOLAHAN: Well, our current schedule is to 21 get it done by the end of the year, end of December. And 22 whether we can beat that schedule or by how much, I think 23 depends upon the nature and the magnitude of the comments.
24 CHAIRMAN APOSTOLAKIS: So there is time then, q
to include any major lessons from the pilots, in your
_) 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
14 1 opinion?
,g 2 DR. HOLAHAN: There certainly is, yes, but if 3 -- my concern would be that, hopefully one would make the 4 decisions on the pilots and be comfortable with those 5 decisions and then not change the rules of the game after 6 you've made those decisions. It's rather a clumsy 7 arrangement.
8 CHAIRMAN APOSTOLAKIS: Well, but that's why 9 they're called " pilots". I mean, it's back and forth, 10 right?
11 DR. HOLAHAN: Yes.
12 CHAIRMAN APOSTOLAKIS: I mean, you set the ,
13 rules of the game based partly on what you learn from the fn)
\/ 14 pilot --
15 DR. HOLAHAN: Yes, but hopefully they're 16 converging in such a way that there's still the 17 consistency --
18 CHAIRMAN APOSTOLAKIS: That's something 19 though, that I would like to discuss in more detail 20 tomorrow perhaps, with the general discussion and where 21 we're going from here, because that's really a key I 22 think, to resolving a lot of the concerns perhaps, that 23 some people have regarding these documents.
l l 24 DR. HOLAHAN: Okay. In fairness to the ex
, ( ,) 25 industry, there are individuals here who might want to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
1 15 i
i speak for themselves rather than have the Staff speaking r~ 3 2 for them on this -
%,/
3 CHAIRMAN APOSTOLAKIS: I don't know. Do you 4 want to say something now Rick, or wait for your turn l 5 tomorrow? It's entirely up to you. l 6 DR. GRANTON: We'll be making our presentation i
7 tomorrow.
8 CHAIRMAN APOSTOLAKIS: Okay. But that's 9 something we want to discuss tomorrow in detail and see ]
10 what your perspective will be.
11 DR. CARROLL: I have one other minor issue l 12 that I'd like to get clarified and that'is, when the Staff i I
_ 13 first got into the PRA business in a big way, I saw 4
( )
\- 14 documents that talked about PRA and I saw other documents l l
15 that talked about PSA. I i
16 I guess I'm still seeing it in this stuff, but 17 I can't remember whether it was Staff-generated or whether 18 it was NEI. Have we settled on this consistency? Being 19 the hob-goblin of small minds, that's why I bring it up.
20 DR. HOLAHAN- Foolish consistency is the hob-21 goblin of small minds. I think in this arena we would 22 allow flexibility.
23 (Laughter.)
24 I s^e no reason to be prescriptive and require
/~T
(_);
25 everyone to call it --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
i 16 1 CHAIRMAN APOSTOLAKIS: Remember the comment,
(~~g 2 the first comment we got at the PSA '96 meeting in Utah Y
3 was from a Japanese colleague that said, the title of the i
4 meeting is PSA '96 and you keep referring to it as PRA.
i 5 Why don't you guys settle and market -- when are you going 6 to market? Somebody said that in America it's PRA.
7 DR. CUNNINGHAM: I don't believe that was me 8 but that's the way I -- !
I 9 CHAIRMAN APOSTOLAKIS: You said it but not 10 that time.
11 DR. CUNNINGHAM: On any number of occasions, 12 but probably not there. I l
13 DR. HOLAHAN: I would say in Rockville it's i'-
14 PRA.
15 DR. CUNNINGHAM: Yes. The Commission has a 16 PRA policy.
l 17 CHAIRMAN APOSTOLAKIS: PRA, not PSA. Although 18 we seem to be regulating on the basis of PSA, Dr. Powers.
19 In fact, on the basis of core damage frequency as Dr.
20 Powers has told us many times.
21 DR. CARROLL: But NEI is still calling it PSA?
22 DR. HOLAHAN Yes.
23 DR. CARROLL: Okay, that's --
24 MEMBER POWERS: A clarification, George.
f~
()) 25 We're regulating based on core damage frequency at power NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
1 17 !
l 1 operations.
gm 2 CHAIRMAN APOSTOLAKIS: At power operations.
3 I'm sorry, power operations.
4 MEMBER POWERS: And due to internal events.
5 DR. HOLAHAN: But we're working on rule-making l 1
6 for shutdown. j 7 CHAIRMAN APOSTOLAKIS: What it comes down to 8 is that PRA and PSA --
9 MEMBER POWERS: But it's not risk-informed 10 rulemaking for shutdown. But it will be.
11 CHAIRMAN APOSTOLAKIS: I think in Europe PSA 12 is much more popular. !
13 DR. CARROLL: Yes, it is, for sure.
,a i
U 14 CHAIRMAN APOSTOLAKIS: So it's PRA for us?
15 DR. CARROLL: I don't know what probabilistic 16 safety is.
17 CHAIRMAN APOSTOLAKIS: Sometimes we don't even 18 know what a s t.,e s s m e n t is. This is slide number one.
19 DR. KING: Yes sir.
20 CHAIRMAN APOSTOLAKIS: You only have 55 21 minutes, Tom.
22 DR. KING: We'll go quickly. We'll be at the 23 appendices very soon. I mean, you've seen the objectives I
i 24 of the Reg Guides and SRPs before.
i (O) 25 The only other thing I want to remind the l l
i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS ;
1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
18 j 1 committee is that this whole process is voluntary on l
,s 2 licensees in terms of using these Regulatory Guides and i !
\
\
'~'/ SRPs, and that the Staff will certainly give priority to 3
4 any application that's a safety improvement. And burden 5 reductions that come in and utilize this process will also 6 get priority. That's our way of encouraging voluntary use 7 of this.
8 CHAIRMAN APOSTOLAKIS: I have a question. I 9 was reading the Special Evaluation Report from the Office 10 of the Inspector General on Better Definition and Planning 11 Needed to Guide NRC's Transition to a Risk-Informed 12 Performance-Based Regulatory System, October 4, '96.
13 And they state that the Agency needs to define Ip)
\> 14 performance-based regulation and described how the Agency 15 plans to implement it. And then later on they say that 16 the maintenance rule is not performance-based. So given 17 your bullet-before-last there, describe performance-based 18 implementation strategy, is that part of this? I mean, do 19 we all understand what performance-based regulation is?
20 DR. CARROLL: Does the Inspector General 21 understand it?
22 DR. KING: You'll recall the SRM we got from 23 the Commission on January 22nd asked for a separate plan l
[ 24 on performance-based regulation. This was one element.
,r-x
( ,) 25 CHAIRMAN APOSTOLAKIS: Have we seen that?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 23M433 WASHINGTON, D.C. 20005-3701 (202) 2344433
19 i
1 DR. KING: Yes, we discussed it at our last t
gx 2 couple of meetings. One element of performance-based
(^ ) 3 regulation is in implementing these risk-informed changes l
1 4 and implementing a performance monitoring strategy that we 5 do that in a performance-orie ated f ashion.
6 In sending the policy issues to the Commission 7 we noted that that was one option for dealing with 8 performance-based regulation. There are other options 9 which just take a broader look at the regulations and see 10 if there were other areas outside the risk arena that 11 could be converted to a more performance-oriented 12 approach.
13 The Commission wrote back and said, that i
\/ 14 sounds good, gives us a plan. I forget what the data is 15 out we owe the Commission a plan for this; to take a 16 broader look at things in the performance area.
17 DR. JONES: Late August?
18 DR. KING: Late August.
19 DR. CARROLL: By 8-29-977 20 DR. KING: Yes.
21 DR. CARROLL: Now, the document you're talking 22 about -- I have something called a Pre-Decisional Draft 23 with no date on it -- has been issued, and that's the one l
l 24 he's talking about?
g^)s
(, 25 DR. MARKLEY: That's the one he's talking NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 2344433
20 1 about.
DR. JONES: Yes, January 22nd is the official jq t 1 2
v 3 one. .
4 CHAIRMAN APOSTOLAKIS: Now, do you agree with 5 this though? I'm curious that the maintenance rule is not 6 performance-based.
7 DR. HOLAHAN: I don't agree with that.
8 DR. KING: I don't agree with it either.
9 DR. CARROLL: Jim Curtis wouldn't agree with 10 it either.
11 CHAIRMAN APOSTOLAKIS: But it has -- well, 12 maybe it's not for the present time to discuss that. I m 13 mean, I think the issue is -- okay, you have performance
/ \
\' ~] 14 criteria but the question is, if you meet these criteria 15 what do you do about other regulations that are below? ,
16 That's really the issue they're raising.
17 They're saying that under the maintenance rule 18 you do nothing. You still have to comply with everything 19 else, which goes against the spirit of performance-based 20 regulation, does it not?
21 I mean, it's okay to say I'll have criteria 22 and I'll meet them, but that's not enough. If you meet 23 those then perhaps you should leave the licensee alone 24 regarding how he got there.
iO)
( ,/ 25 And the complaint about the maintenance rule NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
21 1 is that you don't leave them alone. We still want them to rx 2 follow the process below, but I don't think these Guides V) 6 3 do that, or as I remember. I mean, the Reg Guides that 4 they're about to release or -- that's something now we 5 have tc be careful about.
6 DR. KING: Now, we're going to talk --
7 CHAIRMAN APOSTOLAKIS: I know.
8 DR. KING: There's a slide in here you haven't 9 seen before where we talked before about the questions.
10 Is this just an overlay on the existing process, and so 11 forth --
12 CHAIRMAN APOSTOLAKIS: Yes, yes.
-_ 13 DR. KING: -- and when we get a couple more
pages back we'll get into that.
14 15 CHAIRMAN APOSTOLAKIS: Okay.
16 DR. KING: All right, this is just a list of 17 the Reg Guide numbers and the technical areas that they l l
18 cover. You've seen that before. Again, we've discussed 19 the general guidance previously in several meetings.
20 Today we're going to hit the appendices that go with the 21 general guidance and then all the application-specific 22 material, and then --
l 23 MEMBER POWERS: In your general guidance you 24 indicate that any application of that general guidance
(_) 25 should be quantitative and explicitly modeled in the PRA.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234 4433
l 22 l
l 1 True? I mean, it does say that. I wonder if you could l -
2 tell me how you interpret that, especially with regard to
<j 3 graded QA?
l 4 DR. KING: Clearly, graded QA is more of a l
5 qualitative assessment.
6 MEMBER POWERS: You would say don't apply 7 1061?
8 DR. HOLAHAN: No, I wouldn't say that, but I 9 think your quote out of context doesn't tell the whole 10 story about how to implement that Reg Guide. I think the 11 Reg Guide explains a spectrum of approaches which would 12 run the spectrum from using qualitative insights to p._
13 quantitative analysis.
'- 14 MEMBER POWERS: Well, what I'm trying to 15 understand is how I do QA with a PRA, and why I'd even 16 try.
17 DR. HOLAHAN: I think the approach that's 18 being taken is, the guidance in the Reg Guide and the SRP 19 would establish sort of approaches and quality l
20 requirements which are not actually quantifying the 21 effects of QA, nor quantifying the changes in the QA 22 program.
I 23 But you are using the PRA and other insights
! 24 to decide which systems are more important than others, i O
(_) 25 and which we therefore make the assumptions that those NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C- 20005-3701 (202) 234-4433
23 1 ought -- the more important ones ought to have more 73 2 careful treatment in terms of quality assurance.
t) 3 MEMBER POWERS: And again --
4 DR. HOLAHAN: And that result is not, you 5 know, quantitative output but there's quantitative 6 analysis used as part of the decision making.
7 MEMBER POWERS: It's truly risk-informed --
8 this is information-based that you use --
9 DR. HOLAHAN: Yes.
10 MEMBER POWERS: And then changes in your QA 11 program -- it is not an essential step to get some ;
I 12 quantitative measure of changes in your QA program because !
I I
13 you can't, but it's not an essential step to use risk l
("%
k- 14 information to do that.
15 DR. HOLAHAN: That's right.
16 DR. KING: Yes, you use risk information in l 17 the ranking, in selecting your criteria for deciding, is 18 it high safety significance or low safety significance?
19 The Reg Guide also talks about tying that back to the CDF 20 and LERF guidelines in terms of being able to convince 21 yourself that you're still meeting those when you're 22 bidding things in the categories that you select.
23 So I think there is a quantitative input even 24 though you can't, on an absolute scale, quantify the p) x 25 effects of QA.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433
24 l
1 DR. HOLAHAN: And it seems to me that's a very
- 2 sensible thing to do. If the current state-of-the-art i ('~' /
3 says you can't quantify those effects but you can use some 4 insights, you use what you have. Maybe ten years from now 5 we might be able to quantify QA and its effects. I mean, 6 at that stage you can go to another level and you use that 7 information.
8 MEMBER POWERS: And I couldn't agree with you 9 more, and I guess now I'll come to my final question.
10 Does the general guidance adequately reflect the 11 possibility of using PRA in the way you have outlined its 12 use in QA? Or does this -- having statements like this 13 quantitative and explicitly modeled in the PRA deter k/ 14 people?
15 DR. HOLAHAN: Well, clearly the intent is to 16 use it as I described, and we would like the guidance 17 documents to be sending that message. And to the extent 18 that you can help us make it to that better by finding 19 offensive phrases, we'd be glad to clarify them.
20 DR. JONES: I just would make one comment. I 21 think -- I'm not sure where exactly you're quoting from, 22 but certainly in the forward we have put in words about, 23 licensees can and should identify how chosen approaches l 24 and methods, whether they are qualitative and quantitative 1
(_,/ 25 and traditional or probabilistic data, and are criteria-NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
l 25 1 appropriate for the decision to be made.
r^3 2 And that point of doing an appropriate t 4 Q) 3 analysis for the kind of change and the balance between 4 qualitative and use of PRA and quantitative, is buried 5 throughout the document. You may indeed, as I said, find 6 instances where we may have tried to be a little more 7 specific in a few points, but the intent is to allow the 8 broad brush scope and it would be incumbent on the 9 licensees to justify that what they're doing is 10 appropriate for the application in hand.
11 DR. CARROLL: You know, I think Dana's point 12 is very well taken in that I think there are a lot of 13 places in the document where you really look at -- just 7_
()
14 focus on this paragraph or that paragraph. It asks for a 15 lot more than I expect we're really going to need to do in 16 the real world for a given application. ;
l 17 And I guess I'd encourage the Staff to find !
18 somebody to play devil's advocate with these documents and 19 find those places that somebody, you know, with good 20 intention, just went too far.
21 CHAIRMAN APOSTOLAKIS: I think that's the idea 22 of the pilots, right? Which unfortunately --
23 DR. HOLAHAN: And the public comment 24 component.
r- %
(- 25 CHAIRMAN APOSTOLAKIS: Yes, but I think the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISI.AND AVE., N W.
(202) 234 4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
26 1 pilots are really the real thing.
.ex 2 DR. CARROLL: Before you put this thing out on
( )
\_J 3 the street -- once you put it out on the street you guys 4 get a scrooge mentality and don't want to change anything.
5 DR. HOLAHAN: You must be talking about some 6 other guys.
7 DR. KING: Yes.
8 (Laughter.)
9 DR. CARROLL: I would think you'd want to do 10 it internally and really feel good about what you're 11 putting out for publication.
12 CHAIRMAN APOSTOLAKIS: I think the fundamental I
13 question is, can one do a lot of the things that the Reg 1
e 14 Guides have in there, with less than perfect PRA? That's 15 really the issue. And right now if you read the 16 documents, that's not acknowledged enough.
17 The emphasis is on the good PRA -- what would 18 you have to do to have a good PRA? Which I understand 19 fully because, you know, when you put things down on paper 20 you have to think in a certain way. So I think the 21 relaxation of some of these requirements so the quality of 22 the PRA will come with experience, I think.
l l 23 There's no way you can -- for example, when 24 you use importance measures to classify SSCs as a high
' 1 C/ 25 safety or low safety, do you really need a very good NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433 i
27 1 state-of-the-art PRA to do that? It's not clear to me 2 that you need that.
V DR. HOLAHAN: But I think you do need to 3
4 understand the limitations of the PRA --
5 CHAIRMAN APOSTOLAKIS: Certainly, certainly.
6 Yes, yes, yes. But the question is, do you have to have a 7 state-of-the-art PRA for all of these things? Probably 8 not. But that comes with experience, I think. But that's 9 another issue that we will discuss tomorrow.
10 MEMBER SEALE: And on other occasions, also.
11 CHAIRMAN APOSTOLAKIS: And on other occasions, 12 yes. Now, one bullet here that I'd like to ask you a 13 question about is this -- again, the one before -- the k '#
14 penultimate bullet: pilots provide feedback on issues and 15 guidance. To what extent is that really true? I mean, 16 have you really had real feedback from the pilots, or is 17 it sort of a general feeling?
18 And I'll tell you why I'm asking this.
19 Because a lot of people out there are saying that the 20 pilots really are on hold; nothing has happened in the 21 last few months. So I wondered how you got that feedback 22 if the pilots are on hold? ;
23 DR. HOLAHAN: Well, I don't think the pilots 24 are on hold. I think -- we haven't made any final
()
/3 25 decisions on them, but the experience we're getting on the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
i 28 1 pilots has, you know, has been over a fairly long period j 73 2 of time. I think it's fair to say some of that experience (G \
3 was early-on and was perhaps a while ago. You'll probably 4 get a better feel for that when we discuss, you know, the 5 issues one at a time.
6 CHAIRMAN APOSTOLAKIS: Individual pilots, 7 okay. Sure.
8 DR. KING: I mean, the pilots presented 9 practical examples of what licensees want to change, and 10 we had to make sure that our guidance dealt with those.
11 DR. HOLAHAN: Yes, I think we were tested that 12 way --
13 DR. CARROLL: But doesn't overdeal with them?
[~h kl 14 DR. KING: But doesn't overdeal with them, I l
l 15 agree. And we're going to get into the pilots. As we go 16 into the application-specific activities you'll hear about 17 the pilots and you'll see what ch'anges and the delta CDFs l 18 and all of that that go with it.
19 DR. HOLAHAN: I mean, there's one obvious l
l 20 example which is, if we had started writing these j 1
l 21 documents in an abstract sort of way without any '
i 22 experience, we probably wouldn't have written anything on 23 relative ranking.
24 Your tendency would be to write guidelines on
,,- ~ .
( j) 25 how to do a base PRA and use that information. But it's NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS j 1323 RHODE ISLAND AVE., N W.
j (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 t
29 1 become pretty clear that several of the pilots are really l
i i
\/ 14 they now have a process where they've got risk information 15 to help support their argument; they have a process with 16 which to present that risk information; and some l
17 guidelines on how to use that information in the decision 18 process.
19 We're encouraging licensees to follow this box 20 versus this box, but they're not prohibited from following 21 this box. And if they want to come in over here, the 22 Staf f :.s also free to ask questions in the risk arena, and 23 the licensee can either answer them to choose not to 24 answer them. And if they choose not to answer them, we'll
,m k,) 25 review it on its own merits without risk information.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
39 1 MEMBER KRESS: Do the same people do the
,e 2 reviews in each of those boxes, or are they different c 1 LJ 3 people?
4 DR. HOLAHAN. I think there are additional 5 people in the -- with risk information box.
6 MEMBER KRESS: Would there be any priority set 7 on which would be given preference?
8 DR. KING: Well, like I said earlier, the 9 priority would be -- clearly if it's of a safety 10 improvement either one will get priority, no matter which 11 box. But if it's a burden reduction we would tend to give 12 priority to this box.
13 The question that came up before was over i \
U 14 here. If they don't send in any risk information and we l i
i 15 ask for it, or ask questions regarding risk, is that a j 16 backfit? And we've concluded no, that's not a backfit.
17 We can ask -- when you're down in the arena of doing 18 something new, we can ask questions that we feel are 19 relevant to making -- helping us make a decision, and 20 that's not a backfit.
21 MEMBER KRESS: Is part of the definition of 22 backfit -- isn't it predicated on you guys taking the 23 initiative and making the proposed change? Isn't that the 24 difference?
,m
(,) 25 DR. KING: Well I mean, clearly -- backfit's NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
I 40 I i
1 up here; that's the NRC initiative. But there was a 2 question of, since PRA isn't required for currently 3 operating plants, would the fact that we're acking for ,
4 risk information over here be sort of a de facto backfit?
5 But the conclusion, the consensus is no, that's not the !
l 6 case.
7 And what we've done now is added some words in
-8 the Reg Guide that describe this process and describe ,
9 where this Reg Guide fits in terms of, it's really over l 10 here and it's not a backfit if we ask risk information.
.11 So that's a new set of words you'll'see'when i
12 we get you this revised version next week. I wanted.to :
13 point that out.
O 14 DR. HOLAHAN.: It's also fair to say that all j 15 these boxes are not equally likely to be approved. If you i
16 imagine the center-left where things have been approved 17 before, not only is the licensee allowed to submit such a 18 thing, but I think they ought to have the expectation 19 that, if we've approved this before and it's our standard 20 position, then it will be approved.
21 I think what we're trying to do now is to say )
22 -- to tell people what the rules are of the game so that 23 when they do provide risk information in a new 24 environmant, on a new issue, they also expect to be i
25' approved also.
NEAL R. GFU3SS COURT REPORTERS AND TRANSCRIBERS j 1323 RHODE ISLAND AVE., N W.
(202) 234 4 33 WASHINGTON D.C. 20005-3701 (202) 234-4433
41 1 And I think that's one of the differences between the lower-left and the lower-right box.
()
V 2
3 Hopefully, if we do a good enough job we will be defining 4 the process and the guidelines well enough so that 5 licensees ought to be choosing the risk-informed box 6 because they have a lot higher level of information, a lot 7 greater expectation that what they do here will be 8 accepted versus the other box, for which frankly, at this 9 point, I think we don't have, you know, accepted 10 guidelines.
11 As a matter of fact, almost by definition, as 12 things get accepted they would move from the lower-left 13 box to the upper box and you really wouldn't be exploring 7.~ '
(' ,) new territory any further.
14 .
15 MEMBER BARTON: If I choose to keep the 16 existing process on the lower-left box, I could end up l 17 submitting a tech spec change -- let's take extended 18 outage on a safety piece of equipment --
19 DR. HOLAHAN: Yes.
20 MEMBER BARTON: -- I could end up, by 21 additional requests from the Staff, I could end up with a 22 submittal that almost looks like the right box.
23 DR. HOLAHAN: We've talked about that, and 24 that could be the outcome. But I think we decided that
,9 k ,)
m 25 that's probably not a good practice. That in fact, by NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
42 1 laying out this process and by giving licensees a choice
-)J 2 we're saying, if you chose this box, it's not our job to
\
3 push you into this box.
4 We may ask some questions about the risk 5 implications of what you're doing -- and I think that's 6 fair and appropriate -- but to say well, this is an 7 interesting proposal but I'd really like to see all the 8 level of detail in this Reg Guide, I think as a matter of 9 practice -- well, we'll try not to do that.
10 My preference would be to just say, you 11 haven' t provided enough inforraation for us to support 12 changing what has been our standard practice over these 13 years, so we deny your application. And you always have
/,,h 1
'/ 14 the right to bring it back in with more and better 15 information. And if you want to bring it back in in this 16 box that's your choice to do so, but not my pushing you to 17 do so.
18 DR. CARROLL: I think it would be very l
19 tempting, Gary, to push people towards the box on the J 20 right.
21 MEMBER CATTON: Well it seems to me, it's up 22 to the person who's bringing it in. If you wind up in 23 that right-hand box I think you're going to suffer from 24 conservatism. How can you avoid it? And if they want to Ch
( ,) 25 eliminate some of that they're going to have to move to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2: 4-4433 WASHitlGTON, C.C. 20005-3701 (202) 234-4433
43 1 the other side.
fg 2 CHAIRMAN APOSTOLAKIS: Whose right?
3 MEMBER CATTON: I got up there and looked at 4 you and then said "right". Conservatism is inherent where 5 the pencil 1s.
6 DR. HOLAHAN. Yes.
7 DR. CARROLL: The left, Ivan.
8 DR. HOLAHAN: The left.
9 MEMBER SEALE: Ivan has already changed his 10 point of view.
11 MEMBER CATTON: That's right.
12 DR. KING: Mention one other thing. In laying 13 all this out and putting words in the Reg Guide, you'll
- s ,
f i
back in section 3 on documentation, some new words that
'~
14 I
15 deal with -- when you're in this box how do you document - l 16 - there's some words in that ask the licensees to propose 17 as part of their submittal, how they capture this change l 1
18 in the current licensing basis and make sure that's 19 commensurate with the safety impact of the change.
20 I mean, whether it's a tech spec change or 21 whether they want to document it via change in the FSAR or 22 somehow, that tie it to capturing it in something that's l 23 clearly considered part of the current licensing basis.
24 So you'll see some new words about documentation that deal
, ,3
'N ,) 25 with making sure this change is captured in the current NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS
! 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
44 1 licensing basis.
7x 2 MEMBER POWERS: When you make a change using
- )
3 your normal staff review process, that evolves upwards to 4 a change consistent with approved Staff positions. When 5 you make a change using risk-informed regulation because 6 it has all of this body work that's very, very plant 7 specific and institution-specific, does it also evolve up 8 into the licensee requests change consistent?
9 DR. HOLAHAN: Not as you've expressed this.
10 If it's an example which, in fact, the decision really is 11 plant-specific and based on plant-specific information, I 12 think it doesn't. But I think some of the examples we're 13 dealing with might very well lend themselves to moving up k/ 14 into the middle, left-hand box.
15 MEMBER POWERS: Certainly the idea of a change 16 in that area, moves up into that box?
17 DR. HOLAHAN: Certainly the fact that the 18 Staff is receptive to, if not a specific change at least 19 the methodology for dealing with that issue, could move 20 into the box.
21 MEMBER POWERS: But it will clearly become a 22 different way of evolving from new to business-as-usual?
23 DR. HOLAHAN: Yes. And I --
24 MEMBER POWERS: Does this translate into a
,n
() 25 savings of staff resources and time?
NEAL R. GROSS COURT REPORTERS AND TRAN'iCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
45 1 DR. HOLAHAN: One would hope so, but I think 2 that remains to be seen, really.
7-~\
U 3 MEMBER POWERS: Yes, I think it remains to be 4 seen.
5 MR. ABBOTT: I have a question about the l
6 current licensing basis in the elements and where those 7 elements fall in the boxes on the chart. The current 8 licensing basis, according to the definition at least, is 9 fairly complicated; there's a lot of things in it.
10 If I change the FSAR I can do it under 50.59; 11 if I change a procedure I can do it under 50.59; if I j 12 change a come.itment in a licensee event report or in an 13 inspection report, I can probably do that consistent with
! )
\- / 14 the change process of the tech specs.
15 Isn't the bulk of the work that's done on a l l
16 current licensing basis at the first level, and most of 17 what we're talking about here is a fairly small segment of l
18 the world that ends up being performed on a current 19 licensing basis, and yet we're applying more complex 20 process to that than we would if we were doing a 50.59 21 evaluation? Is that a clear question?
22 DR. HOLAHAN: Yes. Yes, I think I understand 23 the question and I think -- it's exactly appropriate, 24 because if you look at the other change processes, 50.59
/ \
/ \
(_,/ 25 is basically a test to see whether the change is really NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D.C. 20005-3701 (202) 234-4433
._ . . _ . . _ _ _ _ _ _ . . _ _ _ _ _ _ . _ _ _ _ . _ . _ _ _ . _ _ . ~ . _ _ . . . _ . . . _
i 46 l' consistent with the original licensing basis.
1 ~
2 MR. ABBOTT: It's three questions.
{
i v 3 DR. HOLAHAN: It's three-and-a-half questions, e-
- 4 depending upon how you count them. Some people like to i
i 5 count seven, but they're a number of complicated ,
- 6. questions.
7 But the answer to those questions really i
8 confine the licensee I think, rather tightly not to change 4 .
9 things, which would be significant risk changes. .;
i !
4 '10 And so the fact that the licensee can do it on i 1 . l 11 their own and that we haven't written additional 60 pages !
l
! 1 12 of guidance on how to do it, well, maybe there's not a lot .!
= !
13 of risk associated with those changes. At least there o
j 14 certainly wasn't intended to be.
)
g_ 15 And where there are licensee commitments -- !
- i 16 and those are things that are not captured in the FSAR or
4 f 1 i 17 tech specs -- presumably-those are also-lower-level i 18 details for which changes of those commitments, you know, i 19 through the NEI process or you know, some respectable 20 process, isn't likely to produce a big risk effect. So l 21 that's okay.
t 22 Now, here what we're talking about -- and I !
23- think you're quite right that this is numerically a small l
24 fraction of the changes done by licensees -- but it's the l 25 ones that involve the most regulatory decisions and the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234 4 33 2 -. _ . , - _ _ _ _ - - . _ _ _ _ _ , _ . . . . . . . _ , _ , , ~ . - - -
__ _ - ~ .
47 1 biggest potential risk changes, because we will be 3 2 changing things like tech specs and broad programs like 7~A N~
3 IST -- the kinds of things that are too big to be changed 4 under 50.59 and commitment processes and those other 5 things.
6 So this is where the big -- effect-wise, this 7 is where the big changes can take place, not in those 8 other programs.
9 MR. ABBOTT: Let's just try an example. A few 10 years ago a BWR replaced the recirculation piping, and 11 that was done on 50.59.
12 DR. HOLAHAN- The first dozen or so were done 13 under license amendments, and then the later ones were i'--) 14 done under 50.59, 15 MR. ABBOTT: Well I mean, if I did it today 16 I'd probably do it under 50.59.
17 DR. HOLAHAN: Fair, yes.
18 MR. ABBOTT: Now, why is that -- I mean, 19 that's a fairly big change in terms of the physical work 20 that's done --
21 DR. HOLAHAN: Physical change --
22 MR. ABBOTT: -- and effects it has on other 23 systems.
24 DR. HOLAHAN: Physical change is large.
im k,) 25 Economically it may be large -- when I was talking earlier NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
48 1 about big changes, I'm talking about in a safety 2 perspective. From an economic perspective of the l lq '
6 ,
Li '
3 licensees, some of their other commitments in 50.59s may 1 4 in fact, involve a lot of dollars.
5 But if a licensee is doing a BWR recirc pipe 1
6 replacement and it's under 50.59, that's basically -- it j 7 fits under 50.59 basically because they're putting the 8 same design back in. They're doing a replacement-in-kind l
9 or some sort of improvement or -- certainly no eignificant i
10 changes. l l
4 11 And so the risk of running a plant after that 1
12 change is either lower or no different from what it was l 13 before. In which case I don't think it deserves a lot of g)
L V 14 Rockville treatment.
15 CHAIRMAN APOSTOLAKIS: You have five minutes.
I 16 Speed up.
17 DR. KING: Okay. I only put this viewgraph in l
18 to say these are the pilot plants and the technical areas 19 that we'll be talking about later, and we'll be getting 20 into the CDFs and LERFs associated, and then examples of 21 the types of changes they are proposing. So when we talk 22 about pilots there's a number of them that will be coming 23 up as we get into the application-specific stuff.
24 Okay, just two more viewgraphs just to refresh
,/3
() 25 you memory on the overall approach --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
49 1 CHAIRMAN APOSTOLAKIS: We've seen those many l g 2 times now. )
3 DR. KING: Okay. Let me just point out one 4 thing.
l 5 CHAIRMAN APOSTOLAKIS: Oh , he increases the 1
I 6 risk.
7 DR. KING: The fourth bullet here, which is 8 the fourth principle on risk. That's the new wording.
9 CHAIRMAN APOSTOLAKIS: Yes, replaced 10 " insignificant". I i
11 DR. KING: We got rid of the " insignificant" l 12 and we got rid of that word elsewhere in the Reg Guide --
l 13 CHAIRMAN APOSTOLAKIS: So the only troublesome j iv/ 14 bullets now are the second and the third?
15 DR. KING: We're going to talk to defense-in-16 depth and safety margin tomorrow in the wrap-up where we 17 get into the details.
18 CHAIRMAN APOSTOLAKIS: The verb " maintain" 19 implies --
20 DR. CARROLL: The fourth one has some problems 21 too.
22 CHAIRMAN APOSTOLAKIS: Which one?
23 DR. CARROLL: The fourth one.
24 CHAIRMAN APOSTOLAKIS: Increases in risk and
,e V 25 their cumulative effect --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433 1
50 1 DR. CARROLL: I'm sorry, the fifth bullet.
,r s 2 CHAIRMAN APOSTOLAKIS: Oh, the fifth.
]
3 Implement --
4 MEMBER CATTON: Are there any that you're 5 willing to tell him are good?
6 DR. CARROLL: One is okay.
7 MEMBER SEALE: Two is not bad either.
8 DR. JONES: Recognize these are a paraphrase 9 of what's in the document, to some extent.
10 DR. HOLAHAN: Right.
11 DR. JONES: I mean the specific words have not 12 been changed for items 1, 2, 3, and 5 from the earlier 7,
13 draft we gave you following the review, and they're more i
S- 14 explicit. The last one for example, Jay says, 15 performance-based implementation and monitoring strategies 16 are proposed that address uncertainties in analysis models 17 and data and provide for timely feedback and corrective 18 action. So it's different than what's there. This is a 19 summary.
20 DR. CARROLL: Nice words.
21 DR. JONES: This is a summary of what's there.
22 DR. CARROLL: That's a nice platitude; what 23 does it mean?
24 CHAIRMAN APOSTOLAKIS: Okay.
\._,,/ 25 DR. KING: Let's just go to page 8 and talk NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
51 !
i 1 about appendices, because there wasn't anything new on j es 2 page 7. Starting with the General Reg Guide appendices, U 3 there's two appendices now in the General Reg Guide.
l l
4 Appendix A deals with categorization of i t
5 systems, structures, and components. It's really in there 6 to provide some guidance on the important factors that !
l 7 need to be considered when you're doing categorization of l I
8 safety significance. It describes technical. issues that 9 need to be considered by the licensee in' developing and ;
10 using importance measures.
11 It's written in a very general sense, which we 12 feel is consistent with the overall thrust of the Reg 13 Guide in that it provides licensees a lot of flexibility O 14 in terms of: which important measures they choose to use, 15 what are the criteria for being in the high or low 16 category, what kind of sensitivity studies do they need to 17 do in looking at how the risk rankings would change if 13 you'd make different assumptions?
19 And it also references NUREG-1602 which is the 20- supporting document that expands on each of the points 21 that are in this appendix in terms of describing specific 22 risk-importance measures and other things that go into i 23 applying them.
24 CHAIRMAN APOSTOLAKIS: Now, this is just an 25 overview --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
WASHINGTON, D.C. 20005-3701 l (202) 234 4433 (202) 234-4433 l
52 1 DR. KING: Yes. What we suggest to do is, !
,r's 2 'his is closely related to the SRP Appendix C which covers
( )
v 3 the same subject but provides guidance to the Staff, but 4 they parallel in terms of topics they cover.
5 What we suggest to do is wait until we get to 6 SRP Appendix C to go through each of the items in more 7 detail, and let us go right to Appendix B at this point on 8 the General Reg Guide, which is the one dealing with how 9 do you estimate LERF when you only have a Level 1 PRA?
10 CHAIRMAN APOSTOLAKIS: So you're done with 11 your overview if I look back at the agenda? We're now 12 discussing the appendices?
13 DR. KING: Yes. We're ready to go into a more
( )
'/ 14 detailed discussion of Reg Guide Appendix B.
15 CHAIRMAN APOSTOLAKIS: Okay.
16 DR. KING: And what we have for the detailed 17 discussion is, Trevor Pratt from Brookhaven who desaloped 18 the Appendix B, is going to give a presentation providing 19 description of what's in there and why it's in there, as 20 well as the results of some trial applications we've done i
21 where we've taken Appendix B and we've taken some plants 22 that have a Level 2 PRA; that we've ignored the Level 2 j 23 PRA part and just taken the Level 1 and applied Appendix B l 24 and see what comes out, and then compared it to the actual p
(,} 25 Level 2 results.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 2344433
. . . . - . . ~ . . . . . . - . . .. . . - . . . - - - - . - .
53 ;
i 1 So I think this would be a good way to walk i 2 through Appendix B. So at this point I propose we turn it O 3 over to Trevor and go through that.
f 4 MR. PRATT: You should have a package that's 5' been issued to you all? Okay. What I thought'we would do
+
l 6 is.go through a little bit of the background first and .!
j 7 show you why we decided to come up with this approach, 8 then go into the approach in a little bit of detail. And
! [
L 9 then discuss the event trees -- one in particular that we :
i l.
l 10 developed for the various containment types.
11 Basically what we did is generated a couple of l i
12 teams at Brookhaven and Sandia, and so the Level-2 guys 13 did the PWR event tress at Brookhaven and Sandia did the l
[ -
14 BWR event trees. When they.had finished that information 15 and we wrote the guidance document', we gave it to some 16 Level 1 analysts and asked them to apply the results of 17 the trees and see how they came out with the IPE results.
18 MEMBER POWERS: Mr. Chairman, once again I'm 19 reminded that I've got to hae more time at Sandia. It 20 sounds like an interesting place to work. j 21- (Laughter.)
22 Presumably, some fraction of this I have a f 23 conflict on.
l 24 MR. PRATT: What this slide does is refresh 25 your memory in terms of what we're using as the metrics NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISt.AND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
54 1 here: one is the core damage frequency, one is LERF.
2 The core damage frequency of 10-' and the
(~)
V 3 number of occasions we've met with you and mentioned --
4 Rick Sherry has made presentations and I have --in 5 relationship to how we compare the IPEs to the safety 6 goals will let you know that the core damage frequency of 7 10 -
- is more limiting than you would get from the 8 quantitative, latent cancer fatality health objective.
9 I think that's an important point because 10 right under that what it says is, so long as you have a 11 containment in which you are monitoring the core damaging 12 f requency of 10-4, you cculd have that containment open and 13 still meet that particular health objective.
l\ ') 14 So what that means is I don't need to spend l i
15 time in calculating whether I've got late containment 16 failures or whether I don't have late containment I 17 failures, whether it's vented late or not, because those 18 types of failure modes contribute to latent cancer They don't in general, contribute to early i 19 fatalities.
20 fatality risk.
21 So I don't need to spend time in my 22 containment of entries allocating as to whether the core 23 debris is coolable, whether it goes through the basement, 24 or whether it flows late as a result of g~3
(,) 25 overpressurization. So that observation simplifies my NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
55 1 event tree significantly, right there, and it's an
, -~s 2 important point which I think we've made on a number of 3 occasions.
4 So what that means is then I have to focus in 5 on trying to estimate those events which I think are going 6 to result in an early containment failure and which, more 7 importantly, are going to result in early fatality risk.
8 It's really only those types of events that I'm trying to 9 monitor here.
10 What I'm trying to do is take the large number 11 of core damage accident sequences, and if you like, 12 allocate those into those types of events that I think l 13 might bypass the containment building -- in which case I k ')
14 think I have a potential for early fatality -- and for i 15 those cases which might present these large epistemic
]
1C uncertainties that we've talked so much about in the past, 17 where there is a significant uncertainty with regard to 18 containment performance for this particular accident 19 sequences.
20 So if I can identify those accident sequences 21 I can then allocate those to LERF and deal with them to a 22 slightly higher standard than I would the rest of the core 23 damage frequency. And that's really the approach that 24 we're dealing with.
(~\
(_) 25 A little bit more background, and what it says NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l l (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
56 1 is, really the way we use this in the overall Regulatory r" 's 2 Guide. And the objective that's in the Guide and
~
N~ Y .
3 permeates through it, is you use this risk-informed 4 information as the best you can.
5 So what we're trying to do is, we would like 6 to establish a process in which we can quickly decide 7 which accident sequences we want to allocate into this 8 LERF bin, and worry about those in a little bit more 9 detail then we would perhaps, the other accident 10 sequences.
11 Also, if we are close to the decision criteria 12 using this approach, then obviously that would kick us in
,s
- 13 and say, well we need to look at those things in a little 14 bit more detail because the Regulatory Guide says that if 15 we're close to the decision criteria we really need to do 16 a more formal analysis, we need to do uncertainty studies, 17 and we need to perhaps focus in on those types of events 1
18 in more detail.
19 If you are far away from LERF then maybe you 20 have some confidence that this approach would tell you )
i 21 that okay, I'm sufficiently far away from my decision 22 criteria that I can feel comfortable with my -- so it's I
! 23 really the way that we would use this.
24 MEMBER KRESS: Trevor, could you tell me what
,i (Q/_
25 role site-specific characteristics on meteorology and NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l WASHINGTON, D C. 20005-3701 (202) 234-4433 l (202) 234-4433 i
57 1 population show up in this?
,r 3 2 MR. PRATT: In this approach?
NY MEMBER KRESS: Yes.
3 4 MR. PRATT: They do not.
5 MEMBER KRESS: Well, I think they're implied 6 in the 104 7 MR. PRATT: Yes, I mean, in terms of the 8 decision criteria they creep in, but in terms of the 9 approach that we're adopting here we're really not taking 10 into account things like meteorology and so on. We're 11 just identifying those events which we think have the 12 potential for an early fatality -- whether or not they do 13 would depend on the site-specific consideration.
i \
14 MEMBER KRESS: I guess I'll ask the question 15 of the Staff. If the licensee or plant comes in with a 16 proposed change and they're in one of these gray areas 17 that you have to have more information, and if they show 18 you that their site and population distribution is such 19 that they could actually be granted a much higher LERF and 20 still meet the safety goals, would that be part of your 21 decision process on granting the change?
22 DR. HOLAHAN- Yes, it would be part but not l
l 23 the whole of the additional consideration.
24 MEMBER CATTON: Does it say that?
f (3_) 25 DR. KING: Yes.
- NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS i 1323 RHODE ISLAND AVE., N W.
(202) 234 4 33 WASHINGTON, D C. 20005-3701 (202) 234 4433
58 1 DR. HOLAHAN: Yes. If you look at the list of )
f- 2 additional things that the Staff expects us to look at for
\
3 those ones in the management decision -- in this, what we 4 call the gray area -- you'll see it's in the list.
5 MEMBER CATTON: Okay.
6 MEMBER KRESS: Thank you.
7 MR. PRATT: Yes, I think the bottom line is, 8 what this would do, the simple approach will do, is tell 9 you when you're close to that number. If you then wanted 10 to pursue it further you can always do that and 11 demonstrate that perhaps this was a little conservative 12 and for my site, the site conditions are such that it's l l
13 less.
(~)
~) 14 Okay, the approach was to -- and I've said 15 this on the previous slide -- to really focus in on l
16 identifying bypass sequences and identifying those types l 17 of conditions that would result in early containment 18 failures.
19 These are the two categories that we 20 essentially are bending the event tree into. Those that 21 we think have the potential to cause early fatalities and 22 those that don't.
l 23 MEMBER KRESS: You throw away the second --
24 just discard the second group?
25 MR. PRATT: Discard the second group? Well,
( )
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234 4433
I 59 l
1 the second group remains, basically, the CDF which is the i g- 2 thing that you have to ensure is below 10-*. Or, not
('~/
3 ensure that below, but deal with it in terms of the l
l l 4 decision criteria as we have in the guide. So the aim --
l 5 DR. CARROLL: Now, if it is above 10'4, it 6 doesn't apply at all, or - -
7 MR. PRATT: Well again, the way you would deal 8 with that is as we discussed in the Regulatory Guide. You l
9 can compare the core damage frequency to 10'* and with 10 various guidance. If you're above, we would expect you to 11 be risk neutral or go down. If you're below then you 12 would look at the small changes defined as about ten 13 percent of the other number.
(~'N l
r x> 14 So this is not addressing that. This simply i l
15 takes the core damage frequency and out of that, 16 establishes what a frequency of LERF would be.
17 MEMBER FONTANA: Trevor, those things that 18 lead to the containment bypass are pretty straightforward.
19 With respect to early failure, what's your definition of 20 "early"? Related to evacuation time or what?
21 MR. PRATT: Yes. Well, I mean, we had a
- 22 definition on a previous slide, but basically --
l 23 MEMBER FONTANA: Did I miss it?
24 MR. PRATT: -- what we're referring to there 25 for PWR is events that are fairly close to the time of NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS i 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
l l
60 i 1 reactor vessel failure, and for BWR is,within a couple of. l q 2 hours of reactor vessel failure.
%)
3 DR. CARROLL: Why the difference?. j 1
4 MR. PRATT: Because we usually find on BWRs if l i
5 they go early they're going to go close to the reactor l:
6 vessel, and failure time because the failures are usually l 1
7 associated with the events associated with reactor. ;
L 8 pressure vessel failure. ,
9 'On PWRs, because of smaller pressure !
l 10 suppression containments, they can fail by overpressure ,
11 and over-temperature relatively quickly compared to the .
12- large volume containments. And those types of failures 13 which are gradual pressure builder, which could have been 14 defined as late, can-actually lead to early fatalities in 15 certain circumstances. So we didn't want to ignore those.
16 MEMBER KRESS: You seem to have the same value i
-18 approach it considering that a large number of the 19 accidents -- even with early containment failure -- the i
l 20 fission products are scrubbed through a suppression pool?
l l
l 21 MR. PRATT: Well, on the event trees I'm not .)
22 actually going to go through those, but on the event trees 23 described in the Appendix, we do distinguish between Bs ;
i
!' 24 and Ps in terms of the potential for early fatalities by
\'
25 taking into account pool scrubbing.for the BWRs versus the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
61 l l
1 Ps. I i
l 7\ ')
2 So you have an opportunity there if you
%,/
1 3 believe that you have significant pool scrubbing to get 1
4 out of the LERF category for a BWR even if you have an 5 earlier failure.
6 DR. KING: Yes, I think your questions may be 7 more fundamental. I mean, we have one safety goal -- not 8 a safety goal for PWRs and another one for BWRs.
9 MEMBER KRESS: Yes, I understand that.
10 DR. KING: So when you work backwards you come 11 up with one LERF.
12 MEMBER KRESS: Yes, but the LERF you have is 7,
13 basically a containment failure frequency, which doesn't
- r c
)
14 capture the fission product behavior as it moves through 15 the pool -- the BWRs. It doesn't capture that.
16 DR. KING: But what goes into that LERF is 17 ale > , core damage frequency, is part of the consideration 18 in coming up --
19 MEMBER KRESS: I understand that's part of it, l l
20 but -- )
21 DR. KING: And you've got some room to work 22 between the CDF of 10" and the LERF limit of 10"'.
l 23 MEMBER KRESS: But basically you don't treat 24 BWRs any different than PWRs? I 1
f~'s t i
(_,/ 25 DR. KING: In terms of the absolute value of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS j 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
l l
62 1 the guideline, no. But in terms of how they demonstrate how much credit they take for CDF and demonstrating LERF, i
ex 2 V) 3 there is some maneuvering room in there for a BWR to take 4 more credit for its lower CDF.
5 MEMBER SHACK: I thought I read somewhere in 6 the appendix that in fact you redefine LERF for a BWR; 7 that it's only those sequences that aren't scrubbed?
8 MR. PRATT: Yes.
9 MEMBER KRESS: Okay, I'm sorry. If I had 10 heard that I would have been a lot happier. That helps me 11 a lot. Thank you, Bill.
12 MR. PRATT: Yes, I mean, we could have --
13 MEMBER KRESS: So it's basically the bypass i <
14 sequences --
15 MR. PRATT: Yes, I mean, we can go to the 16 actual event trees later on --
17 MEMBER KRESS: Okay, well then that does l
1 18 account for --
19 MR. PRATT: Yes. ,
l 20 MEMBER SHACK: It makes a big difference.
21 MR. PRATT: It makes a big difference. And in j 22 fact, when we get to the results you'll see that quite 23 well in terms of the difference between the frequencies of 1 l
l 24 early failure and bypass in the BWRs and the frequencies i
O)
( 25 of source terms with release fractions larger than, say NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
63 1 .1. There's a big difference in the BWR which is a point i
,-~ 2 you're making, compared to the PWR where there's 3 essentially no difference.
4 MEMBER KRESS: Yes, you make a big difference 5 in your definition of LERF, actually.
6 MR. PRATT: Right, and the way -- and I think 7 the confusion is, we don't define LERF differently in 8 terms of the criteria but in terms of quantifying the LERF 9 number, then --
10 MEMBER KRESS: You can do it either way, and 11 that's a perfectly acceptable way to do it.
12 MR. PRATT: Yes, right.
13 MEMBER KRESS: Thank you. That clarifies it.
\- 14 MEMBER POWERS: I'm still a little bit puzzled 15 why this is a useful definition -- useful to pursue this.
16 I can understand why core damage frequency might be a 17 useful measure, because we can all agree that core damage 18 is a bad thing and it really does not matter to us whether 19 it's a little bit of core damage or a lot of core damage; 20 core damage is a bad thing. )
l 21 MR. PRATT: Right. l l
22 MEMBER POWERS: But now when you talk about 23 containment failure and fissure product release you say 24 that's a bad thing, but isn't it important to know whether p
(_,/ 25 it's two times bad or three times bad or ten times bad?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
64 1 In other words, is it enough to know that I've got an i
-s 2 early fission product release, or do you have to know more
/ \
'x '/
3 about the magnitude of that fission product release?
4 MR. PRATT: Well, I think the way we're 5 dealing with it here is that we have two objectives that 6 we're trying to meet, and we hold one to a slightly higher 7 standard that's about an order of magnitude difference 8 between the core damage frequency and the LERF.
9 What this scheme tries to do is it tries to 10 identify those subsets of the core damage accident 11 sequences that might result in an early containment 12 failure --
13 MEMBER POWERS: I think I understand that.
' - ' 14 MR. PRATT: Okay, and so the objective then, 15 is to come up with a fairly simple approach to identify 16 those accident sequences and then hold them to a slightly 17 higher standard. So for the bypass sequences it's very 18 straightforward. We know those things have a potential 19 for early failure and we know that we need to worry about 20 those.
21 For the rest of the core damage accident 22 sequences we know that certain of those accident sequences 23 have got a lot more at the -- uncertainty than others.
24 For example, if we have a high pressure core meltdown
(~'
( ,%) 25 event we know that's a significant challenge to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
65 I containment integrity, particularly in the BWR, compared
- 2 to a low pressure sequence.
p V 3 So what these trees are trying to do is 4 identify those types of plant attributes that we think 5 would cause more uncertainty, more of a challenge from a 6 containment point of view, and then deal with those 7 separately and hold them to a slightly higher standard 8 than we would the rest of the core damage --
9 MEMBER POWERS: Yes, I think I understand what 10 this particular approach does, and I understand it as an 11 attempt to accommodate the requirement that you maintain 12 defense-in-depth. And you've said, okay, I've got two 13 sets of barriers that are damaged by core damage, and now
(' ) 14 I have another set, a set that I want to be much more 15 consorvative than the first set, that are breached by --
16 that describes it.
17 And that means I've now got one barrier left 18 and that's my evacuation or sheltering kind of programs, 19 and so I definitely want to know when that's my last 20 barrier to this. I guess what I'm asking, is that enough?
21 Or, what I'm asking is, why is that enough? wny is it 22 that I don't have to know more about this final breaching?
l 23 MR. PRATT: Well, I think you may well need to l
l 24 know more, depending upon where your calculated LERF from i
rx I \
i (_/ 25 the system comes in relationship to your decision NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C, 20005-3701 (202) 234-4433
66 1 criteria. The way we have formulated this, we believe r~ N, 2 these numbers to have captured most of the results out I (x/
l l
3 there.
4 They're certainly not going to bound 5 everything that we've seen in the IPEs because some of the 6 numbers were very high and we know why they're high and we 7 did not try to bound those.
8 We have tried to bound a good percentage of 9 the previous studies that we have seen. So if you run 10 through this calculation and process and come up with a 11 LERF that's relatively low, we would have reasonable 12 confidence in that number.
13 If that LERF is very close to the decision 7s i i )
'~-
14 criteria, then I think you want to look at it in a lot 15 more detail, particularly if you'd like to make changes 16 that would affect those accident sequences that are 17 contributing to your LERF.
18 MEMBER POWERS: I guess what I'm still 19 struggling with, and it may be some -- maybe Dr. Kress can 20 help me here a little bit -- that suppose indeed I came up 21 with a LERF of 10" that would always bypass accidents in a 22 high population zone. Isn't that qualitatively different 23 than a 10 4 that's always containment overpressurization 24 events in a low population zone?
f%
k,) 25 MEMBER SEALE: Could I ask the question a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 1
67 1
1 little bit differently? How do you decide that an ERF is
<3 t 2 a LERF?
'O 3 MR. PRATT: How do I decide --
4 MEMBER SEALE: How do you decide that an early 5 release fraction is a large early release fraction?
6 MR. PRATT: This method does not do that.
7 MEMBER KRESS: Dana, on that thing that I 8 wrote up, if you use 10 4 for LERF, there's an implicit 9 assumption in there that the release of the fission 10 products is relatively high. Because if you have a 10-*
11 containment failure and an overall low that's 10", and you !
12 put in a number for the fission product release associated 13 with that, it will be a big number to get the quantitative p_
i
\' 14 health objective on early fatality. ;
15 So there's really an implied assumption that -
16 - and it's a relative large release of fission products in 17 that number that I chose in the first place. And whether 18 it's due to early conta.inment overpressurization failure 19 or bypass, it really doesn't distinguish between them; it 20 just says both of them have a large release.
21 MR. PRATT: I mean I think -- again, the 22 bottom line here is, if you're bypassing the containment 23 with an interfacing system LOCA then we know the release 24 is large unless it's somehow scrubbed through a pool.
p
(,) 25 For the other early failures we're looking at NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHOD 5 ISLAND AVE., N W.
(202) 234-4433 WASH!NGTON, D.C. 20005-3701 (202) 234-4433
l 68
, 1 pretty battered ends. We're looking at rapid l ,3 2 pressurization which causes structural failure; we're
\ t a 3 looking at liner melt through which has been shown if you l 4 don't go through the pool or if there isn't water around 5 to results in large source terms. So we don't quantify )
6 that; we make the presumption that if these failure modes 7 have occurred, that it's going to be a rather significant 8 release.
9 Now, we do in the trees when we get to it, ask 10 the analyst to make a decision about: one, for BWRs is it 11 scrubbed -- if it is scrubbed then we lower it; and two, 12 if the release timing relative to the declaration of the 13 general emergency is such that you can evacuate, then we
/,_T
)
\ / 14 give the opportunity there to not have early fatalities 15 and not have a LERF.
16 So you do have some flexibility. As I say, 17 these are -- I thought it was interesting when you said 18 that based on NUREG-1150. I mean, in NUREG-1150 we had a 19 lot of questions in our event tree. It's not hard to put l l
20 the questions in; it's very hard to take them out and to ,
1 21 get sufficient number of questions in there that you've i
22 captured something meaningful, but yet not made it too l
l 23 complicated as to make it workable. And that's what we're 24 really trying to balance here.
- n
(,) 25 MEMBER FONTANA: Trevor, you mentioned before, NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4 433
69 1 this is primarily a screen and if you get close to the fs 2 decision points, does that mean that you do a more YY 3 detailed analysis? You actually do something more like a 4 Level 2, is that what you're going to end up --
5 MR. PRATT: Well again, I think that's a point l 1
6 that's been made all along. It depends on the application 7 and it depends on what you're wanting to do.
8 MEMBER FONTANA: I do have a little problem 9 with the implication of the last bullet. Is that failure 10 modes could be identified on the status of the plant 11 rather than on outcome of uncertain phenomenological 12 issues. Of course the plant damage data is very important l l
13 and dominate a lot, but I'm not so sure that assessing a I
/~N l l
t I
\' 14 plant damage state is any less certain than some of the 15 phenomenological issues that we deal with. The structural 16 guy over there can tell you some things about that.
17 I think that thread goes all the way through 18 here that the determination of probabilities of failures 19 and that sort of thing and the determinations of plant 20 damage states which involves the loads but also the 21 structural behavior of the plant, is easier and more 22 precise than the phenomenological issues, and I don't 23 think that's true.
l
! 24 In other words, it seems that the kind of
/~N
(_,) 25 analysis that you have in a Level 2 assessment is deemed NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
70 1 to be a lot more uncertain than a CDF kind of' analysis,
'2 and I'm not really-sure that's the case.
3 MR. PRATT: I see the point you're making. I 4 think the main point we were trying to make though, is if 5 we had an area where there was a very large.epistemic 6 uncertainty, like a high pressure core meltdown event --I 7 mean, when I first did these trees a year back I said,-if 8 I'm going to have pressure core meltdown-event I'm going 9 to put that into LERF and I don't even want to discuss 10 whether it's .3, .5, point-whatever.
11 And you've had a presentation from Professor 12 Theofanous saying that, you know, putting conditional 13 probabilities in there are probably meaningless anyway. I ;
14 mean, I had the original -- that if it was a bad set of ]
15 circumstances I was just going to put it into LERF then- l l
16 use it for a screening and if I was close then do a little I
- 17. bit more if I needed to do more.
~
18 We backed down from that a little bit and we 19 have incorporated just one question that we prescribe and 20 that's dealing with the probability of failure at the time 21 of reactor vessel failure. And we've come up with numbers 22 that we think are representative of the types of studies j l
23 we've seen out there, and it give you a little bit finer 24 detail. ]
25 But I certainly wouldn't want to imply that NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
- (202) 234-4433 - WASHINGTON, D.C. 20005-3701 (202) 234-4433
r.
-71
] ,
1 all of'the. areas, you know, in the Level 2 are more 2 uncertain than some of the Level 1. .But certain areas in O 3 particular are subject to very large uncertainties. L 4 MEMBER SEALE: Dr. Kress reminded us, and I l L !
l 5- recall that of course,.that the' basic assumption with {
6 regard to LERF is that it is large release. But {
I !
7 revisionism'is not an unknown practice, and I wonder if a !
t 8 footnote or something that would confirm that assumption, l I
9 might not be helpful to guard against reinterpretation at f l
10 some point in the future. f
.i.
11 CHAIRMAN APOSTOLAKIS: Rick Sherry? !
12 DR. SHERRY: What Trevor describes in Appendix 'l 13 B, simplified analysis, seems to be less restrictive than f
() 14 the definition for LERF that's provided in the body of the 15 Regulatory Guide. In fact, I think if you would use a .
i 16 Level 2 analysis and' apply the' definition of LERF in the 17 Regulatory Guide, you may get.a higher value for the LERF 18 than if you use the simplified methodology.
19 And I believe that the assumptions that Trevor 20 makes in the Level 2 simplified methodology with respect 21 to considering scrubbing so you can have some handle on 22 the magnitude of release, and the timing issue -- timing 23 of release versus timing evacuation -- are important to 24 consider.
25 In fact on a slide that Trevor will present NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4 33 WASHINGTON, D C. 20005-3701 (202) 234-4433
72 1 later, you'll see for his two examples that hic simplified
,y 2 methodology actually gives lower values than the IPE Level
!s.-)
3 2 analysis using the regulatory guide definition for LERF.
4 MEMBER KRESS: It's less conservative, you're 5 saying? .
1 6 DR. SHERRY: I think the methodology he 7 proposes in the simplified methodology is actually more l l
l 8 realistic and gives a truer frequency in relationship to 9 the safety goal QHOs, than the LERF definition in the Reg 10 Guide which is simply essentially, early containment 11 failure frequency.
12 MR. PRATT: Yes, I think up to that point what l g
13 Rich -- this definition that we have here which was on i t
\> 14 your second slide or so, was taken from the Regulatory l
15 Guide. And this is the definition we use in our 16 containment event trees to initially come up with the 17 accident aequences that we believe will cause early 18 fatalities. l l
19 The qualifier that perhaps isn't here that we 20 may wish to add to the guide, is that we go beyond that 21 definition. And if we believe that the fission products 22 are scrubbed and pooled for BWRs, we reduce that number.
l 23 And if we believe that there is ample time for evacuation, 24 we would reduce the number we'd get there.
rm
( ,/ 25 MEMBER SHACK: I could swear that was in the NEAL R. GROSS
- COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l j (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4 33
73 1 definition in one of these versions. There have been so g 2 many of these --
( )/ I agree with you. I thought 3 MEMBER POWERS:
4 it was in there as well.
5 DR. KING: We just read it. It's like word-6 for-word the same.
7 DR. HOLAHAN. That definition?
8 DR. KING: Yes.
9 MR. PRATT: But I think all that Rick is 10 saying and I think it's a good point is, that we need to 11 put an asterisk on that and say, oh by the way, if the 12 early failure is scrubbed then you could reduce --
13 MEMBER SHACK: I think that asterisk was there I';
\ i
~ 14 once upon a time, in one version or another.
15 MR. PRATT: Yes, I thought so.
16 DR. CUNNINGHAM: We've had almost any 17 combination of words you can imagine -- versions of the 18 report. That's something we can go back and look at and 19 make sure that the definition --
20 DR. HOLAHAN: I'd just like to make one point.
21 If you turn the clock back an hour or so, you would see us 22 talking about doing qualitative analysis and not doing too 23 much quantitative analysis where it isn't necessary to the 24 decisions that we're making.
O Q 25 And we all talked about being smart enough to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
74 1 figure out what was the right level of analysis needed for
, ~s 2 the decisions that we were trying to make. That's what
( )
N/
3 we're doing here. Clearly, this appendix is a simplified 4 approach. Maybe it will be conservative in some cases; 5 maybe it will be a little non-conservative in some cases.
6 It's an approximation; it's a simplification.
7 To a certain extent there are qualitative 8 judgments involved. But I think the way it's being used 9 is in fact, not so different from what we talked about a 10 little while ago, about using qualita*.ive and quantitative 11 judgments appropriate to the decision that's being made.
12 So this shouldn't seem very strange. this is 13 very much along the lines about the way the Reg Guide is
(,%
14 structured.
15 MR. PRATT: Absolutely. Okay, in your handout 16 is the PWR tree. We're going to go through the questions 17 in there quickly and just indicate what goes into that.
18 What I would point out in looking at it is !
l 19 that we tried to have most of these branches quantified l l
20 based on the actual accident sequence in the information 21 coming from the Level 1 analysis.
l 22 So the only question that we prescribed in 23 this particular tree is containment failure at vessel 24 breach, which is question 5, and we specify a split
(,) 25 fraction for what that likelihood is. And again, we can NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON D C. 20005-3701 (202) 234-4433
75 1 get into discussions as to what those numbers are when we get to it.
'(~) ,
2
'O 3 One point that was brought up -- Tom, let me 4 show you -- I have backup material which is not in your 5 handout but it is in the actual write-up. That if we look 6 at the BWR tree, the Mark I tree, the point that you were 7 bringing up was the effect of scrubbing. And so there is 8 an opportunity here if you feel that there is adequate 9 scrubbing for that particular failure mode, to take it out 10 of the LERF definition and allocate it to the other one.
11 MEMBER FONTANA: Trevor, while you're'on this, 12 I noticed that the BWR trees -- see where it says, water 13 on the drywell floor? If you were to look at that tree t
\') 14 alone, the water on drywell floor doesn't have an effect 15 on the large early release, because no, yes, no, yes is 16 the same.
17 On the following pages you do have some 18 distinctions.
19 MR. PRATT: I'm sorry, it does not?
20 MEMBER FONTANA: Maybe we ought to go off-line 21 here because it's a little bit complicated what I'm I 22 talking about. The water on a drywell floor is more 23 likely to affect a later kind of release than the early 24 release. Is that a correct statement?
(_,)
25 MR. PRATT: No, this is a Mark I container and NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
. _ . . . _ . _ _ _ . _.. _ _ - ._. - _ . _ .. _ _ _ _ , _ . m._ _ . - _ . _ . _ . _ . _ . _ _ . . - . . . . .
76 i
1 what --- >
l l
- O 2 MEMBER FONTANA
- I know but -- look at the ,
1 3 branch where it says, water on drywell floor ~
i i
4 MR. PRATT: Right.
'i 5 MEMBER FONTANA: Go out to the right -- go up I l
- l. 6 it means you do have water on the right half on the floor, . l t
l 7 right? Then you go look at -- pads 3 and pads 4 and it ;
l i
j 8 says no, yes -- ;
i 9 CHAIRMAN APOSTOLAKIS: Could you point to the t
j 10 screen Trevor, please, so everybody can follow you? !
11 MR. PRATT: Here we -- absolutely that the l 12 split fraction that we prescribed here is quite different ;
i 4
i 13 if you have water in -- l l
i O 14 CHAIRMAN APOSTOLAKIS: Sorry, sorry, you need 1
I l 15 a mike.
1
[ l
' 16 MEMBER FONTANA: Okay, why don't you move on l 17 here --
18 MR. PRATT: Well, but this is an important 19 point. I mean, this is just a structure of the event 20 tree. If you go into the text you'll find that the split 21 fraction in this particular case, the only tree that we -- 1 l
22 the only thing that we specify is seven, and we tell them l i
23 what-the split fracture is there.
24- And if there is water on the floor I think we i
25 have the resolution'of.-- the liner Mark I failure problem
. I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
.(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
1 , .--. .- - -~ - -, .:
77 1 was.that if you have a depressurized reactor vessel, you 2 have water on the floor and you have a pretty good chance O 3 of'not failing. LIf you ever had pressure sequence with no j 4 water on the floor, you've got a bad' situation. So the 5 split fraction that goes in here is quite different.
6 CHAIRMAN APOSTOLAKIS: Here, where? j 7 MR. PRATT: At these two points. ;
'l 8 MEMBER FONTANA: You have to go to other pages 9 in the text to notice that, and then I have concern -- the j 10 split fraction, that's a different argument.
11 MR. PRATT: Yes, and I think those are the 12 things we need to talk about because as I said, it was 13 pulled together by analysts based on their best judgment j 14 and we need a' wider evaluation of those. But just to i
B 15- recap, the point you bring up is important, i
The only question that we specify on this tree j 16 17 is this question, and we would have quite different split 18 fractions for each of these conditions. And those split 19 fractions depend on whether it's at high pressure, whether 20 it's at low pressure, whether you have water, whether you i
21 do not have water, j l
22 Now, if you go to NUREG-1150 you'll see an i
23 awful lot more questions deciding on what these types of 24 split fractions might be. We boiled it down to these
("% i
\m,/ 25 parameters because we were trying to simplify the !
I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
78 1 approach, and that's where, you know, this business of
,r 3 2 trying to go from 80 questions down to seven, you know, t <
'q,I 3 comes in.
4 MEMBER KRESS: That event wouldn't even show 5 up on Mark IIIs or Mark IIs?
6 MR. PRATT: Pardon?
7 MEMBER KRESS: That event wouldn't even be on 8 your tree for Mark IIs and Mark IIIs?
9 MR. PRATT: The scrubbing -- the line of flow, 10 no it does not. It does not; it's in Mark I.
11 MEMBER KRESS: So that has to do with melt 12 through --
13 MR. PRATT: What comes into the Mark --
r'N
! )
V 14 MEMBER POWERS: Trevor, I believe water on the 15 pedestal floor does show up on your Mark II.
16 MR. PRATT: On the Mark II, sorry. But not 17 the Mark III. What comes into the Mark III in the ice 18 condenser is the operability of the igniter system, which 19 is an important element of that containment event tree 20 which does not apply of course, to the Mark Is, Mark IIs 21 of the large volume containments.
22 So there are minor differences between each of 23 the trees; not too many differences. We tried to keep 24 them to the minimum. Basically, the igniter is a major 7s h 25 difference, the drywall flooding was a major difference, NEAL R. GROSS l CoVRT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
l
79 1 the scrubbing for the Bs versus the Ps was a difference.
l
, 2 MR. ABBOTT: You don't have igniters in the
(
)
3 Mark Is.
4 MR. PRATT: Right, they're not up there. The 5 Mark IIIs. Did I say Mark Is? I'm sorry.
6 MR. ABBOTT: I have sort of a simple question.
7 The way to enter this tree is with a Level 1 PRA, right?
8 MR. PRATT: Yes, you come in with -- you take 9 each of the accident sequences for a Ievel 1 PRA and run 10 them through the tree. And what we tid basically is 11 got some Level One people.at Brookhtven and gave them the 12 IPE for certain of the plants, and ave them the Level One 13 IPE and then asked them to run through these event trees 7
- I
'k J 14 and bend the accident sequences either into the LERF or 15 the --
16 MR. ABBOTT: There may not be much in the I
17 Level One PRA that talks about containment performance.
18 So you are extracting this containment performance !
)
19 information from other sources, right? j 20 MR. PRATT: Oh, yes. I mean, yes, absolutely.
21 But what we do expect though, what we do expect is that if 22 there is potential to depressurize that that would be i
! 23 factored in somehow into the information coming out of the 24 --
(D
( ,) 25 MR. ABBOTT: Well there is always going to be l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
80 1
1 a good match between the plant's Level One PRA and these l x 2 other sources of information like NUREG-1150. So this can i l
I
.\_/ l 3 actually be done? i 4 MR. PRATT: I'm -- you don't need any 5 information other than what you've got in the Level One 6 PRA plus --
7 MR. ABBOTT: So regardless of what's in the 8 Level One PRA, I can always do this mapping.
9 MR. PRATT: Well, I mean, yes, that was the l
i 10 objective.
i 11 MR. ABBOTT: Right, I just asked.
12 MR. PRATT: Yes, you take a Level One PRA and 13 then you sort of proud around it through the tree. Now,
/ '
i
~ 14 some of the interesting things that come out when you get 15 Level One people trying to do this is that they ask 16 questions specifically about depressurization. And for 17 particular accident sequences it would look at what the 18 time available was and whether it might happen or not.
19 Another one that was kind of tough for them to 20 do was restoration of ECC. What they tended to do if if 21 we had high pressure sequences with loss of power, it 22 would look at power recovery codes and see if we had 23 enough time. I think we specified 30 minutes or something 24 like that, in order to quantify that tree.
(_/ 25 Another interesting one was the operability of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
81 J
1 the ignitors. They tended to say okay, if I have power e s, 2 than I have the ignitors. If I don't have power, I don't
\_/
3 have them. But yet they kept thinking well maybe they l 4 should develop a fault tree to see whether the operators 5 would actuate them or not. And that would be something 6 that would go in addition to what you would normally find 7 in a Level One PRA because it really doesn't affect the 8 core damage frequency.
9 So there is a little bit of extrapolation.
10 But the intent was not to get them into discussions of 11 phonemonology and epistimec uncertainty because they don't 12 do that. And to only prescribe those types of events for l
j 13 them, in fact it's only one tree where we do that. ,
7 l i \
r <
- ' 14 And again, this tree is a lot more simpler 15 because we are not dealing with all this stuff that goes 16 on for late containment failures. We kind of stopped it.
17 You know, does it fail early? Does it got a potential for 1
1 18 a large early release? If it does, we worry about it, if 19 it doesn't, forget it. Well, not forget it but --
I 20 MEMBER FONTANA: No, all I was going to say is l
l 21 that all he has got to do is say, all you've got to do is 22 say aliatory, and he is home free. He has already used 23 epistemics.
24 MEMBER POWERS: Trevor, you may not be the t
r'N
(_) 25 person to ask about this, but I'll ask you anyway. And if NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
82 1 you are not the right person to answer, maybe you can tell 73 2 me who is. In the Appendix dealing with this material
_] that you are presenting, it says frequently at each box 3
4 that gee, if an applicant or -- wants to change some of 5 these probabilities he is free to do so if he can justify 6 them. And I was wondering how you would go about doing 7 that justification if you did not have somewhere the 8 details of how you arrived at the probabilities that you 9 do arrive at?
10 MR. PRATT: Well, I think one of the comments 11 at the end is that we may wish to add a little bit more 12 discussion to the Appendix in terms of where the numbers -
13 - I mean, I know where the numbers came from. But it's 14 not obvious from the discussion as to what the incremental 15 contribution of the various failure modes --
16 MEMBER POWERS: Well, I wonder -- I wonder if 17 --
18 MR. PRATT: If we quote a number --
19 MEMBER POWERS: I wonder -- I think I would 20 despair of ever having enough information in this 21 particular, these particular documents. I'm wondering if 22 this isn't one of those things that just cries to be 23 documented very carefully in a topical report or something 24 that gets referenced. Because I've sat there and said, n
( ,) 25 gee, I would have no idea how to go about justifying a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISt.AND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
83' 1 change since I don't know where the numbers came from.
, ~3 2 And I don't know what you were thinking about when you --
( )
\ /
3 you know, you've already for every junction on there 4 that's been brought up today you go through a fairly 5 detailed discussion of what was in your thinking. But 6 that's not in the document. And --
7 MR. PRATT: That's a fair point.
8 MR. KING: Yes, I think that's a fair 9 statement. But I think an example of where a licensee 10 could come in and propose a different number would be the 11 DCH issue, where Sandia has gone through plant by plant 12 and come up with a probability --
13 MEMBER POWERS: Once again you are telling me a
's 14 I've got to spend more time there.
15 MR. KING: No, I think that's a good example, 16 where this method uses a .1 probability for containment 17 failure if you get to a situation that leaves the DCH in a 18 large dry containment, for example. That number is a lot 19 lower for a bunch of plants. And that's documented in 20 some NUREG/CR reports that happen to be done by Sandia.
21 I think a licensee could come in and say, for 22 my plant, that number is .01 or whatever it is, and use l 23 that, and that would be a perfectly valid thing to do.
24 MEMBER POWERS: You handcuff me immediately
("N
'q,) 25 because I cannot respond to anything that was said by l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
[
84 1 Sandia.
,3 2 MR. KING: That's an example, that's an
( 1
'%J 3 example.
4 MEMBER POWERS: I can respond to the extent of 5 the University of California Santa Barbara, but not 6 Sandia.
7 MR. PRATT: Yes, I think that's an important 8 point. They can make those changes and we do now, but 9 also the advantage of this as I see it is if they do run 10 through this, and this is a very quick process. I mean, 11 the Level one people did this very quickly. And if they 12 don't need to look at the LERF value that they calculate l 13 in more detail, if it's enough for their purposes, then I\->h 14 they avoid having to go through all the details of the 15 justifications of the numbers.
16 MEMBER POWERS: Yes, I think it's --
17 MR. PRATT: It has an advantage, and they 18 would only want to go into the detail, I think, if they 19 felt the need to do that.
20 MEMBER POWERS: I see there is a fine way to 21 address the practical problem. And the practical problem 22 is that people don't have good technology for doing Level 23 Two and you want some Level Two-ishness in your risk 24 information, and it's a very useful thing. My concern O
(_,g/ 25 here are really two fold. One I can't change it if I NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 2344433
85 1 don't know what you were thinking about, and I would like r~N 2 to see document very carefully, very thoroughly. So that U 3 somebody can go through and says "I don't like this 4 number, how did they get it?" And they make that change.
5 And my other concern is understanding exactly 6 why LERF is an adequate measure of for defense-in-depth.
7 I don't know that I think it isn't, but I need to 8 understand why it is.
9 MR. ABBOTT: I have a quick question. It's 10 fairly easy using a Level One PRA to judge the effect of 11 changes on the plant systems core melt frequency because I 12 you have got a plant specific model. But, you know the !
I
-s 13 containment and its associated systems are all part of the
'-'t 14 licensing basis. So if licensing makes a change to that l l
15 portion of the FSAR or whatever you want to call it, how 16 do you use this system to make judgements on that in the 17 absence of site specific information on containment i i
18 performance?
19 Say, for example, I want to change the timing 20 in the MSIVs in a containment isolation -- in a Mark I 21 BWR, for example, which is part of the containment 22 isolation function, part of the containment function. How
[
23 would I do that?
24 MR. PRATT: Well, I think --
(_, 25 MR. ABBOTT: I mean how would I use this i
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
l (202) 234-4433 WASHINGTON D C. 20005-3701 -(202) 234-4433
~ . . - - . . . - . - - . - - ~ - - - . - . - . . . . . - - . - . - . . . .
l l l
86 ;
1 process to judge whether or not that change is within.the l 2 safety goals and the other things we've talked about?
3 MR. PRATT: Well I think what we can be 4 assured of is by looking at the numbers that we came up l l
5 with, of 10-' for core damage and 10-5 for'LERF, that if you 6 are below those numbers then there is quite a bit of
.7 margin in-the way we derive those numbers for the safety
}
8 goals. So if I ran through a containt tent of entry and i l
9 came in with a LERF number that was below the 10-5 from 10 this approach, I would have a lot of confidence that it I
?
i l 11 would be below the safety goals.
12 So in that sense, this method would at least l L {
13 give me the level of margin that I have to reaching those O 14 two safety goals. But I don't think you would want to use l
l )
15 this method in terms of some of your detailed -- I mean, 1 l
t you still have to go through the normal deterministic 16 17 analyses and so on to show that things were okay. ;
- 18 What this really does is given that we want l 19 to, one of the objectives there was to make sure we don't L
20' exceed the safety goals and we stay within those things, 21 this gives you, I think, a very good monitoring of just 22 how close to them you are.
23 MR. CUNNINGHAM: Another way to put it might 24 be the example that you used, this basic information would L
25 tell you that that risk is not very sensitive to the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
87 1 change that you propose. Because it doesn't show up even 2 as a primery question. If you were coming in and trying g3 s 1 v to deal with something that would affect one of these 3
4 questions, then you've got a different situation and you 5 have to think much more seriously about whether chis is an 6 adequate screen --
7 MR. PRATT: Oh, okay. Yes, I mean we are 8 issuing a report, I guess, where we looked at some of 9 those types of issues, some of those containment issues 10 and showed how the impact did risk, you know, as a 11 separate identity. And in that sense you can see that 12 there was not a strong influence on these numbers.
13 But I mean I think the point we are making is i >
' - - 14 that certainly if you were, had a very high LERF number 15 and you were going to start and mess with some of the 16 containment systems, you would want to look at that.
17 MR. ABBOTT: Okay, and then this is an open 18 season on containments? This is an open season on 19 containments?
27 MR. PRATT: Oh no, not at all.
2 ". CHAIRMAN APOSTOLAKIS: Shall we go on?
22 MR. PRATT: Yes, well I think we can go l 23 through the questions pretty quickly because I think we've 24 addressed most of those in our discussion.
/3
( ,) 25 MEMBER POWERS: Trevor, could we come back to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 88 1 your PWR slide?
73 2 MR. PRATT: My PWR slide? ,
1 ~- \
3 MEMBER POWERS: The tree. l 4 MR. PRATT: Yes. Okay.
5 MEMBER POWERS: If I come to your last two 6 questions, no containment failure at vessel breach and 7 then no potential for early fatalities. Aren't you guilty 8 here, well, that's a loaded question, isn't it?
9 MR. PRATT: Absolutely.
10 MEMBER POWERS: Don't you run the risk here of 11 making a decomposition in rare events to get your 12 probabilities down when you ask those two questions?
13 Isn't there just one question there? Shouldn't the r~N !
I
- ) l N' 14 question be 6 with no reference to 5? l 15 MR. PRATT: The reason we separated them out 16 was with the thought that we would be extending this 17 perhaps to a wider range of potential accident sequences. l I
18 For example, on some of the shut down work we did, we had 19 large early releases relative to the time of the accident 20 initiation, but because it had occurred so late there was 21 a significant decay of those radionuclides that caused 22 early fatalities. So we had a very low early fatality 23 calculation for those numbers. And also quite a lot of 24 time available for evacuation.
(,) 25 So I think you -- and how this was applied by NEAL R. GROSS CoVRT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
89 1 the Level One analysts most of the time if they had a l l
,e3
, 2 large early release at the time of reactor vessel failure, l
\ ) 1 s_- '
3 then they virtually went straight through and said, "Yes I 4 have potential for early fatalities." So the two were 5 highly coupled for those types of events.
6 But there are events where you have a long 7 time to these events where you might want to take into 8 account --
9 MEMBER POWERS: I still don't underet ' . thy 10 there should be two questions here. I mean, if I have no 11 containment failure at vessel breach, then I have an 12 answer to question 6.
13 MR. PRATT: Right, and I don't have a split
/,,D s
\- 14 there. It's a double -- no is yes. I mean, yes, I have 15 no containment failure, j 16 MEMBER POWERS: I understand and I cursed you 1
17 mightily when I went through the trees and said I hate l 18 this. But I understood why you did it.
19 MR. PRATT: Yes, I mean if you have no 20 containment failure, you whole thing you don't worry. But 1
21 if you have a containment failure, the only reason we i 22 asked the question about early fatalities was, you know, 23 gee, there may be circumstances where you do have a 24 significant release which is relatively early, but because l
p) i 25 of other things we think that we could get people away.
- (_
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l i 90 1
1 Another one was the thought was for seismic l 1
p 2 events, there is a little note in the Appendix saying gee, i ;J t
}
3 you know, if you have a seismic event you could have 4 evacuation impeded. And if you think you've lost l
5 containment heat removal on the path, even though you l 6 think you've survived containment failure of vessel 7 breach, you should perhaps think about late failures more 1
8 in-depth. )
l 9 MEMBER POWERS: Yes, I took it as your time I 1
10 scale just changed on an external event.
11 MR. PRATT: Right.
12 CHAIRMAN APOSTOLAKIS: Just a minor detail, 13 would you delete the word " frequency" from the first box? !
t i
'"/ 14 Just listing events there, not frequencies, core damage.
15 MR. PRATT: Yes, right. ;
l l
16 CHAIRMAN APOSTOLAKIS: Okay, nine minutes. i 17 MR. PRATT: How much?
18 CHAIRMAN APOSTOLAKIS: Nine minutes. ,
1 19 MR. PRATT: Okay. We'll go through -- then ;
20 maybe I think we can go through the questions real fast --
21 CHAIRMAN APOSTOLAKIS: Sure.
22 MR. PRATT: -- because we've covered them --
23 CHAIRMAN APOSTOLAKIS: You can have more time i
24 after the break.
g
() 25 MR. PRATT: Oh, okay. Question 2 --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
91 l
1 CHAIRMAN APOSTOLAKIS: If the Secretary l
l ('S 2 approves.
I V Question 2 simply pulls out those 3 MR. PRATT:
l 4 events that are going to bypass containment. So that's 5 very obvious and we want to know that.
l 6 This question 3 deals with the ability to 7 depressurize. And so what we did most of the time was 8 identify transient, small break LOCAs to the high pressure 9 sequence, and small break LOCAs to the low -- large break 10 LOCAs to the low sequence. And then any credit that might 11 be taken for depressurization would, this is with the 12 opportunity to introduce that consideration.
s 13 Question 4, the Level One guys had a little
( )
14 bit of trouble quantifying this number. We know that it i
15 was important from a lot of the studies that we've done, l
)
16 and also from some of the IPE results. But just being 17 able to go in there with a Level One analysis and decide i 18 which of those sequences, you had to go chance the 19 depressurize -- to restore ECC was kind of a tough one.
20 We've discussed the numbers that we used here 21 in terms of the conditional probability of containment 22 failure given whether or not it's a high pressure or low 23 pressure. And as Dana said, what we need to do there is 24 perhaps go in and write up a little bit more of a p
l
(_ 25 justification of where those numbers came from.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
1 92 l
1 To give you an idea, I don't have this in the
,s 2 handout, but it is simply a reproduction of the 3 information in NUREG-1560. I think this gives you a sense l
4 of the range of numbers that we've seen in the IPEs. So i 5 you have seen this, I think, on a number of occasions 6 where what this does it gives you the range of conditional 7 probabilities of early containment failure and bypass as 8 reported in the IPEs, and also the frequency of these 9 events. And you can see the range is quite large.
10 What we obviously didn't do -- well actually 11 this, let me give you another one. This gives you just 12 the Bs, the Bs and Ps, okay this gives you the information 13 separated out in terms of the boiling water reactors and
\
\- 14 the pressurized water reactors. So obviously when we went !
15 into pick these numbers, we did not obviously go into the )
16 highest extreme of these ranges. For example, this is a 17 PWR, these are the conditional probabilities of early 18 failure. We have numbers in the range of .3 and so on.
19 We obviously didn't pick that number, and part 20 of the reason why we didn't do that is that some of the 21 reasons for why we ended up with those very large numbers, 22 we were very plant specific, not applicable to a wide 23 range of plants. Some of them are modeling assumptions 24 where we felt they were somewhat conservative in terms of gg is,) 25 what they did. So, the ranges that we were trying to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
93 1 encompass were down in this particular range. )
2 MR. CARROLL: Now suppose this plant specific r
73 :
'O 3 problem containment wants to use this methodology. Does 4 staff know which plant, or is the plant that's up at .4?
5 MR. PRATT: Oh, yes. I mean this is fully 6 documentea in the -- I mean, we have reviews and SERs on 7 all of this stuff.
8 MEMBER KRESS: I'm struggling with what you 9 said Trevor. The reason being that these are real results 10 for real plants. And they are developing a set of 11 acceptance criteria based on safety goals for early 12 fatality. I don't see that they have any relationship to 13 each other at all. I don't understand you statement.about f%
i i V 14 selecting a number that represents some portion of these.
15 I don't understand why they are related at all.
16 MR. PRATT: Well, no, I mean, they are related 17 only in the sense that what we were trying to do in this 18 approach is was to capture what we thought were reasonable 19 numbers for the plants out there. So we looked, the team 20 that came up with the numbers came up with the numbers 21 looked at --
22 MR. KING: We are getting confused here. The 23 acceptance guidelines are not based on these numbers.
24 MR. PRATT: Oh, no.
(~g 25 MR. KING: What Trevor is saying is these --
()
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
94 1 MR. PRATT: It's a simple rationale --
2 MR. KING: -- these fractions or probabilities 7g b 3 of containment failure that he is putting in is generic --
4 MEMBER KRESS: Oh, he is using actual data --
5 MR. KING: He is using actual data --
6 MEMBER KRESS: -- to select the split 7 fractions.
8 MR. KING: To get a reasonable number for 9 that.
10 MR. KING: Okay, that, now I get your 11 connection. Thank you.
12 MR. PRATT: I'm sorry. Maybe --
13 MEMBER KRESS: I must have lost something
/,_ \
\ /
N' 14 inbetween. But that's reasonable.
15 MR. PRATT: This was alluding to what Dana was 16 saying earlier, you know, where did these numbers come 17 from? Well these numbers came from a compilation of 18 looking at NUREG-1150 and all of the IPEs, because the 19 people who did the trees have reviewed most of these Level 20 Two IPEs. And so --
21 MEMBER KRESS: Now I understand the 22 connection.
23 MR. PRATT: Yes, I didn't want people to come 24 in and say " Wait a minute. We've got this reported result C)h
(, 25 which is way higher than this, why didn't you take that NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 2344433
95 1 into account?" Well, there are reasons, and as Dana said fS 2 we need to document some of that.
- i LJ I wonder why those reasons, 3 MEMBER POWERS:
4 why you adopted that strategy? Maybe you didn't, but why 5 that strategy was imposed on you?
6 CHAIRMAN APOSTOLAKIS: Which strategy is it?
7 MEMBER POWERS: To go through and find 8 reasonable numbers. Why not say, gee, what I would really 9 like in this world is a plant specific analysis. And some 10 people just do not do that. So, why not put in the 11 highest value that you've ever had reported to you and say 12 well I know you are going to be less than this, and if you
~s 13 want to justify being less than this, do so.
"' 14 MR. PRATT: Well, I mean, I think we did 15 pretty good in that regard with the exception of a couple 16 of outlyers that I think weren't appropriate to be counted 17 in. I mean, I think that's the answer to your question.
18 In other words --
19 MEMBER POWERS: Okay.
20 MR. PRATT: -- what one, in one particular 21 case it had a particular design feature that caused that 22 early fatality which we don't see in any of the other 23 large volume containments. So that was truly an outlyer.
! 24 And in another case the modeling assumptions that went l
t r^x
(_,) 25 into the failure distribution on the containment were l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
96 I i
1 really simplified and I think would skew the general results that we had. I
(~% $
2 l
t N_/ ;
3 So I think we kind of enveloped -- we l l
i 4 certainly didn't take all the results we'd ever seen and ]
I 5 averaged thera. We tried to envelope but we did not want 1
6 to includt outlyers that we thought would skew the 7 results.
8 MEMBER FONTANA: There were several places 9 where you say the plant IPEs cannot be used for this 10 purpose. What did you mean by that?
11 MR. PRATT: In the Appendix?
12 MEMBER FONTANA: Yes, Appendix B in a couple 7-13 places it says a plant IPE should not be used, if I could 14 find it here, but it's -- l i
15 MR. PRATT: You'd have to refresh my memory. j 16 I think that was towards the end as I recall. !
i 17 MEMBER FONTANA: I'll find it. I'll look for 18 it. Here we go. Page B-11, bottom of the tcp paragraph, 19 last sentence. It says the Level Two IPE information is 20 not generally considered adequate to use for these 21 probabilities.
22 MR. PRATT: Oh, yes, again, that may be a 23 little bit of a strong statement. I think this is, 24 generally what we are saying is that we consider, this is
,ry kl m 25 the one case where we specify the split fractions. And NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
( 97 l
1 the only purpose of mentioning that here is that this, i I
,s 2 this is done because we believe that certain of the (v) 3 information was, shall we say a little optimistic. So, 4 the intent was to try to come up with something that we 5 felt comfortable with.
6 I don't know that --
7 MEMBER FONTANA: This only, that's repeated on 8 page B --
9 MR. PRATT: Yes, I notice --
10 MEMBER FONTANA: -- fourteen -- but that's 11 intended really only for that particular question?
12 MR. PRATT: Yes, absolutely. And again, it 13 maybe inappropriate I think in that particular case. We
{}
14 don't have it in the rest of the event trees. It was part 15 of the thought processes when the guy was coming up with 16 the numbers. Well I don't necessarily trust all the 17 results I've seen, therefore, this is the number I'm going 1
18 to recommend. But you have cases where, as I said before, l
19 the IPEs actually came in with higher numbers than this in 20 certain cases. So, I think that's part of the original 21 rationale which probably needs to be cleaned up.
22 CHAIRMAN APOSTOLAKIS: Shall we break now and 23 then there is plenty of time after the break for you to l
[ 24 complete your presentation. Okay, thank you.
(3 25 (Whereupon, the foregoing matter went off the
(_,/
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
98 1 record at 10:35 a.m. and went back on the 7s 2 record at 10:55 a.m.)
I )
3 CHAIRMAN APOSTOLAKIS: Okay, Trevor, are you 4 ready?
5 MR. PRATT: Yes, where are we at? We had left 6 off at the, I think the last question on the tree which I 7 think we've covered in a lot of detail anyway. Basically, 8 it just recognizes those early failures and bypass 9 sequences that might result in the potential for early 10 fatalities. So --
11 MEMBER FONTANA: I did -- Trevor, I did have a 12 question on question 5.
13 MR. PRATT: Oh, sorry.
) That's the prior slide. No
(~/ 14 MEMBER FONTANA:
15 credit is given for the operator manually depressurizing 16 the reactor and using a low pressure system to inject 17 water between core damage and vessel breach. Is that a 18 kind of a general thing? It seems --
19 CHAIRMAN APOSTOLAKIS: What is this now?
20 MEMBER FONTANA: On question 5 the words say -
21 -
22 CHAIRMAN APOSTOLAKIS: Where are you now? Oh, l
23 in the Appendix. Oh, oh, I think it relates to --
24 MR. PRATT: I think it's uniform across the --
(O,) 25 MEMBER FONTANA: It relates to question 5.
NEAL R. GROSS l COURT REPORTERF AND TRANSCRIBERS 1323 RHODE SLAND AVE., N W.
(202) 234-4433 WASHINGTO'J. D C. 20005-3701 (202) 234-4433
99 1 CHAIRMAN APOSTOLAKIS: Can you find the actual 2 page again? Because it seems --
3 MEMBER FONTANA: We don't really have to do 4 that, but on page B-5 --
5 MEMBER BARTON: B-5?
6 MEMBER FONTANA: Yes, down on question 5, four 7 lines from the bottom. "For example,-no credit will be 8 given for.an operator manually depressurizing the reactor 9 and using a low pressure system to inject water between 10 core damage and vessel cooling -- vessel breach"'la.de la 11 de la. Is that kind of a general approach? It appears 12 that no credit would be given for operator actions of that-13 kind, and no credit for any severe accident management?
14 MR. PRATT: No, I think the instructions.that 15 we gave, and this is consistent with this question across 16 each of the containment types. These were the ones that 17 we would expect them to take credit for easily. I mean, 18 they can always come in with the case'and. point where they 19 believe that they have additional actions that would 20 reduce that number.
21 -- MEMBER FONTANA: As far as the screens --
22 MR. PRATT: Yes, as far as the screen --
23 MEMBER FONTANA: As far as the screen is 24 concerned, you are not going to let --
k 25 MR. PRATT: Then this is okay to us. Now you NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
'(202) 234 4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
100 1 want to go beyond that --
,q 2 MEMBER FONTANA: And you have to come in and N] 3 justify what you are going to do?
I i
4 MR. PRATT: Yes, yes. l l
5 MEMBER FONTANA: I see. I see. !
l 6 MR. PRATT: Probably the words there sound a 7 little harsh. I mean the intent was that this is ,
8 something that we think we can accept fairly easily and i
9 it's fairly straight forward. If you are going to get 10 into more exotic procedures than we want to look at those.
11 MEMBER FONTANA: Oh, okay.
12 MR. PRATT: So, you know, take a note to look 13 at the wording in there.
/ )
\d 14 Okay, had we a question on 6 or were we done?
15 CHAIRMAN APOSTOLAKIS: Mario, was there 16 anything on 6?
17 MEMBER FONTANA: I don't think so.
18 CHAIRMAN APOSTOLAKIS: Okay.
19 MR. PRATT: Okay, so that really goes through 20 the approach that we adopted. And we touched on a little 21 bit on each of the event trees. What we then did was to 22 take the write-up that you see in the Appendix along with 23 the guidance and have some Level One people run it through 24 the tree. We have a number of plants. The ones I've got
,\
l )
\, ,/
25 results for today are four, and we have a few others in NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
101 l
l 1 the mill just to spread the experience around a little l n 2 bit.
iv) 3 What you will see in the results table are l
4 four sets of calculations. The first calculation that we 5 did was what we got from the guidance provided in the 6 event tree. So I came up with a frequency of LERF.
7 That's one number, 8 We then went into the IPE results, because 9 they had all done Level Two PRAs and calculated the 10 frequency of those source terms that were early and that 11 resulted in releases of the more volatile fission products 12 greater than .03 and .1. And then we also just counted
,, 13 from the IPE results those bins, if you like, that were in i \
t V) 14 the category of early containment failure and bypass.
15 Okay? So those are the calculations that were done and 16 the results are shown in the table.
17 What I've also put in there, the first row is 18 the core damage frequency, just for reference. That was 19 the reported core damage frequency. And these are all for 20 internal events, full power, as reported in the IPE 21 submittals.
22 And what you see below that are the 23 calculations that were derived from the IPE results. And 24 you see there the frequency of bypass and early
(/^% ,) 25 containment failure as being the top row. Then as we go NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. O C. 20005-3701 (202) 234-4433
102 j j
I 1 down we-would expect the frequency to drop because we are
- 2 looking at only those subset of the early failures and 1
)
.3 bypasses that resulted in source terms larger than .03 and l 4 .1. At the bottom you see the number that came out from- :
5 the Level one people from their use of the event trees.
6 I think if we look at those results, there is ;
7 some interesting things that.I should point out. For the i 8 PWRs which is Surry & Sequoyah, you see there that the 9 agreement is pretty good and that all the numbers are very 10 close, even the large source term numbers. And I i
l 11 apologize for Mark for putting so many decimal places l 1
12 here. We put, you know, this is 1x10-5 but I put the )
1 13 difference in there just to show you that there was 14 individual calculations that went into those numbers. !
15 The interesting thing here is this is probably .;
i .
16 not a really good test for our approach. We are going to l 17 do other types of PWRs to show it because these LERF in 18 both of these plants is almost. totally dominated by bypass 19 events. So we would expect the agreement to be very 20 close. I mean, if you are not, if it captures anything it 21 should capture bypass events. And also if you have a 22 bypass event, this is the point you made earlier when we 23 talked to Tom about the magnitude of the' source term.
24 Basically all of those bypass events resulted in 25 significant source terms with a release of larger than .1 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
103
.1. for the iodine and cesium groups, f-s. 2 So I think there is a case where, you know, t
3 the LERF was calculated based on bypass sequences, and 4 that, you know, we captured that in the event tree fairly 5 well. So that the discussions on the .1 in terms of 6 containment performance really didn't factor into this 7 estimate of a LERF terribly much at all.
8 Now for the PWR results, we-have:a slightly 9 different picture there. For the Mark I, Peach. Bottom 10 analysis, you will see that the number we came up with was 11 more conservative than the utility's estimate of.LERF by a 12 factor of about 2 or so. When you go in and look at that, 13 you find that the main reason for it was in the !
.( l 14 conditional probabilities that we attached to some of 15 these various accident sequences.
16 In the Peach Bottom'IPE most of the core
.17 damage frequency here, bypass doesn't play a role at all.
18 Most of the LERF frequency came from high pressure melt 19 down sequences. And a very large fraction-of them -- a 20
- very large fraction of that was sequences. And for.those 21 events, if you look at the Appendix, you will find that we 22 gave a very high conditional probability or loss of 23 containment integrity for those sequences. Okay? And so 24 that's why our numbers come out a little bit higher than 25 their estimates.
NEAL R. GfN3SS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 - . WASHINGTON. D.C. 20005-3701 (202) 234-4433
i 104 i t
! 1 Also, what you find also in that approach is !
2 that the method for scrubbing, which is the point you were ;
f-s 4
' b 3 making Tom about those sequences that we felt should.get.
-i captured. There is a part in that containment event tree 4
- E 5 that should have captured some of those scrubbing versus t
' non-scrubbing.
6 ;
4 !
' So not only was the judgement on the part of I 7
f l
' 8- the Level One person who did that quantification that it :
2 a f
S 9 was a higher conditional probability of failure because of 1
I !
10 the high pressure sequences, it was also a judgement made~ j a i i 11 that.very little scrubbing occurred for those sequences.
+
{ 12 So that's an interesting point that we need to look at and i 13 see whether'or not the guidance given to the Level One 3 i 14 people were appropriate. Whether we need to qualify that, !
['
, i 15 whether it was appropriate. So I think there is' good i 16 feedback coming from some of these, you know, kind of 17 blind applications.
18 Grand Gulf is another interesting case because 19 we have the opposite effect. There we have an estimate
-20 from the simplified approach'which is actually lower than 21 what you would have got if you had looked at the IPE 22 results. And interestingly enough there, the reason for 23 that is that this Grand Gulf IPE came in with a very high 24 contribution to operator actions to vent the containment .
l 25 building.
i NEAL R. GROSS I
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
. _ _ _ . _ _ -_-_._ _ _ _ ._ _ _ .._ _ . ._ _.. . _ . , _ _ _ . ~ _
105 1 Imd I think in our presentation to you on the
-s g 2- IPE insights program, this is something that we actually J
3 mentioned where if you luse level indication and you go 4 through a flooding procedure in these designs, you open up 5 a vent. And they actually found that by opening this 6 event,Eit was a very significant contributor to early 7 releases. And they were actually thinking about what the
-8 appropriateness of those particular procedures in this 9 case.
10 So that's a case where we certainly didn't 11 capture that in our estimate of early containment failure.
12 That was not one of the elements that went'into the 13 conditional probabilities that we put into the tree. So O
~%) 14 it's a case where we actually were a little bit below what 15 the applicant came in with. For a very specific reason.
16 And so.that's, I think, instructive.
17 MR. CARROLL: Do you know how-they ultimately-
'18 resolved that issue?
19 MR. PRATT: I'm not sure. I'm not sure. I 20 know it was flagged by one, by the utility and also by us 21 in the insights report. It had been certainly something 22 that they need to look at.
23 MR. SHERRY: Trevor, is this the procedure for 24 containment flooding which requires you eventually to vent
() 25 your' reactor vessel?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
I think the thing that comes out of it is that the results are ;
4 ,
5 pretty close. They are not half bad. And the deviations !
( 4 6- that we are getting are understandable and explainable.
l -
I 7 That's really all that we had to share with 8 you. The summary slide, I think, says stuff that we've P
- 9 said already many times. We think this is a pretty good !
10 approach for giving you an initial impact of what we think [
11 the LERF frequency might be. And it allows'you to see l 12 where you are in relationship to these decision-criteria i
13 and then you can use this information as we intended in 14 the Reg Guide as a means of giving you risk based 15 information for this work. -
'16 We have applied it, we didn't show the fifth i
17 containment, but we've applied it to all five containments f i
18 and we are getting reasonable results which are in the ,
19 ballpark. And I think as a result of these we-are getting
{
20 some very good feedback from the Level One people in terms ;
i 21 of how we might improve the guidance and also, I think, .
I 22 Dana made the point-that we need to, perhaps, have a 23 document.that justifies in more detail where we got the l 24 numbers from.
25 MR. KING: I think we agree with that and we l
NEAL R. GROSS l
[ COURT REPORTERS AND TRANSCRIBERS !
1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 ;
i
i 107 )
i
! 1 talked at the break. I think we will go ahead and put l
fx 2 together a NUREG/CR that provides the basis for this and l
(s- ) l 3 then it can ultimately be a reference in this Appendix in 4 the Reg Guide.
5 MEMBER FONTANA: That would resolve a lot of l 6 the questions that I had and the thing that Dana said. l 1
7 When I first heard about, at prior meetings, about 8 estimating LERFs from Level One, I was pretty skeptical.
9 But it turns out you are not really using only Level One, i 10 you are taking a lot of past information on Level Twos and 11 that sort of thing.
12 I feel a lot better about it now. I think as 13 a screen it's pretty good.
(m.
-' 14 MR. PRATT: Good.
15 MEMBER FONTANA: And you know, when it 16 triggers, you have a capability of triggering a more 17 detailed analysis if it's deemed to be appropriate. So, I 18 think it sounds no; bad.
i 19 MR. PRATT: Yes, I think it fits in well with 20 the way the Guides -- l 21 MR. KING: We think it's at this point good 22 enough to go out for public comment. We recognize there 23 is some improvement, some additional documentation needs 24 to be done. But our view is let's get it out, let other
/~N
! _) 25 people look at it and we will make the improvements in NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHOCE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
. _ . . _ _ _ . - . . - . . _ . . _ . _ _ . . _ . _ _ _ _..._._ _ _..- ~ _. - _ _ . . . _ _ _
, 108 i l
, i l; 1 parallel with'that process.
i 2 MEMBER FONTANA: Yes, especially if we can substantiate it with a NUREG substantiates where these 3 j I
4 things came from like Dana was saying. It's good.
L 5 MR. KING: Okay, that completes what we wanted l 'I
! 6 to say on Appendix B to the General Reg Guide. What we j 7 want to do now is shift to the Appendices for the SRP and l
l 8 then as part of that, that will come back and also pick up 1.
-. 9 Appendix A of the General Reg Guide.. Mike Cheok was going l 10 to do.this presentation. That will take us through to l
i 11- lunch time. Do you have yocr own view graphs? Mike is '
l l
12 from NRR, the developers of the SRP. l l
13 MR. CHEOK: Okay, the SRP basically has got O-14 three appendices. Appendix A is guidance to the review as 15 to how we can do a focus scope review of the PRAs that is
- j. 16 sent-in with the application. Appendix B is basically 17 details what we need to look at in the integrated decision l
, 18 making process' . And the last appendix, which is similar l
19 to Appendix A in the Reg Guide, deals with component 20 categorization. ;
21 Okay, I'm going to start off with Appendix A.
22 We expect that when licensees send in an application in l
l- 23 risk informed regulation,'that the analysis is based on a r
24 PRA of adequate quality. The licensees are expected to G. show us why they think the PRAs of a quality sufficient to
_V 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
109 1 support the application.
( 73 2 The staff is expected to do a review of the .
l !
l
(_ ,/
3 PRA quality in a focused scope sense and the fact that we j 4 can look at the PRA and see if it supports the conclusions l 5 that is being made for the application. The industry, we 6 have lots of industry initiatives now to do PRA reviews 7 before they send the applications in. Our examples that l
8 are there now include the peer review, the PRA I
9 certification and a cross comparison that is going on 10 within various industry groups.
11 CHAIRMAN APOSTOLAKIS: Is a peer review that 12 is organized by the licensee an acceptable way of 13 guaranteeing quality? Or you guys review the PRA anyway?
(b 14 MR. CHEOK: We will look at the results of the 15 PRA review and we will also look at a PRA anyways.
16 Depending on what we think we see in the peer review l
17 report. l l
18 CHAIRMAN APOSTOLAKIS: Will it shorten the 19 time that is required to do the review if they submit it, 20 the review of a peer group of--
21 MR. CHEOK: Yes, we expect that it would 22 shorten our time if the licensees do present documentation 23 as to why their PRA is of adequate quality.
24 CHAIRMAN APOSTOLAKIS: Now the words adequate
( ,) 25 quality are really the ones that require further NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
110 1 operation, I guess. It's a value judgement, yes. Because (3 2 right now, in the documents, adequate quality tends to be Y
3 towards state of the art PRA. Perhaps with the pilots we 4 would be able to be a bit more specific as to what is 5 adequate for certain applications. Unless you try it, in 6 other words, you can't really tell in advance what is 7 adequate for certain things. So, you tend to go with the 8 best possible, right? So that's something that we all 9 agree on?
10 MEMBER POWERS: George, it's interesting that 11 we --
12 MR. HOLAHAN: Well, I think you state it 13 strongly.
7 14 CHAIRMAN APOSTOLAKIS: Pardon? I 15 MR. HOLAHAN- I think you state it a little 16 too strongly. I think that the Reg Guide and the SRP l
17 don't really call for the highest quality available. I 18 mean, you certainly acknowledge, you know, that that's one 19 possibility. But, I think they do allow the licensee to 20 choose something less and to justify that based on what 21 particular issue and decision is trying to be made.
22 CHAIRMAN APOSTOLAKIS: That's right. But if i 23 you look at the various discussions there in the sections 24 and subsections, I mean there is a discussion on common n
( \
(,/ 25 cause failures, there is a discussion of human error and l
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
111 1 so on. And the result was a statement that you can do j 2 less. But what is described is really above quality i
r~N s ,
\_/ l 3 stuff, so -- and I'm not saying that's not the way it 4 should be at this stage, because at this stage maybe we 5 don't know better. But Dana wanted to say something now.
6 MEMBER POWERS: Well, it's interesting that 7 you had that interpretation of what they meant by quality.
8 Because when I read the words I said, well, now, how do I 9 apply Appendix B? And didn't come up with a very good way 10 to apply Appendix B. It's the quality assurance --
11 MR. HOLAHAN: Well I think in many cases 12 Appendix B does apply, just like it applies to ECCS and --
,_ s 13 MEMBER POWERS: It applies to everything.
~
I )
14 MR. HOLAHAN: -- calculations --
15 MEMBER POWERS: And claims to apply to 16 everything.
17 MR. HOLAHAN: Well, it applies to safety 18 related activities. But I think that just means that the 19 licensee needs to have a OA program for those activities.
20 And that might involve peer review or it might involve 21 certification or it might involve, you know, something 22 else.
I 1
23 CHAIRMAN APOSTOLAKIS: Bob? l I
24 MEMBER SEALE: Well is it the reality that I (Q
(/ 25 many of these pilot studies have been based on some NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
112 1- upgrade, if you will, of the results of the IPE that was
,f S 2 performed for the particular plant that's involved?
i, s_
3 MR. HOLAHAN: Yes.
4 MEMBER SEALE: And I think it would be very I
5 revealing to look at which upgrades the utility judged 6 were appropriate to discuss the issues that are involved 7 in this particular pilot study. Because in a very real 8 sense you get an opportunity to judge what they think are 9 the appropriate issues in risk space, in probabilistic 10 space, to the problem that they are addressing.
11 So there is really an awful lot of information 12 from a lot of different points of view if you consider all 13 the different utilities involved as to what is adequate.
('-') 14 CHAIRMAN APOSTOLAKIS: No, I agree.
15 MEMBER SEALE: At least in theit first blush. l l
16 CHAIRMAN APOSTOLAKIS: I agree.
17 MR. JONES: If -- Bob Jones from staff. Just j 18 following up on your comment. One of the activities we 19 are about to initiate in the next month or so will be to 20 go back and look at the PRAs done as part of the pilot 21 activities and send that out to a contractor to do exactly 22 what you are talking about. We want to understand what 23 was done, how it was done, how does it stack up to the 24 Guide? What is a practical way to review it? We do want (3)_
25 to do that test and that is our intent.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4 433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
..__._.__.._q 113 I, i
l 1 MR. HOLAHAN: _I also think it's going to take !
2- us some time to understand the issue fully. Right now l
I .
3 when we talk about the pilots, I think you have a biased ;
L .!
-4 sample, okay? It's some of the licensees who put a lot of j l-5 effort into their PRAs and probably have the closer to the :
i 6 state of the art than others, are the ones that have come i l
l 7 forward and volunteered to be in the pilots. So I think j l ,
l l 8 there is more to be tested out there --
i ;
9 MEMBER SEALE: Sure.
10 MR. HOLAHAN: -- than the examples we have.
11 MEMBER SEALE: But we are probably glad that ,
12 the people who have chosen to use the PRA methodology to ,
13 do this are the people who really perhaps understand j ;
14 better what's involved in doing it well. j 15 MR."HOLAHAN: I think that makes sense.
1 I
16 MEMBER SEALE: Yes.
17 CHAIRMAN APOSTOLAKIS: Well I think this. issue l 18 of adequate quality is a key one. We are getting some 19 comments on the Guides that they represent onerous burden 20 on the licensees. So I think it is important to really l
21 make sure that the quality that's demanded of the PRA will l
22 be commensurate with the use, yes, the importance of the I
23 decision and so on. So I think it will be to our benefit 24 to acknowledge that up front that there is work to be done 25 along these lines and --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W. l
. (202) 234-4433 WASHINGTON, D.C, 20005-3701 (202) 234-4433 I _ . _ __, . - _ . . . _ __ _ _ . . _ _ _ _
114 1 1
1 MR. ABBOTT: Can I ask a question?
2 CHAIRMAN APOSTOLAKIS: It's hard to predict in 1 l
(Oo) 3 advance. Yes?
l 1
4 MR. ABBOTT: It's just a process question. In I l
5 order for this Appendix A to kick in, the utility sends 6 you the PRA for review? That is to say if he has 7 something upcoming -- no -- I mean if you are going to get 8 into this business here --
9 MR. HOLAHAN: Yes.
10 MR. ABBOTT: And you are going to start 11 looking at his PRA you are going to use Appendix A to 12 review it for quality.
13 MR. HOLAHAN: Yes.
,e
! l
'v' 14 MR. ABBOTT: And how does he get it? How do 15 you get it?
16 MR. HOLAHAN- You have to look at the 17 documentation section which identifies -- the licensee 18 doesn't send us the whole PRA.
19 MR. ABBOTT: Does it end up on the docket?
20 MR. HOLAHAN: Not the PRA. A summary of the 21 PRA ends up on the docket, and the extent of the summary 22 is what is discussed in the documentation section.
l l 23 MISSING:
24 MR. JONES: Section 3 of the Reg Guide
/
n I
() 25 describes.
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N.W.
j (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
i 115 1 MR. HOLOHAN: And when we are trying to figure
/^x 2 out what does this really mean, I think it is analogous to w] 3 things the staff has done for many years.
l i
4 It is very rare that a licensee would actually 5 give us the computer output form an ECCS calculation. In l
6 that case, I think the computer output is analogous to the
! 7 PRA.
8 MR. ABBOTT: Gotcha.
9 MR. HOLOHAN: What you get is a summary 10 description of what it is, what the results are, maybe 11 sample graphs, sufficient to explain what was done but not 12 necessarily -- I wouldn't expect the actual whole analysis
,_ 13 to be on the docket.
1
)
'y 14 It is a quality calculation kept with the 15 licensee to be inspected or questioned if that is needed.
16 MR. ABBOTT: Okay. And then you in turn look
! 17 at thais documentation and what do you issue? Some sort 18 of stamp of approval? )
19 MR. HOLOHAN: License amendments.
20 MR. ABBOTT: License amendments. ;
21 MR. HOLOHAN: And also recognize that what we 22 are doing, this is not a two step process, it is a one 23 step process. We are not saying the PRA is good and the 24 decision to change the tech specs is okay. It is only one
,o
(,) 25 decision which says this whole presentation of information l NEAL R. GROSS
' COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433 I
116 1 supporting this decision is acceptable.
l The PRA is not given any quality N stamp as tA)
N_/
2 3 part of that process. It just says it adequately 4 supported this judgement.
5 MR. ABBOTT: Okay. So the summary information i
6 is going to be reviewed by the staff each time an 7 application is made?
8 MR. HOLOHAN: Yes. To the extent that that is 9 necessary to support the change that is being proposed.
10 MR. ABBOTT: So, you would have to have 11 different summary information for each change?
12 MR. HOLOHAN: Certainly, you would want the 7,s 13 summary information to reflect what is being changed.
k'# l 14 Now, there may be some base information. But 15 it seems to me that the most important part is, the PRA is 16 supposed to shed some light on what is being changed.
17 MR. ABBOTT: Okay.
18 MR. HOLOHAN: And that is what you would want 19 to focus on.
20 MR. KING: Let me follow this thermal 21 hydraulic analogy a little bit because at least it helps 22 me. That is quality, assurance applies to a lot of 23 thermal hydraulic calculations. That doesn't mean in the 24 engineering sense they are all equally state-of-the-art A i i
() 25 calculations.
1 l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
! 117 It may be that if you pick up an FSAR and it 1
I
,eg 2 has a loss of feedwater transient in it, that is done with l V l 3 a relatively simple code. If you have a detailed large 4 break LOCA analysis or something more complicated that 5 might have some more complicated analysis.
6 You might have a TRAC analysis in one case and l
! 7 a very simplified in the other case which says the 8 licensee is picking the tool which is appropriate for the l
9 kind of analysis and decision that they are making. Both 10 of those fit some QA program.
11 You don't want people to make mistakes even 12 when they are doing relatively simple calculations. I 13 So, sometimes when we say appendix B people p
t I k/ 14 get scared. And sometimes when we say we want quality 15 PRA, people get scared. But this is not different from 16 what we have done for the last 30 years.
17 MR. ABBOTT: So, carrying the analogy a little 18 further, at the end of each cycle I do a reload analysis; 19 I don't submit all the back-up documentation, just a 20 summary.
21 MR. HOLOHAN: That's right.
22 MR. ABBOTT: So, the analogy would be that if 23 I was going in for a change I would tweak the PRA to 24 accommodate that change, I would generate a summary
/~T
(_) 25 document for your review as part of my license submittal, I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS
( 1323 RHODE ISLAND AVE., N.W.
l CJ) 234-4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433 l
118 l
l 1 and then you go ahead and review that.
! ew 2 MR. HOLOHAN: It is a little hard for me to j 3 say yes because I would have to know what the application 4 is. In theory, I think you are right. But in some 5 specific applications we may need to know more.
6 MR. ABBOTT: No, I understand that. But, in 7 general, that is the process.
8 MR. CHEOK: And there is a warning in the SRP l
9 that says that if you have looked at the PRA before, the 10 staff is guided to go look to see what was done before to 11 see if that is applicable to the present application. ,
1 l
12 MR. ABBOTT: Okay.
13 MR. CARROLL: Now, one thing I came across was
( l
'/
'- 14 the concern about the synergistic effect of a sequence of 15 changes.
16 How are you going to deal with that? Or how 17 should the licensee deal with that?
18 MR. CHEOK: Basically, what we are asking is 19 for licensees to come in with a PRA that reflects the 20 plant as it is. In other words, if he had come in two 21 years ago and had asked for an IST change, he now has to 22 reflect the IST change in his PRA.
23 So, when he now asks for a change in QA, he 24 now has got the base PRA that reflects the change in IST
,(, ~,) 25 space.
NEAL R. GROSS i
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 2344433 j
119 I 1 In addition, we have guidance to the reviews l
2 to look for things like you said, the synergistic effects l
' (7 7
v i
3 that in IST space we had assumed reliability of a 4 component, maybe based on the level of QA it was 5 receiving.
6 Now, if you are going to less QA on it, how 7 does that affect your previous application in the IST?
8 And words like that are in the SRP right now.
9 But basically, we expect them to have a base 10 PRA that reflects the plant as it is.
11 MR. CARROLL: That is automatically taken care 12 of when they do that. l 1
1 13 MR. CHEOK: Each one of the bullets in this p 1 1
\ j v 14 page basically is a different subsection in Appendix A.
15 These are the areas that we anticipate will be affected by I
16 applications.
17 We realized that in the overall review of the i 18 PRA we need to look at elements like the fault trees and 19 event trees. But we do not anticipate at this point that 20 the fault trees or event trees will be affected by the 21 applications.
22 So, at this point we do not have sections in 23 the appendix as to how we would look at changes to the 24 fault trees or event trees.
( The elements that we have, starting from the
( ,/ 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l i 120 1 top, is the use of appropriate data, whether the data that p., 2 is used, whether it be plant specific or generic, whether L) 3 it is applicable to the plant and whether it is applicable 4 to the change being proposed.
5 We are looking to see if the licensees have 6 done a search for, let's say, poor performing SSCs and if 7 those poor performing SSCs have been lumped into a group 8 and given a generic failure rate, how that would affect 9 the application.
10 We are looking to see if changes to the 11 parameter values in the data, whether these changes in the 12 data are relevant to the application.
13 Finally, I think we wrote int he appendix that
>V 14 we would like to see the performance monitoring program 15 from the licensees to be a follow-up for the data that is 16 used to support the application.
17 MR. PARRY: In the second bullet, which is the 18 effect of the application on initiating events, we 19 basically have some items in there, the most prominent 20 being, does the application introduce a new initiating 21 event or does it result in the change of his screening 22 criteria? Whether an initiating even that was screened 23 our previously, but could no longer be screened out 24 because of your application, or if it increased the i A i l
U/ 25 frequency of the initiating event that was included in i
NEAL R. GROSS j COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
121 1 there already.
f' 'g 2 The next bullet is quite application specific.
O 3 If we take the conclusions of an application are dependent 4 on the choice of the. mission success criteria chosen, we 5 would take a closer look at how the success criteria was 6 determined.
7 But we would probably only look at this if we 8 thought the conclusions were quite dependent on the choice 9 of the success criteria.
10 The next bullet is on common cause failures.
11 The guidance to the review is how does this application 12 affect my current common cause modeling and how the
,, 13 application affects the modeling and how the modeling, in
- 4 14 turn, affects the application.
15 CHAIRMAN APOSTOLAKIS: That is going to be a 16 difficult thing to do, I think. It's not on the same l l
17 level as the other bullets in the sense that they are 18 models of common cause failures are what they call 19 parametric models where you really lump classes of 20 potential failures together.
21 So, you propose a change now and try to figure 22 out what is a change in the alpha factor; that is hard.
23 That is really hard, especially if the basis for the alpha 24 factor is some past incidents that may have nothing to do
!m i
i
(_) 25 with the change.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
! 122 l 1 MR. CHEOK: Yes. I think we realize that it )
,rg 2 is difficult, but I will let Gary here address this a i C) 3 little more.
l 1
4 MR. PARRY: Thanks, Mike.
5 This is Gareth Parry.
6 It may be difficult, but I don't think, in a i
7 sense, that it is any more difficult than judging the 8 impact of changes on failure rates. Because, even a 9 simple pump fails to start, is a collection of 10 contributions from lots of different causes.
11 So, to that extent, in principle it is the 12 same. The techniques are available for doing it. It 13 involves a little bit more work probably.
A ;
! k 14 CHAIRMAN APOSTOLAKIS: Well, say bullets one l f
15 and four are in a class by themselves because looking for l l
16 effects on initiating events would probably be more j l
17 straightforward, wouldn't it? i 18 MR. PARRY: Well, actually, one of the points I 19 that I think Mike brought up about initiating events is 20 that in that case, one of the things that you are going to 21 do is look for things that have been screened out of the 22 model.
23 So, that involves a little bit of digging back 24 into documentation.
rh,
) 25 CHAIRMAN APOSTOLAKIS: Yes.
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 1
1
123 1 1 MR. PARRY: I think the intent of all these
,3 2 bullets effectively are helping you to really understand
( )
3 the assumptions behind the model and how they are being 4 tweaked to reflect the change.
5 CHAIRMAN APOSTOLAKIS: If we take the pilots, 6 do you see any impact of any of these on graded quality 7 assurance or IST and so on? And what would you do. I 8 don't know what you would do to see the impact of common 9 cause failures on data.
10 MR. JONES: Let me try one on the common cause 11 failures from the pilots, and then somebody can correct me 12 later, I guess.
13 MR. PARRY: In the greater QA on common cause
,- m
? a
\/ 14 failures, one of our key concerns is that to reduce QA 15 requirements on low e.afety significant components, exposes 16 you to potential common cause failure mechanisms across a 17 sea of components and systems, arguably.
18 To some extent we have backed up and looked at 19 system functions, more deterministic system review kind of 20 items to help us understand and assess what that may be.
21 Another way that you can use it is to, 22 understanding that common causes may be impscted, it can i 23 help guide the development of your monitoring program and 24 performance monitoring.
( ,) 25 So, if your concern is that common cause NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 ;202) 234-4433
124 1 failures could increase as a result of applying greater QA
,S 2 and you cant' really quantify the risk impact, then what N.) 3 you probably need to do is assure that you have an 4 effective monitoring feedback program so you catch it 5 before it becomes a significant problem.
6 So I think a lot of these elements are 7 questions that focus on what is the change, what might it 8 do to the application, how might it impact risk? Some may 9 be readily answerable within the risk model. Some may 10 not.
11 Based on that, you may trigger other elements 12 of the overall integrated decision-making process, 13 including performance monitoring, to feed back into your
! ^)
' 14 ultimate decision on why to grant the change and what 15 monitoring or other activities might be necessary.
16 CHAIRMAN APOSTOLAKIS: But, could I come to 17 you and say, look, I can't really quantify all these 18 impacts. There may be some changes in the numbers, but 19 really, I cannot and I don't think anybody else can, tell 20 you how much that is.
21 But here is my monitoring program, okay? I 22 will be monitoring the failure rate or something 23 egr.ivalent like number of failures over a period of time.
l 24 And I am going to give you some criteria.
l
() 25 As long as the number of fai'ures over a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE. N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
125 1 period of time is smaller than the criterion, leave me rS 2 alone; I don't even want to look at the effects on common UI 3 cause failures. Because I don't see any failures.
4 And that I think will also answer the question 4
5 that has come up as to what is performance-based 6 regulation. Because.you can't really look at these things 7 and also have the monitoring program, right?
8 You look at them when you make a decision, but 9 if you accept the monitoring program, you are saying, 10 well, maybe I will go with real life results. Unless, you 11 feel that something is so rare that you will see it under 12 normal conditions, but it may materialize under accident 13 conditions, in which case, of course, it is a different 14 story. You can't monitor, in other words.
15 MR. JONES: I think in the abstract I would 16 tend to agree with you. I think one of the questions in a 17 lot of the applications that we are dealing with is you 18 hear about ranking components and then reducing 19 requirements or oversight on low safety significance 20 components.
21 But one of the things that we want to 1 22 understand is that when you reduce that oversight, is that l
23 reduced oversight going to have a risk impact.
24 We may use some sensitivity studies to
,/
As ,/ 25 quantify or understand how the monitoring program relates NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433 l
_ .._. . - = _ . . . . _ . - _ . . . . . - . . . , . _ . _ - - . . . - - .. - ._ _ _ . . - _
126 1 to a risk impact to help guide us in accepting monitoring 1
.r]. 2- programming.
d 3 Where we wouldn't necessarily be quantifying, 4 specifically, the effect on graded QA on the components, 6 5 but we may be able to say, look, if I have 20 failures and 6- increase the failure rate by orders of magnitude for the f 7 top 20 components,-its impact on risk is insignificant.- l f
8 So, I am going to have a monitoring program i If I start seeing QA-related
~
9 that counts failures.
10 failures at 5 or ten of them, I am going to start getting !
i c
11 worried that something is significantly broken and maybe I I 12 need to re-look at my QA program and enhance some specific j
i 13 aspect of it. ,
i 14 So, it.is hard in the abstract-to answer it, l
15- but that is how we do see it; as some combination thereof 16 might be appropriate.
17 MR. MARTIN: Lawrence Martin. That's a 18 performance'-based issue which the licensees have been '
p 19- working with over the years. You have components that 20 have been used in safety related applications and non 21 safety-related applications.
- 22 You recognize that those are common j
23 components, you have been looking at those all along. (
i l 24 But we also don't want to make a mistake and .
l l '
- 25 talk about failure rates, because really what we are l !
l NEAL R. GROSS
!. COURT REPORTERS AND TRANSCRIBERS l l 1323 RHODE ISLAND AVE., N.W. !
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 I
127 i
i talking about here at this particular time happens to be l
,,3 2 number of failures, because we are not counting success. l t a \'
k s!
3 And that is a big difference in what we are 4 looking at and what the capabilities of the licensees are.
5 CHAIRMAN APOSTOLAKIS: Well, and the failure 6 rate is not observable any way. So you would have to have 7 a statement in terms of actual failures.
8 I guess, there is a minor concern that, in my 9 opinion, it would take a very experienced and confident 10 guy to say that yes, there is an impact, but nobody can 11 quantify it; let's go on and do something.
12 I think that people who are not confident will 13 be very reluctant to say this and this may drag on (V) 14 forever. You know, let's look at the data, let's look at 15 this or that.
16 But that will probably come out again, from 17 the pilots and real life applications. I don't think that 18 somebody who really doesn't understand the models will 19 sign off easily.
1 20 MR. HOLOHAN: There is another piece to this l 21 that you are not looking at and that is because it is not 22 in the Reg Guides or the SRP.
That is, there are other elements that make l 23 24 this program work. For example, it not only takes a
?" guidance document, but it takes responsibilities, training 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
l 128 1 program, and an approach, a process approach to doing 2 these reviews that isn't reflected in the guidance 73
( )
3 documents themselves, because they are only a part of the 4 process.
i Mr. Jones and I have been discussing how he is ,
5 going to do these reviews. I think that is something that l 6 l 7 we need to learn.
8 CHAIRMAN APOSTOLAKIS: Yes.
9 MR. HOLOHAN: I think that starting out we are 10 going to have some of our most senior and experienced 11 people doing these reviews in a sort of team atmosphere 12 and that will make these possible to do.
13 Then I think we have to lay out a process so
,m i t C/ 14 that not only can we accomplish four pilots over several 15 years, but so we can do this in a much more routine basis.
16 That might mean that we have to establish some 17 team work approach that we haven't done before because in 18 some sense we are doing more complex and more judgmental 19 reviews than I think we have done in the past. j i
20 MR. CARROLL: It is going to be hard to !
21 maintain consistency unless you do have sort a team of 22 four.
23 MR. HOLOHAN: But I am patient for Mr. Jones 24 to tell me how he is going to do it. I am sure he will D
! \
Q) 25 tell you too when he figures it out.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
r 129 -
1 MR. JONES: I have been hit with that stick a '
2 few times; I know I have that one. t
-Q O i r
3 MR. ABBOTT: I have a question on the first ;
4 bullet. Tliere is an awful lot of dead data taken in a~ j 5 plant. There is surveillance test data,' work requests, 6 post-maintenance testing and the maintenance rule data [
t 7 that is being collected on performance of systems. My [
i 8 understanding is that.you expect that information to'be 1
l 9 fed back into the PRA.to reflect the actual, so-called u i
- 10. actual working condition of the plant, for lack of.a j e
6 11 bet'ter term.
12 Is that true? Is that what you mean by use of l 13 appropriate data? ;
O 14 MR. CHEOK: Yes, that is basically what we i
15 mean.
i 16- If you have sufficient. plant-specific data, l l
i 17 they expect you to use it and apply it. ;
i 18 MR. CARROLL: You are talking about failure i
19 rate data or what he was talking about?
20 MR. CHEOK: We are pretty much talking about 21 failure rate data at this point.
i: l 22 MR. ABBOTT: Okay.
23 MR. CHEOK: And I guess the last ;wo bullets 24 on this slide, modeling of human performance. In the f
25 human performance area you are looking to see if the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISt.AND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
1?0 1 modeling is appropriate, to see if the reactions that are l f- x 2 proposed are defensible, and if any human actions that
( )
%./
3 they can credit for to justify a change in the licensing 4 basis, that the licensee has appropriately captured this 5 commitment in the training or procedures for human action.
6 The last bullet on truncation is not as 7 important as the first five bullets.
8 It is put in there because there would be 9 applications that would come in that would use pre-solved 10 cut set equation. We would normally expect that all 11 applications would come in and re-quantify the PRA to 12 obtain a bottom line result.
13 But when pre-solved cut set equations are 1
\ '
\') 14 being used, we expect to look to see what truncation 15 limits used were and how they effect he results.
16 MEMBER POWERS: It seems to me that when you 17 authored Appendix A, you were very careful to avoid 18 putting in quantitative guidelines.
19 For instance, use of appropriate data when you 20 discuss generic versus plant-specific data and you say 21 there is not enough plant-specific data so you will have 22 to use generic data.
23 But you didn't say what enough was.
l l 24 You put in some quantitative numbers on
(
- (,/ 25 truncation limits, but it was four orders of magnitude NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
131 1 range in there, so it was like not having any guidance.
,r' 2 What was the motivation in avoiding
()g 3 quantitative guidance here?
4 MR. CHEOK: I guess we didn't want to be too 5 prescriptive and to tie the staff into certain common 6 cause models or human reliability models or certain data 7 values.
8 We would have to review each one of its 9 elements on its own merit and how it applies to the 10 application.
11 MR. JONES: I think, Dana, on data values, j 1
12 while we didn't put specific numbers in the appendix, we 13 did note that what we would anticipate the reviewer would !
I,,h
\# 14 de would be to compare the basic assumed numbers to 15 relevant data. And basically, if looks within a factor of l
16 three, we were comfortable that the licensee had likely 17 used data that made sense.
18 MEMBER POWERS: You are quite right, Bob. But 19 --
20 MR. JONES: And what we'were primarily looking 21 for was strange numbers or assumptions to trigger a more 22 detailed review to assure that the assumptions made sense.
23 So that if they had assumed a very low failure rate for a 24 component which might even be driving the risk impact, we
(~ 'N,
( ,/
25 wanted to make sure that we focused on that.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4 433 WASHINGTON, D C. 20005-3701 (202) 234-4433
132 f 1
1 MEMBER POWERS: I think I understood that fw 2 intent. And you are right; there is that factor of three
! )
\J 3 guidance on parameters in there.
4 What I was asking is why there wasn't more 5 that told me what a strange number was. And maybe it is 6 premature to put in guidance on what strange numbers are, 7 are you going to have some sort of repository of non-8 strange numbers that I can use for comparison against?
9 MR. HOLOHAN: I doubt that we will have them 10 at this 1. vel.
11 MEMBER POWERS: But somewhere?
12 MR. HOLOHAN: But if you think about the scope 13- of applications for the general Reg Guide and the general l )
\# 14 SRP, I think it covers so many issues that can be 15 addressed with so many different types of decisions to be 16 made, that more dretailed guidance isn't going to be in 17 this document.
18 Now, when we talk about the application-19 specific documents, an IST or a graded QA, then maybe with 20 more experience we will begin to settle on comfortable 21 values of models that are acceptable, where we have l
22 larrowed ourselves to a certain type of decision on a 23 certain type of issue.
24 But at this level, I don't expect a lot more j
( ,/ 25 detail.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
WASHINGTON. D C. 20005-3701 (202) 234 4433 (202) 2'A-4433
. . _ _ . _ .. _ . - _ _ . _ . _ __ ._ _ . . . _ _ . _ . _ _ _ - . _m -... _ _ _ _ _ _ .. _
133
.1- MEMBER POWERS: Because when you get to the L- 2 application-specific, there is some nice guidance in there p 3 that'says be real careful because you can't ignore the 4 rest of- the - PRA. People can dial numbers around the rest 5 of the PRA so that your application-specific doesn't look l.
6 so-terrible.
! 7 So, what you are saying to the. guide is, you l l s i
8 are.a specialist in this particular specific application, !
9 but he has to go look at this other stuff. ,
l 10 That is when he needs the. guidance most.
11 Because he is not a specialist in those areas. He doesn't 12 have numbers on other component failure probabilities off !
i i 13 the top of his head.
14 MR. HOLOHAN: But I expect Mr. Jones to solve ,
l 15 that problem in'a.different way.
I 16 MEMBER POWERS: Ah.
t 17 MR. HOLOHAN: Because I think probably the
, i 18 mechanical engineer who'is worried about the IST valve :
19 testing, who is a valve expert, the way to solve that 20 problem is not to give that analyst more details about how 21 the rest of the PRA ought to be done, but to give him some 22 help by someone who does understand the whole risk-23 assessment.
- j. 24 So, I don't know how Mr. Jones is going to 25 solve this problem, but I imagine that it is going to be NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
l
134 1 more than either guidance or training for the technical n 2 topic expert so that they can deal with it on their own.
- t LJ 3 MEMBER POWERS: You think that it is going to 4 be worthwhile then to include in this discussion of a 5 focused scope review to inform the reviewer that you are 6 going to, almost ipso facto, have to work in a team 7 environment here?
8 MR. HOLOHAN: Well, I don't want to be too 9 prescriptive in telling Mr. Jones how he is going to solve 10 this problem.
11 MEMBER POWERS: I understand; it is a 12 performance-based thing.
13 MR. HOLOHAN: But your suggestion is not
)
.i' 14 unreasonable.
l l
15 MR. CARROLL: Didn't I see a section some l
t 16 place that pretty much suggested who the primary reviewer 17 ought to consider?
18 MR. CHEOK: For the body of the SRP basically l 19 says who the primary reviewers are which is the technical 20 groups. Then they are supposed to go to the PRA branch or 21 whoever they see fit.
23 MR. CHEOK: That is correct.
24 MR. HOLOHAN: Yes, but how that would actually
,e y
) 25 be accomplished has not been set out in detail.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
135 1 MR. JONES: It's not been set out. You start
,e s 2 with a flow chart that was put up this morning about the N '/
3 licensing process, and we are going to have to start from 4 there in developing the guidance.
5 These things ccme in to PMs, they get 6 distributed to staff to do reviews, how do you make that 7 decision? Who is the primary review branch? When 8 should they seek help? What are the type of guidance and 9 questions they should ask themselves as to the level of 10 help they will need? That will probably include some 11 understanding of what previous applications have been 12 generated. What SERs are available on our internal home 13 pages, for example that could be referenced and should be
( )
N> 14 searched?
15 I anticipate that kind of a process document 16 being developed. Then I anticipate that we are going to 17 go out and train everybody on it.
18 Some focused initial training on it to get 19 feedback from key people, especially the people who have 20 been involved in the pilots. And once we get the feedback 21 and modify it, I anticipate that we will move it out to 22 the organization as a whole.
23 Just to give you the top level where I think l
24 we are going with this thing. As you can see, I have
(~N l (q,) 25 heard that I got the job. j l
NEAL R. GROSS 0 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433 0
l 136 l 1
l 1 (LAUGHTER)
,f ~] 2 MR. ABBOTT: I have a question on the next-to- i La' 3 last bullet on modeling human performance. j 1
4 In general, when PRAs are done, the human I 5 performance models are fairly narrowly focused on the i l
6 operator and the maintenance people who do the work.
7 But if you go back and look at the industry 8 over the last few years, the real screw-ups have occurred, )
9 by and large, because the organization itself that the j 10 individual resides in, isn't performing as well as it 11 should. It doesn't have the appropriate safety culture l l
l 12 and all the rest of it.
p 13 MR. CARROLL: Sounds like that letter we i )
',Vf 14 wrote.
15 MR. ABBOTT: How are you going to factor j i
16 organizational effects into the human performance when you i
17 are looking at a plant-specific PRA? In 30 words or less.
18 MR. CHEOK: In 30 words or less? Right now, I 19 guess we couldn't; not very easily.
20 But organizational factors is one of the 21 elements that we would look at in the grey area in our 22 decision-making process.
23 MR. ABBOTT: What specifically would you look 24 at?
/%
f i
, ( ,/ 25 MR. CHEOK: We would look at --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
137 1 MR. KING: I mean you'd look at things like
,rm, 2 recent events, performance indicators.
( !
v 3 MR. ABBOTT: Okay.
4 MR. KING: There are a number of things that 5 you could look at. They are all qualitative at this 6 point.
7 MR. JONES: We in fact have in the SRP 8 specifically when we get into the gray area, the kinds of 9 information we should pull together for management. This 10 is in the main body, not in this appendix.
11 MR. ABBOTT: Oh, okay. I just didn't see it.
12 MR. JONES: It was in the earlier document 13 that was transmitted.
7 i ~' )
14 And included as part of that is to consult 15 with the regions. People have more direct interface with l
16 the licensees on a day-to-day basis to understand what l
l 17 they see as the licensee performance out there so that can 18 be factored in by management.
i 19 MR. ABBOTT: You said that was in an earlier 20 document that we got?
21 MR. CHEOK: That is in the body of the SRP.
22 MR. HOLOHAN: But let me add a little 23 something on that subject.
24 We all know that we don't have models for
,r3
(_,) 25 organizational factors and those sorts of things. So, I NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 ,
l
.._ - .-- ~ . . _ . ~ _ . . _ . - . - ..... - - .. .. . _ . - . - - _.-..-.- - -
, 138 i-i 1- don't expectLstaff to be running off and looking up SALP-1
+
2 reports and performance indicators and this sort of 3 information, in trying to judge whether this licensee is 4 50 per cent better than the last licensee. !
I i
5 'It just doesn't work that way.
6 In effect, I think this subject'probably falls b
l
. 7 in the area of a search for outlyers. That is to say, if I
8 there is nothing special about this licensee who has this [
9- application, then you expect that there operators are not b 10 so different from other operators and their valves are not l 4
I 11 so different from other valves. ;
12 If there is some reason to believe that there j- - 13 is something very broken in this licensee's program,.and-i t l 14- one place to look is.on the so-called watch list, --
- j. 15 MEMBER BARTON: Negative trend letters. )
J c 16 MR. HOLORAN: Negative trend letters which say 17 what it is about their program that you are concerned 18 about.
19 You may then say that maybe, if I have been 20 watching their maintenance program and they have a SALP-3 21- in maintenance, this may not be the time to reduce the 22 requirements on their maintenance program.
23 So, I sort of see it as the exception to the 24 rule. You just watch out for doing dumb things.
( 25 otherwise, I just think that everybody else is just within NEAL R. GROSS !
COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D.C.- 20005-3701 (202) 234-4433
139 1 the bulk of -- you can't tell the difference between the 2 operator performance at one plant versus another.
3 That operator might be hired by another plant 4 two days later. You just can't control these things.
5 Hopefully, in the bulk of these activities, it 6 is not a major issue. But once in a while it might come 7 up as a special case.
8 MR. CHEOK: The wording you are looking for is 9 in the second paragraph of page seven of the main body of 10 the SRP.
11 MR. CARROLL: I would comment that your 12 reliance on the watch list probably should be tempered by 13 what Arthur Andersen said about the process development in
- _x
)
-N/
14 their report which questioned the way plants get on the 15 watch list.
16 MR. HOLOHAN: I hope that what I said wasn't 17 interpreted as reliance on the watch list.
18 MR. CARROLL: Okay.
19 MEMBER FONTANA: One thing that doesn't seem 20 to come through in this SRP -- and I don't know that that 21 is the place for it. That is a lot of the benefits from 22 the PRA is the rigor that enforces the people who are 23 doing it to really address their plant.
24 Thinking along those lines, it doesn't seem to
(_,/ 25 encourage the licensee to have his own people doing the l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
! (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 140 1 PRA rather than some Beltway Bandit.
,s 2 In addition, there is a lot of benefit for the (v) 3 licensee people to be in the expert panels just to get the l
4 give and take of the discussions that take place.
5 Is that kind of out-of-scope for something 6 like this?
7 MR. HOLOHAN: I think it is sort of out-of-8 scope in the sense that there are things that we might 9 like that don't quite fit in with what we might require.
10 I think it is good to have licensee 11 involvement in this sort of activity. Certainly the IPE 12 program is one of its benefits; the voluntary program.
13 I am sort of reluctant to write that into the f
'A 14 guidance documents to say that we reject Beltway Bandit 15 work.
16 MEMBER FONTANA: No, you wouldn't want to do 17 that.
18 MR. HOLOHAN: I am even reluctant to say that 19 we are going to give more priority to a licensee who did 20 more of their own analysis, I think.
21 MEMBER FONTANA: Okay.
22 MR. HOLOHAN: I think that the licensees at 23 this stage are recognizing the value of these things. If 24 they don't recognize the value then we probably can't
(-)
(_) 25 shove it down their throat.
l l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 l
l 141 i 1
l 1 MR. CARROLL: Didn't you pretty well make that
]
l rx 2 point in the IPE process? l l ) l
\) '
3 MR. HOLOHAN: I think we certainly made the l
4 point during the IPE process. But it seems to me, back in 1 5 1988, maybe the industry needed some guidance and i 1
I 6 encouragement to get involved. That was nine years ago.
7 If at this stage, they really didn't find that !
8 having their own staffs involved with these is a valuable i
9 thing, I don't think I could convince them. i 1
10 MR. CARROLL: Although, if what we are now 11 embarked on doesn't look like it is beneficial to them, 12 you will probably see utilities backing away from the PRA.
l p_ 13 MR. HOLOHAN: That is a possibility and a i 1 1
' ',\ 1 14 concern of ours. I 15 MR. CHEOK: Okay. The next slide is on 16 Appendix B: Integrated Decision Making. I 17 Here, basically we have guidance to the 18 reviewers on what they should be looking for when the j I
19 licensees put together the information from the PRA and 1
20 from the traditional engineer analyses.
21 We expect them to use all the information they 22 have. The process to use for this integration is expected l
23 to be systematic and defensible. And we expect that l 24 documentation will be available for staff review when we
/~
k m,)1 25 come upon results that we don't think are what we expect.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
i 142 1 So, we do not anticipate looking at all expert l
l fm 2 panel type results, but we reserve the right to ask for
\ -
-] 1 1
3 them. 1 l
4 We also give guidance as to what the technical 5 information basis is. In other words, we do not expect l l
6 the process or the expert panel to be making too many i
7 decisions based on what feels good. We basically g.tve j 8 guidance as to what we expect as the technical information 9 basis.
10 CHAIRMAN APOSTOLAKIS: Now, those guys, during i
11 the integrated decision-making process can say, look, I 12 looked up your slide 11, Use of Appropriate Data; Effects 7 13 of the Application; blah, blah, blah. ;
i\ < i
'-'/
14 We discussed it, we recognized that there is l l
15 some impact, we don't want to bother quantifying it so we I 16 are going to be little conservative in our assessment and 17 here is what we did. I 18 So, with one stroke, that would be acceptable. l 19 MR. CHEOK: Yes, sure, that would be 20 acceptable.
21 CHAIRMAN APOSTOLAKIS: I guess it doesn't come 22 through in the appendix that you can do that.
l 23 I think what we are having here is a failure 24 to communicate.
pm 25 A lot of people who have seen these things i
NEAL R. GROSS l
COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
! 143 l
1 feel that the burden on the industry will be such that
. t 2 they will not even bother.
f-)
1 Q) So, I think it is very important to emphasize 3
4 that there are some ways of handling these concerns that 5 are not as burdensome as they appear. For example, 6 Appendix A, Use of Risk Importance Measures, tends to be a 7 little mathematical in the sense that whoever wrote it 8 really wanted to make sure that the reader understood that 9 the importance measures used could be affected by the 10 truncation limits, whether you should do sensitivity 11 analysis or what.
?
12 This is okay if you are interested in 13 importance measures and their properties. But that is I,_h 1
'N / 14 different from saying that yes, these may be affected by 15 these, by my integrated decision-making phase, I will 16 account for these in another way.
17 So, I am not going to spend three months 18 trying to figure out how truncation limits might affect my 19 -- Do you see what I am saying?
20 MR. CHEOK: Right.
! 21 CHAIRMAN APOSTOLAKIS: If someone says, as you 22 say here, that the risk -- I remember seeing that 23 someplace.
24 MR. PARRY: For the Fussell-Vesselly you can
(/S) 25 use a .02 or something. Whatever. Then I come in and say l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l
I 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
l 144 l
1 I don't want to bother with all of this, I will be f3 2 generous and use .00 something.
x) 3 So, I don't want to spend more than five 4 minutes studying the impact on data.
5 I would be conservative. It is my money; I 6 can be as conservative as I want. If I am conservative, 7 it costs me money, right? I don't want to have to do a 8 whole study on the properties of importance measures.
9 MR. CARROLL: For a particular application.
10 CHAIRMAN APOSTOLAKIS: For a particular 11 application, I am willing to spend a few extra dollars na 12 be a little conservative, but do it.
p, 13 MR. PARRY: But George, I think that guidance
('s ) -
14 is there for those people who want to use it. I 15 CHAIRMAN APOSTOLAKIS: Yes, but what I am i 16 saying is that it is not clearly stated.
17 MR. PARRY: Then the reference should be i
18 somewhere int he body of the Reg Guide. I believe it does 19 say that where you want to use risk importance measures 20 for categorization, then Appendix A gives some guidance.
21 CHAIRMAN APOSTOLAKIS: Yes, but --
22 MR. PARRY: Maybe it needs to be stronger.
l 23 CHAIRMAN APOSTOLAKIS: Yes, that is what I am 24 saying. It should be emphasized that doing these kinds of p
( ,) 25 things is not a must. There is a great part of the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4 433 WASHINGTON, D C. 20005-3701 (202) 234-4433
145 1 process called the integrated decision-making where you
,e y 2 can come up with either qualitative arguments or I t
\_/)
3 recognize there is a problem, and there is the solution.
4 I don't really want to study this problem any more.
5 MR. MARKLEY: George, I think what you say is 6 to some extent true, but when you are preparing a 7 submittal of some sort, you have to consider how many 8 requests for additional information you might want to 9 tolerate int he process.
10 CHAIRMAN APOSTOLAKIS: Yes, and that is again, 11 as a licensee, my business. It's my money, right?
12 MR. CHEOK: I mean, I guess we agree with you, 13 but I guess the message is being lost.
(~,
\- 14 What we are trying to say here is we are 15 trying to steer the staff to areas where there might be 16 non-conservatisms. Put if the licensee is going to come 17 in with a conservative analysis, we will obviously accept 18 that and do less review work.
19 CHAIRMAN APOSTOLAKIS: I think a general 20 comment about these documents, and I think that is why 21 people have this impression that we are asking the 22 licensees to do major studies every time they want 23 something for us.
! l
! 24 We don't emphasize enough, I think, the l
(\ \
- () 25 significance of the monitoring program and the integrated ;
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
146 1 decision-making part. That you can do a lot of things 1
rx 2 there that will eliminate the need of detailed l t ) I
\_/ l 3 investigations up front. l 4 Now, that is something that will, again, come 5 with experience.
6 I am really bothered by these comments that l
7 nobody would ever do it because it is too much and this l 8 and that. Of course, in all fairness, a lot of people 1
9 have not read them. l 10 MR. HOLOHAN: That is right.
11 CHAIRMAN APOSTOLAKIS: But also, I got some 12 comments from people who have them and they are supposed
,, 13 to have read them. Members of this committee; this is too i \ l
\ /
14 much.
15 So, I think it is perception. And I am not 16 sure that we can correct that now before release for 17 public comments, but at least we can identify that.
18 I mean, Appendix A is a good example. If you 19 really read it and you don't know anything else, you would 20 think that what is required is a detailed mathematical 21 analysis of the properties of importance measures.
22 And I am sure that that is not the intent 23 because I have been talking to you for a while is why I am l
24 sure.
3 w_,/ 25 MR. CARROLL: Only a very small percentage of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 147 1 the applications that come in really require --
.rN 2 CHAIRMAN APOSTOLAKIS: All this. i
\s_j!
3 MR. CARROLL: Yes, that thing in Appendix A.
1 4 CHAIRMAN APOSTOLAKIS: So, it is just comment.
1 5 MEMBER FONTANA: Along those lines, several I 6 times ago we saw a document and I can't remember the l
7 number, but it went into great detail of what a quality fl PRA should have in it.
9 I think that if one read through it I think 10 they would really be discouraged; it just calls for so 11 much.
12 MR. HOLOHAN: 1602.
-s 13 MR. KING: It was 1602.
/ \
\ )
14 MR. HOLOHAN: Okay, 1602.
15 CHAIRMAN APOSTOLAKIS: 1602 which we were 16 talking about review at the end.
17 MEMBER FONTANA: Yes, I read it all the way 18 through.
19 CHAIRMAN APOSTOLAKIS: I still admire you for 20 it.
21 (LAUGHTER) 22 MEMBER FONTANA: It is still a reference in 23 the general reg guide.
l 24 MR. KING: And it's okay.
,,- \
t t
, i._,/ 25 MEMBER FONTANA: It could be kind of off-NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 148 1 putting for some.
l f3 2 CHAIRMAN APOSTOLAKIS: I think Mario, it is l ;
NJ 3 okay, to say that this is what we think a quality PRA is, l
4 as long as you say at the same time that you don't always )
l 5 have to do that.
6 MR. HOLOHAN: And I think we have said that.
7 If we need to say it a little better, we will.
8 There is no doubt in my mind that people who 9 don't read it will be frightened by it.
10 CHAIRMAN APOSTOLAKIS: That is what I think.
11 MR. HOLOHAN: And that will be the case by 12 many of the people who comment on these documents will not 13 have read them.
(\
b )
's / 14 CHAIRMAN APOSTOLAKIS: This reminds me of what l 15 Professor Rasmussen said once. For years after the 16 reactor safety study, people were scared of it but very l l
17 very few had actually read it.
18 MR. HOLOHAN: Yes.
19 CHAIRMAN APOSTOLAKIS: In fact the thinks that 20 he and Levine probably the only ones that read the whole 21 document. I 22 MR. HOLOHAN: It was completely inscrutable to 23 those who didn't read it.
l 24 (LAUGHTER)
(_j\
25 MR. CARROLL: Seriously, that's the basis for NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
149 l l
1 my recommendation that you find one wise man and say you
.g 2 are the devil's advocate within the agency here; you read i j v
3 this pile of paper.
4 MR. HOLOHAN: One of the reasons that we put 5 some of the flavor into the forward is that some people 1
6 might not get much further than that in the document.
i 7 Also, recognize that if we put lots more !
8 caveats and explanations in a document that isn't being )
9 read, it isn't going to change people's perceptions.
10 CHAIRMAN APOST0LAKIS: Yes. But I think that 11 is something that we will discuss tomorrow afternoon.
12 I have to stop at this point because we have 13 some interviews that we have to do. But we will certainly (v' )
14 come back to that, if not right after lunch then maybe ,
l 15 later.
16 There is plenty of time tomorrow for I 17 discussion.
18 So, we will come back around 1:00 p.m.
i 19 (Whereupon, the proceedings were recessed at 20 12:08 p.m.)
21 22 l 23 24
(%
() 25 l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
150 l
l l
1 A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N l
[ g~ 2 (1:08 p.m.)
l N~]) 3 CHAIRMAN APOSTOLAKIS: Okay. Ready to start?
4 Technical specification, right?
5 MS, GILLES: That's right.
6 Good afternoon. I'm Nanette Gilles. I am the 7 NRR team leader for the risk-informed technical 8 specifications team. This is my research co-leader, Roy 9 Woods, and this my Millard Wohl, from the PRA Branch in 10 NRR who is our PRA tech. spec. expert. And we are going 11 to talk today about the technical specification reg. guide 12 and SRP and pilot program.
13 Much of the information on the reg. guide and
(\--) 14 SRP you heard in the November meeting. So if there's 15 anything in particular you would like us to focus on, we 16 will do our best to accommodate that. Of course, the 17 tech. spec. documents follow the same format as the 18 general reg. guide and standard review plan -- the four-19 element approach, the first element being define the 20 proposed change.
21 As Bob Jones mentioned earlier, PRA has been 22 used in tech. spec. changes for a number of years, and 23 specifically, we are concentrating on changes in allowed 24 outage times and surveillance test intervals and --
(,) 25 MR. CARROLL: How about other changes?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
-151 1 MS. GILLES: The guidance doesn't address
- 3< 2 other changes at this time. I would expect that at a
\
v l
3 later date we may choose to expand the guidance to cover 4 things like change in safe end states for particular tech.
5 spec. action statements.
6 MR. CARROLL: Admunstrative control?
7 MS. GILLES: Where they're amenable to PRA 8 use.
9 MR. CARROLL: Okay.
10 MS. GILLES: In this element, we not only ask 11 the licensee to define the change, but also to demonstrate 12 the need for the change. And the typical acceptable 13 reasons would be things like, obviously, improvement in (v ;
14 operational safety, consistency of requirements either i
j 15 across the industry or across the group, an industry l 16 group, and reduction in unnecessary burden, such as 17 eliminating unnecessary shutdowns associated with non-risk l
18 significant equipment.
19 MEMBER KRESS: Those are four statements?
20 They're not "and"?
21 MS. GILLES: They're "or," correct.
22 Element 2 is to conduct engineering 23 evaluations, and we address some of the key principles i
! 24 here, the first principle being that the change meets the c s
(,/ 25 regulations, and the regulations of interest here are the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
152 l
1 tech. spec. rule and the rules that apply to the license f'
3 2 amendment process. Since we are concentrating on allowed i
I '%Y 3 outage times and surveillance test intervals, those are 4 really the majcr rules that would be affected, major 5 regulations that would be affected.
6 We also address the principle of maintaining 7 defense in depth, and we try to give some specific 8 examples of what a licensee might look for in regard to a 9 tech. spec. change. For example, in the area of 10 maintaining redundancy and independence and diversity, we 11 ask that they look at appropriate restrictions to preclude 12 simultaneous outages when they are entering these extended
,-, 13 allowed outage times for preplanned maintenance, and that
(
14 they also could identify compensatory actions that could 15 be taken.
16 For example, if you're doing preplanned 17 maintenance on a diesel generator, you would want to 4
l 18 preclude any switchyard work during that time. And also, j 19 don't remove equipment from service if adverse weather 1
20 conditions are predicted. Obviously, you have a greater 21 risk of loss of off-site power and such things during l l
22 those conditions. l l
23 MR. CARROLL: Do you really mean that, with l 24 the first triangle, that you shouldn't -- shouldn't there 1
- (_,/ 25 be some more qualification to that? I guess " appropriate" NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
1 r
153 1- does it, but it almost sounds like you shouldn't have.
2' pieces of totally unrelated equipment outages.
I 3 MS. GILLES: No. Obviously, we don't mean ;
4 totally unrelated. We would mean simultaneous equipment ;
i 5 outages, whether the two pieces of equipment would cause-a -l 6 risk spike, so to speak.
7 MR. CARROLL: Okay. !
8 MS. GILLES: Some other elements.of defense in j 9 depth that are outlined in the general reg. guide and some 10 specific examples given in the tech. spec. documents -- ;
11 avoid overreliance on programmatic activities to 12 compensate for weaknesses in plant design. And you'll see i 13 a little later on in the presentation that a piece of the ;
O- 14 tech. spec. implementation is a configuration control 3
15 program, and we just caution licensees, don't try and rely. i 16 on such a program to account for a large risk increase . !
?
17 that you might encounter for certain tech. spec. changes.
18 CHAIRMAN APOSTOLAKIS: What's large? I mean, 19 large is not allowed anyway. Isn't that one of the 20 principles of the thing, that the risk increases should be 21: small?
22 MS. GILLES: I guess what we're saying is that 23 we'would look.at -- if they came in with a number that was 24 above the guidelines, and then said, "
But using the
( 25 program, we believe'that the actual number could be i
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
'(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
-mn- .-m , -- , -----7yw .
154 l 1 brought down below the guidelines," we would be suspicious l
l 2 of it.
\)
3 MR. WOHL: For instance, some of the safety 4 monitors, when they are used, have a trip point at a 5 factor of 10 above baseline in real time, so that that
- i 6 would be considered a large increase at that point 7 management has brought in through the discussion in terms 8 of real time aspects, which are taken into account in 9 configuration risk management programs, not just the 10 average over time.
11 Is that --
12 CHAIRMAN APCSTOLAKIS: Yes. Again, isn't this 13 too detailed? Unless my fellow committee members think l
,\
\-- 14 otherwise. I thought the idea of risk-informed and l I
15 performance-based regulation was to minimize this kind of ,
i 1
16 guidance. I mean, we are asking them not to rely overly 17 on programmatic activities, but we seem, as a regulatory 18 agency, to rely on prescriptive guidance a lot.
I i
19 I mean, what if somebody came and showed, on 1 20 the basis of probability calculations, that a lot of this 21 stuff really doesn't matter? I mean, obviously, they have 22 to avoid to have simultaneous equipment outages, and so l
23 on. But that will show in the PRA, won't it? Say 24 establish appropriate restrictions. Is --
t'%
( ,/ 25 MS. GILLES: The PRA could identify them, but l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D.C. 20005-3701 (202) 234-4433
155 1 that wouldn't -- the PRA wouldn't provide the means for
,g 2 them to -- you know, they would still need to put some
/
3 kind of procedures in place to avoid those types of 4 outages when they are intentionally entering their allowed 5 outage time.
6 CHAIRMAN APOSTOLAKIS: But why do we care 7 about that? I mean, you're going to review these 8 appropriate restrictions and approve them?
9 MS. GILLES: We are going to ask -- we are 10 going to ensure in our review that they have taken steps 11 to identify appropriate restrictions and have procedures 12 in place for those.
13 CHAIRMAN APOSTOI AKIS: But you will not pass
\- 14 judgment regarding the reasonableness of these procedures?
15 You will just want to know what are the --
16 MS. GILLES: I would say that we would 17 question if we felt that they had missed some major risk 18 significant equipment configurations. We would question 19 that. What we have not questioned is the exact criteria 20 necessary that they are using to identify those 21 configurations.
22 MR. CARROLL: Have you historically been doing 23 that when people have come in with risk arguments to 24 change a tech. spec.?
(~h
(_,) 25 MS. GILLES: I'd say in the recent history we NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
156 1 have been.
i 2 MR. WOHL: Yes. A good example is South Texas ;
7S
\ /
3 plant, which received AOT relaxations for diesels in ECW l
4 last Thanksgiving. No, it was Halloween Eve I think the 5 SER went out.
6 But at any rate, we have ongoing dialogue with 7 all of these people before any of these relaxations are 8 granted, and they had a very strong configuration risk 9 management program, which was referenced in the bases of 10 the specs. and described further in administrative 11 controls. Furthermore, they had some other equipment 12 outage prohibitions that are in procedures.
,_. 13 So we watch this very carefully, and we do it I i J 14 in a dialogue situation. And when somebody says to me, 15 "Our trip point is a factor of 10 higher than baseline,"
16 we discuss things like that. One reason for a factor of 17 10 is that a single piece of equipment going down can give l l
18 you a risk spike as high as a factor of seven or eight 19 occasionally. So it's a convenient kind of a number.
20 MR. CARROLL: Okay. Now, suppose Podunk 21 Light & Power came in with a similar proposal, but it went 22 to somebody else in your group. Would they take exactly 23 the same --
24 MR. WOHL: We talk within the group a lot, and
(~'T
( ,) 25 John Flack is our group leader. He makes sure that NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON. C C. 20005-3701 (202) 234-4433
157 l l
1 there's a lot of conversation coordination going on and we l
2 were doing things in a similar fashion. In fact, we have,
,r x, ,
\_/
3 you know, items that come in that do not look very strong.
4 And when they don't, we so state. I mean --
5 MS. GILLES: And obviously, one of the reasons 6 we're trying to write the guidance down is to achieve some 7 of that consistency, both on the part of the industry in 8 their submittals and on our part in our review. And I ;
1 9 would point out that, you know, we are trying to provide l
10 some of those guidelines here.
1 11 MR. CARROLL: I guess my trouble has been in 12 reading all of this stuff. It's more in the nature of p.- 13 nice platitudes than it is in terms of guidance that t\) 14 doesn't require an awful lot of interpretation. That's 15 the nature of the best I guess.
16 MEMBER BARTON: For example, if I submitted a 17 tech. spec. change now for extended outage times for a 18 piece of safety equipment, and I look at the -- I'd still 19 have to document an engineering evaluation and 20 justification for doing that. If I looked at the current 21 process as opposed to what is in the reg. guide for 22 conducting engineering evaluations, you know, it seems to
! 23 me that the process today is less onerous than what I see 24 in the reg. guide in Section 4 on conducting an
/% s
\._,/ 25 engineering evaluation.
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS
! 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
158 l
1 So why would I -- well, let me ask you !
gS 2 something. In the pilots -- and I know the pilots haven't t, !
s"/
3 seen the latest reg. guides -- but as part of the pilot 4 program, do you expect to see a submittal similar to 5 what's in the reg. guide before we go public with the 6 document?
7 MS. GILLES: We actually had the pilot 8 submittals before they were even identified as the tech.
9 spec. pilot. So we are a little unique in that respect.
10 And the pilot submittals, which I will get to here in a 11 moment, are joint application reports, which are topical 12 reports from the CE owners group.
13 And those have sections that address both and
-) 14 what they use called deterministic considerations, where l
l 15 they address, you know, what is the design basis of the 16 system, what kind of design basis regulations was it sized 17 to meet, and they address tnose issues, and then they 18 additionally address the risk significance. So I don't 1
19 think that the pilots had any problem, you know, l 20 addressing the types of issues that we have outlined in 21 the guidance documents.
22 And to be honest, I don't believe that there l 23 has been a parallel process in the past for doing the 24 types of things we're doing with PRA, as far as extending (j 25 allowed outage times and surveillance test intervals.
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
l
159 1 Almost all of the submittals we've seen to date in these 7m s 2 areas have had some risk or reliability piece to them.
(V )
3 MEMBER BARTON: Yes, some risk. But it's less 4 than what you're asking for now, I believe.
5 MR. HOLAHAN. This is Gary Holahan. Let me 6 see if I can answer your question a little more directly.
7 And that 33, I think the word " onerous" is not the one I 8 would have chosen. I think we are talking about guidance 9 documents that do ask more of the licensees than they 10 would have sent in the past.
11 MEMBER BARTON: Okay.
12 MR. HOLAHAN However, the staff is also 13 willing to grant more based on that additional information
(~~)\
\- 14 than it would have granted in the past. You can have a 15 very simple example that says -- I mean, would the staff 16 likely have approved a 21-day diesel generator outage time 17 in the absence of, you know, substantial support from a 18 risk analysis? And I think that never would have 19 happened, or was asked for but a lot more was granted.
20 That's the bargain.
21 MR. CARROLL: That's a real case? That's --
22 MR. HOLAHAN: That's a real case.
23 MS. GILLES: Actually, the final South Texas 24 was 14 days, f3
(_,) 25 MR. WOHL: But the 21 days was justified on a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
160 1 risk basis.
7~ 2 CHAIRMAN APOSTOLAKIS: Is there anything
'w] 3 magical, by the way, about these times being multiples of 4 seven? Why do they always seem to be multiples of seven?
5 MEMBER CATTON: Because Monday is a good day 6 to start things.
i 7 (Laughter.)
8 CHAIRMAN APOSTOLAKIS: It's a magic number.
9 MR. HOLAHAN: A magic number. i 10 CHAIRMAN APOSTOLAKIS: I think you're going to ;
1 11 tell us to read that paper again.
12 MR. CARROLL: It's what the lawyers call l
l 13 arbitrary and capricious. l
/ )
v'
/ 14 CHAIRMAN APOSTOLAKIS: I mean, why couldn't it 15 be 17?
16 MR. CARROLL: Arbitrary and consistent, 17 actu..lly.
18 CHAIRMAN APOSTOLAKIS: Okay. Let's go on.
19 MS. GILLES: Okay.
20 MR. WOHL: Would you like it better if we 21 called it one week?
22 (Laughter.)
23 MS. GILLES: There you go.
24 CHAIRMAN APOSTOLAKIS: Maybe.
,r3
) 25 MR. WOHL: I'd like to add one thing here --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS .
1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
161 1 that as a PRA tech. reviewer, quite often project managers 2 get requests for relaxation that don't even have any risk-7x
() 3 based content. And they come to us and say, "Well, what 4 do you think of this, you know? I don't think it's too 5 bad, but what do you think of it?"
6 And quite often, that will result in some 7 phone calls to the licenser and getting extract things on 8 PRA from the PRA staf f rat her than through licensing 9 people, and so we can get a gauge for what kind of risk 10 impact this proposed action might have. It goes on all of 11 the time, even though we doi:'r document everything.
12 CHAIRMAN APOSTOLAKIS: I suppose what makes me 13 uncomfortable with this is the word " maintain." I think t t U 14 the word " maintain" -- defense in depth and safety 15 margins.
16 MEMBER SHACK: Do you want to increase or 17 reduce?
18 (Laughter.)
l 19 MS. GILLES: The actual words, I think, might ;
l 20 be " maintain sufficient," not --
21 CHAIRMAN APOSTOLAKIS: Well, wouldn't it be 22 reasonable to say --
l 23 MS. GILLES: Not necessarily current level.
24 CHAIRMAN APOSTOLAKIS: -- " assure" or --
I \
(/ 25 MEMBER SHACK: Assure adequate safety margin?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 162 1 MR. JONES: The specific words are " sufficient j]
2 safety margins are maintained."
l U What?
3 CHAIRMAN APOSTOLAKIS:
4 MR. JONES: " Sufficient safety margins are 5 maintained" is the principle. And the key phrase there is 6 obviously " sufficient."
7 CHAIRMAN APOSTOLAKIS: Yes, right.
8 MR. JONES: And then, if you look in the 9 general reg. guide, there is some words relative to safety 10 margins and what sufficiency means, and find it real 11 quick.
12 MR. WOODS: We've changed that word from
,._.s 13 " adequate" to " sufficient" and back two or three times.
(#)
'- 14 We could change it again, I suppose.
15 MR. HOLAHAN: No , please don't offer.
16 (Laughter.)
17 CHAIRMAN APOSTOLAKIS: Just saying maintain 18 safety margins. I mean, if you say " sufficient" --
19 MS. GILLES: Sufficient -- this is, I guess, 20 an encapsulated version of what's actually in the reg.
21 guide. It does say " sufficient."
22 MR. HOLAHAN: The reg. guide says " sufficient" 23 for safety margin.
i 24 MEMBER SHACK: Are maintained. That's a good i
n I '
( ,e) 25 statement.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
I 163 l 1 MR. HOLAHAN: Right.
l
,cz 2 MR. HOLAHAN: The word " maintain" should not l t }
1 v \
l 3 be read to imply it can't change. !
\
4 CHAIRMAN APOSTOLAKIS: Yes. Yec. l 1
5 MEMBER SHACK: Just coming back to George's 1 6 general comment, too, on these things, none of these 1
7 statements are necessarily meant to be layered on top of 8 the PRA. That is, PRA may well show that this is all 9 true.
10 MR. HOLAHAN: Right.
11 MEMBER SHACK: I mean, this is just kind of an 12 engineering statement of how you get to good PRA numbers.
13 But I don't think these should be interpreted as something
- s U 14 that, you know, they're going to walk in with a PRA that l l
15 shows the risk is very low, and then say, okay, you've got j 16 to do more, necessarily.
l 17 CHAIRMAN APOSTOLAKIS: Well, the way I see it 18 is the PRA has the answer as: this is the probability of And you look at i 19 core damage, or whatever the metric is.
20 how they did the analysis, and you realize that there is 21 no probability that two pieces of equipment will be out at 22 the same time. Right? Because they say, you know, "We're 23 not going to do that."
l 24 Then, you come back and eay, "Well, convince n
l C) 25 me that that will be the case." That's really what you're NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
164 1 saying here -- establish appropriate restrictions to 73 2 preclude simultaneous equipment outages.
l !
L/
3 MS. GILLES: Yes.
4 CHAIRMAN APOSTOLAKIS: Convince me that this 5 assumption is, in fact, true. That's what you're really 6 saying here.
7 MEMBER SHACK: Yes. But, I mean, it was 8 already built into the PRA.
9 CHAIRMAN APOSTOLAKIS: It was already built 10 into the PRA, yes.
11 MEMBER SHACK: So it's not an additional 12 restriction in any sense.
13 CHAIRMAN APOSTOLAKIS: But you will do the p
! I
'd 14 same thing with all of the assumptions in the PRA, won't 15 you? I mean, that's what we're discussing this morning, I 16 mean, the quality of the thing and -- so you're just 17 singling out this one because --
18 MEMBER SEALE: Here's one quality measure that 19 they want to be --
i I
20 CHAIRMAN APOSTOLAKIS: Well, it's closer to 21 their interests, so they really want to know about it.
22 MR. JONES: I think to a certain extent you l 23 can't, through the PRA -- well, you could through the PRA 24 if you wanted to analyze it to death. But you cannot C/ 25 preclude multiple equipment being out of service just by i
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
, (202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
165 1
l l 1 doing the PRA analysis.
3 2 CHAIRMAN APOSTOLAKIS: No, no, no. The PRA i )
LJ 3 analysis --
4 MR. JONES: So the PRA analysis has made 5 certain assumptions relative to the AOTs and surveillance 6 test intervals, what's proposed. And we look at that on 7 an average basis. Part of the process, or the second part 8 of tier 2 of the process, is to say, okay, when you're in 9 those configurations, identify what appears to be the most 10 significant -- risk significant equipment that if taken 11 out of service would have a potential large risk impact.
'12 Pardon the word "large," but a significant risk impact. l 13 Having that understanding, you then make
.m
'> 14 judgments, do I or don't I need to ensure that assumption 15 by incorporating it as a requirement in the tech. spec.,
16 so that when I'm in an AOT and doing work on a diesel, I 17 can't take out this other piece of equipment. And then 18 that becomes part of the action statement that would go 19 into the tech. spec.
20 Now, we also recognize, however, that we're 21 not going to catch them all. And, in a similar vein in 22 the maintenance rule, when you go out to do maintenance, 23 you're supposed to evaluate what your system looks like l
24 and assess its risk before you do maintenance activities.
l r~T
, { ,) 25 So tier 3 is basically make sure you have a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
166 1 configuration control program, which is embodied basically i
em) 2 in the admin. control sections of the tech. specs., which
'V 3 make sure you look at that -- those things that you 4 haven't specifically identified -- that you look at those 5 before you do maintenance and make sure those are not 6 causing particular risk vulnerabilities that you shouldn't 7 have.
8 And that's kind of the framework, if you wish, 9 of the whole risk informed tech. spec. process. So, you 10 know, it's in that framework -- to try to identify what's 11 most important, put the appropriate controls, and then 12 have configuration management underneath that to help do 13 that, which is in a sense, arguably, what the maintenance
/si,
'd 14 rule already requires.
15 MEMBER KRESS: Can I read into the word 16 " preclude" in this, where you want to preclude 17 simultaneous failures, can I read that as reduce the 18 probability of simultaneous failures to an acceptably low 19 level? Is that --
20 MR. HOLAHAN: To a negligibly low level 21 consistent with the assumptions in the PRA, presumably 22 which assumes zero.
23 CHAIRMAN APOSTOLAKIS: Without quantifying it.
24 Which we don't want to get into.
't.,/ 25 MR. HOLAHAN Right.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
167 l
1 CHAIRMAN APOSTOLAKIS: We don't want to
/~N 2 quantify.
()
3 MR. HOLAHAN: Right. I don't think we need to 4 quantify the relationship between the procedure and the 5 likelihood. l l
6 CHAIRMAN APOSTOLAKIS: Okay. I l
7 MS. GILLES: This is the three-tiered approach 8 that Bob just described to you, which I guess he basically i
9 just described tiers 2 and 3, and tier 1 is obviously 10 where we look at the quantitative numbers and try to apply 11 some guidelines to those analyses.
12 CHAIRMAN APOSTOLAKIS: And that applies to low
,s 13 power and shutdown operations, too, right?
h]'"
14 MEMBER KRESS: Well, it has to. Talking about 15 outages.
16 CHAIRMAN APOSTOLAKIS: Avoidance of risk l
17 significant plant configuration. Does that --
l 18 MS. GILLES: Well, it depends on what the 19 particular piece of equipment --
20 CHAIRMAN APOSTOLAKIS: ,Yes.
21 MS. GILLES: -- for which the change is being 22 sought would be used for.
i 23 CHAIRMAN APOSTOLAKIS: But you're not looking 24 only at PRA at power, are you?
i n i
(_) 25 MS. GILLES: Not necessarily, no.
NEAL R. GROSS l
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
168 1 CHAIRMAN APOSTOLAKIS: But the CEO owners
,s 2 group did al.1 of them, didn't it?
- t 3 MS. GILLES: They did --
4 MR. WOHL: They looked at shutdown risk and at 5 transition risk also. And I might add that in talking to 6 plant people, many of them are much more concerned about 7 transition risk than any other kind of risk.
8 CHAIRMAN APOSTOLAKIS: Yes.
9 MR. WOHL: Having to shut down and then coming 10 back to power, and all of the realignments that have to be 11 done, at hot standby for instance, cause people a lot of 12 grief. And some of the BWR owners are very concerned 13 about a lot of the transition risk problems, too.
! )
'/ 14 And we have only addressed this in one place 15 that I know of, which is the little handbook on risk 16 applications to tech. specs., work bv Thomas Montcomo in i 17 Finland who was a subcontractar to s chaven in that 18 area. But it has been relatively unexplored. l 19 And some of the original estimates from the CE 20 owners group indicate that transition risk can be as large 21 as operational risk. So --
22 CHAIRMAN APOSTOLAKIS: By the way, do you have 23 an extra copy of that report that Montcomo did?
24 MR. WOHL: As a matter of fact, yes, I --
/ \
/ i
( ,/ 25 CHAIRMAN APOSTOLAKIS: Not right here.
l NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
l
169 l 1 MR. WOHL: -- have a copy of the handbook in j
)
1 l
l; n i 2 my briefcase.
%._)
3 CHAIRMAN APOSTOLAKIS: But if I could get a 4 copy by tomorrow --
5 MR. WOHL: Yes.
6 CHAIRMAN APOSTOLAKIS: -- that would be nice.
7 MS. GILLES: Again, this is going along the 8 structure of the reg. guide in the section that talks 9 about principle 4 and the risk evaluation. We have 10 sections that cover each of these topics, similar to 11 topics discussed in the general reg. guide, which talk 12 specifically about tech. spec. issues in relation to 13 quality, scope, modeling, assumptions, sensitivity, and p,
i !
14 uncertainty analysis. And then we have sections on each 15 of the tiers 2 and 3. !
16 CHAIRMAN APOSTOLAKIS: Let's look at quality. ]
l 17 What would you look, in your extending the allowed outage 18 time -- I don't have to look at the failure rate, do I? I ;
19 can say the failure rate that was using when I was out for 20 seven days, ought to be the same as if you're out for 1
21 seven or 14 days, right?
22 MS. GILLES: Well, typically, in many cases, 23 the reason for the extension is to increase preventive 24 maintenance on line, so the failure rate wouldn't f~)
(_) 25 necessarily be looked at in those cases. What we did --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W-(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
! l 170 i
i 1 CHAIRMAN APOSTOLAKIS: Common cause fallure l g3 2 potential? I mean, there was some potential at seven days i
- V' l 3 -- the same at 14? In other words, can I dismiss that in 1
4 five minutes?
5 MR. HOLAHAN: No.
6 CHAIRMAN APOSTOLAKIS: No. Okay. Why not?
7 MR. HOLAHAN: Well, because you probably 8 didn't -- well, it depends on the case. But if you had 9 reviewed the base case and felt comfortable with that 10 analysis, that's one situation. But the staff may not 11 have seen the analysis previously, so just looking at the 12 implications of the change I think is not enough.
13 CHAIRMAN APOSTOLAKIS: Well, I think Mike, in o
i'"} 14 his presentation earlier, used the words " focus scope 15 review." So that's what I'm trying to do. I'm trying to l 16 see which parts of the PRA I have to look at.
17 MR. HOLAHAN Yes. But, for example, if the 18 diesel generator reliabilities were chosen unrealistically 19 high, it might suppress the importance of a station 20 blackout sequence and. lead you to some wrong conclusions 21 about tech. specs, on maybe even auxilialf feedwater 22 system. So I think you can't just say, well, the change 23 that's being envisioned --
24 CHAIRMAN APOSTOLAKIS: No. But, okay, let's
?
V(
I 25 take that example and push it to the extreme. I am NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
171 1 requesting AOT extension to the diesel generator. So i rT 2 you're saying I have to look at the reliability of the YI 3 diesels that was assumed to begin with, and to understand 4 that better I have to -- and the accident sequences were 5 the diesel reliability of this. And then I have five new 6 questions regarding other elements of these accident 7 sequences. Before I know it, I'm reviewing the whole PRA.
8 Could that happen, or there will be something that will 9 preclude doing so?
10 MR. HOLAHAN: Well, perhaps we ought to have 11 -- since some of these reviews have already been done, 12 maybe you'd like to hear how they were done.
13 CHAIRMAN APOSTOLAKIS: Yes, I would like to l7_ \
)
14 know more about it. Yes.
15 MR. FLACK: John Flack with NRR.
16 Yes. I mean, as these reviews begin to 17 unfold, they do get somewhat extensive and major, and of 18 course you have to try to bound them to some regard.
19 Now, we have various AOT extensions in house, 20 some of which are very limiting. In other words, for 21 example, if you're going to take out the SITS tank, or the 22 accumulators, it's limiting it to LOCA, large break LOCA.
l l 23 So you really don't have to go and review station 24 blackout. Okay?
i
(\ <
V 25 If you're reviewing diesels, then it's going NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 1
172 1 to be station blackout. You don't have to review LOCA.
,/S 2 So, in a sense, it's limiting to what you're focused on.
NY 3 What is the AoT? What does it mean in the context of the 4 analysis? And that's what you follow through on.
S' With respect to quality, scope, and modeling, 6 it's hard to distinguish what we mean by any one of these.
7 It's all in the context of the review -- begin to look at 8 the cut sets, you look for completeness, you look for 9 certain things that are important in those type of 10 sequences, are they there, what are they using to quantify 11 it, how much credit are they taking for human reliability, 12 are they using cross ties, and that sort of thing.
13 But you follow the review as deep as you have 7,, .
'"] 14 to go before you realize that this is a reasonable process 15 that they did in a quantification. But it's not to the 16 extent where you have to review the whole PRA, at least 17 not to this point it hasn't.
18 MR. HOLAHAN: But it is to the extent that you 19 can understand the change that you are approving, and that 20 you're prepared to go to a public hearing and justify why 21 you made those decisions.
i 22 CHAIRMAN APOSTOLAKIS: Would it help, in this 23 case, us to have a library of results, from the IPEs, say, 24 so when you look at the station blackout sequences you
.O 25' will be able to very quickly say, "Yes, these numbers are
'(s/
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
173 1 reasonable. I don't have to check them"?
, em 2 MR. FLACK: Well, we do look in the contractor i s 3 reports and the IPEs that have been reviewed in-house for
- 4 weaknesses or issues that might have come up that we want 5 to understand if it really is an issue, has it been 6 resolved, and that sort of thing. We don't really rely as 7 much on those results, though. It's more for insights 8 that we use it.
9 CHAIRMAN APOSTOLAKIS: Well, what I'm saying 10 is that if it takes you too long to complete this review, 11 that is also a negative.
12 MR. FLACK: Sure.
13 CHAIRMAN APOSTOLAKIS: Disincentive for you to (mv) 14 even try. So if you can make that process faster, 15 shorter --
16 MR. FLACK: I think --
17 CHAIRMAN APOSTOLAKIS: -- by using -- again, 18 let's say that I do a very superficial analysis for a 19 plant somewhere which, however, gives me the number that 20 everybody else is using. It doesn't matter that I did a 21 superficial analysis. The reviewer should know that. I 22 mean, all of the IPEs for similar plants have come up with l 23 that number, or in that, you know, ballpark. So I can 24 shorten the time it takes to complete these reviews. Of
(^)
(_) 25 course, you have to focus on the actual diesels in this NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
174 l
1 case.
.s 2 MR. FLACK: No, I agree. I think we do
)
'w/ 1 3 capitalize on a lot of that. I mean, nobody takes it from i 4 square on. And a lot that can be written here, a lot j l
5 depends on the reviewer themselves and how much experience !
6 that reviewer has. But as we do more and more of these, 7 there's going to be certain things you expect to find, and 8 a certain structure to these, that will carry through on 9 the other reviews.
10 It's a little bit more difficult on the first 11 one, because everybody wants to get it right, you know?
12 They look at a lot of things.
13 CHAIRMAN APOSTOLAKIS: But I have seen p
V 14 tabulations like that. There was one paper I remember 15 from Brookhaven several years ago, you know, for these 16 kinds of systems, this is the kind of unavailability you 17 expect to see, for these kinds of systems something else.
18 I think you should have tables like --
19 MR. HOLAHAN- Yes. Beyond tables, I mean, 20 there is a database that has sequences from the IPEs.
21 CHAIRMAN APOSTOLAKIS: Well, I mean, you need 22 some sort of synthesis. Is that what you mean, that they l
23 have taken, say, all of the high pressure injection l 1
l 24 systems?
1 !
, h,m 25 MR. HOLAHAN: There is a --
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
175 1 CHAIRMAN AP^STOLAKIS: That this is the range 2 of unavailabilities?
m) l O 3 MR. HOLAHAN: I don't remember that it has a 4 numerical value.
5 CHAIRMAN APOSTOLAKIS: I mean, that kind of 6 thing. I'm sure the raw information is there. ]
7 MR. HOLAHAN: Yes.
8 CHAIRMAN APOSTOLAKIS: But just putting it 9 together in a useable form. That would be helpful --
J 10 MR. HOLAHAN: Yes.
11 CHAIRMAN APOSTOLAKIS: -- in the review. l l
12 MR. HOLAHAN- Which I would have been a little i 13 more comfortable if you described that as a simplified I ')
\
') 14 analysis rather than a superficial one.
15 (Laughter.)
16 CHAIRMAN APOSTOLAKIS: I really want you to be 17 comfortable, so I --
18 MS. GILLES: As a matter of fact, the process 19 you're describing is what the CE owners group did in their 20 reports. They took all of the CE plants and had each 21 plant run the analysis, and then put all of the numerical 22 results together in a table so that we could do just that; 23 we could do the cross comparison.
24 And obviously, our intention is, yes, we did a On much more detailed review on the first one, the lead V 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
1 176 l 1 plant, but we are hoping that those insights will spill i l
g~) 2 over to the remaining plants and that we will be able to
%)
3 do them with a much shorter review time.
4 MR. CARROLL: George, you spoke of things 5 being a disincentive to the utilities to utilize this 6 process. I think you also have to broaden it to say that 7 it is overkill, may be a disincentive to the agency to use 8 this process.
9 CHAIRMAN APOSTOLAKIS: Yes.
10 MR. CARROLL: In the sense that it's going to 11 take a lot more manpower to --
12 CHAIRMAN APOSTOLAKIS: But given that the
_ 13 licensee really initiates the process, it's really more 14 important.
15 MR. CARROLL: Maybe that's their strategy.
16 CHAIRMAN APOSTOLAKIS: See, I think maybe what i
17 Gary said earlier this morning that something you don't l 18 read is inscrutable lies here. But if you look at these !
l 19 bullets without the discussion that followed, I think 20 people do get scared for anything they want to request 21 they have to go down this list. And I don't know how you 22 correct that, by the way. I'm not saying there is an 23 obvious solution. People should read the guides; there's 24 no question about it. But still --
i ,-
[ (_ / 25 MEMBER KRESS: Sometimes we need to have a i
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
177 l 1 little balance here. If you look at all of those bullets,
(~3 2 which one of them would you not do?
V 3 CHAIRMAN APOSTOLAKIS: See, if you put it that 4 way, you are putting me on the spot. But each one is, Jay 5 correctly has said several times, is really a platitude.
6 Make sure your PRA is a quality PRA. The devil is in the 7 details. How you --
8 MEMBER SHACK: Platitudes are true.
9 (Laughter.)
10 MR. CARROLL: Not always.
11 MR. JONES: I mean, I think there is a 12 relatively extensive discussion on quality of the PRA and 13 attributes in the tech. spec. reg. guide, including the 7_.x t-) 14 emphasizing that the quality should be compatible with its 15 intended use, scope and level of PRA necessary to support 16 the type of change. I mean, those words are --
17 MR. CARROLL: Whatever those words mean.
18 MR. JONES: But, I mean, I can't say, "Give me 19 an AOT." I can't describe it for every possible change 1 20 that may come in, so I can say you need to do it 21 appropriate for the type of change you want.
22 MR. ABBOTT: Absolutely.
23 MR. JONES: I mean, we're not saying you have l
24 to do that.
/\
f
(_/
1 25 CHAIRMAN APOSTOLAKIS: You're right. We're NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
178 i
l 1 not arguing with that.
l
,- 2 MR. JONES: I'm just saying that we have done N,- ]
3 that and put them forward in the guides as are. Just 4 wanted to make sure that it was clear.
5 MR. ABBOTT: I have a question. I mean, I 6 have read this as scrutable. The question I want to ask 7 though is, is this process -- as described in the reg.
8 guide and the draft SRP -- the same process that has been 9 used in the past for processing these same kinds of 10 changes to the technical specifications, much like the one 11 you mentioned on South Texas?
12 MS. GILLES: It is the same process.
13 MR. ABBOTT: This process has been tested?
( )
14 MS. GILLES: Yes. As a matter of fact -- I 15 was going to make this point later, but this seems like a 16 good time. In regard to feedback, it's kind of 17 interesting. As I mentioned, the CE owners group embarked 18 on this process before we were even into risk informed 19 pilots. And what they used was the handbook that Millard 20 mentioned earlier, the handbook on risk-based tech. spec.
21 improvements. And they developed their reports, and then 22 they wrote an industry guide for EPRI on how to do it.
23 And then we took that industry guide, and we 24 took the handbook, and we wrote the reg. guide. And so I
/h
(,) 25 would say that there has been a considerable amount of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
179 1 feedback, and I don't think anybody is going to be
,~ 2 surprised at what they see in our guidance documents.
'^'
]
3 MR. ABBOTT: Okay. Now I have -- well, I 4 guess for lack of a better term -- sort of a mirror image 5 question. Technical specifications and the use of risk 6 assessment in configuring the plant in a safe way, 7 regardless of what's going on, is all a good thing. And 8 if someone comes in with an application to change an AOT, 9 he will -- again, for lack of a better term -- be 10 subjected to all of this stuff, and all of this stuff is 11 good. You know, get a longer AOT, you'll have 12 compensatory measures and other things, and, in general, 13 the overall safety of the plant will probably be better.
,cy
)
x/ 14 On the other hand, if I don't come in, none of 15 this is going to get done. And if I have two plants that 16 are exactly the same, one comes in and one doesn't come 17 in, it may well be that the one that doesn't come in will 18 be less safe than the one that does.
19 Now, as a regulator, how do you handle that?
20 MS. GILLES: I guess at this point in time, I 21 don't have a way to handle that, because both plants are 22 meeting the regulations. Both plants are, you know, l
23 adequate protection and all that.
24 MR. CARROLL: The figure we saw this morning
,~N I ) 25 would suggest that after plant A got the change it would NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
180 1 be very simple for plant B to come in and say, "We do."
gx 2 MR. ABBOTT: Only if that plant had a quality 3 PRA, which it may not have.
4 MR. MARKLEY: Or if they don't have other 5 performance issues that would cause them to have a cloud 6 that woald disdain --
7 MR. ABBCTT: All plants aren't equal.
8 MS. GILLES: That's true.
9 MR. CARROLL: You just said that -- that had 10 me worried, because that's what --
11 MR. ABBOTT: No. Equal from the point of view 12 of design, not equal from the point of view of the level 13 of effort and time that's been spent putting the PRA
(' \
'-- 14 together. I mean, some plants have taken the back of the 15 hand to the IPE process, and they've put it on a shelf and 16 that's where it sits, because they've answered the 50.54 17 after it was done. !
l 18 Other plants, on the other hand, have taken it ;
I 19 more seriously. They get subjected to this process. They 20 end up being safer in the end, whereas the other plant 21 that didn't isn't as safe. And that's the dilemma we 22 face.
23 MR. JONES: I'm not sure I would agree with 24 your premise that the other plant has to be safe. It may
(\
(_) 25 be that what a licensee is doing is deferring maintenance NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-370) (202) 234-4433
181 1 or moving maintenance from shutdown --
f- g 2 MR. HOLAHAN.: Since he assumed that --
D 3 MEMBER CATTON: I think you ought to read the 4 transcript.
5 (Laughter.)
- 6 MR. HOLAHAN
- Well, let me say two things. l o
4 Okay?
7 The first licensee was not subject to this process.
4 8 They were offered this new opportunity.
I' 9 MR. ABBOTT: All right.
1-l 10 MR. HOLAHAN: Okay? And the second plant --
b 11 if, in fact, this improvement would be a substantial
- . 12 improvement and could be justified in some way,.the staff
)
13 might, in fact, pursue that issue on its own. Remember,
-O 14 the chart this morning said we can backfit things which 1-
~15 are substantial improvements.
I
$' 16 But my guess would be in most cases these sort i
- 17 of changes would sort of be in that middle range where-18 maybe there is some safety benefit, but I doubt that it-is
- i. -
1 19 a huge safety improvement.
l 20 MS. GILLES: O' cay . Let me talk a little bit 21 here about the acceptance guidelines. Obviously, the 22 tech. spec. changes are subject to the guidelines in the 23 general reg. guide. And in addition to that, we have 24 added additional guidelines in the tech, spec. area
]
25 because of the nature of tech. spec. changes -- change to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
, 182 1
- 1. any one allowed outage time or surveillance test interval
[ .
2 will give you a negligible increase in the types of 1
i 3 guidelines that are outlined in the general reg. guide. j 4
4 And, therefore, we have added a guideline in
{ 5 the tech. spec. area which we have presented to you j t .
! 6 previously, which is a very small change in the j s
t; 7 incremental conditional core damage probability and the
_8 incremental conditional large early release probability.
t g.
4- 9 MEMBER KRESS: That's conditional on the event ;
3' i 10 happening, or a series of events happening, or --
.!' 11 MR. WOHL: That's conditional on a piece of ,
4 12 equipment going down for a certain length of time. l I
] . qs 13 MS. GILLES: This may help.
\s / 14 MR. WOHL: And when that happens, she has I
5.
[ 15 defined the ICCDP here -- incremental conditional core 16 damage probability -- which is the difference between the i
i 17 conditional CDF and the baseline, multiplied by the time i g
] 18 of outage for the equipment.
l
{ 19 MEMBER KRESS: What do you mean by an 4 l i 20 instantaneous core damage frequency?
i
- 21 MR. WOHL: If a piece of equipment is taken A
22 out of service, the core damage frequency changes, and the 23 new core damage frequency is called the conditional core 24 damage frequency. And at that instant, it is 25 instantaneous that if that equipment is put right back NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS j 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON D.C. 20005-3701 (202) 234-4433 I
183 1 into service, then, of course, it doesn't much matter.
g~3 2 But if it's out for seven days, for a planned 4
.%.,)
3 preventive maintenance, then the instantaneous just refers 4 to a real time situatioli wiiere the equipment has come down 5 out of service. That's all.
6 CHAIRMAN APOSTOLAKIS: In reliability, I think 7 they use the word "pointwise."
8 MR. WOHL: "Pointwise" is just as good. I 9 mean, we could --
10 CRAIRMAN APOSTOLAKIS: As opposed to the 11 average unavailability where you average over time.
12 MR. WOHL: Yes.
13 CHAIRMAN APOSTOLAKIS: Do we have this?
,y 4 e
14 MS. GILLES: No, I'm sorry. This is a backup 15 slide. This definition is contained in both the reg.
16 guide and the SRP, though.
17 MR. WOHL: Yes.
18 CHAIRMAN APOSTOLAKIS: Is this the only place 19 where you have additional guidelines?
20 MS. GILLES: I believe so. I don't think 21 there are any in any of the other --
22 MR. WOHL: Yes.
23 MS. GILLES: And again, this is because of the 24 temporary nature of these types of changes.
(_) 25 MR. WOHL: It controls the risk during this I l
NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS i 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
184 1 outage which is going to be extended. And it's a prudent
, ~)
, 2 thing to do. And the number SE minus 7 that's suggested
~
3 really already exists in the tech. specs. It's made from 4 an estimate of the following type. We have a safety goal 5 level plant, a major piece of equipment goes down, and it 6 can be repaired in five hours. And if you take a factor 7 of 10, increased from 10-4 per year to 10~2, times five 8 hours -- that fraction of a year, it should be 5 over 9 8,760 in years -- you get 5 times 10-'. This is already l
10 implicit in the way corrective maintenance is done now.
l 11 And what we're saying, since we don't have a 12 breakdown in our specs. the way the Swedes do of 13 corrective and preventive maintenance type specs., we're t
\ -) 14 saying that it's all right for any kind of maintenance, 15 corrective and preventive. And it really already exists.
16 So we haven't laid an extra trip on anybody here. We've 17 just said, "All right. This is about what's in there. We 18 think it's a reasonable thing. We think you ought to 19 maintain that kind of a risk envelope in doing these 20 extended AOTs." So, in a sense, it isn't new.
21 MR. CARROLL: Would this be a place where a 22 footnote in the reg. guide, in the interest of public 23 relations so to speak, would really explain to people, 24 hey, we're not doing --
p
( ,/ 25 MR. WOHL: That's fine. The industry is NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
1 185 1 suggesting a number of 10-' It's a factor of two higher I p 2 than this, but we could certainly put some kind of t 1
\
3 footnoting in.
4 MR. CARROLL: Well, I'm just saying that I 5 think a lot of this -- if you'll just read the kind of 6 stuff that's up there, they'd say, " Gee, whiz, it's a i 7 whole new concept." From a public relations point of 8 view, if you will, you might help yourself by explaining 9 it.
10 MS. GILLES: Another option is to -- one of 11 the things we have talked about is, in the Federal 12 Reaister notice going out for public comment, to explain 13 some of these things and specifically ask for comment on ln\
V' 14 items such as this, which is going to appear in the 15 guidance document for the first time.
16 CHAIRMAN APOSTOLAKIS: Is it too late to 17 change the word " instantaneous"? Because that has caused 18 problems all along, ever since it was used for the 19 shutdown PRAs. You know, they also use it a lot there.
20 The same question always arises: what does it mean? If 21 it's not too late, I would suggest you -- "pointwise" 22 would do, or something else. I don't know. But that's 23 what they use in reliability theory to mean it is a 24 function of time.
/n ; i
( ,)
(
25 Now, isn't this -- Rick, maybe you remember. l l
i NEAL R. GROSS j COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. I (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l i
186 1 Isn't this related to one of the importance measures, g3 2 where you set the unavailability equal to one? One of
+
\s>1 3 those does it?
4 MR. WOHL: Yes. The RAW is very close --
5 CHAIRMAN APOSTOLAKIS: Which one? RAW?
6 MR. WOHL: Risk achievement worth.
7 CHAIRMAN APOSTOLAKIS: Yes, risk achievement 8 worth.
9 MR. WOHL: Right. You take a piece of 10 equipment out --
11 CHAIRMAN APOSTOLAKIS: They don't --
12 MR. WOHL: The RAW is basically the taing that 13 you multiply the baseline, whatever it happens to be, by
/ \
-' 14 to get the pointwise risk achievement worth.
15 CHAIRMAN APOSTOLAKIS: Now, these figures of 16 5 times 10'7, 5 times 10-8, they're justified somewhere?
17 MR. WOHL: The SE minus 7 exists in a NUREG/CR 18 of 1991, some work done for the tech. spec. branch at that 19 time by SAIC. And I can reference it. I don't have the 20 exact number of that NUREG here with me.
21 MEMBER POWERS: Why isn't that tied to the 22 small increment criteria in the general guidance, rather 23 than to the experience with the regulations? That is, why 24 isn't the delta CDF the same as the delta CDF that they ID
(_,/ 25 talk about in the general guidance?
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N W.
( (202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
187 1 MS. GILLES: You're saying the terminology is
('N)
+
2 different in the two --
\_/
3 MEMBER POWERS: The numerical value here of 4 the allowed change between the CDF with the equipment and 5 the CDF without the equipment. You multiply that times 6 the number of years that this thing is going to go on, and 7 you have this numerical criteria -- did it that way. And 8 I wondered why did you do it that way and not the other 9 way around, where they say, okay, you can make a change of 10 -- a small change, and for sake of argument it 's 10-5 -- in 11 CDF and work the other way around on this thing? I mean, 12 you took another number.
13 It turns out that for many of your allowed
- ,,_ i
14 outages and things like that your numbers are almost the 1 15 same, but they're not -- I mean, I can find other cases 16 where they're not almost the same. And I wondered why you 17 didn't just make it consistent, so there's just one l l
l 18 philosophical approach here instead of two.
19 MR. HOLAHAN: It seems to me the reason is 20 because you need to do two things because of the unique 21 aspect of the way tech. specs. are structured. You want 22 the total analysis -- that is, including an estimate of 23 how long outages really are and how frequently they occur 24 -- to give you the CDF, the delta CDF.
/
(..) 25 But because of their nature, tech. specs.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
186 f
1 cover individual times for equipment being out. They j
, i l7 -2 really don't cover the number of times it's used in a ;
.(
'I t
- 3 given year and that kind of activity. So they are really .
r 4 -- the way a tech, spec. is structured, we'd also like to !
'i 5 control the individual effect on risk. .,
l
'6
! 6 MEMBER POWERS: So you felt like you just had ll i
! 7 to have a different structure here because of the l I 8 necessity, r l
l-I i 9 'MR. HOLAHAN: Because we started out with a' L
10 different structure. If we would have said, "Well, I want -
l 11 to start over again writing tech. specs.," I think I'm 12 more interested in the accumulative amount of time ;
1 13 equipment is out, rather than-each individual one. And if l
.O t 14 we had tech. specs, written on that basis, then'I wouldn't 15 need this individual control either.
l
'16 MEMBER POWERS: Just a question of_a
]
17 philosophical approach, and when you start mixing 18 philosophies I get confused easily.
19 MR. WOHL: Well, in a sense, the area of tech.
20 spec. relaxation is much older than the other areas. It 21 was developed a lot earlier, and the others are relatively l
l 22 latecomers in comparison. We're talking about nine or 10 t
i 23 years of risk applied to tech. spec. relaxations and some 4
4 1 24 tightenings, and our thinking has evolved also during this 25 time. But we have always looked at relative changes in '
s, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
- . (202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
,-y ,- -
- -- n . -- -
--.-,e e
i 189 l 1
1 core damage frequency, and for outage times and, you know, l l
,3 2 the probabilities involved there. So -- :
l t ! )
N'^'/
3 MEMBER POWERS: There's a certain advantage to !
4 having precedents in the field, then. )
5 MR. WOHL: Well, I don't consider it sacred.
6 It's just that it happened that way. It happens that 7 there is higher motivation for tech. spec. relief at 8 plants than for a lot of other things, because it means 9 dollars. I mean, if you can do careful, thoughtful 10 preventive maintenance at power, you save outage time.
11 And outage time can be many, many dollars per day. So 12 that's not our concern, but it is a motivation on the 13 industry.
7g
( }
'~ / 14 CHAIRMAN APOSTOLAKIS: Now, at the last 15 meeting of the full commi.ttee actually, February, I think 16 it was Tom King or you, Gary, who showed some guidelines ;
i 17 regarding core damage frequency as a function of time.
l 18 That during outages or during some special periods it l
And I remember there was I 19 should be less than a number.
20 something, and then Tom and I started arguing about it.
l 21 MR. JONES: Yes. Yes. At the last meeting --
22 and, in fact, I think tomorrow you'll see basically the 23 same slide.
l 24 CHAIRMAN APOSTOLAKIS: Temporary changes in
,9 i (,,) 25 risk, right?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASH:NGTON, D C. 20005-3701 (202) 234-4433
190
! 1 MR.~ JONES: Right. You'll_see the same slide j l
l j ,
2 tomorrow for tomorrow's discussion. I'm just pulling it 3 out. 'But it basically says that -- [
1
)- 4 CHAIRMAN APOSTOLAKIS: It's a presentation by- ,
L 5 the'four of you? .
6 MR. JONES: Right. !
7 CHAIRMAN APOSTOLAKIS: On February 7th, f l :
j 8 page 14. .
9 MR. JONES: Well, I have tomorrow. i 10 CHAIRMAN APOSTOLAKIS: Okay.
11 MR. JONES: In my head. But anyway, yes, and ,
l 12 at that point we pointed out that these kinds of numbers !
t 13- were what was in the' tech. spec. program, and that these t'
are -- as an example, what these'mean are effectively l 14 l
15' temporary'CDF of about 10'3 per reactor year during a five-16 hour AOT period, just to give a feel'as to the relative l
l 17 risk exposure over that-kind of timeframe that we were-i l
l 18 talking about.
l * '
19 CHAIRMAN APOSTOLAKIS: But they are not the i
20 same probabilities. I mean, there you say these l
21 guidelines are stated in terms of probability for the AOT 22 period. Core damage probability, less than 5 times 10-'.
23 Now,~here you are talking about incremental conditional-24 core damage probability. Right?
25 MR. JONES: The change. ,
? l NEAL R. GROSS <
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
- - . . . - - . .~. __ _
191 1 CHAIRMAN APOSTOLAKIS: And it's also 5 times
(] 2 10-'. So which one is it? l
( , ,
V 3 MR. HOLAHAN: It's the same.
i 4 MR. JONES: It's the same. i l
5 MR. HOLAHAN: Because incremental in this case j 6 means you are adding one piece. That's the increment.
1 7 It's this one experience of having some additional amount i
8 of risk for some fixed period of time. I i
1 9 CHAIRMAN APOSTOLAKIS: No. But in the l l
l 10 definition that we just -- if you take the pointwise l l
11 conditional CDF, it seems to me that that, times the 12 duration of a single AOT, according to the earlier j i
13 presentation, is the core damage probability should be l l
U) 14 less than 5 times 10-7 Now you are saying I am ,
l 15 subtracting the baseline and then multiply by the 16 duration, and that also has to be less than 5 times 10-7 17 MR. WOODS: The baseline is only 10 percent, 18 so it makes hardly any difference, really. You're talking 19 about .9 versus 1. Inside the parentheses, you've got 10-3 20 minus 10-4 It would hardly matter if you subtracted the 21 10-4 22 CHAIRMAN APOSTOLAKIS: Well, the question is 23 conceptual, though. We should be consistent, I think, in 24 what we're saying. Whether numerically it turns out to be
()
() 25 almost the same is a separate issue. So --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N.W.
l (202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
192 1 MEMBER KRESS: Yes. I would have multiplied
(~~] 2 the instantaneous --
N.] \
3 CHAIRMAN APOSTOLAKIS: Times the AOT.
4 MEMBER KRESS: -- times the AOT. I would have 5 multiplied the CDF times the -- here.
l 6 CHAIRMAN APOSTOLAKIS: Yes, that's what we ]
7 were discussing last time. But that would be consistent 8 with everything else we've been doing, which is really 9 regulating in terms --
10 MEMBER KRESS: This seems like mixing apples ,
1 11 and oranges.
1 12 CRAIRMAN APOSTOLAKIS: -- of the core damage l
-s 13 frequency. Here we are --
I )
14 MR. WOHL: We're looking at a plant here j 15 operating in normal fashion, with normal unavailabilities l 16 and normal operation. And then we're saying, "All right.
17 I'm going to take this piece of equipment down. How much 18 risk do I have?" That's all we're saying.
19 CHAIRMAN APOSTOLAKIS: I understand. But my 20 question is: we have two different interpretations of the 21 thing in two different documents. I mean, either one will 22 do, because the principle that you just stated is correct.
23 You want to know what is happening, and it's just a matter l 24 of consistency.
('m.
(_.) 25 Either we go with the total core damage NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
193 1 frequency, or probability during that time, or we go with
,s 2 an increment. Now, whether numerically, it doesn't G matter, it's a separate issue.
3 4 MEMBER KRESS: Sure. It's not important. And 5 I would have left that baseline out of --
6 CHAIRMAN APOSTOLAKIS: I would leave it out, 7 too. I would say, you know, the instantaneous times the 8 duration, and that has to be less than 5 times 10-7 9 MEMBER KRESS: Yes.
10 MR. WOHL: Well, what the baseline does for 11 you is, normally, these mr.enines like to just keep 12 running. And it's perking along, generating megawatts, 13 and then you're changing something, and you're asking, l
)
'/ 14 what is the perturbative impact of that change? That's 15 what this is. Okay.
16 So you are going from a baseline risk level to 17 some higher risk level for some period of time which you 18 are permitting.
19 CHAIRMAN APOSTOLAKIS: You are skipping a step 20 then, because eventually you are looking at the delta CDF.
21 Right? I mean, the guidelines we have are in terms of the 22 delta CDF, and those guys are calculating it directly.
23 That's what it is, delta CDF, isn't it?
l 24 MEMBER KRESS: No, because (f 25 MR. WOHL: CDP.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
. - . - - . ~ . - - - . - . - . . ~ . . ~ . . . . ~ . .
194 i- 1 CHAIRMAN-APOSTOLAKIS: CDF.
r 2 MR. WOHL: Yes.
O 3 CHAIRMAN APOSTOLAKIS: Times the -- yes,.if 1
f l- 4 you divide. }
s !
1:
L 5 MR. WOHL: Yes. It's an added chunk of risk !
.. 6 due to the fact that I am taking this equipment l 7 deliberately out of service, or if it's broken I am fixing ,
L 8 it. .And that's a useful thing to know. 'If I'm a plant 1
!- 9 manager, and I know that my -- I have a tremendously low 10 baseline, I have worked for:five years to make this plant .
11' very safe, and now somebody wants to take an important .;
! 12 piece of equipment out of service for a PM.at power that i
!. 13 we haven't done before. I want to know what the risk i/
i $ .
L 14 increment of doing that is, and that's all this is 15 supposed to represent.
l l 16 : MEMBER CATTON: Tom, I think they agreed. But 17 aren't the units a little bit goofy?
!' 18 MR. WOHL: In what sense?
l l
! 19 MEMBER CATTON: Well, baseline CDF. Isn't 20 that'per year? l I
21 MR. WOHL: Yes. But I'm multiplying by time, !
22 du::ation.. ,
, 23 CHAIRMAN APOSTOLAKIS: The units will be j i
1 --
24 right.
25 MEMBER CATTON: So the units come out -- AOT NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS lL 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
\J
_ _,__,a - , . , , - . - . . ,. _ . .
195 1 will be fractions.
cy 2 MR. WOHL: Pure probability.
b 3 MEMBER CATTON: So you take the 4 instantaneous --
5 MR. WOHL: Yes.
6 MEMBER CATTON: Okay. Okay.
7 MR. CARROLL: It isn't allowable outage time.
8 It's the proposed outage time divided by the --
9 MR. WOHL: Yes, if it's in hours. I mean, --
10 MEMBER FONTANA: To put this into perspective, 11 10 ~ 8 is -- you can leave a containment open for .15 12 seconds. And if you put that on top of 10-4 of core damage 13 frequency, then you've got two and a half hours. Could be
!7 ,)
C' 14 wide open. It's a meaningless statement, because it shows 15 what 10-8 is. You know, it's --
16 CHAIRMAN APOSTOLAKIS: Yes. I'd like to see 17 the SAIC work that came up with these numbers.
18 MR. WOHL: Well, this is just a very general 19 estimate. We were just looking for an allowable increment 20 of risk, and they didn't go through a great detailed 21 analysis to get this. They looked at some crude industry 22 data, and they looked at the crude mean time to repair for 23 major equipment, and went from a safety goal to a factor 24 of 10 higher than safety goal, which is slightly greater C, 25 than any single piece of equipment I'm familiar with going NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234 4433 ,
1
196 f
l 1 out of service would produce, because even large diesel l f s, 2 normally would give you a factor of no more than about i 1 l 's_/'
3 seven, sometimes eight.
4 So it's a very approximate kind of thing, but 5 we were trying to get a handle on what is already 6 tolerated in risk space, and how can we use that in a more 7 constructive way and make sure that people aren't going 8 outside of some risk envelope. Because the specs., as 9 they're constructed, do not control risk. They control 10 availability, and risk is controlled in the 11 configurations. And, you know, two or three things going 12 out at once, not too much of that is covered in the tech.
,_s 13 specs., so we wanted to make sure we had a risk handle on I l I
'\ # 14 any changes we made.
l 15 MR. HOLAHAN- We will look at the definitions.
16 I'm partial to this one, but we'll look at them and be 1
17 sure that we have only, you know, one consistent set of l 18 definitions.
19 CHAIRMAN APOSTOLAKIS: Also, I don't think we 20 have settled on this number 5 times 10-' But --
21 MS. GILLES: We felt it was a place to start 22 for public comment purposes.
23 MR. WOHL: This is not interpreted to be a 24 stop sign at all. But it's recommended guidance.
(,/ 25 CHAIRMAN APOSTOLAKIS: The thing is that it's NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
I 197 ;
1 kind of new. The Commission did not set the policy on 1
l r~T 2 instantaneous, and so we have --
t )
%.J 3 MR. HOLAHAN: Given this simple equation, I 4 could plot the allowable instantaneous conditional CDF as 5 a function of outage time. And I just did a quick 6 calculation, for 3/10 of a minute I can raise it to one.
7 You've got something in there to stop that, right?
8 MR. WOHL: Yes. Yes. Our tier 3 requirement, 9 okay, which is configuration control --
10 MEMBER CATTON: Okay. Just asking.
11 MR. WOHL: -- controls that.
12 MR. HOLAHAN- And I hope no one forgot that
- 13 this is not risk based, but that they've got other issues
'- 14 involved. And nobody can meet defense in depth by having '
l 15 core damage frequency of one.
16 MEMBER CATTON: Just checking. I 17 MR. WOODS: Well, you'd raise it to one per 18 year, Dr. Catton.
19 MEMBER CATTON: Yes.
20 MR. WOODS: One per year. And if you only 21 expose yourself to that for one second, well, obviously, 22 public relations you wouldn't want to do that, but the 23 number comes out the same.
I l 24 MEMBER CATTON: Well, actually it turns out t eN
- \_s 25 that it's a little more than one.
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D.c. 20005-3701 (202) 234-4433
198 1 MR. WOODS: It's not one, period -- one. It's 7
-~ 2 one per year. You're only supposed to --
i i 3 CHAIRMAN APOSTOLAKIS: Isn't the incremental 4 conditional core damage probability, the thing you have in 5 the brackets, times the change in the duration of single 6 AOT? Is that what you mean?
7 MR. WOHL: That would be a different 8 definition. We're not defining it that way.
9 CHAIRMAN APOSTOLAKIS: Yes. But that's what 10 it should be, though, shouldn't it?
11 MR. WOHL: Well, that would be incremental 12 with respect to the existing AOT. But that isn't the way 13 -- we're just looking at --
r~Nt Y/ 14 MR. HOLAHAN: They're two different things.
15 MR. WOHL: -- what is consumed in risk space l l
16 by what is requested. That's all. You could also do the 17 change you've suggested.
18 CHAIRMAN APOSTOLAKIS: Wait a minute. The 19 whole approach we have taken so far, not just here, is 20 that change from the existing situation.
21 MEMBER SEALE: That's right.
22 CHAIRMAN APOSTOLAKIS: Right? We have a CDF 23 and we change it.
24 MR. HOLAHAN: Well, no.
f3 This is no change from
(_) 25 CHAIRMAN APOSTOLAKIS:
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
199 1 anything.
,- 2 MR. HOLAHAN: That's true. No, you
%J 3 remember --
4 CHAIRMAN APOSTOLAKIS: This is completely 5 artificial, because I was allowed to be out se mi days.
6 Now I'm allowed to go out for 14 days. What I really want 7 to worry about is these extra seven days.
8 MEMBER SEALE: Or whatever.
9 MR. HOLAHAN: Remember that our guidelines 10 include both CDF and delta CDF, not just delta. Now, this 11 is a measure of -- I think the physical analogy is impact.
12 This is force times delta T. Okay? This is the impact of 13 the outage time of the outage. It's not the impact of the
/s
! )
'/ 14 change. Okay? You could construct -- either one of those 15 could have controls on it.
16 MEMBER KRESS: You should have already had 17 that in your PRA.
18 CHAIRMAN APOSTOLAKIS: But the PRA itself 19 already has some of it. That's what I'm saying.
20 MR. HOLAHAN: But what we're saying is -- what 21 this says, the 5 times 10"' is one-half of one percent of 22 the guideline. You can think of it that way. Okay? So 23 in addition to having looked at the effect of the whole l l
t 24 change, okay, what we're saying is you don't want to have l
("N l
(,) 25 too much of the risk associated with any one given 1 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
200 1 activity all tied up at one time.
r's 2 MEMBER KRESS: That's sort of the defense in I i i %J 3 depth.
4 MR. HOLAHAN: It's a defense in depth risk 5 allocation issue. Okay?
6 MEMBER KRESS: Okay.
l i
7 MR. HOLAHAN. So it's one-half of one percent 8 of what I'm willing to live with ought to be associated 9 with things that my tech. specs. tell licensees they're 10 allowed to do. It's not an accident. It's not a piece of i
11 equipment that happens to be unreliable. This is 12 something trat is allowed to be used on a normal routine 13 recurring basis. What we're saying is we think it makes l
/ T
14 sense to control the amount of risk associated with those 15 activities, each one.
16 CHAIRMAN APOSTOLAKIS: I mean, what are you 17 saying to amend them? Do you mean there is something 18 going on now, if I make a change, I want to know by how 19 much the condition --
20 MR. WOHL: Baseline now, and I'm going to take 21 this equipment down, and how does that impact it? We're 22 not really interested in change. I'll tell you why I 23 don't like the change from the old numbers, because a lot 24 of the old numbers were not even risk informed at all.
l l
p)
(_, 25 CHAIRMAN APOSTOLAKIS: That's true.
i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON. D.C. 20005 3701 (202) 234-4433
201 1 MR. WOHL: So I don't want to compare a risk
,g
< 2 informed number with a non-risk informed --
V) 3 CHAIRMAN APOSTOLAKIS: But the baseline CDF 4 that you put in there has in it the fact that the thing is 5 already out for seven days.
6 MEMBER CATTON: Well, it shouldn't.
7 MS. GILLES: No.
8 CHAIRMAN APOSTOLAKIS: No?
9 MEMBER CATTON: If they're going to write this 10 equation, it shouldn't.
11 MEMBER KRESS: If Dana Powers has his way, 12 it --
13 CHAIRMAN APOSTOLAKIS: That's why you 14 calculate it. l l
l 15 MEMBER CATTON: If it is there, then George is l I
16 right.
17 MS. GILLES: It doesn't necessarily reflect i
18 the AOT. It reflects the expected or nominal 19 unavailability of that piece of equipment, which may or 20 may not be related to the AOT.
21 CHAIRMAN APOSTOLAKIS: Okay. So how much is 22 that nominal?
23 MR. WOHL: We don't force that change. The 24 nominal unavailability will probably change with time if a
(, 25 new AOT is granted. But up to this point in time --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
202 1 MEMBER CATTON: I just lost it. Does the
,r] 2 baseline CDF include the seven-day outage?
'G 3 MR. WOHL: It includes the nominal 4 unavailability.
5 MEMBER CATTON: Does the nominal ;
6 unavailability include the seven days? ,
1 7 MS. GILLES: No.
8 MEMBER CATTON: It's as if it runs all year 9 round and --
10 MS. GILLES: No. It's --
11 MEMBER CATTON: Well, how come everybody says 12 no?
fm 13 CHAIRMAN APOSTOLAKIS: It's the distribution i
( ")
14 issue of the outage time.
15 MEMBER CATTON: But built into that is the --
16 whatever they're allowed to do.
l 17 CHAIRMAN APOSTOLAKIS: Exactly. j l
l 18 MEMBER CATTON: So there is --
19 CHAIRMAN APOSTOLAKIS: Then is the fact you 20 cannot exceed seven days.
21 MEMBER CATTON: So if you change, it should be 22 -- a delta has to be the change.
23 MR. CARROLL: But it isn't seven days per 24 year. It may be three occasions where --
/^s I U 25 MR. ABBOTT: It's based on the performance NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4433
203 1 data of the plant.
f3 2 MS. GILLES: Right.
i \
Q) 3 MEMBER CATTON: Whatever it is, the units 4 ought to be consistent. That's the least we can do as l I
5 engineers.
I 6 MEMBER POWERS: The units are okay. ;
l 7 CHAIRMAN APOSTOLAKIS: The units are okay.
8 MEMBER CATTON: You've got to pick the right 9 values to go along with the multiplier.
10 MR. FLACK: I think what you have to do is 11 separate out corrective from preventive maintenance here.
l 12 You know they have corrective maintenance going on at a j 13 plant. What they're asking for is to do something i
(~\
\ !
\
' - ' 14 different and that's preventive maintenance. They're 15 going to take a piece of equipment out, and that's going 16 to add. So we're not double counting here, I don't think.
17 This is the question I think you raised, George, whether 18 we're double counting unavailability.
19 Yes, establishing a baseline meaning we're 20 taking out what the likelihood of the equipment being 21 unavailable i s at that time, putting in the fact that it 22 is unavailable, calculating that over a period of time, 23 and that's the piece that you're adding because they are 24 now doing preventive maintenance. But still, there is a
(_j 25 baseline that has unavailabilities due to corrective NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
204 1 maintenance which equipment can still fail randomly. So g-) 2 there is that difference between the two pieces.
LJ MS. GILLES: I guess the point we're at now is 3
4 we've been using this as an informal guideline for a 5 number of years, and we feel we're about to go out for 6 public comment. And we know that there are a lot of other 7 ideas of possible guidelines that could be used. But this 8 is the one that we have been using and we have some 9 confidence in, and we feel it's appropriate to include it 10 in the documents at this time.
11 CHAIRMAN APOSTOLAKIS: Is there a report some 12 place that derives this or explains its use?
13 MR. WOHL: It's this NUREG/CR that I mentioned
/~x >
Y 14 before that -- 1991, which I can make available.
15 CHAIRMAN APOSTOLAKIS: Thank you.
16 MR. ABBOTT: I want to get this straight. l 17 There is --
I 18 CHAIRMAN APOSTOLAKIS: Unlike other things. '
19 (Laughter.)
20 This, in particular, you want to understand.
21 Okay.
22 MR. ABBOTT: The AOT is typically is not l 23 included in the failure rate data for equipment related to 24 the technical specifications. For example, the diesel
/%
( ,) 25 generator failure rate is based on actual plant data, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 l
l
205 1 which may include taking the diesel out for a period of 2 time for maintenance or whatever. And that time, or f3 N,.,A 3 whatever it is, and how you end up getting a failure rate l
4 of that, has little or nothing to do with the length of 5 the AOT. Correct?
l 6 MS. GILLES: Yes. '
7 MR. ABBOTT: Okay. i 1
8 MR. JONES: And part of what we do in this 9 process is if a licensee proposes to change the AOT, and 10 as a result of change in the AOT chooses to change its l 11 practices related to that diesel, then, if that is an i
12 impact on the unavailability of that diesel, he has to 13 factor that into his estimate of the increase in the p_ '
(N-) 14 overall risk over the year, the delta CDF, as well as the 15 instantaneous risk which we're controlling by this 16 incremental, or pointwise as George said it. You have to l
17 look at that also; we're trying to control that element i 18 also.
1 1
19 So, in a sense, when they request an AOT, we 20 believe there is some motivation for that request. It 21 will likely have an impact on risk. We expect that to be 22 estimated consistent with the general reg. guide and its 23 expected impact on reliability of that equipment.
24 MR. ABBOTT: If he's applying for an extension C\
( ,) 25 of the AOT to do preventive maintenance, you're assuming NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
206 1 that the preventive maintenance he performs will end up gS 2 reducing the failure rate of the diesel generator at some
() 3 time in the future. 4 1
4 MR. JONES: I'm not assuming one way or the 5 other. He has to -- he is intending to do something. He 6 may be doing it for purely economic reasons. He may be l 7 doing it to improve the diesel generator reliability. I 8 don't know his motivation. He needs to assess what he's 9 going to change, why he's going to change it, and then 10 determine the impact. And if it is an increase in the 11 reliability, he can account for that in his analysis.
12 MR. ABBOTT: Okay. All right. That's all I'm
,_s 13 asking. So there is a feedback there?
/ i 14 MR. JONES: Yes.
15 MS. GILLES: Yes.
16 MR. ABBOTT: Okay. All right. That's all I'm 17 asking.
18 MR. CARROLL: Okay. I had a pending question.
19 What are the units of duration of single AOT under 20 consideration?
i 21 MEMBER KRESS: What do you want it to be? l 22 (Laughter.)
23 MR. HOLAHAN: It's the same units of time as i 24 the --
("T ,
(_,) 25 MR. WOHL: And if it's 14 days, it's 14 over NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 2344433
207 1 365. That's the number of years.
2 MEMBER FONTANA: I hate to be kind of slow
('"x) x-3 here. But take that second paragraph there on 4 instantaneous conditional core for the subject -- out of 5 service. If that's conditional to the AOT -- and let's 6 take a very simple case -- let's say you've shut down and 7 containment is wide open. Now, if the containment is wide 8 open for any period of time, then the release -- the 9 probability of the containment being open over that period 10 of time is one.
11 Now, is that what you mean by -- see, the 12 frequency there would be a little misleading in that case, 13 because if that containment is open, and the probability 7_ 4 14 of it releasing fission products in case of an accident is i
15 one, then when I do that calculation I end up with an 16 extremely tiny amount of time in which the AOT is allowed l
17 to progress.
1 18 MEMBER POWERS: Mario, the probability of a i 19 fission pre 3 tct release during a shutdown is not one.
20 MEMBER FONTANA: I know. But I am 21 oversimplifying the case. I'm just trying to understand 22 the formula.
23 MEMBER POWERS: Well, I don't think -- I am 24 sure that you can extrapolate any equation in to the point p
(s ) 25 where you get a logical conundrum.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
208 1 MEMBER FONTANA: No, I'm doing reducto ad g 2 unc.rtium just to understand the equation. Where you got
\m_/
3 LERF really isn't LERF. It's a probability of that ,
l 1
1 4 particular equipment not being functional, if you're i
5 performing the maintenance, or whatever it is, on it.
l 6 MEMBER SEALE: No.
7 MEMBER FONTANA: It's not spread over the 8 whole year, is it?
9 MEMBER SEALE: No.
10 MEMBER FONTANA: Frequency is per year.
11 MR. HOLAHAN: No, frequency is per time. You 12 can put in any time unit you like.
13 MEMBER FONTANA: Okay. I got you.
t V) 14 MEMBER SEALE: And you integrate. And 15 calculating the area of the brick on top of the column you ,
16 get if you take the outage -- the first probability, l l
17 frequency times the time interval, and when you take the 18 nominal baseline number times the outage time, you 19 calculate the part under it. In essence, what you're I 20 doing is calculating the incremental difference in that l
21 time interval because you're not in nominal expected l l
22 operation space. You're in instantaneous out-of-service 23 space. )
l i
24 MEMBER FONTANA: Okay. Now, if the (O) 25 containment happens to be wide open, is that frequency NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
209 1 one? No?
gy 2 MEMBER POWERS: No.
( )
\J 3 CHAIRMAN APOSTOLAKIS: But the point earlier 4 about the delta --
5 MEMBER POWERS: Because of some sort of damage 6 that gets you into a situation where you can release 7 fission products. Now --
8 MEMBER FONTANA: Well, that's what I was 9 asking, if that --
10 MEMBER POWERS: Presumably, initiators really 11 don't happen every time we go into shutdown.
12 ME>1BER FONTANA: I understand that. Yes.
13 MEMBER POWERS: It seems to me that the two
/~h
( )
<' 14 concepts that need to be struggled with here are the one 15 that we brought up before is, is the safety goal, and 16 consequently, do the surrogate regulations constitute --
17 or are expressing averages over a year, or, in fact, are 18 they expressing accident rates, in which case they apply 19 at every given second? That's one concept that has to be 20 accepted.
21 And the next concept that has to be accepted 22 is that there is a somewhat different criterion here for 23 these allowed outages than there is for the delta CDF for 24 a continuing operation sort of thing.
(m
() 25 MEMBER KRESS: It's a question of not being NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
210 1 stochastic.
2 MEMBER POWERS: And, you know, I become very 3 comfortable with having this different criterion for just 4 the issue of precedence. Hey, it's been used in the past.
5 Everybody is happy with it. The regulator is happy with 6 it. The applicants are happy with it. I don't see any 7 reason to change it, and especially when I run through the 8 back of the envelope the calculations for what I believe 9 to be typical new allowed outage times. I get the same 10 number. I mean, it's a coincidence of numbers, but it's 11 about the same, and so I become very comfortable with it.
12 I am much .ess comfortable with the idea that 13 CDF and LERF have to be integrated over a year's operation
\/ 14 and can't take into account swings. But I get persuaded ,
15 by Gary's argument that, hey, you've got to have an <
l 16 initiator, and initiators don't occur very often, and that I
17 inherently capture -- you know, that becomes a very i
18 difficult argument to refute.
19 CHAIRMAN APOSTOLAKIS: A PRA produces an 20 average frequency.
21 MEMBER CATTON: Over a year.
22 CHAIRMAN APOSTOLAKIS: The aliatory variations 23 with time are eliminated or averaged out.
24 MEMBER CATTON: Yes.
(-)
(, 25 CHAIRMAN APOSTOLAKIS: For example, a trivial NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
211 1- example, if you have one component that is inspected and 2 renewed every tau hours, then the unavailability.that we }
7 i
3 will use in the PRA is one-half lambda tau. The half 7 4 comes from that integration. You are averaging over time. !
t 5 So the goals, then, must address averages ;
l 6 because that's what you produce from the PRA. I think
- 7. what happens with shutdown --
8' MEMBER POWERS: I would point out that the 9 safety goals don't mention PRA. ;
10 CHAIRMAN APOSTOLAKIS: They don't mention.it, :
11 but they must. That's what I'm saying. I mean, there is t
12 no way we can say, "This refers to the pointwise," because 1
13 F..e RA does not produce pointwise frequency. What !
l 14 happens I think with shutdown, and possibly with AOTs, is 15 that there is special interest in these configurations, 16 especially during low power and shutdown, because they are 17 so different. Okay? So we are'not willing to average.
18 But even within those intervals, though, where 19 we know that the configuration is X, individual --
- 20 MEMBER CATTON: Or still average. )
i 21 CHAIRMAN APOSTOLAKIS: -- ali ntory changes 'are l
1 22 also averaged. . We are always working with averages.
i 23 Otherwise, PRA cannot be done. You can't keep track of 24 the history of each piece oE equipment in time and then do 25 the probability. So you are saying, look, if you are NEAL R. GROSS :
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
212 ,
l 1 inspecting these two trains every month -- you know, it's !
2 a staggered test, so one train on the 1st, the other train j 3 on the 15th, the average unavailability is -- and there is ;
- . 4. a. bunch ~of terms. You never look at' Monday, the 15th, you 4
5 know, and say this is the unavailability. You are l 6 averaging over time. .So all of these are averages.
~
! 7 MEMBER CATTON: So they should take the word 8 " instantaneous" out of the equation, and then it makes 9 sense.
10 CHAIRMAN APOSTOLAKIS: Well, they mean '
11- something specific. They mean under this condition -- i 12 MEMBER CATTON: Well, but-that condition is
- 13. stated. It says,."With the subject equipment out of O 14 service." l 15 CHAIRMAN APOSTOLAKIS: Okay. They may call it 16 pointwise, or something.
17 MEMBER CATTON: They don't need the word 18 " instantaneous."
19 CHAIRMAN APOSTOLAKIS: But it's not truly 20 instantaneous. It's not truly instantaneous. Everything.
21 else is average. But as long as we understand what you-22 mean. I don't have a problem.with that.
23 .I just address the first question that Dana 24 raised, that even though the safety goal statement is 25 silent on this, I mean, there is no way it cannot be NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l \
r 213 1 applied on a pointwise basis, because we don't have i ni 2 those --
i
- G l 3 MEMBER CATTON
- Not only that -- {
i 4 MEMBER POWERS: But be more idealistic, l 1
5 George, and don't accept the failings of probabilistic 6 risk assessment as an excuse.
7 CHAIRMAN APOSTOLAKIS: That's not a failing.
8 No. You are dealing with a very complex situation. And l
9 first of all, it's not clear that it would be worthwhile 10 doing it. I mean, the complexity in the calculations 11 would be astronomical.
12 MR. HOLAHAN: I'm not sure that it has any 13 statistical meaning.
/ i 14 MEMBER SEALE: That's right.
15 CHAIRMAN APOSTOLAKIS: Yes. I mean, the 16 variations will be relatively small, given the 17 uncertainties we have, and so on.
1 18 MEMBER CATTON: So why not take the word 19 " instantaneous" out of there?
20 MR. HOLAHAN: Exactly.
21 MEMBER CATTON: And hacc it mean --
22 CHAIRMAN APOSTOLAKIS: Take it out.
23 MEMBER CATTON: -- what it's supposed to?
24 MR. HOLAHAN. It's still conditional.
,r
(,)3 25 CHAIRMAN APOSTOLAKIS: Just call it NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W, (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
._. __m . . . __ _ .._. _ _.____m.. _ _ . _ _ . _ _ _ . _ _ _ _ _ _ . _ _ _ _ _ . _ _ . _ _ _ _ . . . _ _ . _ _ _ _ ,
214 1 conditional CDF with'the -- yes, sure. I will go for 2' that. ]
1 3 MEMBER CATTON: And that makes sense. Then 4 everything fits.
'5 MEMBER POWERS: Well, the next question you .
6 have to ask I think is they've set up a criterion here for i l
7 an unusual situation, equipment out of service. There are 1
8 lots of other unusual-situations that are going to come l 9 up. Is this the same criteria we're going to use for !
10 those? And if it is -- and off hand, I don't see why it' 11 should not be -- shouldn't it be elevated forward.into the 12 general guidance? .
i !
13 CHAIRMAN APOSTOLAKIS: That's why I_want to (3
h 14 know where the 5 times 10-' came from, and you promised to
)
15 send us the NUREG where this is explained, because this is 16 really an important question.
l' 17 MR. WOHL: Okay.
18 CHAIRMAN APOSTOLAKIS: Are we going to use the 19 same criterion elsewhere?
20 MR. HOLAHAN. Well, I agree with Dana, but I 21 was thinking of it 'n a little different way, which is it 22 seems to me that all of our guidelines in the application 23 specifics, and even in the more detailed ones that are in 24 the general ones, really need to derive from the 25 principles that we've articulated. And if you can't.say
- . NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
?
215 i 1 why it derived or how it derived from your principles, f3 2 then either you don't need it or you need more principles.
%./
3 Okay? !
4 Now, in my mind, this is an application of the 5 defense-in-depth principle, because what it's doing is 1
6 it's controlling the way you allocate risk. It doesn't l 7 control the total. We already have controls over CDF, j 8 delta CDF, and those things. This just tells you 9 something about the distribution that you're going to 10 allow yourself to put in place.
11 MEMBER SEALE: So what you're saying is that 12 for one request for an allowable outage extension, you 13 will allow an incremental change of one-half of one
/,,s .
) ) l
/ 14 percent.
15 MR. HOLAHAN: No. What I'm saying is each 16 time it is actually used, one-half of one percent would be 17 accrued.
18 MEMBER SEALE: All right. Now, how many times 19 can you use that in a year?
20 MR. HOLAHAN: Well, we're not controlling 21 that, but we can estimate it. Okay? And its estimate has l
22 already been folded into the delta CDF/ delta LERF 23 calculation. And if we estimated that number was too big 24 -- for example, if it were more than 200, all right, if it l')\
\ ,
25 were more than 20, you begin to get up to --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
216 1 MEMBER SEALE: You would have a 10 percent e's 2 increase in the overall --
3 MR. HOLAHAN: That's right. That's right.
4 And that is already addressed by the other delta CDF and 5 delta LERF guideline.
6 CHAIRMAN APOSTOLAKIS: But isn't also another 7 principle, Gary, that in a risk informed framework, our j l
8 risk calculation should be as realistic as possible?
9 MR. HOLAHAN: Yes.
10 CHAIRMAN APOSTOLAKIS: I mean, you guys are l l
11 saying they should reflect what's happening on the plant.
l 12 MR. HOLAHAN: Yes. l 13 CHAIRMAN APOSTOLAKIS: And I submit 7
- s i
\'- ,/
14 multiplying by the duration of a single AOT does not do 15 that. It is an artificial thing to do. So we have to 16 think about that.
17 Now, which brings me to another point. After 18 you get all of the public comments, what is the 19 involvement of this committee, and when will we have 20 another chance to come back and look at this? Can you 21 tell us what you've got?
22 MR. HOLAHAN: Oh, absolutely. The schedule 23 that we laid out in January of '96 included -- '96 --
24 included, after public comment, the staff will resolve
/ )
(_/ 25 public comment r.nd take the proposed final versions of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
217 1 those back to both'the CRGR and the ACRS before it goes to r~'
c 2 the Commission for final disposition.
3 CHAIRMAN APOSTOLAKIS: But when we will have 4 an opportunity to tell you that, you know, maybe we don't 5 like this -- too;f?
6 MR. HOLAHAN: No. Well, what you really what 7 to know is: when is your last chance.
8 CHAIRMAN APOSTOLAKIS: My last chance so.that 9 you will have time to do something-about it.
l 10 MR. HOLAHAN: Right. Well, if these documents ;
i 11 go out for comment -- l 12 CHAIRMAN APOSTOLAKIS: 90 days? !
13 MR. HOLAHAN: Well, if they go out in, let's j
(:) 14 say, April, we'll deliver them to the Commission by.
15 March 31st. But it's not clear exactly when the 16 Commission will tell us -- begin the public comment. But 17 let's just say since most of the documents stay with the 18 Commission in April -- May, June and July would be public 19 comment. Somewhere like August and September, presumably, the' staff is working very hard' figuring out what those 20 21 public comments mean and what to do with them. And then 22 in the fall, I think September / October / November, we'll be 23 back explaining to the. committee, you'know, what we 24 propose to do in the final version.
25 CHAIRMAN APOSTOLAKIS: Now, what if nobody NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
218 1 raises a question about this, but we want to raise the zm 2 question. That's when we're going to do it? Are we
/ )
'J Can we make comments during the public --
3 public?
4 MS. GILLES: Sure.
5 MR. HOLAHAN. You have to change your name.
6 Otherwise, I -- well, first of all, it isn't necessary 7 since both -- I notice Mr. Jones and I both wrote that 8 down as things we have to go back and think about.
9 CHAIRMAN APOSTOLAKIS: Okay.
10 MR. HOLAHAN. Okay?
11 CHAIRMAN APOSTOLAKIS: So what you're saying 12 is it's happening right now?
13 MR. HOLAHAN: Sure.
, ~3 14 CHAIRMAN APOSTOLAKIS: But I want to think 15 about it and read those NUREGs.
16 MR. HOLAHAN: Yes. ,
i 17 CHAIRMAN APOSTOLAKIS: It's just a matter 18 of -- j l
19 MR. WOHL: If you were to strike the CCDP, my 20 question to you is: how do you make a judgment over what 21 is safe in terms of an AOT?
22 CHAIRMAN APOSTOLAKIS: I am not striking it, 23 no. I agree with you about that. We need a measure. I'm 24 just looking for -- I'm scrutinizing every term. I am not
() 25 questioning the existence of this thing. I'm questioning NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 2344433
219 l l
1 the individual terms. l c 2 I think we all agreed -- and you can disagree t 3 with me -- that the " word" instantaneous probably has to 4 go, because " conditional" is good enough. l l
5 And my problem right now is the last 6 parenthesis -- the duration of a single AOT -- because I .
i 7 think that introduces a new practice now of doing ;
8 conservative calculations within a risk informed i
9 framework, which I don't -- ,
10 MR. HOLAHAN: But I think to address.it 11 logically what you have to do is -- start out to'say is, 12 what am I trying to achieve --
13 CHAIRMAN APOSTOLAKIS: Exactly. ;
14 MR. HOLAHAN: -- with this item?- .
I 15 CHAIRMAN APOSTOLAKIS: And I' keep asking 16 myself that question. And what I'm trying to achieve is i 17 to limit the conditional probability, but I think I should 18' have an'as realistic as possible estimate of that 19 conditional probability.
-20 MR. HOLAHAN: I think that depends on what 21 you're trying to achieve. If you are trying to limit the 22 maximum amount that you approve for any given decision, i
l 23 then this is the way you should calculate it. !
24 CHAIRMAN APOSTOLAKIS: Well, that's what I'm -l 25' saying. I want to think about it. I mean, you may have a i
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. -
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
- . _ . . _ - . - . ~ . . . ~ . . ~ ~ . . . . . . . . - . . . - . - - - . - . . . . . - . _ . . - - . -
220 1 point. I'm not --
I don't think we can resolve it in 2 three minutes, so that's why I ask you, what is the 3 process. I mean, how much time do we have to think about 1
l 4 it?
i i
5 MR. WOODS: I can tell you the process. I ;
6 have charge of the reg. guide because I put together. And i i
7 when I get back to my office this afternoon, I'm going to 8 take the word " instantaneous" out. And unless somebody i
9 puts it back in, it's going to stay out.
10 (Laughter.) ;
11 ' CHAIRMAN APOSTOLAKIS: That takes care of one l 12 of them.
13 The. question is: should it be delta duration-14 or duration? That's really the issue.
.15 MR. WOODS: This thing has been' discussed 16 several times, and each time one particular word -- and 17 it's been different words at different discussions. Each 18 word has come into question.
19 MR. HOLAHAN: We discussed it a little last l
20 time at the full committee.
21 But I think, George, your comment is, should i
22 it be maximum. duration allowed of the AOT,'or should it 23 be --
24 CHAIRMAN APOSTOLAKIS: The change.
25 MR. HOLAHAN: -
the typical one used? Or NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4 433' WASHINGTON, D.C. 20005-3701 '(202) 234-4433
221 1 should it be the change?
o 2 CHAIRMAN AI'OSTOLAKIS: I think it should be i \
't.'
3 the change. Based on what I know now, I think it should 4 be the change.
5 MR. HOLAHAN: But each of those has a 6 different meaning.
7 CHATRMAN APOSTOLAKIS: And the last one --
8 M. HOLAHAN: And controls a different -- and 9 each answers a different question.
10 CHAIRMAN APOSTOLAKIS: I know. I know.
11 That's why --
12 MR. WOHL: We also look at other numbers. We 13 also look at estimated time of usage given that so and so i
)
d 14 would be the AOT. And maybe that's a good mean value to 15 use, if it turns out to be all right.
16 CHAIRMAN APOST0LAKIS: Coming back to the 17 earlier comment, the baseline CDF already has some, outage 18 for these. Not the allowed outage, but the estimated --
19 MS. GILLES: Some unavailability.
20 CHAIRMAN APOSTOLAKIS: -- which is influenced 21 by the allowed time. So there may be some double counting 22 here, but I haven't thought about it. I don't want to 23 insist right now on this position. I want to think about 24 it.
,~ l
) 25 MR. ABBOTT: I mean, in an ideal world, you NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS j 1323 RHODE ISLAND AVE., N.W. ;
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433 !
1
222 1 wouldn't have ACTS. But we have to deal with them. And I r~N 2 guess this is an artifact of that. I mean, in an ideal la3 3 world, if you had e quality risk assessment, and you were 4 -- you know, you were --
5 CRAIRMAN APOSTOLAKIS: The Platonic world.
6 MR. ABBOTT: Right. Well, you know what I 7 mean. I mean, you would determine the length of time that 8 you would take equipment out of service based on the 9 increased risk you would incur in accordance with some 10 process which would take account for that increase in 11 limiting it in some way. You wouldn't have an AOT.
12 MS. GILLES: Right. In an ideal world.
13 MR. ABBOTT: Yes. Now, is there any thoughts 7,
(
) 14 on the part of the staff that go into that sort of concept 15 in the event that someone proposes it in the future? Or 16 is that impossible?
17 MS. GILLES: I don't know that it's 18 impossible. I think we're a long way from there right 19 now, and I think this is just the first step maybe toward 20 that.
21 MR. WOHL: Even the four-train British plants 22 have AOTs, and Hasham 2 has a risk monitor. But they also 23 have a lot of 30-day AOTs, so they still --
24 MEMBER KRESS: But let me throw something at n
)
\. / 25 anybody. Let's sort of divorce this and look at CDF NEAL R. GROSS l
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4 633
223 1 that's due to a number of accident sequences. And I want 2 to control risk with my regulatory process so that it's 73 t 4 G below acceptable values. One thing I could do is say each 3
4 sequence should not contribute more than a certain amount 5 of risk, which is kind of a defense in depth concept.
6 However, when I make that kind of incremental 7 risk partitioning among sequences, I can't give a number 8 to it, because I don't know how many sequences there are 9 going to be for a given plant or for a given set of 10 conditions. So what I do is I say each of these sequences 11 can only add a certain percent, and that percent is the 12 total divided by the number, or something like that, is 13 the way one generally handles that.
O r I
/ 14 Now, here we have an analogous situation, 15 really. You are trying to take an incremental part of the 16 risk and put a number on it, and the thing that seems to 17 be missing to me from that is you are putting a number on 18 it, which I don't think you can really do unless you know 19 something about how many of these are going to go on, how 20 long they're going to be, and how they're related to the 21 total overall risk for given plants.
22 And that's what is I think causing us to be 23 confused. You're mixing a method of controlling based on 24 an incremental risk compared to an overall risk acceptance I
r~N
( ,) 25 criteria. And if you could straighten that out somehow, NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 224 1
1 that confusion out, it would help me a lot I know.
s 2 Did I muddy the water up with that a lot, I )
LJ 3 or --
4 MEMBER CATTON: You're kind of afraid they're 5 going to do that once a month maybe.
6 MEMBER KRESS: Well, I don't know what -- they 7 may very well.
8 MEMBER CATTON: -- without correcting the 9 equation to account for it.
10 MR. HOLAHAN: They may very well.
11 MR. WOHL: The maintenance rule has some 12 control over that.
13 MEMBER KRESS: Yes. Well, there are other
/
14 things in this consideration.
15 MEMBER CATTON: If they had a risk meter and 16 kept a running year average, they'd have what you need.
17 MEMBER KRESS: Perhaps, yes.
18 MEMBER SEALE: Well, let me ask another 19 question related to that. Let's suppose we have a 20 plant --
21 MEMBER KRESS: It would be easy enough to do.
22 MEMBER SEALE: -- that has a core damage 23 frequency of 104 And they have a series of these kinds 24 of outages that they want to do, and the sum of all of the
(_) 25 incremental risks associated with that series of outages NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON. D C. 20005-3701 (202) 234-4433
225
! 1 -- they still meet your individual outage ground rule, by l
l ,s 2 the way, of one-half of one percent. But let's say they i
% j) 3 wanted to do enough of them to raise the total yearly core 4 damage frequency of 5 times 10-4 5 So, in essence, there would be 1 times 10-5, 6 which would be due to the normal operating, and 4 times 7 10'S due to these particular things, these kinds of 8 operations.
9 MEMBER CATTON: Somebody is shaking their head 10 no.
11 MS. GILLES. Keep in mind, we still have to 12 meet the criteria in the general reg. guide.
13 MEMBER SEALE: But still, that's 10'5
? )
\-' 14 MEMBER KRESS: This is in addition to that.
15 MS. GILLES: In addition to that, right.
16 MEMBER SEALE: Yes, but it's still less than 17 10-4 l
18 MR. HOLAHAN: But the delta is bigger than 19 10'5 I
20 MR. MARTIN: If I could, let me address that.
21 MEMBER KRESS: Identify yourself, please.
22 MR. MARTIN: Lawrence Martin, South Texas 23 Project.
24 We really worked through this, and I don't
(,) 25 know where the staff is today, but I know where they were NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
- . . ~ . - . - . . . . . - . - - . - - - .-. .-......- -. . . _ . _ . - . . . . . . .. ..
226 1 on. Halloween Eve. Okay? j I
2 (Laughter.) l D When we started developing our ITS with the 3
'{
4 risk informed ITS, we originally proposed not putting AOTs 5 in there. We really wanted to go with PRA numbers, and we {
6 looked at each individual outage and what that l 6
7 contribution was, and then not just.that one but any I
-8 others that were going on at that time for where we are in 9 risk, and then cumulative over the total thing. Okay?
l 10 We were advised by the staff, and Bill Russell ;
i
.11 in particular, okay, that that was not an acceptable :
1 12 approach, that the staff was not" ready to deal with that l 13 at that time. Okay? So we went-back and we agreed we'd i 14 put AOTs in there, or backstops, if you wanted, okay, and 1 15 we'll discuss the diesel generator one. Okay?
16 . We came out at a crossover. point on 10~6 at 17 about 17.8 days. We were advised by the staff it was 18 really maybe a little bit more comfortable if we kept that 19 in increments of 21 days. When we took a look at what the 20 extra three or four days was, we went with the 21-day, 21 one-time exemption each fuel cycle, okay, that we could go l
22 into a 21-day AOT on a diesel.
i 23 MR. CARROLL: Fuel cycle being? l I
24 MR. MARTIN: 18 months. Okay? And that was j O
V 25 an exemption that we were going in for.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
227 j 1 As we went through that, when we really
,S 2 started out with this tech. spec. submittal a couple of l
!w,- /
3 years ago, it was taking us 10 to 17 days to do our diesel 4 outages, because of the scope of the outages and the 5 proficiency we had in them. ,
I 6 As we went through this tech. spec. change, j 7 though, low and behold it came up that we no longer really 8 needed 21 days, or even 17 days, because our maintenance 9 -- the scope of it had changed, and our ability to perform 10 that significantly more effectively and efficiently. And 31 we can now do all of that maintenance in under 14 days.
12 So what had came back to us is we resubmitted
,_ 13 it as a 21-day, or started to, and we were advised that it 14 was probably better to go for 14, because the staff i
15 advised us at that time that, "None of your maintenance is !
l 16 taking more than 14 days, so why should we grant more than j 17 14 days?" Okay?
18 We really wanted to operate in the CDF and I 1
19 LERF and take the cumulative of where we were on a given 20 day, and the cumulative over a year, but they just really l 21 weren't ready to deal with that, okay? So the ideas that 22 you're proposing is a good thing; we just couldn't quite 23 get there at that time. And that may be the next 24 incremental step.
(N
(._., I 25 MR. CARROLL: What did Bill Russell offer as 4
l NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
228 1 the problem that this created? Well, that's not it.
2 Gary, you --
(n)
V 3 MR. MARTIN: The problem was that he was not 4 going to approve a tech spec that had --
5 MEMBER POWERS: Factor that into your PRA.
6 MR. HOLAHAN: Well, I think it was a bit too 7 rrevolutionary for an evolving technology.
8 MEMBER KRESS: Not ready to do it.
9 MR. WOHL: I think that Mr. Martin's stars are 10 interesting in that they have had a configuration risk 11 management program, a very thorough natured plan, all 12 their maintenance for quite a while, at least the three 13 years that I'm familiar with. And that may have fed back
( !
k/ 14 into making the decent maintenance more efficient. I l
15 don't know that to be the case. ]
16 MR. HOLAHAN: But back on this point of why 17 we're deciding what we're deciding, in the long term 18 there's no reason why the agency can't go to a more 19 risk-informed configuration control approach, which means 20 making decisions based on their effects on core damage 21 frequency, large early release, and maybe figuring out the 22 implications with respect to safety margins and l 23 defense-in-depth and the whole arrangement.
24 That's probably not where you want to start.
C
(,_,\/ 25 I think that's a big enough change so that you want to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4 433
I 229 l l
1 practice on other things first. I think that to that ;
l 2 extent, Bill Russell was right, not to say this is not the 7-
\)
3 promised land, but I think we have a few more simple 4 things to work through before we get to that point.
5 MR. ABBOTT: Let me make the following point.
6 Under the South Texas process that did not get approved, 7 there would be a clamp on doing diesel generator testing 8 by stacking up -- I mean, diesel generator maintenance.
9 Let me back up.
10 Say I've got a bunch of maintenance to do.
11 It's going to take me longer than the 14 or 17 days that I 12 need. It's going to take me three months. So I take the 13 diesel out of service. And I do the first four PMs in O
14 that time period. I declare it operable. I wait a day. ,
l 15 I take it out of service. And I do the next four. I wait l
16 a day, j 17 So, I mean, under the technical specifications 18 as they're currently written, that's permitted; whereas, 19 under his construct, it wouldn't be.
20 MR. MARTIN: It isn't under the maintenance 21 rule limits.
22 MR. HOLAHAN: And, in fact, this is not 23 entirely hypothetical. I think the staff has gone back 24 and did some analysis to look at the relationship of n
( ,/ 25 plants with short allowable outage times and longer NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
l 230 ;
}
i l- 1 allowable outage times and see which ones had the most ;
i 2 total reliable equipment. 1
- 3. In fact, there's no correlation. That is, l i
- 4 having short individual AOTs does not guarantee or even l 5 strongly influence reducing the unavailability over the j
' i 6 course of a year.
7 MR. ABBOTT: So it's a non-problem. j l
8 MR, HOLAHAN: Well, what it says is so long as l 9 you have tech specs as your structure for controlling l i.
'I
. 10 configurations, you need to put some numbers in it. Okay? l 11~ Now, in the stage that we are of development, 12 what we want to do'is we want to put in reasonable numbers l 13 based on risk insights and reasonable controls associated j 14 with it.
15 But I think we have to recognize that the ,
t l 16 maintenance rule and other things are really what's' ;
)
( 17 influencing the overall effect. And, in fact, in other i 1
- 18 cases, South Texas has volunteered to do some things, but 19 there are other plants who haven't had their tech spec 20 changed.
21 And probably the most significant thing about
- 22 doing maintenance is not the fact that -- the tech specs
, 23 but the combination of activities going on. And so with i !
i 24 either the maintenance rule or a configuration control-or l 25 a risk meter or whatever, however the licensee is !
(
i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
- , -m ,
231 1 approaching this issue from all the analyses that we've l fx 2 done, this is more important than whether it's 7 days or i i K.)
3 14 days.
4 MR. CARROLL: On the subject this is 5 discussing, I find in these documents a lot of discussion 6 of configuration control. And, of course, that's a key 7 element of the maintenance rule.
8 Why do you need to talk in detail about 9 configuration management in these SRPs and reg guides?
10 Why don't you just say, "Thou shalt follow the maintenance 11 rule," period?
12 MR. HOLAHAN: Well, in fact, I think in some 13 place we say, "You can use the maintenance rule if it 7_
14 suits the situation that you're proposing."
15 But let's face it. The maintenance rule 16 doesn't say that you're going to verify or validate the 17 assumptions done in any one of these analyaea. The 18 maintenance rule is very general in what it requires 19 licensees to do. And in some cases we might need a 20 tighter connection between what needs to be verified and 21 how that's being monitored. It's not clear that the 22 maintenance rule covers all those situations.
23 MR. CARROLL: I hope we're not getting into 24 another situation where we have three or four rules, all O
(.s/ 25 of which you have to follow. We had that with license NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
232 l
1 renewal and the maintenance rule at one point 2 operationally.
v 3 MR. HOLAHAN: Well, the maintenance rule 4 doesn't guarantee the assumptions in these PRAs. Okay?
5 However, if a licensee were to put into effect some level 6 of monitoring like we address in here, I think that level 7 of monitoring certainly meets the maintenance rule.
8 So you don't have to do two things. But your 9 implementation of the maintenance rule might be more than 10 it would have been before.
11 MR. CARROLL: Did that say the maintenance 12 rule was effective at the time it was promulgated?
,_s 13 MR. HOLAHAN: The question is: What were they i )
'# 14 trying to achieve? It wasn't the risk configuration 15 control rule. To some certain extent, we're trying to 16 make it do things that it only dealt with peripherally.
17 CHAIRMAN APOSTOLAKIS: You have 15 minutes.
18 MR. WOHL: Paragraph (a) (3) in the maintenance 19 rule says you have to be able to determine the impact on j 20 safety function of equipment out of service. And we feel i
21 that our Tier 2 and Tier 3 really deal with that more 22 explicitly than the rules stated.
l 23 MS. GILLES: I'll quickly finish up the reg 24 guide and SRP so we can get to some of the pilot results.
~
(_,) 25 I know you're interested in that.
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
233 1 MR. CARROLL: So, again, here's the case where 77 2 you could say to people who are reading this document, o >
C/
3 " Hey, if you're really doing a good job of implementing 4 the maintenance rule by doing these things, this isn't 5 anything new," more PR.
6 MR. HOLAHAN: I think we say something like 7 that, not quite in those words.
8 MR. CARROLL: Okay.
9 (Slide) 10 MS. GILLES: These are just the acceptance 11 guidelines for the Tier 2 and Tier 3. And it's 12 demonstrate appropriate restrictions on dominant 13 risk-significant configurations and implement a 14 risk-informed plant configuration control program and 15 procedures for that program. And, again, as we said, that 16 could be your maintenance rule program provided it fits 17 the bill.
18 (Slidei 19 MS. GILLES: And the Element 3 in Principle 20 Number 5 had to do with performance-based implementation 21 and monitoring strategies. And for tech specs, those are 22 basically the three-tiered implementation approach. And 23 we do say that the maintenance rule monitoring would 24 suffice for monitoring effects of tech spec changes.
n C) 25 (Slide)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS
- 1323 RHODE ISLAND AVE., N W.
I (202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234 4433 l
234 1 MS. GILLES: The last element in the process 1 x 2 is just to submit the proposed change. And there is a
- )
\ __/
3 list of the types of things you could expect to find in 4 the risk-informed tech spec submittal. And those are 5 typically the same types of things you would find in any 6 license amendment submittal in addition to the additional 7 analysis you would need to do for a risk-informed 8 submittal in the PRA information in the process to meet 9 the three-tiered approach.
10 (Slide) 11 MS. GILLES: A little bit of information on 12 the pilot plant activities. I mentioned the industry 13 submittals were CE Owner's Group joint application reports
/ \
> 14 for three systems: safety injection tank, low-pressure 15 safety injection system, and the emergency diesel 16 generators. And they requested allowed outage time 17 extensions for all three of those submittals.
18 In addition, this should now say, " License 19 amendment requests from nine of the ten CE plants" to 20 adopt one or more of those joint application reports and 21 --
22 CHAIRMAN APOSTOLAKIS: I would like to see the l
l 23 documents of the first three bullets if possible. Do we 24 have those, Mike.
rm
_) 25 MS. GILLES: We provided them to the Committee NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISt.AND AVE., N W l
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
235 1 about six months ago, I believe.
l fx 2 CHAIRMAN APOSTOLAKIS: Okay.
- )
L.)
3 MS. GILLES: If you cannot locate those, 4 please let us know. We'll get them to you.
5 (Slide) 6 MS. GILLES: The interactions we've -
7 MR. CARROLL: Why nine of ten? Who's the 8 holdout?
9 MS. GILLES: Palisades. I believe at the time 10 they had some other large design modifications going on 11 that they were -- it's a resource --
12 MR. CARROLL: It isn't their disagreement with 13 the other --
,r N i )
N/ 14 MS. GILLES: No. They provided information 15 for the joint application reports. They just have not yet 16 submitted any amendment requests.
17 MR. HOLAHAN: I believe they were thinking 18 about the reactor vessel at the time.
19 MR. CARROLL: Could be, yes.
20 MEMBER SEALE: That's probably a little 21 preoccupying.
22 MS. GILLES: We've had three requests for 23 additional information go to the owner's group. We 24 visited three of the pilot plants. We've probably had at p least half a dozen meetings with the CE Owner's Group, PSA 25 l ()
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
236 1 working group. And they have made one presentation to
,q 2 this Subcommittee.
N.Y 3 So there have been extensive icteractions over 4 a period of almost two years now.
5 CHAIRMAN APOSTOLAKIS: Can we see those 6 requests? I mean, what you --
7 MS. GILLES: Certainly, certainly.
8 CHAIRMAN APOSTOLAKIS: I'd like to see them.
9 MR. CARROLL: Before tomorrow?
10 CHAIRMAN APOSTOLAKIS: No.
11 MS. GILLES: I can get them to you tomorrow.
12 I have them.
13 CHAIRMAN APOSTOLAKIS: Oh, you can get them,
( )
'd 14 of course, but I'm not sure I can look at them. I mean, 15 are we talking about large documents or --
o MS. GILLES: No.
17 CHAIRMAN APOSTOLAKIS: Oh, okay.
18 MS. GILLES: You want to see the responses, 19 the questions and the responses?
20 MEMBER BARTON: What you requested for this --
21 (Slide) 22 MS. GILLES: This is just some idea of really 23 how the issues of defense-in-depth and safety margins were l 24 approached. And, really, you could say both of these were ry (j 25 tackled through the three-tiered implementation process.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
237 1 The restrictions on the simultaneous equipment outages, r~T 2 and the plant configuration control program really take
() 3 care of almost all the issues under defense-in-depth and 4 safety.
5 CHAIRMAN APOSTOLAKIS: So the real meat is 6 your next viewgraph? Negligible ICLERP?
7 (Slide) 8 MS. GILLES: And this is just the lead pilot 9 plant, which is the only one we have totally completed the 10 review on and are currently writing the safety evaluation.
11 As we mentioned, for the two systems we 1csok 12 at, safety injection tanks and low-pressure safety
,_ 13 injection system, there was almost no discernible change i I 14 in the CDF or the LERF.
15 So, really, the tech spec criterion was the 16 one that we focused on. And we came out in a range of 17 2.3E-8 to 2.9E-7 for the incremental conditional core i 18 damage probability.
19 CHAIRMAN APOSTOLAKIS- So you made a decision 20 with that?
21 MS. GILLES: Yes, to grant those.
22 CHAIRMAN APOSTOLAKIS: Oh, great.
23 MR. CARROLL: Now, if this request had come in 24 ten years ago, when you first started doing those r
I
(_) 25 risk-informed tech specs, can you give me some sense of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 238 1 what the amount of effort on your part and on the part of !
73 2 the utility would have been then, as compared to what it 4
.L_) I 3 is now? 1 4 MR. WOHL: I can tell you what we would have l 5 done ten years ago is look at the change in the average ,
6 annual CDF, which is almost nothing here, and ask for 7 proposed changes in unavailability due to additional 8 maintenance that could be done at power, how this would 9 affect plant operations and unavailability, but it would 10 have been an easy call.
11 I think in a sense, these two systems could be l 12 subjected to a deterministic call, even today, because 13 they really represent only 1.6 and 1.2 percent of the
- 14 total CDF, respectively. I 15 So they're low-risk contributors. So that's 16 why you have the very low CCDP numbers.
17 MS. GILLES: And they were chosen for that 18 reason by the owner's group because they thought they 19 would be easy candidates for approval.
20 MR. CARROLL: Now, when you did it in the 21 relatively recent time, how would you compare the effort 22 level on your part and on the part of the -- order of l
i 23 magnitude more work or --
24 MR. WOHL: This is a tremendously larger
/^N
(_) 25 effort because it's a pilot effort, because we really dug NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 l
._ m,. . . . _ . _ _ _ _ . . _ . _ . . . _ . _ _ _ _ _ _ _ - . _ ._ _ _ _ . _ . _ _
l 239 ]
i ' into every aspect of the work presented to us ' and hoping l
that because these CE plants are a very similar design 2
L 3- with respect to these systems and particularly that.
t ,
I 4 getting a pilot approved for these systems would'be not I J
I 5 too ' difficult.
1 6 So'we had an extensive contractor effort. We 7 had an extensive in-house effort. But this is the first j 8 of a-kind, as I say. So - .
i 9 MR. CARROLL: Now the second of a kind.
10 MR. WOHL: The second of a kind. Now we are ,
11 developing methods for cross-comparison of the results of I
12 the. pilot with the other plant results. And-we expect to l l :
i 13 have that. developed inside of six to.eight weeks from now- )
(d i
! 14- and be able -- to see whether we can ' e:ttrapolate the pilot '
l 15 results to the other plants. and if some don't pass ,
i 16 muster, there would have to be some good reasons for that.
17 We're not trying to guarantee in these group 18 submittals that on the basis of_ pilot results we can h 19 approve everything requested, but we're trying to save our l
l 20 own people power because we have limited amounts of that.
l-21- And this seems an efficient way to go.
l-l' j
22 MR. CARROLL: Okay. Now combustion comes in l
l- 23 three months from now and says, "Okay. Now we want to do
- 24 the same kind of thing on" --
25' MS. GILLES: HPSI. That's next,
- i. .
I l- NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
240 l
1 MR. WOHL: Yes.
l f-'y 2 MR. CARROLL: Okay. Do you think that effort t 1 V
! 3 is going to be much less than what you've put into this?
4 MR. WOHL: I think because of what we learned 5 here, it will be substantially less but of the same 6 general scope. When I say "substantially less," we had to 7 find out things like Excel software is terrific for 8 looking at cut sets. And we had to stumble around to find 9 that out in the first place. And we don't have to stumble 10 anymore. So we have certain lessons learned.
11 MR. FLACK: John Flack, NRR.
12 We have to remember that we're looking at SITS 13 and LPSIs and LOCAs. So it's not as expansive as some of
/,,__)
' 14 the other submittals that were involved, like diesels and 15 station blackout.
16 So there is difference of levels depending on 17 how much we need to look at. These were pretty 18 straightforward because the cut sets were somewhat 19 limiting to an extent, larga-braak LOCAs and this sort of 20 thing.
21 MR. CARROLL: So the same is true of HPSI.
22 MR. FLACK: Well, I don't know if it's that 23 limiting. I agree that it's a different set of accidents.
l l 24 You have a broader range of LOCAs. You have steam i
( ( ,/
25 generator tube ruptures. You have a lot of other things NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
241 l
1 that you may not have to worry about. So we can't rec.lly l
i 2 generalize at this point.
'V ])
7
- 3 But the comment that George was saying earlier 4 is well-taken. I mean, there's a lot we're learning as we 5 go that we can capital
- .ze on. We're not starting from 6 square one every time we do them, but recognize that some 7 of these are unique going in, that we have to also be a 8 little more careful about it the first time through.
9 CHAIRMAN APOSTOLAKIS: If you put back in your 10 Slide 8?
11 (Slide) 12 CHAIRMAN APOSTOLAKIS: Coming back to Mr.
13 Carroll's question, would you have -- how much of this
(' '
)
14 that you have already done do you think that you will be 15 able to use in the next application? j 16 MS. GILLES: Well, that's a good question. In j l
17 the area of PRA quality, one of the things that the 18 owner's group has dona 4s this cross-comparison process.
19 And we are hoping that we e going to be able to use that 20 in all of the submittals because they have carried it 21 through from one system to the next.
22 CHAIRMAN APOSTOLAKIS: So they have a PRA for 23 transition node?
24 MS. GILLES: They have a method for cualyzing
,a \
's / 25 transition risk. They included the very beginnings of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
242 I
They have sense done even
~
'1 that'in.this first submittal. !
2 more work, which they will be including in future 1 -
l-l 3 submittals.
I i 4 CHAIRMAN APOSTOLAKIS: 'I see. But you have i 5 not seen it? ,
i 6 MS. GILLES: Not the new work. !
7 CHAIRMAN APOSTOLAKIS: By the way, is that :
8 available to us?
y 9 MR. WOHL: They have developed a full set of 10 standards, which is under discussion right now for release ]
11 to us. So we have seen bits and pieces in discussion.
1 12 But the owner's group actually has a set of l l
~
13 standards for each application that are not presently I
14 docketed.
15 CHAIRMAN APOSTOLAKIS: I see.
l But some of the members think it
~
l 16 MR. WOHL: 1 17 would be a good idea for them to. So they'have a very 18 organized, unified activity going. ,
i 19 CHAIRMAN APOSTOLAKIS: What we had, Mike, is 20 the comparison but not the actual PRAs. And those were l
21' available to you, how these numbers were derived? I mean, ;
22 when the owner's group came here to make a presentation, i
23 they gave us a document that says, "This is the risk you
- 24 transition. This is the risk shutdown." And then they 25 work with that.
?.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 I
243 1 The question is: How were these numbers j 2 derived? Did you have access to the actual PRAs? Because 7x e
G 3 otherwise how could you check on PRA quality?
4 MR. WOHL: For the transition and shutdown 5 risk, we didn't have access. We did not include those as 6 part of the judgment being made here either. That's still 7 under development.
8 Some of the people that have the safety 9 monitoring, like San Onofre, of course, is developing a 10 full-scale shutdown risk model with 15 plant operational 11 states, the same was done in our shutdown risk studies by 12 the Commission.
13 And others haven't developed models. So they
/ s
'- 14 use approximate spreadsheet approaches. So we're not -- l 15 CHAIRMAN APOSTOLAKIS: It's interesting that 16 you say that you didn't use those because I think their l 17 argument was that, even though there is an increase in 18 risk during power, there is a significant decrease during 19 the transition and shutdown.
20 So the total, then, was a decrease in risk.
21 But your criteria for making a decision is different.
22 You're just looking at the incremental.
23 MR. WOHL: By averting the transition and 24 shutdown risks, you have a substantial risk aversion. I 7m
( ,) 25 am not in a position to say since we didn't study these NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D.C. 20005-3701 (202) 234-4433 1
244 l l
1 analyses thoroughly whether they're totally credible at f- 2 this point in time.
(s) 3 CHAIRMAN APOSTOLAKIS: I understand.
4 MR. WOHL: But certain shutdown modes can have l 5 very high risk contribution, others not so high. And when 6 the shutdown rule is in effect, all of that will change 7 radically, too, because the amount of power available on 8 site during shutdown will be larger than we now require.
9 MR. CARRULL: No. The regulations will ,
1 10 require it to be what it is. Your shutdown rule is behind 11 the times.
12 MR. WOHL: Well, I'm not an expert on shutdown 13 rule. I am quoting secondhand what was told to me about 14 what the intent of the reg guides for the shutdown --
15 CHAIRMAN APOSTOLAKIS: But as I recall, that's 16 an interesting detail.
17 MR. WOHL: Yes.
l 18 CHAIRMAN APOSTOLAKIS: They argued that they 19 should get a requested extension on the basis of the total i
20 risk reduction, which included transition and shutdown.
i 21 And you granted it on the basis for something else, your 1
22 criteria that you presented today.
23 MS. GILLES: We felt that the transition and 24 shutdown risk analyses were somewhat crude and not
,n
( ,) 25 comparable to the analyses done at power, not directly !
i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
1 245 1 comparable, because of the quality and scope of the f3 2 analysis.
I i ,
' l 3 And so I believe that we felt it was 4 appropriate to make a qualitative statement that we 5 believe there was risk aversion, that there was some risk 6 aversion, due to avoiding the transition and the shutdown )
7 risk but that we didn't feel that the quantification of 8 that was appropriate for this analysis for this --
9 CHAIRMAN APOSTOLAKIS: But it's interesting 10 that you granted it for a different reason. You granted 11 the request but for a different reason. They satisfied i
12 the criteria you presented to us today. j 13 MR. WOHL: Yes.
(~ s
- 14 MS. GILLES: Yes.
15 CHAIRMAN APOSTOLAKIS: But they didn't argue 16 that way. They argued in terms of a total --
17 MS. GILLES: They argued. Oh, yes, they did 18 both. They presented this information, and they also 19 presented --
20 CHAIRMAN APOSTOLAKIS: Even though the quality 21 of those two other modes of the PRA for the two other 22 modes was not the same as power, I think the argument that
! 23 they make that the total risk is going down is a very 24 important one in terms of the concerns that have been
,f%.
( ,) 25 raised regarding this agency allowing the licensees to NEAL R. GROSS
- COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 2J005-3701 (202) 234-4433
246 1 increase the risk to the public because that may not be
,o 2 true, in fact, if you look at the total, the whole
.Y 3 picture.
4 MS. GILLES: It is important. And we are 5 hoping that in some of the future submittals we will be 6 able to see some of the more detailed analysis of 7 transition risk because there hasn't been a lot of work 8 done there. And it could be important to a lot of areas, 9 more than just tech specs.
10 CHAIRMAN APOSTOLAKIS: Now it's a1 ready past 11 3:00 o' clock. Any questions?
12 MEMBER SEALE: Could I ask a question?
13 CHAIRMAN APOSTOLAKIS: Sure, sure.
r 14 MEMBER SEALE: I apologize for not being here 15 when you started. But did you hear anything? Can you 16 tell us whether or not there is a series of additional 17 requests along the lines of tech specs that are waiting in 1
2 18 the wings?
l l
19 MS. GILLES: Yes, there 3.s.
20 MR. WOHL: Yes.
21 MS. GILLES: The CE Owner's Group has several 22 lined up that they have told us about. And both l 23 Westinghouse and the B&W Owner's Groups have told us that 24 they plan to follow with similar systems after the first (3
'\._,/ 25 CE amendments are approved.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
247 1 MEMBER SEALE: And these are generally rm 2 PRA-based? ,
l v I
? '
3 MS. GILLES: Yes, generally following the same !
4 process that the CE Owner's Group follows.
5 MEMBER SEALE: It sounds to me like PRA is 6 alive and well.
7 CHAIRMAN APOSTOLAKIS: And here to say.
8 MS. GILLES: Could sure use a shot in the arm 9 right at this moment, though.
10 MEMBER SEALE: As opposed to what we heard 11 last time.
12 CHAIRMAN APOSTOLAKIS: Well, in light of this,
- 13 though, in light of the fact that you have already closed
( )
'"' 14 one pilot, you made the decision --
15 MS. GILLES: Haven't gotten it past the 16 Commission yet.
17 CHAIRMAN APOSTOLAKIS: Well, still, though, 18 it's a major thing. I really don't understand why the 19 industry has such a negative view towards this whole 20 effort. I just don't. I mean, comparing now what I heard 21 today with what Mr. Pietrangelo told us last time he came, 22 I'm really at a loss. Does anybody understand --
23 MEMBER BARTON: I think it's because the l 24 industry has got a method to deal with the Commission that
(
V 25 they feel comfortablc with. And this is something new NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
248 1 that they look at as additional stuff that I've got to do
,3 2 to get license changes. I think that's a big piece of it.
( )
3 CHAIRMAN APOSTOLAKIS: So it's the newness of 4 it.
5 MEMBER BARTON: I think that's a part of it.
6 MR. ABBOTT: I think reading the documents and 7 even viewing the slides, which are available to the 8 public, do convey the feeling that there's a substantial 9 extra burden being placed on licensees when they apply for 10 amendments.
11 Now, that's in the absence of understanding 12 what's been done in the pilot program. At least what you 13 said today with respect to the tech specs, I did read the
(_\i
\# 14 one on the tech specs. And I said this isn't too bad, but 15 there seems to be some added burden in it, which I would 16 find difficulty dealing with if I worked for a utility 17 again.
18 CHAIRMAN APOSTOLAKIS: But remember --
19 MR. ABBOTT: On the other hand, the 20 discussion, it seems like it went fairly well.
21 CHAIRMAN APOSTOLAKIS: Yes.
22 MR. ABBOTT: It gets back to your point l
l 23 earlier that I think there's a failure to communicate 24 here. Now, that may not necessarily be true with the
(,) 25 other items. I don't know. But this one seems to be I
! NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
i (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
249 1 fairly good.
2 MR. CARROLL: I think this is the easy one.
O 3' MR. ABBOTT: Yes. Actually, that is the easy 4 one.
5 CHAIRMAN APOSTOLAKIS: But remember what Gary 6 said this morning, that, okay, there is extra burden, but 7 they get something extra, too.
8 MR. ABBOTT: I don't believe the extra burden 9 is.all that big.
10 CHAIRMAN APOSTOLAKIS: So it's probably a 11 failure to communicate.
12 MR. ABBOTT: Right. I think it is.
13 MEMBER BARTON: That's a piece of'it. In 14 order to try.to. keep things simple, you might want to look
'15 at Section 5, Element 3, " Developing implementation and 16 monitoring strategies." It just gets wordy, and you l
17 repeat. i 18 It' starts out, " Application of three-tiered- J 19 approach described below." You go to the next paragraph, 20 "As described in 4.3." You do a very good job of 21 describing the three-tiered approach. Then you go on and 22 explain it again.
23 So, in the sake of simplification, you may 24 want to look at Page 5.1 of the reg guide and kind of
() 25 skinny down some.of those words.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS ;
1323 RHODE ISLAND AVE., N.W.
(202) 234 4 33 WASHINGTON, D.C. 20005-3701 (202) 234-4433
. - ~ . . . .
.1 250 l 1 MR. CARROLL: Well, all of the supplemental q '2 reg guides I guess I'll call.them for the pilot programs 3- are very redundant to the --
4 CHAIRMAN.APOSTOLAKIS: Yes.
5 MR. CARROLL: -- the main one. I'm not sure 6 why you couldn't do more --
7 CHAIRMAN APOSTOLAKIS: Even within themselves.
l
-8 MR. CARROLL: Just to cut the paper down. '
]
l
- l. 9~ CHAIRMAN APOSTOLAKIS: Yes, yes. l l
r 10 MR. CARROLL: When I see a stack of' paper that 11 high, I'm'less inclined to read it than when it's that j
- l. . .
l 12 high.
' r 13 CHAIRMAN APOSTOLAKIS: .And the burden will not :
O 14 appear to be so onerous. {
15 MEMBER BARTON: That could be true, George. ;
i.
16 CHAIRMAN APOSTOLAKIS: That could be true.
I '
17 MEMBER SEALE: Somehow if we could --
- 18. CHAIRMAN APOSTOLAKIS: Do you have anything 19 else that you really feel you ought to tell us? .l i
20 MR. CARROLL: Have you considered hiring )
21 Dilbert as a consultant to help on'this effort?
22 MS. GILLES: The only point I was going to )
i
)
23 mention is I think one reason the industry is a little :
1
- 24 discouraged is, as you said, tech specs is the easy one.
l 25 And it's'been almost two years since they originally 1 1
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. ]
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
251 1 submitted these requests. I think that's where some of i
em g 2 the discouragement is coming from.
f O 3 CHAIRMAN APOSTOLAKIS: Okay. I'm not saying 4 that one side is right and the other is wrong. I'm just 5 puzzled by this. Judging on what we heard the last hour 6 and a half, I would say it would be great until you said 7 it took two years.
! 8 (Laughter.)
9 MEMBER SEALE: They learn s1.ow but well.
10 CHAIRMAN APOSTOLAKIS: Okay. Thank you very 11 much.
12 (Whereupon, the foregoing matter went off the i ,_
13 record at 3:10 p.m. and went back on the l )
\- 14 record at 3:28 p.m.)
15 MR. FISCHER: The floor is mine?
16 CHAIRMAN APOSTOLAKIS: The floor is yours.
17 (Slide) 18 MR. FISCHER: Good afternoon, gentlemen. My 19 name is Dave Fischer. I'm with NRR's Mechanical 20 Engineering Branch. And I'm here to describe for you the 21 efforts of the risk-based in-service testing team, our 22 efforts to develop reg guide and standard review plans, 23 and a couple of other things which I'll mention.
24 (Slide)
(~N
( ,) 25 MR. FISCHER: With me today is my co-team i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(232) 234 4 433 WASHINGTON. D.C. 20005-3701 (202) 234-4433 l
252 1 leader Brad Hardin, on my right, with the Office of
's, 2 Research and Mike Cheok, on the right, who you heard from 7
Ll 3 earlier today. He's with NRR's PRA Branch. And on my 4 left is Joe Colaccino. He's also in the Mechanical 5 Engineering Branch of NRR.
6 Together we developed these two documents and 7 some other things. And we tried to be as consistent as 8 possible with the general reg guide and general SRP. And 9 I'll describe that a little bit later.
10 I will apologize at the beginning for my 11 focus. My focus was on the risk-informed IST SRP. So 12 when I'm talking about things, I'll probably refer to
,_ 13 sections in the SRP, but the wording that I refer you to I )
\
'/ 14 should be identical to the wording that's in the reg 15 guide. And it should also be consistent with what's in 16 the general reg guide and the general SRP.
l 17 So I'm just pointing you to where it is in one l 18 of the documents. Hopefully they're all consistent. And 19 if they're not, I'm sure you will bring it to my I
20 attention.
21 If you need a cite for where it is in the IST 22 reg guide, Brad may be able to help us with that because l
23 Brad's main focus was on the reg guide and mine was on the 24 SRP.
( ,/ 25 (Slide)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433 j
l
l 253 1 MR. FISCHER: This is nothing more than an
-~ 2 outline of my presentation. I plan to spend most of my I L ,i 3 time on the third and the last bullets; that is, the 4 risk-informed IST approach and on the status of pilot 5 reviews.
l 6 (Slide) 7 MR. FISCHER: This slide shows the team's 8 three major objectives, all of which relate to some 9 products that we are producing. The first one on the 10 slide is the reg guide. The second one is guidance to the 11 step, which is called an SRP. The last bullet is the 12 safety evaluations that we plan on preparing for the two 13 pilot plants, which are Comanche Peak and Palo Verde.
t
\ ') 14 There are also other objectives and activities l I
15 that the team is working on. We are reviewing ASME code 16 cases related to risk ranking and related to test 17 strategies for motor-operated valves, for check valves, i 18 for pumps, for air-operated valves, and different 19 components. So they're developing code cases related to 20 each of these components as well as a general risk-ranking 21 code case. And the team is working to review those.
22 We're also working, as need be, with NEI. We l 23 prepared a draft guide on in-service testing. So when we 24 get something in from industry, the team also reviews
~
'q,,/ 25 those. Our focus isn't just on those three documents, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., NW.
l WASHINGTON, D C. 20005-3701 (202) 234-4433 (202) 234-4433 t
_ _ . . - _ _ _ _ . ._ _. _ . _ . . . . _ . _ _ . . _ _ _ _ _ _ . _ ~ . . . . _ _ _ . _ . _ _ . _ _ . - _
254 !
1 although that is what we're here to talk with you today, 2 about che reg guide, the SRP, and the pilot. That's why !
O 1
those are the only three listed. ;
i 3 4 Implementation of the risk-informed IST i
5 program, as discussed in both the reg guide and the SRP, l 6 is optional for licensees. .They can choose to maintain ,
7 IST programs in accordance with-the current. code or.they. ,
8 could propose an alternative that's different than that ,
9 described in the reg guide. So this is just one ;
10 alternative that the staff feels is acceptable.
11 The staff anticipates going to rulemaking to ;
l 12 codify the use of risk-informed testing. In the meantime !
13 and for the pilot licensees, we plan to authorize an r _
i.
l 14 alternative to the ASME code test requirements, including {
15 the scope requirements, the specified code, pursuant to a j 16 section called 50.55a (a) (3) (i) , which says if the staff 1
17 can make a finding, that-the proposed alternative provides !
18 an acceptable level of quality and safety. And that's as
- 19. specific as it gets. If we can make that finding, then we 20 can' authorize an. alternative.
21 And that's the provision under which we 22 anticipate approving the pilot and possibly other 23 licensees between now and when we get the rulemaking j L l 24 completed. That's just kind of a framework.
s 25 Next slide.
1 r-l NEAL R. GROSS !
I- COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
, - - . ..- - _ _ _ -. - - _, . - . -...1
- 255 l
l 1 (Slide) l
<-~ 2 MR. FISCHER: The team's accomplishments to l
l j
O 3 date, we do have, as you can see in front of you, draft l
4 reg guides and SRP. We also have completed an initial l
l 5 review of the pilot applications at Comanche Peak and Palo 6 Verde.
l 7 We have prepared one set of RAIs based on
- 8 their initial submittals. We have had one set of meetings 1
9 with the licensees. And our future plans are to continue 10 to work with the pilot licensees.
11 We have an additional set of RAIs that's about 1
12 ready to go out the door. We don't think it's the last 13 set of RAIs that we're going to have for the two pilot i
(' -) 14 licensees. I will discuss those RAIs further in my 15 presentation. I can delay talking about what's in those 16 RAIs, but I will get to that.
17 We hope to get the pilot plant review 18 completed by June of '97. And we will, as I mentioned 19 earlier, continue to work with ASME on the code cases.
20 (Slide) 21 A risk-informed approach starts by ensuring l
l 22 that the plant is designed and operated in accordance with i
23 this current licensing basis and by ensuring that the PRA 24 reflects the actual plant. These points are worked into
,m k_,) 25 the reg guide and the SRP.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
256 1 A risk-informed IST program should be l
<)
l LJ 2 developed using traditional engineering evaluation, PRA l 3 insights, and an integrated decision-making process.
l 4 In the reg guide and the SRP, components are l
l 5 categoriced as either high or low safety-significant, l 6 which we have acronyms for, unfortunately. The IST l
7 program scope may include non-code components that are 8 high safety-significant or low safety-significant 9 components if the licensee's integrated decision-making 10 process or their expert panel suggested those components 11 should be included in the program.
12 So we're not only going to change the test
,._ 13 requirements, the frequency of some components, might i )
! 14 extend the interval for some components. We might be ;
l 15 doing different tests. And we might have different I
16 components in the program. j 17 So this is a proposed alternative to both the j l
18 test requirements themselves and to the scope requirements 19 as stated in the code. It's a very broad alternative.
20 It's much broader than the typical authorized alternatives 21 staff has done in the past.
1 22 Changes to the current licensing basis should l
23 be identified and justified. I can go into more detail on 24 that.
,a
(._) 25 MEMBER SEALE: By that, do you mean that any NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON D.C. 20005-3701 (202) 234-4433 l
257
.1 changes in the licensing basis that would result from the
- 2 proposed risk-informed in-service testing program are ;
3 identified in this application?
4 MR. FISCHER: Yes. That's what I mean. Just i i
5 as an example, if you were taking an MOV and you had a f 6 stroke time test and now you weren't going to do the 7 stroke time test, you were going to do some other test 8 that was deemed to be a better test and the licensing 9 basis may say for this system that valve needs to operate 10 in a certain amount of time and the staff relied, at least i i
.11 in part, on the f&ct that it was going to be tested every -l 12 quarter, stroke time-tested, and gaining the confidence ;
13 that they had in order to grant the license, then we would l0 L
14 want them to identify that change to the licensing basis i l
15 because we relied to some degree on that valve operating 16 at a certain time if that's what happened and then tell us :
17 why it's okay to do that.
18 MEMBER SEALE: Okay. Now, is there any'other 19 paper that they would have to submit in order.to get 20 approval to depart from their present commitment?
21 MR. FISCHER: Maybe I think is the way I'd
(
22 like to answer that. This review and approval of a 23 risk-based IST program doesn't take away the need for the I 24 licensee to submit paperwork, say, if it needed a tech 25 spec change to change.a test on a particular component.
l .
NEAL R. GROSS
. COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4 33
258 !
1 .If that test is specified in the tech specs, 2 then they'd have to'come in and get an amendment to the 3 tech specs. If it were required for some regulation, then l,
! 4 they would have to come in and get an exemption from that ;
i 5 regulation. ;
i I I
- . 6 What we have tried to do in the reg guide and t 7 SRP in to make it clear that in their program submittal, 8 if their proposed in-service test change necessitates a .;
I f- 9 change to the tech spec or an exemption, that they l l !
10 identify that fact and submit the appropriate paperwork at j
.i 11 the same time so that we c.an just process it, review and ;
I 12 approve it.
i 13 MEMBER SEALE: Okay. !
14 MR. FISCHER: Okay. So there may be some !
15 other paperwork they need to do, possibly not. j l
15 MEMBER SEALE: But presumably the entire i
17 submittal, including the approvals on those changes, would 18 be part of the initial submittal and would be made to one 19 central manager, if you will?
20 MR. FISCHER: That is what we would like. In 21 the case of the pilot licensees, what they have elected to 22 do is where there was a-tech spec change that would be
)
j.
l 23 impacted, they have elected not to change the testing on 24 that particular component. So they have tried to avoid l
25 additional --
1 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
-(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
259 1 MEMBER SEALE: Yes. That's kind of what I fS 2 thought might happen.
Y_
3 MR. FISCHER: But the reg guide and the SRP 4 want them to identify on a component-specific basis how 5 the change might affect the licensing basis and what the 6 justification is. And then the staff will review and 7 approve that.
8 (Slide) 9 MR. FISCHER: We make it clear in the reg 10 guide and the SRP that test strategy relaxations and 11 possible improvements should be considered. What we're 12 saying is that not only should the licensees use the PRA 13 determine the acceptability of proposed relaxations, but i )
'/ 14 also they should use the PRA to identify important failure )
l 15 modes and investigate the failure causes and design the 16 component test strategy accordingly. So we want them to 17 look both ways.
18 So I said two things to you so far about 19 things that get added. Some components may get added to 20 the scope, and some test strategies may need to be 21 adjusted based on review of the PRA information.
22 The overall effect on plant risk should be 23 estimated. And we rely on the criteria that are in 1061 24 to determine the acceptability. And we refer to that in
/')
(,,/ 25 the reg guide and the SRP.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBb?M 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l i 260 1 MEMBER POWERS: When you say you rely on those fm 2 criteria, that meanc that --
'%.,) 1 3 MR. FISCHER: We reference them. l l
4 MEMBER POWERS: -- when a plant comes in and 1
5 it says, "My oCDF or my CDF exceeds 1 times 10-4, I will 6 not make changes that increase risk"? Is that what you 7 mean?
8 MR. FISCHER: Please, Mike.
9 MR. CHEOK: That's what we mean, correct.
10 MEMBER POUERS: And when they come in and say, 11 "My CDF is" such and such a number, they include in that 1
12 their core damage frequency due to accidents at operation, l 13 accidents initiated by external events, and accidents D)' 14 during shutdown?
l 15 MR. CHEOK: I guess I wasn't sure what the 16 question was. ,
l 17 MEMBER POWERS: That's true; right? Okay. l l
MR. CHEOK:
18 Yes.
l 19 MEMBER POWERS: So everything is lumped into 20 that CDF? ,
1 21 MR. CHEOK: Correct.
22 MR. HARDIN: I was just going to say this is i
23 one case where we have not repeated the same criteria in l 24 the specific guides. Some of the things you saw for tech
() 25 specs were specific to that, but for IST, we just refer to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
261 1 the general reg guide. So what you have been hearing ex 2 about that applies to this as well.
(~. / ) 1 3 MEMBER POWERS: Well, let me not be i l
4 mysterious. You're going to show us some examples of some 1
5 CDFs where I insist that if you put the right CDF in, you 1
6 would not allow the change because it's an increase in 7 risk that you tend to use the operational CDF, rather j l
8 than using what the real CDF is.
9 MR. CHEOK: I guess we'll discuss this when we 10 get to the -- i 11 MEMBER POWERS: We'll get to it?
12 MR. CHEOK: Yes.
13 MR. CARROLL: Is that right, Gary?
I-
'\- ] 14 MR. HOLAHAN: I guess I'd like to wait until I
15 we get to it also. J 16 (Laughter.) I
'7 MR. CARROLL: Well, it's a very simple 1
18 question. When you say 10", does it include --
19 MR. HOLAHAN: Yes. The 10" is intended to be 20 for full scope. Now, we recognize that in some cases 21 there will be some judgment involved because you won't 22 have analysis for all those circumstances.
23 But the answer is yes, it's intended for 24 comparison with full scope.
n \
I
(_,! 25 MEMBER POWERS: And you try to handle those NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON D.C. 20005-3701 (202) 234-4433 ,
1 262 1 things where people don't have full scope by putting in
(S 2 this 3.5 times 10~5 boundary.
Nj 3 MR. JONES: No. That was removed.
4 MEMBER POWERS: Oh, that boundary is gone?
5 MR. JONES: That was an earlier draft.
6 MR. FISCHER: Okay. I would like -- this 7 probably is going to open up a little discussion, I would 8 think -- to turn the subcommittee's attention to a section 9 in the SRP. On Page 16, there's a little bit of a 10 discussion. You don't necessarily have to go to there, 11 but it is on Page 16 of the SRP version that you had.
12 It talks about trade-offs when evaluating a
,_ 13 licensee's proposal when they come in and they say, "We're 14 going to improve the test strategy on this group of 15 components. We're going to extend the frequency, extend 16 the interval on this group of components," so you can i
17 trade off on that, i 18 The paragraph as it's currently written starts l 19 with talking about relaxations. And that is relaxing the 20 test interval for components where these relaxations are i 21 offset by alternative measures; for example, additional 22 monitoring or different test procedures or training, the 23 licensee should identify and quantify to the extent 24 practicable the effects of the alternative measures.
(~)'\
(_ 25 Similarly, if there are benefits associated NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 734-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
263 1 with the proposed relaxation; for example, reductions in l 1
2 initiating event frequency or reductions in system l k ,s) m 3 misalignment or reduction in radiation exposure, the 4 licensee should identify and quantify to the extent 5 practicable the benefits that should be realized from 6 that. l l
7 And then it says, "As a general rule, the 8 alternative measures and the benefits should be directly 9 linked to the system or components associated with the 10 proposed relaxations."
11 On a case by case basis, the staff may assess 12 the licensee's proposed improvements made to the test 13 strategy for a group of components against the proposed i
'- 14 relaxations and test requirements for another group of 15 components in assessing the overall acceptability of the 16 proposal.
I 17 So, as currently written, it says generally we 18 want you to stay on a component basis on a case-by-case 19 basis, we'll consider trading off improvements on these 20 components against increases in risk on those components.
21 MEMBER FONTANA: How do you work in the 1
22 reduction in radiation dose by the staff, going from one l l 23 testing procedure to another? Is that in the numbers 24 anywhere or is it mostly judgment?
,r ~g
(,,) 25 MR. FISCHER: It's definitely not in the l
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
I (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) P34-4433
264 1 numbers right now and --
- 2 MR. HOLAHAN: Well, I was going to say it's 3 part of the integrated decision-making. And I don't think 4- you have to convert everything to numbers to do that. J 5 MR. FISCHER: The pilot licensees just 6 described it in general terms and said, "Here's something 7 that compensates in some way for the increase in risk." ]
8 MR. CARROLL: Now, you had the words in there 9 in a couple of places, " quantify to the extent 10 practicable." It seemed to me the things you were talking r
11 about are not things you're going to be able to quantify.
12 MR. FISCHER: I think some of them may not be ]
13 able to be quantified and others may be able to some ;
O 14 degree to be quantified.
)
15 MR. CARROLL: Has that been the experience i'
16 with the pilot program? People have been able to --
17 MR. FISCHER: I'm kind of going to rely a 18 little bit on Mike, but I think if they have_ established 19 procedures or compensatory measures, they have tried to 20 quantify it a little bit.
21 Mike, do you'want to elaborate on this a 22 little-bit?
23 MR. CHEOK: For any increases in test l
.i 24 intervals, data-has been quantified in the PRA, basically ;
) 12 5 multiplying your reliability by your testLinterval to get l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
__ - ._ _. __ , -. _ ,_ . - ~-
265 1 your --
rx 2 MR. CARROLL: Okay.
IV) 3 MR. CHEOK: So that's been quantified in that 4 sense. The effectiveness of the test, on the other hand, 5 right now is non-quantified.
6 MR. CARROLL: So all you're quantifying are 7 the negative. You don't really know to what extent it's 8 offset by the improvements that are alleged.
9 MR. CHEOK: To quantify the improvements, we 10 basically have to go back into the PRA and say that the 11 tests we have now are not as effective as we think they 12 should be and maybe take less credit for those tests. I'm 13 not quite sure that we will require that at this point.
I,%
14 MEMBER FONTANA: Let me ask a hypothetical 15 question that turns that around. Lat's say you have a new 16 testing technique which is not intrusive and it allows 17 testing of a piece of equipment in such a way that the 18 equipment doesn't have to be shut down and it's not l l
19 damaged, umpty, umpty, ump, but the testing technique 20 isn't quite as good as what it's replacing. ,
1 21 If you make the argument that, even though the .
l 22 probability of catching something is not quite as good, 23 the fact that you have less down time and less chance of 24 damaging the equipment is better, how would you do that?
7-( ,) 25 How would you make that argument?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
266 1 MR. CHEOK: I think if the licensee would come fs 2 in and try to quantify that in some sense, we would look V 3 at it.
4 MEMBER FONTANA: We could run the numbers both 5 ways.
6 MR. CHEOK: I'm sorry? I missed the question, 7 I guess.
8 MR. FISCHER: He said he could run the numbers 9 both ways.
10 MR. CHEOK: That's correct.
11 MEMBER FONTANA: Yes. But the prior question 12 looked like you can only get the negatives but not the 13 positives. But the answer to this question --
p,
( )
Y2 14 MR. CHEOK: We will entertain any positive 15 type numbers that the licensees would submit. l l
16 MEMBER FONTANA: Okay.
17 MR. FISCHER: And the staff in the past has 18 without quantification tried to entertain proposal 19 alternative test strategies like that. We have been faced 20 with some of that kind of decision-making, but we haven't 21 really asked the run of the PRA.
22 MEMBER CATTON: It sounds like in the future 23 it will be without quantification, too, except you'll try 24 a little harder to understand it.
p-(._,) 25 MR. FISCHER: Well, to the extent practical, NEAL R. GROSS COURT REPOPTERS AND TRANSCRIBERS 1323 RHODC ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON D C. 20005 3701 (202) 234-4433
267
}
4 1 we'll try to quantify it, like you said, in those words.
2 MR. HARDIN: The industry has been thinking
, t 1 .
)
i 3 about some ways to do some of this quantification that the j i i 2
4 ASME has included in some of their discussions. And I 5 think it's just a new area that we really haven't gotten l l
6 into too much. So maybe in the future there will be some 7 things that they will propose and that we'll review.
8 MEMBER CATTON: If you have a highly reliable f l
9 piece of equipment and you're. going to make it more 10 reliable, it's going to take you forever to get the !
11 statistics that you can use. So what do you do? You're 12 just going to have to make a judgment that it's going to 1
- 13 be more reliable.
(
14 MR. HOLAHAN: And sometimes you can --
.15 MEMBER CATTON: So it's not quantified.
16 MR. HOLAHAN: Well, but in --
17 MEMBER CATTON: With a crappy. piece of 18 equipment, it's easy. But if you have a good piece of 19 equipment, how do you do it, something that has an j l
20 unreliability of what? I don't know. Pick some small 1 21 number. And if you do the statistics, it takes forever.to 22 get what you need to incorporate it into a number-based 23 PRA. Judgment-based is not a problem.
24 'MR. HOLAHAN: And sometimes you can base that 25 judgment on monitoring some condition of the equipment NEAL R. GROSS' COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 23 4 433 WASHINGTON D.C. 20005-3701 (202) 234-4433
- . . -. _ - - .. . . . - . - . . - . - - . - - - . - _ . - _ . ~ .
268 l i
.1 that's less than failure sometimes.
i y 2 MEMBER CATTON: I don't think so, but may be. f r i 3 It gives you a good structure to work in. I think that's l
I 4 clear.
5 MR. HARDIN: One of the improvements that ;
I i
6 wa're mentioning here which we haven't gotten to the point 7 where we see much of it yet, though, is the improvements .
8 in test methods.
9 MEMBER CFM'ON: Improved test methods give you ;
r 10 more information. And maybe with help of other !
t 11 information, you can then relate to the change. But you 12 can't do it in a specific plant on a specific piece of f 13 equipment, I don't believe. ]
14 We had that problem with the diesel engines.
15 You just can't do the. statistics. It's too good.
16 MR. FISCHER: I think somewhat related to'this 17 is the fact that we are in looking at some of'these 18- improved test strategies, -- I bring up I guess only one a 19 little bit so you know where I'm headed, Mike.
20 We are promoting the use of some improved 1
21 testing strategies that would test in this particular case I
22 MOVs less frequently than every quarter using a test 23 methodology that-the staff believes is a more effective 24 test in= determining whether the component would be capable 25 of performing'its function if it were required as a ;
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 - WASHINGTON, D.C. 20005-3701 (202) 234-4433
269 1 performance function.
,-~ 2 And trying to quantify that, there's a
\J 3 negative incentive there. I mean, the PRA says you're 4 extending the interval. It sounds like you're increasing 5 risk. But, in point of fact, we think that it's a net 6 benefit.
7 MEMBER CATTON: Well, if you're not finding 8 anything out'when you do these tests close together, why 9 not do them far apart? And your PRA certainly will give 10 you a positive response to that.
11 MR. FISCHER: Well, PRAs right now --
12 MEMBER CATTON: I'm not trying to argue 13 against it. It's just that you've got to keep a good (x -) 14 solid amount of common sense in what you're doing.
15 MR. FISCHER: I think the reg guide and the 1
16 SRP have a lot of common sense woven into it. And it has i l
17 some features that, in particular, the performance 18 monitoring and corrective action features that we have 19 embedded in these reg guides and SRPs, that will make sure 20 that if we do something that isn't in the best interest of 21 safety, if we change a test strategy and it's not 22 effective, if we extend the interval and we're getting 23 more failures and the test interval is the right test 24 interval, that we will adjust the test interval fairly
() 25 promptly.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
270 1 MEMBER CATTON: But, again, you can only do 2 that if you have a lot of failures. If it's something 73 i
3 that has 10-4 unreliability, you're never going to know for 4 sure. l 5 MR. FISCHER: We're not saying that when you l
1 6 get a failure, examine the statistical significance of 1 7 that failure and you don't have to change anything until 8 it's statistically significant. What we're saying is if 9 you have a failure, do an assessment to see whether you 10 think you ought to adjust the test strategy right then.
11 MEMBER CATTON: Absolutely.
12 MR. FISCHER: Okay? So we're not waiting to 13 see what the effect on the generic failure rate is going
/,N .
l
( ;
\/ 14 to be. We want licensees to consider it right then and i 15 there.
16 MR. CARROLL: Isn't what you just described 17 the maintenance rule?
18 MR. FISCHER: There is a lot of similarities 19 between what we have described in here and the maintenance 20 rule. And it was intentional.
21 MR. CARROLL: It ought to be.
22 MR. FISCHER: It was. Another key feature of 23 our --
24 MR. CARROLL: Have you looked at what you're 7
(3 ,,
) 25 proposing and convinced yourself that you're not
! NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
271 1 compromising the maintenance rule and any of its rx 2 philosophy?
( )
L/
3 I see nothing wrong with amplifying on what 4 the maintenance rule requires, but I see a lot wrong with 5 having two maintenance rules.
6 MR. FISCHCR: We didn't talk about the 7 maintenance rule specifically in too many sections here, 8 but it is mentioned. And I'm afraid I may give you a bum 9 page number. It's about Page 17 or 18 of the SRP.
10 It tries to say, " Don't do anything in 11 conflict of the maintenance rule and try" -- we're not 12 trying to do additional stuff here. Take advantage of
,s 13 what the maintenance rule monitoring is available.
14 MR. CARROLL: Okay.
15 MR. FISCHER: In addition to that, I work 16 closely with Rich Correia and some of the people in the 17 Maintenance Quality Assurance Branch to try to make sure 18 that what we're doing is consistent with what they have in 19 their mairtenance rule and the reg guide.
20 So I think we're on pretty firm ground. I 21 can't say we're perfect, but I'm pretty confident.
22 MR. CARROLL: Okay.
23 MR. FISCHER: Another key aspect of our l
24 risk-informed IST approach is that it be implemented in a )
/')'i
(_ 25 step-wise manner. The pilot licensees have described to NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., NW.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 i
~ - - . - .
- r. i 272 I l
1 us based on PRA that they think they can extend the test I3 2 interval to six years or the low safety-significant 3 components. And what we are asking the licensees to do in .
l 4 one of-our RAIs is: Please describe the step-wise j 5 implementation approach where you don't go from quarterly i
6 to six years. !
7 And what we want to do is to try to extend it i
8 gradually,. gain data, gain experience with the extension, 9 feed that information back into the overall process so 10 that we make sure that we're not doing something that's i i i l' 11 not in the best interest of safety. j l
12 So we're slowing the industry down a little ;
l i 13 bit, but we think it's a wise approach to go in a ;
14 step-wise manner. l 15 MR. CARROLL: Did you look at what was done in )
16 the containment leakage rule, which is sort of that same ,
1 1
17 concept? j l
18 MEMBER BARTON: Are you talking about Appendix 19 J stuff, Jay?
20 CHAIRMAN APOSTOLAKIS: Right.
l' 21 MR. FISCHER: I can't say I looked at it 22 explicitly, Mr. Carroll. .
23 There are also very strong performance 2
- 24 monitoring and feedback sections in both the reg guide and i
! 25 the SRP. And it goes beyond just asking licensees to i
L NEAL- R. GROSS I
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. j I. . (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 ;
1 i
, . . - , , , - - - . . . - , a
273 1 develop the root cause of a failure and correct it. It rx 2 says very explicitly that you ought to consider whether or c s
(_/
3 not the failure assumptions of your PRA should be 4 adjusted. ;
5 It says that you ought to reassess the test 6 strategy, meaning the test methods and the interval that 7 you're on and, if necessary, adjust it. And it also 8 requires that the licensees monitor the performance of 9 components that are in the risk-informed IST program so 10 that at periodic intervals, they can go back and look at 11 the data that they have acquired to see if it's consistent l 12 with the assumptions that are in the PRA to see if things 13 have changed, to see if they need to change the test !
(-
1
\- 14 intervals or change the test sr.rategy or even change the 15 PRA, the data used in their PRA.
16 If any of the other team guys have things they 17 want to add on any of these bullets, please feel free.
18 Again, as stated earlier, the regulatory 19 approvals would be an alternative to the code requirements 20 in 50.55a (a) (3) (i) . I guess that's that slide.
21 (Slide) 22 'MR. FISCHER: The essential aspects of the reg 23 guide and the SRP that are on this slide --
24 MR. CARROLL: How do you do that without a
(_)
s 25 rule change?
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
274 1 MR. FISCHER: Do which? The current 2 regulations have a provision. 50.55a has a provision that 3
Nj 3 the staff, specifically the Director of NRR, can authorize 4 an alternative to the code requirements. And it's a 5 little unusual because of the placement of this section 6 (a) (3) (i) .
7 It says you can authorize an alternative to 8 the whole section on IST, to the whole section on ISI, 9 which means that you can take basically an authorized 10 alternative to any piece of the code. So it's very broad.
11 MR. JONES: I would also state, Jay, that that 12 was a policy issue raised to the Commission in an earlier
,~,
13 update, which came back in that same SRM that said the
( ) The Commission approved 14 staff was authorized to use this.
15 the staff to go this way.
16 MR. MARKLEY: Dave, is this going ahead of the 17 code case, then? What's the time line here when the code 18 cases are going to be wrapped up and available for 19 consideration? This is going forward without the code 20 cases, then. Is that correct?
21 MR. CARROLL: It's currently pending.
22 MR. FlSCHER: We hope that the code cases have i 23 gone through before we -- we thought they would be through 24 before the pilots got approved. It's kind of like a horse
/~'T !
() 25 race.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
i l
275 i l
1 I think what we're asking licensees to do, I 2 can't say how we're ultimately going to end up, Mike. But 3 where I think we're headed is that we'll ask the licensees 4 to consider the code cases when they get reviewed and 5 approved by the staff as they're endorsed by the staff and 6 consider' going to those test strategies.
I J 7 So whether we have a commitment to go to the j l
.I 8 new test strategies during the code cases, I'm not sure i 9 whether we'll do that or whether we'll let them stay with I L
10 the current code.
l- 11 Does that answer your question? -l l
12 MR. MARKLEY: Yes. ,
13 MR. FISCHER: Right now the pilot licensees :
14 have said: For the high safety-significant, we want to i
15 stay with the current code. For the less l l
16 safety-significant, we want to stay with the current code 17 but on an extended interval. 1 18 Now, we may be able to prove that, but we're 19 also working with ASME to develop these revised, we think 20 improved test strategies, particularly for the high 21' safety-significant components, because we. feel that if 22 they make improvements to the_ test strategies on the high 23 safety-significant components, then it'will more than 24 likely compensate for any negative effects of extending l-() 25 the interval on less safety-significant components.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N.W.
l (202) 234 4433 WASHINGTON, D.C. 2000 # 3701 (202) 234 4 433 i
276 1 Did I convey that right?
(3 1 2 MR. MARKLEY: I guess my concern is: What t
w/
3 happens if the code cases come out and they take a 4 different position than what you're proposing here within 5 the context of all of this? I realize you're doing a lot 6 of collaborative work, but they haven't taken final 7 positions on these things.
8 MR. FISCHER: I can't conceive of the code 9 cases coming out and offering a test strategy that's less 10 effective than the current code. I mean, if it's --
11 MR. MARKLEY: That's not what I was 12 suggesting.
13 MR. FISCHER: Okay. Maybe I don't understand
/f_T 14 you.
15 MR. MARKLEY: The code cases are something 16 which in lieu of revising the code is considered to be an 17 acceptable method; correct?
18 MR. FISCHER: That's correct. l l
19 MR. MARKLEY: And the regulations say you're 1 20 required to meet these standards. So now the code cases 21 are a way of doing that. What you're proposing with this, l
22 then, is come out with an NRR Director's policy position 23 saying that this is okay in lieu of another review on or 24 position from the ASME group.
p
( ._) 25 MR. FISCHER: I think we'd have to endorse the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 200054701 (202) 234-4433
277 1 reg guides in the standard manner in which we endorse --
2 endorse the code cases in the manner that we normally
- O 4 i
l f
3 endorse the code cases vis-a-vis by a reg guide or --
4 MR. CARROLL: Sure. This is a voluntary 5 program.
6 MR. FISCHER: Right. I don't think there's 7 anything different than a normal way of doing business on 8 that. This particular slide I don't think will tell you a 9 whole lot other than the reg guide is going to provide 10 guidance to the licensees and the SRP will provide 11 guidance to the reviewers, review procedures and 12 acceptance guidelines.
13 I might mention that these same acceptance
/ %
(, )
'~
14 guidelines here that are mentioned in the SRP, those 15 acceptance criteria, we also put them in the reg guide so 16 that the licensees -- when we described an acceptable 17 alternative, we also told them what the acceptance 18 criteria would be.
)
19 MEMBER CATTON: Could I sort of raise a i 20 troublesome question? You actually started it by 21 mentioning the MOVs. I don't recollect when it was, Jay, 22 when we looked at the MOVs. A few years ago?
23 MR. CARROLL: Yes.
24 MEMBER CATTON: And found that the failure to g s.
(._,) 25 close was much higher than people had thought. Most of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
278 l 1 the PRAs haven't changed. They still use this sort of 1
,3 2 common database. So along comes this utility who's going
()
3 to do it right. And they come up with real numbers.
4 So if they fixed the problem and improved what 5 they do with the MOVs and use the real numbers, they're 6 worse off as far as the PRA is concerned. What are you )
7 going to do?
8 MR. FISCHER: I want to give two people the 9 opportunity to answer this question.
10 MEMBER CATTON: I just threw it out there. I 11 don't c.now what you can do.
12 MR. FISCHER: We remember your question. I 13 have a piece of paper in hand that I was going to discuss I T Y/ 14 it with you after this meeting, talk about this. And Tom 15 Scarbrough is the staff's MOV, has a need for MOVs with 16 depth here, talk about that specifically.
17 MEMBER CATTON: Oh, I've raised this before?
1 18 MR. FISCHER: You have raised this before.
19 And we have an interim answer for you --
20 MEMBER CATTON: Wonderful.
21 MR. FISCHER: -- we hope will be acceptable to 22 you. Maybe we can talk after the meeting and --
23 MEMBER CATTON: Well, I don't know.
24 MR. CARROLL: Some of us would like to --
i /~~'
() 25 MEMBER CATTON: I think some of the people NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON. D C. 20005-3701 (202) 234-4433
279 1 here would like to hear it, particularly those of us who
.f 2 --
(m)
'O 3 MR. FISCHER: Do you want to say a word?
4 MR. SCARPROUGH: This is Tom Scarbrough.
5 Yes. We got your message. And this issue on 6 the PRA, the reliability of the PRA failure rates for MOVs 7 of several years ago from Carlisic Michaelson and 8 discussions with the ACA3 at that time, we discussed this.
9 Back about '92 we started this question about:
10 What is the real failure rate for MOVs because obviously 11 it's not .003, which is what was used in the PRAs at the 12 time? So --
13 MR. CARROLL: And still being used.
('
'-) 14 MR. SCARBROUGH: Right, right.
15 MEMBER CATTON: And Kraftwerk Union would 16 argue that it's 20 times that.
17 MR. SCARBROUGH: It might be. It might be.
18 MEMBER CATTON: Based on the structure of the 19 valves as they exist.
20 MR. SCARBROUGH: Now, what we were doing with 21 89-10, because we didn't go back right immediately to look 22 at what the failure rate really is because 89-10 was 23 bringing everything up hopefully to a much better .; vel 24 than it was before, those programs are basically complete. j (h
(_,) 25 A hundred three plants have told us they're down to 89-10 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISt.AND AVE., N W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 ;
)
280 1 or processing that now. l l
2 So what we plan to do over the next year or so l
l lq
\ /
l
I 3 as all the plants have finished by looking at NPRDS, 4 looking at what the failure rates are, looking at plants j 5 that start to implement O&M1, which is going to be 6 monitoring degradation rates for diagnostics over time, f l
7 looking at that and trying to come up with a decision as i
8 to what would be a more appropriate failure rate put into 9 the PRAs.
10 Right now they use what they use, but over 11 time those numbers may change. And they'll have to 12 reflect those proper numbers in their PRAs and use them.
13 MEMBER CATTON: I'm surprised that we don't
\ 1 C' 14 have enough information now. I mean, you had major 15 studies done by Kraftwerk Union. There were a couple of 1
16 studies done by NRC. I 17 My recollection is that for valves that have 18 to close against the full system pressure, the Germans 19 said that the number should be like eight percent, which 20 is even more than 20 times, because the structure of the 21 valves was never designed to do this.
22 MR. SCARBROUGH: Yes. They're ahead of us.
f 23 The Germans did 89-10 many years ago.
24 MEMBER CATTON: Well, that's because they were
() 25 a turnkey system.
l NEAL R. GROSS
! COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l l (202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 281 i l
1 MR. SCARBROUGH: Right. They had the problem i
2
!n --
N/ 'l And Kraftwerk Union had to eat 3 MEMBER CATTON:
4 it early on.
5 MR. SCARBROUGH: What we're doing is looking 6 at it now. But right now, see, we don't have the data for 7 U.S. plants because they just finished. In 1996 is when 8 they were finishing up 89-10. Most of the plants finished 9 up in '96.
10 We don't have the experience base of the 11 performance of the valves post the 89-10. And that's what 12 we need to --
, .s 13 MEMBER CATTON: I think if you're waiting for
(" ) 14 the experience basa, you're in trouble because the problem 15 comes about when the valves are exercised in an arena that 16 they were never designed for. And you can't test them.
17 MR. SCARBROUGH: Well, we do have quite a bit 18 of testing that's been done so far, flow testing and, i
19 plus, over the next five years, the industry developing a 20 dynamic testing program to look specifically for j l
21 degradation post-89-10.
22 MEMBER CATTON: Okay. So eventually you may 23 get to the bottom of it, but in the interim, what do you l 24 do with your risk-based regulation if MOVs come up?
(~y ,
V; 25 You're in trouble.
NEAL P. GROSS COURT REPORT..P'.; AND TRANSCRIBERS 1323 RHCO2 ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
282 1 MR. CHEOK: To be fair to some licensees out g-] 2 there, not all the PRAs assumed 10-3 failure rate for MOVs.
l 3 MEMBER CATTON: Okay.
4 MR. CHEOK: I mean, that's allotted.
5 Actually, it assumed 10-2 numbers.
6 MEMBER CATTON: What's the number?
7 MR. CHEOK: Four times 10-2 I recall 8 MEMBER CATTON: Well, the 4 times 10-2 is still 9 way too small. And, to be fair to the utilities, at the 10 outset, before there was any program, the estimate was 11 that it was 22 percent. It was very high.
12 MR. CHEOK: Right.
,, 13 MEMBER CATTON: I think getting it down to the l ') I
# 14 eight percent was just the kind of thing that you're i
15 doing. At the other end, I'm not sure you can do l
16 anything. But it should be reflected in the PRA, and it's i 17 not.
18 MR. CARROLL: Where all this started with you, 19 Ivan, was that somebody comes in and says, " Hey, I want to 20 change my testing."
21 MEMBER CATTON: That equation doesn't work.
22 MR. CARROLL: Yes. Do you understand --
23 CHAIRMAN APOSTOLAKIS: What equation? Which 24 equation?
( ,/ 25 MEMBER CATTON: Well, the one where they take NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
283 j 1 whatever egnation they're going to use for this. I g3 2 thought I knew for a moment, but it didn't last long.
(
~'
)
3 MR. CARROLL: No. But I thought your point i
4 was that you need to show that if you're going to make a 5 change in testing, that the PRA has improved.
6 MEMBER CATTON: Yes. But, you see, unless you 7 do the PRA --
8 MR. CARROLL: Right.
9 MEMBER CATTON: -- right the first time, you 10 can't show that it's improved.
11 MR. CARROLL: Right.
12 MEMBER CATTON: So maybe you want to have 13 another way to deal with problems like this where you
( )
\' 14 treat this device by itself.
15 CHAIRMAN APOSTOLAKIS: They do have -- there 16 was a bullet that they will check the quality of the PRA.
I 17 MEMBER CATTON: Well, that's right. I 18 CHAIRMAN APOSTOLAKIS: And there they will say j l
19 you didn't use the right probability of failure to close.
20 MEMBER CATTON: Then maybe they have moved it 21 up to the eight percent.
22 CHAIRMAN APOSTOLAKIS: Well, we've seen that 23 bullet in all presentations. Do you guys have it, too?
24 Do you check the quality of the PRA?
(3 See, they could have answered
- ()_
25 MEMBER CATTON:
i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
284 l
1 --
.y 2 CHAIRMAN APOSTOLAKIS: You don't?
N 3 MEMBER CATTON: -- the question that way at 4 the beginning, George.
5 CHAIRMAN APOSTOLAKIS: Yes.
6 MR. CHEOK: We do check the quality of the 7 PRA, but I guess I'm not sure that the eight percent is 8 the number that you're looking for at this point yet.
9 MEMBER CATTON: Yes. The 4 times 10-2 is what 10 you can use historically or something smaller.
11 MR. CHEOK: I mean, the other point is that 12 raw number you use for that failure rate would only affect 13 your base number.
p_ )
(d
14 CHAIRMAN APOSTOLAKIS: Yes.
15 MR. CHEOK: It probably will not affect your 16 change in risk. j i
17 CHAIRMAN APOSTOLAKIS: Well, I don't know.
18 Don't make statements like that. Just "might."
19 MR. CHEOK: I'll take that might.
20 CHAIRMAN APOSTOLAKIS: It might.
21 MEMBER CATTON: Good, good.
22 CHAIRMAN APOSTOLAKIS: It might. But it seems 23 to me that there is a mechanism for catching it. First of 24 all, is this eight percent cast in stone?
(_,) 25 MEMBER CATTON: No.
NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
285 1 CHAIRMAN APOSTOLAKIS: I mean, is it really --
2 MEMBER CATTON: No. The Germans actually-did 3 some interesting studies. They did full calculations of 4 the structure of the valve and everything else.
5 CHAIRMAN APOSTOLAKIS: And they calculated the 6 unreliability.
7 MEMBER CATTON: And they concluded that about 8- the best you could do is get it down -- I think the number 9 is eight percent. It might have been six.
10 CHAIRMAN APOSTOLAKIS: It must be significant 11 on something, though, --
12 MEMBER CATTON: Oh, I'm sure there is.
13' CHAIRMAN APOSTOLAKIS: -- if it's calculated.
14 MEMBER CATTON: I'm sure there is.
15 CHAIRMAN APOSTOLAKIS: So there is a place in 16 this process to catch --
17 MEMBER KRESS: And I would assume that, 18 somewhere along the line, this PRA group will have a, 19 quote, " acceptable database" for things like that.
20 CHAIRMAN APOSTOLAKIS: Yes, yes.
21 MEMBER KRESS: And they just compare that with !
22 what's used.
23 . MEMBER CATTON: I think you're being a little ..
24 naive.
25 CHAIRMAN APOSTOLAKIS: No. Come on. {
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. j (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4 33 1 i
i
! 286 1 MEMBER KRESS: Surely it would do that )i i
2 somewhere down the line. j 1
. \. l
'~
3 6 reality?"; that happened.
i 7 MEMBER CATTON: Could anybody in this room 8 name some PRAs that have changed the numbers they use to P
9 reflect what we now know or we have known for several l 10 years about these valves? And I'll bet you'll find the !
11 answer is probably no. ;
12 CHAIRMAN APOSTOLAKIS: Yes. But that's not 13 the issue here. The issue --
i 14' MEMBER CATTON: Yes,.it is.
- 15 CHAIRMAN APOSTOLAKIS: -- is: Will this guys
'16 when they review it say, " Hey, the quality of your PRA 17 here is not very good"?
18 MEMBER CATTON: But there were only two guys ;
19 in this room that knew what I was talking about.
l 20 CHAIRMAN APOSTOLAKIS: I don't.know. Can we l
l 21 have a statistic? How many NRC people know about this? j 22 MR. HOLAHAN: Everybody knows that the 23' reliability of valves is influenced by the binding that's 24 led to 89-10. Sure.
4 i ( 25 CHAIRMAN APOSTOLAKIS: All right.
?
! NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE. NW.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
.. - . , . . - - .- -.~.- _. - -..- . - .-. . . . . . . .. . -.
287 1 MEMBER CATTON: I give up for the moment.
2 CHAIRMAN APOSTOLAKIS: So after these guys 3 reject it two or three times, then the industry will not 4 bind it on its face. 'But, see,.I suspect the reason why
'5 it's not'in the PRAs is because when they use the 6 so-called plant-specific data, they really look at. j
)
1 j
7 statistics. And you're saying this is a result of a 8 calculation._ j 1
9 MEMBER CATTON: Well, no, no, no , no, George. ]
10 You would only gain the statistics if you had a lot of -- j 11 CHAIRMAN APOSTOLAKIS: Yes. 1 1
12 MEMBER CATTON: -- ISLOCAs or something.
13 CHAIRMAN APOSTOLAKIS: Right, right. That's O 14 what I'm saying.
15 MEMBER CATTON: Then you would be in~the 16 statistics.
17 CHAIRMAN APOSTOLAKIS: That's right, 18 MEMBER'CATTON: And God help us if that's the
.19 case. They need to get the numbers somewhere else.
1 20 CHAIRMAN APOSTOLAKIS: But they don't have 21 that.. So the next thing is to'do the calculation. i 22 MR. CARROLL: No, no. They've done --
c 23 MEMBER CATTONi They've done some testing and
- 3. ;
l 24 some. calculation.
- 25 . MEMBER SEALE
- I think Gary's got a comment.
NEAL R. GROSS i' COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISL AND AVE., N W.
(202) 234-4433 WASHINGTON, D C.' 20005-3701 (202) 234-4433
288 1 CHAIRMAN APOSTOLAKIS: These are not n 2 statistical.
! ( )
3 MR. CARROLL: And the testing is -- the 4 statistics were almost 100 percent problems with the 5 valves.
6 CHAIRMAN APOCTOLAKIS: No. But what I'm 7 saying is: In a typical PRA, unless an issue is created, 8 like there was with the isolation valves of the 9 interf acing system LOCA or.es, and the utility spent money 10 to find out what exists, unless such an issue arises when 31 they say " plant-specific data," they go to the plant logs, 12 collect information, so many tests and so on.
13 You're absolutely right. You don't have
/ \
14 information given accident conditions. So they're 15 probably using some sort of generic number which is 16 off-base.
17 These guys will never look for a German study 18 that will --
19 MEMBER CATTON: Georg3, it was far more than 20 that, far more than that.
21 CHAIRMAN APOSTOLAKIS: I'm telling you the 22 practice.
23 MR. CARROLL: It's the Idaho testing.
24 CHAIRMAN APOSTOLAKIS: Well, when these guys
(/ 25 raise the issue a couple of times, then things will NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
l 289 1 change.
,c y 2 MEMBER CATTON: If they do.
. : \
3 MR. CARROLL: No. I think --
4 MEMBER CATTON: That's good. You're pushing 5 to make sure that the PRA has the right numbers for this.
6 CHAIRMAN APOSTOLAKIS: Sure. And all the 7 guidas say that. I mean, for the last six months, we've 8 haen seeing this. In fact, we've been trying to --
9 MEMBER CATTON: Every time I've asked 10 somebody, they have not -- the numbers are not in there 11 yet. You're telling me they are. Okay.
12 MR. CARROLL: What the staff on the end is 13 saying to themselves is: We're going to get good i
- \
i
'd 14 diagnostic tools that will allow us to determine whether 15 this valve or that valve is really going to work. To the i 16 extent we can show that this particular valve just doesn't 17 have a big enough operator on it or whatever, we're going 18 to change it out.
19 So I think thay're saying that the situation 20 is going to improve, but we need a little weight getting 21 the diagnostic tools in place. And now we're pretty much 22 there.
23 MR. SCARBROUGH: Right.
l 24 MEMBER CATTON: And nobody would disagree with f
(}) 25 that, Jay.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS I
1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
290 1 MR. CARROLL: And people have been changing 2 out operators, haven't they?
.O 3 MEMBER BARTON: Yes, quite a bit. Yes, they 1
4 have.
5 MEMBER CATTON: But it's not the operators.
6 It was the' locking up of the gate with the high pressure.
7 That was where the problem was, in part. l I
8 MEMBER BARTON: I think you're talking a :
9 different part. You're talking pressure locked versus --
10 MEMBER CATTON: There are two parts. ;
11 MEMBER BARTON: There are two differe".t issues 12 here.
13 MEMBER CATTON: One is if you put bigger ]
l 14 operators, you could jam it in anyway. And then there was. ;
i 15 a question that: Even if you put the bigger operator on, l
16 are you going to get it done?
17 I don't know where the bottom line came out, I 18 but the Germans.kept saying no matter what they do, it's a 19 number that's about 20 times what they had be using. It 20 would be. interesting to know what they're doing now.
21 MR. CARROLL: Well, they were redesigning 22 their valves.
23 MEMBER CATTON: Yes. With the redesigned ]
I 24 valve, it's not a problem. You can design it to close 25 under'these loads. And ABWR promised that. We've got a l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS I
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 . WASHINGTON, D.C. 20005-3701 (202) 234-4433 1
291 1 lot of existing plants.
,g 2 CHAIRMAN APOSTOLAKIS: Do we have anybody from i
,\ / l m
3 the reviewers of the IPEs here? I 4 MEMBER CATTON: The IPE doesn't even consider 5 this.
6 MR. CARROLL: No, that's not the place to ;
1 7 look.
8 MEMBER CATTON: No. And I don't think this is 9 the place to pursue it anymore. I just bring it up as an 10 interesting issue that at some point --
11 CHAIRMAN APOSTOLAKIS: And my answer to that 12 is that --
l 13 MEMBER CATTON: So what?
p-
? )
14 CHAIRMAN APOSTOLAKIS: -- in principle, there 15 is a place in this process where this question would take l
16 place as long as somebody is aware of it. And evidently 1 17 the NRC is --
l 18 MEMBER CATTON: Well, I listened. And I l 19 didn't really hear it.
20 MR. FISCHER: All right. The next -- l 21 MEMBER CATTON: Nor did I find it.
22 MR. CARROLL: Carl's gone. I'm gone. You're 23 about to be gone. Who's going to be the nagging 24 mother-in-law?
pq
() 25 MEMBER CATTON: The conscience.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
292 1 MR. FISCHER: We thought about waiting until l
,' ' s 2 next month, Ivan. i l
%/
3 (Laughter.)
4 MEMBER POWERS: I believe I'm the designated 5 junkyard dog. ,
1 6 MR. CARROLL: That's different than a nagging l l
7 mother-in-law.
8 CHAIRMAN APOSTOLAKIS: Can you send me a 1
l 9 little note of what the issue is? And I'll make sure that 10 for the next two and a half years --
l 11 MEMBER CATTON: Somebody here should be able 12 to get you all the information. There have been a couple
, 13 of reports by a study in Idaho. There was all kinds of i i
' . But those people are gone, too. They've 14 stuff done.
15 retired, the ones who were responsible for that.
16 CHAIRMAN APOSTOLAKIS: So there's no issue 17 anymore.
18 (Laughter.)
19 MR. CARROLL: That's not necessarily true.
20 (Slide) i 21 MR. FISCHER: Coming back to this presentation 22 here, the next six. slides in your package describe the i
23 risk-informed IST program that I just described a little 24 while ago except these slides describe it in terms of the gx
(_) 25 four-element process that the staff uses for all of the l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
l l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
,. - . . _ . . _ . . . . ~ . _ _ ~ . _ . _ _ . . - . _.__-_ _ ._ _ _ _ _ _ . _ _ _ . . . . _ .__._
293 i
1 risk-informed initiatives where you define the proposed change and you conduct an engineering analysis, 2-7-s
'\s,)
l 3 implementation and' monitoring, and documentation. l 4 I would propose that, with the Subcommittee's' l 5 indulgence, if you flip through those slides and see ;
6 anything that you want to talk about, please bring them l l
r 7 up. Otherwise I'll skip to the two pilot licensee slides that I have, which are Slide 15 if you can see.the numbers a t 8 9 at the bottom. Fifteen and 16 are on the pilot licensees.
10 MR. CARROLL: On Page 11,.second bullet, ,
11 that's just what we finished talking about. In the case l t i
12 of motor-operated-valves, changing the test method would 13- give a -- '
l [\
i
\- '
14 MR. FISCHER: Yes, it is. It's imbedded in i
15 there,. absolutely. And we have words in the reg guide. -
i
.s 16 that talk about assessing'the effectiveness of the test. l 17 CHAIRMAN APOSTOLAKIS: So where do you want us ,
t 18 to go?
r 19 MR. FISCHER: I want you to go to Slide 15 if 20 you can see the little 15 at the bottom. I tried --
i 21 CHAIRMAN APOSTOLAKIS: Okay.
i l 22 (Slide) 1 23 MR. FISCHER: The initial submittals from the i
24 two pilot licensees came back in November '95. Since 25 then, as I mentioned earlier, we have sent them one set of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4 33 l'
I
294 1 RAIs. We've had one set of_ meetings. And we have written 2 a response to'the RAIs -- or excuse me. We have gotten 3 one written response.from the licensees on the RAIs.
4 Some of those written responses that we got 5 from-the pilot licensees were partial submittals. For 6 instance, Palo Verde.said: Well, we can't answer this 7 yet. We want to redo our expert panel. We need to do 8 some sensitivity studies yet. So we'll get you that 9 later. We do have another set of RAIs. It's about to go 10 out to the licensees.
.11 Based on their. initial submittal, you can see 12 that there is a significant number of valves that would be 13 categorized as LSSC for both Palo Verde.and Comanche Peak.
O> 14 And there is the savings that they're lookinglat. )
)
15 Comanche Peak decided to try to categorize ;
16 pumps. Palo Verde elected not to. The numbers that 17 you're looking at for both Palo Verde and Comanche Peak in 18 terms of numbers of components is again what was in their 19 initial submittal.
20 I think the numbers are going to change. The 2 1- numbers are going to change for a couple of reasons. One 22 is they have since the initial submittal decided to add 23 about 40 safety relief valves to their IST program, what l 1
24 the effect of that addition to their IST program is going i 25 to be.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 2344433
295 1 And, by the way, that is based on our review ;
i 2 of their submittal, based on our review of their O -3 risk-based submittal. We say, "What about these valves? h i
4-lHowcomethey'renot.inyourprogram?" and we had a 3 discussion. !
0 'And they've-since come in and added 40 valves -
7 to their program. But they have not.yet classified them 1
8 as LSSC or HSSC. So we don't know how that breakdown is going to affect these. numbers.
9~
10 In addition, in one of our RAIR -- j 11- MR. CARROLL: Are any of these 12 balance-of-plant valves? I guess what I'm trying to j 13 distinguish is from Q list valves, if you will.
./'T :
- 14 MR. COLACCINO: The relief valves that they l 15 added are not. They should have been in the IST program ;
16 in the first place.
17 MR. CARROLL: Thank you. ,
18 MR. COLACCINO: That's good.
19 MR. FISCHER: Another reason these numbers 20 would change is that some of our RAIs are aimed at: What 21 can the pilot licensees do to reduce the increase in risk 22 associated with the proposed change?
1 23 So if you were to look at the numbers down 24 below here and what their estimated changes in core damage f
() 25 frequency are and what their estimated changes in LERF --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C, 20005-3701 (202) 234-4433
296 1 MEMBER KRESS: Is there a reasan you didn't r3 2 include a baseline LERF? l i l' !
%J 3 MR. FISCHER: Because I just forgot to do it.
4 MEMBER KRESS: Okay.
I 5 MR. CARROLL: Do you know what the number is?
l 6 MR. FISCHER: We thought you might ask that. l 7 So it's an oversight that we realized when we were sitting 8 in the back of the room a little earlier.
9 MEMBER KRESS: That's all right.
10 MR. CHEOK: For Palo Verde, the baseline LERF 11 is 2.1-6.
12 MEMBER POWERS: That's for operations, 13 external events, and shutdown; right?
! \
14 MR. CHEOK: That's just for the internal 15 events.
16 MEMBER POWERS: And what good is that number?
17 MR. CHEOK: Well --
18 MEMBER POWERS: That number doesn't have any 19 significance here at all, none whatsoever; right?
20 MEMBER KRESS: Well, it has some.
21 MR. CARROLL: How about the baseline CDF? Is 22 it --
23 MR. CHEOK: The baseline CDF is also for the 24 full power Level I, internal events, yes.
l (8 MEMBER KRESS: It gives you a lower bound Ou l (,/ 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
297 l 1 the --
2 MEMBER POWERS: Well, once you look at it and
('~}
E) It doesn't have external events, 3 you say, "Well, okay.
4 but we've got a lot of PRAs on that, roughly double it for i
5 external events. And we've got a little information on 6 shutdown, not too much. But we know that 60 percent of j 7 the shutdown risk is about 4 times 10-5 So that means 7 J
l 8 times 10'S for total shutdown added to that," we're over l
1 9 the CDF limit. We can't tolerate any increase in risk at {
10 all.
l 11 MEMBER KRESS: You may very well be right.
1 12 That doesn't mean -- i 13 MR. FISCHER: You may very well be right.
l
\ !
'~' 14 (Laughter.)
15 MR. CARROLL: Except I'd question your 16 conclusion about shutdown risk. I think we are all 17 quoting numbers that are --
18 MEMBER KRESS: That are old.
19 MEMBER CATTON: But this is part of Dana's 20 argument for a full-blown shutdown risk.
21 MEMBER POWERS: You forget, Jay, what you --
l 22 MEMBER CATTON: You perhaps don't --
23 MEMBER POWERS: You forget which way I'm 24 coming from on this question.
l He wants you to formally
'(,/ 25 MEMBER CATTON:
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
i (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
1 298 l 1 reduce it. l l
s 2 MR. CARROLL: Well --
)
3 MEMBER POWERS: Well, it seems to me it can't 4 he --
5 MR. CARROLL: I guess even understanding what 6 these numbers are, how do they get to the oCDF?
7 MR. CHEOK: They have their base case. And 8 they did a conservative calculation for all the pumps or 9 valves that they can extend the test frequency to. They 10 basically assume a six-year test interval and a 11 recalculated component unavailability based on a six-year 12 test interval and recalculated the new CDF and took the l l
13 difference. !
(. !
('
14 MR. CARROLL: And how did they go about --
15 MR. CHEOK: They did not take any credit for 16 the savings, so to speak, when they go to longer test 17 intervals as far as unavailability is concerned.
18 MR. CARROLL: And how did they decide what the 19 increased unavailability was going to be of the test 20 interval?
21 MR. CHEOK: They assumed a linear model. They 22 assumed actually a constant failure rate over the six 23 years. They just increased the test intervals. So they 24 basically used a Q equals to lambda T, where they just f
n 1 s j 25 varied the T.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
299 1 MEMBER POWERS: Not a Wyboldt distribution or y 2 something like that?
b 3 MR. CHEOK: No.
4 MR. CARROLL: Not very sophisticated.
5 MEMBER POWERS: No.
6 MEMBER CATTON: In the PRA, what did they use 7 for the basic valve data: Just the EPRI data or something 8 or the same as 1150 or what?
9 MR. CHEOK: I believe that both plants used 10 generic data.
11 MEMBER CATTON: Okay. That's what I was 12 afraid of.
13 MEMBER POWERS: Well, again, I'm --
I 'I
'd 14 MR. CARROLL: Well, now that we have decided I
l 15 we're going to ignore those numbers --
16 MEMBER CATTON: It doesn't matter. l l
17 CHAIRMAN APOSTOLAKIS: Is there a document ;
18 with the details of these calculations?
19 MR. CHEOK: We have the submittal with us that 20 we could --
21 CHAIRMAN APOSTOLAKIS: Yes. I would like to 22 see that.
23 MR. CARROLL: But you're telling us they are 24 revising this. But I guess only to reflect the additional
,r'\
j 25 valves or what?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
300 1 MR. HARDIN: They are taking another hard look ex 2 at their whole PRA and trying to see where the things
, r LJ might be fine-tuned. And we'll see that reflected in 3
4 their RAI responses, I suspect.
5 MEMBER POWERS: Brad, it seems to me that 6 you've got a real problem here. It's not a matter of 7 fine-tuning the PRA. It's a matter that the PRA simply 8 does not reflect the number that you need to -- I mean, on 9 the face of this, this application has to be rejected out 10 of hand.
11 MR. FISCHER: If it were purely based on the 12 bottom line PRA, I would agree with you. If we were going 13 to base our decision totally on the number at the change
'_ \
14 in CDF or change in LERF, I might agree with you.
15 But I think we're basing it on an overall 16 program that has some risk insights and has features such 17 as performance monitoring, feedback and corrective action, 18 that I feel we could reasonably come to the conclusion 19 that the alternative will provide -- and including 20 improved testing strategies and some additional components 21 that we decided are highly safety-significant that aren't 22 in the current IST program.
23 I think all of that together kind of would
- 24. allow us to make a finding that the licensee's proposed p)
(_ 25 alternative would provide an acceptable level and quality NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS !
1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
301 1 of safety, even though the bottom line PRA number may not ry 2 be -- I mean, this isn't a risk-based program, a i )
<j 3 risk-based program. It's a risk-informed, 4 performance-based program.
5 CHAIRMAN APOSTOLAKIS: So what you're saying 6 is that this requirement that plants with core damage 7 frequency above the goal shouldn't even bother to submit a 8 request for relief? That's not really a strict 9 requirement because based on other considerations, you can 10 just allow --
11 MR. CARROLL: Theoretically requests for 12 relief could be on an improvement.
_ 13 CHAIRMAN APOSTOLAKIS: No. But the statement t )
'~' 14 now in the documents is that you assure above the goal 15 either on the CDF or the LERF the only -- oh, you are 16 saying theoretically --
17 MR. CARROLL: Could be risk-neutral or --
18 CHAIRMAN APOSTOLAKIS: So there are 19 unquantified benefits here that overwhelm this -
20 MR. FISCHER: If you looked at this next slide 21 here, the very first --
22 CHAIRMAN APOSTOLAKIS: Well, then let's see 23 that.
24 (Slide) 1 i
( n
- ( ,/ 25 MR. FISCHER
- -- the very first bullet on NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
302 .j there is the staff is going to ask the licensees to try to ;
1 ,
l' 2 describe methods that they could improve their L~ 3 determination of the baseline CDP and core damage l
4 frequency.
5 We would like to get some improvements to 6 reflect some of the improved testing th_at they might do; 7 to take credit, even if it's not quantified but we need ;
i i
l 8 some more information from the licensees, how things will' ;
9 improve; and then also to identify the' major contributors 10 to'the change in core damage frequency and LERF and maybe ;
~!
not include some of those as candidates for the extension. 't l 11
!' 12 So we're hoping that we can work with these pilot licensees to kind of document-some of-the l
j 13 l
14 improvements and also maybe not extended' intervals on-some !
15 of the components that.are the more significant ones that ;
! l 16 could affect the bottom line as much. J
~
-17 CHAIRMAN APOSTOLAKIS: But then'it seems to me L 18 that the basic documents, like the regulatory guide 1061, 19 should state this fact and that this is a risk-informed l-l 20 program and there may be unquantifiable benefits that can' L
21 overwhelm the calculated increase in risk. So, even for 22 those plants that are above the goal, they may have a l l
, 23 case.
- 24 That's really what you said. And that's not 25 what the document says'right now. And I'm not saying that
-l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W, (202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
__ .._ _ . . . _ . . . . _ . . .m. .. . . - _ . _ _ . - . _ . _ . _ . _ . _ . . . . _ _ _ _ _ . , _
j 303
- i 1 this.is not a reasonable position. All I'm saying is that 2 we're finding an inconsistency here between what the
!( 3 document says unless there is no inconsistency.
[
.4 MR. KING: No. I think the document does t 5 touch on-that point and does say, even if you are above ;
6' the goal, there are still good reasons that you may want 7 them to be able to allow some risk increases in -- !
i 8 CHAIRMAN APOSTOLAKIS: Well, let's find that, i
,6
(-
l 9 Tom, because I'm not sure it's there. ;
l ?
t 10- MR. KING: Let me find it in the latest i l
L 11 version. We'll see if it's in your version.
I
! 12 MR. FISCHER: And I have to apologize a little i-l 13 bit because I was reacting to Dr. Powers' inference that :
l 14 everything had to be -- you know, they had to have a PRA !
l
[. '
i 15 that was a full-scope PRA in every' detail and his 16 extrapolation of, well, then it.'s twice as much and it's I
17'way the heck over your safety goal limit. I was just --
18- my answer was somewhat of a reaction to that.
19 CHAIRMAN APOSTOLAKIS: No. But the point is
-20 that we have been operating under the assumption that this 21 was the total LERF and the total CDF from all modes, --
l-22 MEMBER CATTON: It is.
l 23 CHAIRMAN APOSTOLAKIS: -- all initiators.
24 MR. FISCHER: It is supposed to be.
25 CHAIRMAN APOSTOLAKIS: It is?
I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 I ,
304 1 MR. FISCHER: Yes. But we also recognize that m 2 --
/ \
() 3 CHAIRMAN APOSTOLAKIS: Therefore, Dana is 4 right.
5 MR. PISCHER: Yes.
6 MR. HOLAHAN. Right.
7 MR. FISCHER: Everything may not be 8 quantified.
9 CHAIRMAN APOSTOLAKIS: That's a different 10 story. That's a different story.
11 MR. CARROLL: -- quantify some of the things 12 that are left out of the full-power question, George.
13 MR. FISCHER: Or everything may not need to be
'd 14 quantified for every application.
15 CHAIRMAN APOSTOLAKIS: Let me read the thing ;
1 16 because I found it. It's on Page 2-7 of the January 17th 17 ledger, "For a plant with a mean core damage frequency at l 18 or above 10 per reactor year, the Commission's subsidiary 19 goal, or with a mean LERF at or above 10-5 per reactor 20 year, it is expected that applications will result in a 21 net decrease in risk or be risk-neutral."
22 Now, does that imply quantify the 23 non-quantified risk?
24 MR. HOLAHAN: It implies everything you can
,r m
) 25 bring to that judgment, yes.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
1 305 1 MR. CARROLL: I guess it would interpret it as w 2 meaning what you --
1
\ ) '
3 CHAIRMAN APOSTOLAKIS: The number, yes. The I 1
4 number.
l 5 MR. JONES: And if you go to the beginning on l
6 Page 2-6, I think it deals with -- the first paragraph l 7 talks about the fact that some applications might use risk 8 importance bounding estimates. Qualitative assessment may 9 be sufficient. So I --
10 CHAIRMAN APOSTOLAKIS: Where is this?
11 MR. JONES: First paragraph in 2.4.2 on Page 12 2-6.
13 CHAIRMAN APOSTOLAKIS: Right. What line?
i :
' 14 MR. JONES: Right at the bottom of that first 15 full paragraph.
16 CHAIRMAN APOSTOLAKIS: "A qualitative 17 assessment of the impact on the plant risk may be 18 sufficient."
19 MR. JONES: So, I mean, we left --
20 MR. CARROLL: Yes. But then when you go on l 21 and read this and it's talking about the concept of 22 risk-neutral or --
23 MEMBER BARTON: Ten sixty-one.
I 24 MR. FISCHER: Yes. The real risk page --
,m l \
_j 25 MR. CARROLL: -- risk-neutral and so forth, it NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
-306 L 1 seems to me you're not talking about qualitative. You're l
l 2 talking about quantitative.
l i-3 CHAIRMAN APOSTOLAKIS: Bill?
4 MEMBER SHACK: Page 2-11, Paragraph 2.4.3, 5 where you come to the great integrated decision-making i
! 6 paragraph, "In this way, it can be seen the decision is l 7 not driven solely by the numerical results of the PRA."
8 CHAIRMAN APOSTOLAKIS: "None of the individual
( 9 analyses is sufficient in and of itself." I like this.
10 MEMBER POWERS: But, on the other hand, let me L 11 point out that this is all just fooling yourself, that I'
12 you're saying, "Okay. These are these unquantified, t
! 13 unknown benefits that are going to come to me that are 14 going to counter-weigh the fact that I'm exceeding the i
15 safety goal."
i i 16 You don't even know how much you're exceeding l
17 the safety goal.
1 18 MR. FISCHER: It doesn't. matter.
19 MEMBER POWERS: You have no idea. I mean, 20 you're not even using the numbers. You've got no attempt l
L 21 to quantify what you can quantify. So how can you bring 22 that -- you bring erroneous information into the
- 23 decision-making process and try to counter-balance that
. 24 with this qualitative information.
!A j h 25 MEMBER POWERS: I'm not sure that you're NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS i
1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
\
307 1 helping yourself here.
2 CHAIRMAN APOSTOLAKIS: That's a key issue and
/]
(O 3 maybe we can discuss it more.
4 MR. PRATT: Well, remember how the guideline 5 works. It doesn't matter how far over the safety goal 6 guideline you are, so long as you are over or you're not 7 sure whether you're over or not, but if you think you're 8 over, you ought to be looking for risk neutral or an 9 improvement.
10 CHAIRMAN APOSTOLAKIS: Okay.
11 MR. PRATT: And if you can make a judgment 12 that this is a net improvement, I think it doesn't matter 13 how far over you are, because what it says is your
()
14 judgment is you are moving in the right direction.
15 CHAIRMAN APOSTOLAKIS: And that's what you 16 told us a few minutes ago?
17 MR. PRATT: Yes sir.
18 CHAIRMAN APOSTOLAKIS: Let's hear from Rich 19 Granton.
20 MR. GRANTON: Rick Granton, South Texas. Just 21 I heard a couple of things here. I want to try to give 22 you a little utility perspective here on these base CDF, 23 delta CDF numbers here.
24 When you look at the base CDF number, we're C) 25 usually talking at power mode 1, core damage frequency on NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. ,
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 )
308 l l
We're talking shutdown risk. It depends j 1 an annual basis.
2 on what you're talking about. Are you talking about the 7-3 incremental core damage frequency during shutdown or are ;
t 4 you talking about an average annualized value?
5 Because if you are, then you have to take a ,
6 l'ook at the amount of time you're at power in a year, plus 7 the time that you're shut down in a year which is a small ,
~
8 fraction and if you talk abut transition mode, there's l l 9 that little teeny fraction of. time that you're in a
10 transition over the year.
11 When you add all that up, and if you're a good f
12 operating plant, you're dominated by the at power CDF.
i 13 These other numbers that you're throwing in here on the ,
14 shutdown risk and the transition risk are there and !
l 15 incrementally when you happen to be in those !
i 16 configurations, they're real and they need to pay more I 17 attention, but when you look at it on average annual l-l- 18 basis, they're very, very small amounts of the total CDF.
i 19 So I think you can make good baseline L 20 decisions-on the baseline CDF. You do have to take a look l 21 at considerations when you are shut down because there 22 will be some differences.
23 And with regard to the MOV reliability issue
- 24 there, when we model the PRA, we take a look at what does 25 the. system have to'do, does the check valve have to open?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 FHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, O C. 20005-3701 (202) 234-4433
\ \
309 i i
1 Does the MOV open against the pump pressure that's l
\
2 supposed to open against?
I 3 And under severe accident conditions, there l 4 are issues about a valve that maybe has to go close under l 5 full reverse flow, for which it wasn't designed. We give l 1
6 very little credit, if any at all, that a valve succeeds 7 under any of those kind of conditions, but when we talk l
l 8 about MOV failure rates we're looking at what does the l
' 1 i
9 component have to do to achieve its function and that's
! 10 where the failure rates come from, from the generic data.
11 Where surveillance data test intervals are 12 looked at, at a similar way, you have gains and l l
13 unavailability that are associated with that by the lack
% 14 of down time due to testing and/or maintenance associated
'15 with that. You relax the surveillance test intervals, you 16 usually see some gain in there. Whether you see an 17 increase in the failure rate or not,.it's something that 18 as was brought out earlier, only_ time and data will tell 19 you that those kind of things, but that's why you collect .)
20 plant specific data and update your PRA, 21 ~ 'Now you look at other MOVs that are normally 22 open. They only have to stay open during accident ;
i i
23 condition. What's the MOV failure rate I've got to use on l 24 that? What's the proper test for that in those kinds of ;
25 situations?
NEAL R. GFU3SS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 - WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 310 l
1 So PRA information needs to be looked at, not 7x 2 only for establishing what the testing surveillance l :V T 3 frequency is, but what's the appropriate test strategy.
4 There's no point in testing or dynamically or statically 5 testing of MOV, that all it has to do is remain as a i
6 passive piece of the pipe during an accident. I'm not 7 gaining anything for that.
1 8 MR. CARROLL: The staff argues that the Davis- l l
9 Besse incident, I guess it was, suggests that you are 10 gaining something if you inadvertently close that, the 1
11 Davis-Besse and then try to open it.
12 MR. GRANTON: If you look at inadvertent 13 situations that's a human error type of situation of those
\ 14 points.
15 MR. CARROLL: It's also a valve problem. If j 16 you have to worry about it, if the valve works.
17 MR. GRANTON: Granted, there may be 18 situations, like I say, unique conditions and situations 19 where a valve has to do something for which it wasn't 20 designed for. And the way I'm trying to explain here, the 21 way we're modeling things at PRA, we're looking at what 22 the system is supposed to do. There are many many other 23 conditions that can occur to a valve under certain 24 specific conditions for which it's not going to be a
- ,m\
/
( ,/ 25 performance intended function.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
.__._..m_, . m._. _ . _ _ _ . _ . _ . _ _ _ . _ . . _ . _ . _ _ _ _ _ _ . . _ . _ _ _ . _ _ _ _ _ . _ . .
I 311 1 CHAIRMAN APOSTOLAKIS: Let me understand p 2 something here because I don't remember. I should 3 remember, but I don't. When people do a shutdown PRA, not 1
4 too many do, but when they do, and you see in the papers,
! 5 for instance of the confidences that the core damage l'
6 frequency they calculate is comparable, you know, not 7 exactly the same, but comparable with the core damage. .
8 frequency for power operation. .;
1 9 MR. CARROLL: Per hour. l i
l 10 CHAIRMAN APOSTOLAKIS: Per year. Per year, i
I 11 that's what they report. Is that assuming that you are on l 12 a shutdown mode for the whole year? Or has it been 13 weighted down already by the fraction of time you are in 14 that mode?
15 MR. GRANTON: I think what happens at times it i
! 16 looks like it's reporter for the whole year, but t's not 17 weighted for the fraction of time that you're actually 18 shut down.
19 CHAIRMAN APOSTOLAKIS: So for the plant --
l 20 MR. GRANTON: For that entire year. I believe i l';- I
- 21 that's the way.it gets reported a lot of times. !
1: i 22 . CHAIRMAN APOSTOLAKIS: Okay.
23 MR. GRANTON: But you're not calculating what 24 the true fraction is.
25 MR. PRATT: I think that's right but it's not i
I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
, (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 I
t
312 1 really as though the plant were the whole year. It's l
2 really as though the plant were seven days into an outage
(-]
U' 3 for the whole year, something like that.
l l 4 MR. GRANTON: Right, what mode you're in 5 CHAIRMAN APOSTOLAKIS: What, what, what?
6 Seven days?
7 MR. GRANTON: Because of the transition mode.
8 CHAIRMAN APOSTOLAKIS: Oh, oh, oh.
9 MR. GRANTON: You know, they're in stable RHR 10 mode cooling.
11 CHAIRMAN APOSTOLAKIS: But it's essentially an 12 unrealistic number.
13 MR. GRANTON: Yes, that's why I say you have 7
/
\V) 14 to take a look at the fraction of time you're in those 15 conditions. l 16 MR. PRATT: It's a rate. It just happens to 17 be --
18 CHAIRMAN APOSTOLAKIS: And that's the l
19 frequency. You just --
20 MR. PPATT : -- the frequencies just happen to 21 be reported in the units per year.
22 CHAIRMAN APOSTOLAKIS: For a long time, by the 23 way, some theoreticians in reliability and decision 24 analysis have been arguing that you can't regulate on the o
l I )
C/ 25 basis of rates. And we have ignored them. But finally, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
313 l
perhaps, we are reaching the point where we're recognizing l
l 1 But I think, in my mind, I
(-)
, a 2 that they may have a point.
U' 3 maybe not in yours, in my mind what's still unresolved l
t 4 here is how do we combine these unusual configurations l
5 with the normal configuration, power.
6 Now one proposal last time and I think we had l
l l 7 a little debate, but we had to finish it, was to use l 8 simply the fraction of time that you are at this mode and c
9 multiply that by the calculated frequency, then add it to 10 the others.
11 Another approach which I think is what you are 12 doing today, what was presented earlier today was to 13 actually regulate the core damage, based on the core o
i 1
14 damage frequency in LERF during that phase. That's what 15 you did with the allowed outage times, right?
16 MR. PRATT: In addition to looking at the --
i 17 CHAIRMAN APOSTOLAKIS: I think we need to )
l 1
18 resolve that before we -- because I think Dana's point 19 will be -- if we agree to weigh these things according to J 20 the fraction of time that the plant is in that mode --
l 21 MEMBER POWERS: George, I believe the numbers 22 that have been shown you, have been annualized, that is, l
l 23 they are already weighted.
24 CHAIRMAN APOSTOLAKIS: They're a1 ready
,m h 25 weighted?
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 23 4 433 WASHINGTON. O C. 20005-3701 (202) 2344433
(
1
314 1 MEMBER POWERS: We still have significant m\ 2 fractions here, both for CDF and for real risk measures.
'J 3 CHAIRMAN APOSTOLAKIS: And the extent risks 4 for sure are definitely numbers that should be added so --
5 MEMBER CATTON: They can wait if they've been 6 additive and Dana is right.
7 CHAIRMAN APOSTOLAKIS: No the extent of 8 events, there's no question about it, the other --
9 MEMBER CATTON: The other modes. If what 10 we're hearing is correct, then it should be divided by a 11 factor of 10 before it's added.
12 CHAIRMAN APOSTOLAKIS: No, no. The other 13 point that -- the unquantifiable benefits may be such that
,/m 14 even a 5 point something, 10 to the minute 6 core damage 15 frequency really is not a negative impact because you have 16 these other benefits.
17 MEMBER POWERS: But George, you've got to 18 change substantially the way you're describing this 19 process. You've got to say do your calculations any way l 1
20 you want to, it doesn't matter, as long as you can come in l 21 with some argument that I will buy that says that these l l
22 mysterious unquantified benefits outweigh -- I mean, I I 23 don't know even how you're going to write the document.
1 It's going to be an interesting document to read.
24
(,) 25 But it's going to be a very different document NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433 J
315 1 than what we have.
m 2 CHAIRMAN APOSTOLAKIS: And a difficult one to f
I \
'L.)
3 attack.
4 MEMBER POWERS: That's right because it will j 5 rely -- it becomes strictly a judgment-based process. It i
6 doesn't matter what number I calculate. It doesn't even l 7 matter if I calculate, because all the judgments made 8 based on how persuasive you are in changing your language.
9 CHAIRMAN APOSTOLAKIS: Are you taking the 10 position though that unless we do a complete risk .
I 11 assessment for all the modes, we can't do anything else?
1 12 That we can't implement --
13 MEMBER POWERS: No. I think we need the kind 1
,m I
('l 4 14 of understanding that we have with power operation that 15 allows us to do things like Mark discussed with us this 16 morning, where we can do LERF, even if we can't do a Level 17 2 PRA.
18 We don't have that level of understanding 19 right now.
20 MEMBER FONTANA: Anyway, delta is more 21 credible than an absolute number anyway, isn't it?
22 They're talking about difference --
23 CHAIRMAN APOSTOLAKIS: No, no, but you start 24 with an abselute number.
q
(,) 25 MEMBER FONTANA: I know.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 316 l
1 CHAIRMAN APOSTOLAKIS: You know the goal, then
,e-~x 2 --
! )
<.s 3 MEMBER FONTANA: Here, you might be above the 4 goals, but they're doing a calculation to show you what 5 the difference is of the change that you're going to do 6 and I would think that the calculation from a difference 7 would be more credible than calculation from a bottom 8 line.
9 CHA1RMAN APOSTOLAKIS: I never bought that --
10 MEMBER FONTANA: A bunch of mathematical 11 calculations. They're all the same except for this one 12 over here.
13 CHAIRMAN APOSTOLAKIS: The absolute value does
- 7. .s i \
14 count.
15 MEMBER FONTANA: I didn't say it didn't count.
16 I'm just saying that delta is more --
17 CHAIRMAN APOSTOLAKIS: No, I don't believe --
18 first of all, any time you take the difference of random 19 variables, there are certain increases. So how can the 20 difference be better?
21 MEMBER FONTANA: That's part of the argument.
22 I'll argue with you later.
23 CHAIRMAN APOSTOLAKIS: We don't have that I
24 level of understanding right now. If that's your point,
~
i \
C/ 25 Dana.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
_ _ _ _ . . _ _ ~ _ . - _ _ . . _ _ _ _ . . _ . . _ . _ , _ . - . _ _ _ . . _ _ _ _ . . _ . - _ _ . .
317 i 1 My question is until we reach that level of ,
i 2 understanding, we should put everything we've been 3 discussing today and we will discuss tomorrow on hold. :
4 MEMBER CATTON: I don't think so. We saw a i 5 chart this morning that gives you the structure to go ,
6 ahead, because'you have the right hand side which is more 7 judgment and you have the left hand --
l 8 CHAIRMAN APOSTOLAKIS: What structure is that? ..
t
?
9 MR. CARROLL: What is left and what is right. ;
10 MR. JONES: The review process, one that has ;
11 50.59 and whether you use risk informed, current licensing l
12 basis, that kind of -- ,
13 CHAIRMAN APOSTOLAKIS: These particular risk- !
- O 14 informed approach will have to be placed on hold.
r 15 MEMBER CATTON: Why?
16 CHAIRMAN APOSTOLAKIS: I think numbers are not 17 representative.
18 MEMBER CATTON: I was out there and the people 19 clean it up and really do a good iob. You sort of move a 20 little bit to the left.
21 MEMBER POWERS: What we have to do, George, is
~
22 real simple. Just admit that you're doing core damage l.
l 23 frequency at power operation informed regulation.
l-l 24 CHAIRMAN APOSTOLAKIS: Yes. !
25 MR. JONES: I will comment that we do have -- j NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
318 1 CHAIRMAN APOSTOLAKIS: Dana's point is that j
,e-s 2 that the numbers you are using for CDF, base line CDF and i >
'x.~/ LERF are not the numbers, are not realistic numbers 3
4 secause they don't represent the full modes of --
5 MEMBER CATTON: Yes.
6 CHAIRMAN APOSTOLAKIS: But the moment you say 7 that, all this guidance here about the neutrality and 8 where you are and where you are not, goes out of the 9 window, because the number you have is not realistic.
10 MR. ABBOTT: All I'm thinking at this point is 11 that if I look at the Comanche Peak pumps and there are 12 33. Twenty-one at one category and 12 on the other, as ;
13 part of that process, I look at the significance of those ,
I t
) l b '/ 14 pumps and the overall risk in the PRA. I may not have 15 looked at the total number, but I understood the 16 importance of those pumps from a risk standpoint by 17 looking at the PRA.
18 It may not have involved any number crunching 19 in the end. It's going to have to be a combination of 20 both. The risk assessment tells you something and it 21 doesn't necessarily mean you do something based on the 1 22 numbers that the thing cranks out. I mean it tells you --
23 it's the only integrated model of the plant you have. It 24 tells you where the sump pumps are more integrated --
(,,I 25 CHAIRMAN APOSTOLAKIS: But it's not integrated NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N.W.
(202) 234-4433 WASHINGTON, D.C, 20005-3701 (202) 234-4433
319 1 enough. That's his point m 2 MR. ABBOTT: I don't know.
1 3 CHAIRMAN APOSTOLAKIS: Take it from 0 to 5 4 when it's supposed to go to --
5 MR. ABBOTT: The only point I'm trying to make 6 is the concentration of the no risk at this point is 7 skewing the argument away from risk-informed decisions.
8 MEMBER POWERS: You see the alternative that 9 you could approach is to say show me your PRA numbers for 10 that part of the plant's operational mode that's 11 pertinent. I mean if this pump is not involved and the 12 shutdown scenario is et al., then I don't care what its 13 risk contribution is.
> l V 14 MR. ABBOTT: I agree with that.
l 15 MR. CHEOK: And that's what the pilots do l
l l
16 already.
l 17 MEMBER POWERS: That is not what you're doing 18 in your guidance.
19 MR. JONES: And I would argue that we are 20 doing that in the guidance document, Dana, specifically on 21 page 211 on lines 10 to 19, talking about what you do when 22 the PRA is not full scope, what do you do with modes of 23 operation that hasn't been analyzed, how you treat th,em, 24 qualitatively, quantitatively, whether they're impacted or 7
(,/ 25 not.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l l (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
t
320 I 2
1 Obviously, the onus is on the licensee to 2 develop this and it's also obvious from the discussion l
3 here that we haven't gotten that far in the pilots. I 4 mean, some of this framework, as I said earlier today is l 5 developed over the last basically three to six months and 3 L 6 we recognize as we look back at the pilots that_we had to j 7 deal with how we were going to deal with 5 times 10 in the
[- 8 minus, 6 10 to the minus 5 plant baselines.CDF, how are we 9 going to deal with the one time 10 to the minus 4 issue on i 10 CDF? How are we going to go through that decision making !
r 11 process? We haven't done it yet. It's one of the items i
12 that has to be dealt with. i 13 MEMBER POWERS: Well, the other thing I ;
t
% 14 encourage you is don't force me to become a Talmudic 15 scholar to understand what you're saying in your guidance. I 16 MEMBER FONTANA: Well, the extent to which the t
- 17. external event or a shutdown operation affects the 18 behavior in that particular component,_then you have to '
19- consider it. Right? Because you're considering a chain.
20 You have considered the possibility of an external event.
21~ That's the only reason, the bottom line of whether you've 22 got external events is not that important, I don't think, ;
i 23 to start with.
24 MR. CARROLL: But in the context of the 25 document, at least some are important and all above.
i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
321 1 MEMBER FONTANA: Yes. Now, I remember reading l
,r's 2 some words. I don't know exactly where that it gives a f I
%.)
3 little guidance what to do with respect to seismic 4 considerations somewhere.
5 MR. KING: General guidance?
6 MEMBER FONTANA: Somewhere in this stack of 7 papers.
8 MR. FISCHER: It's not in the IST one?
9 MR. KING: I don't recall that in the general 10 guicance.
11 MR. JONES: I think there's only one comment 12 on seismic in the general dealing of how we model seismic 13 --
( )
14 MR. KING: NUREG-1602 which has a separate 15 section on seismic and a section on fire and so forth.
16 MR. DINSMORE: This is Stephen Dinsmore from 17 NI. It might -- what you're thinking about might be in 18 the graded QA where we asked them to review their such 19 seismic and shutdown studies, qualitatively, to identify 20 equipment which might be credited to respond to these 21 sequences, but which is not in the numerical part of their 22 submittal.
23 MEMBER FONTANA: Okay, so you did have some l
l 24 words in there?
/O Yes, there's a paragraph.
C) 25 MR. DINSMORE:
I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
! 322 1 1 CHAIRMAN APOSTOLAKIS: So where does this 2 leave us?
(~]
V MR. HARDIN: I was just going to make another 3
4 comment that as Dave said earlier, these are preliminary 5 numbers up here and it wasn't intended that you think that 6 we were presenting these as if they were final and that we 7 had already made some decision about them. The licensee 8 is probably going to change the numbers and as Bob just 9 mentioned a minute ago, the very complex issues of dealing 10 with the incomplete scope is something that we will have 11 to deal with when we write an SER for these two pilot 12 plants and as Dana has mentioned, there are a lot of 13 things that represent loose ends that we're going to have
[\-
14 deal with and we'll have to come up with a defendable 15 argument for whatever position we take.
16 CHAIRMAN APOSTOLAKIS: But isn't it true 17 though that certain things are -- you look at them, you do 18 some quick calculations and you see that's a good thing to 19 do, even though I don't have all the calculations. I 20 don't have the complete picture. How do we define those 21 situations? How do we make sure that -- and these 22 decisions make sense, right?
23 MEMBER CATTON: It's going to be judgments, 24 George.
i%
k- 25 CHAIRMAN APOSTOLAKIS: Even though you don't NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
323 1 have the complete picture, es 2 MEMBER CATTON: You don't have the complete f )
v' That's engineering.
3 picture. It's judgment.
4 CHAIRMAN APOSTOLAKIS: But that really changes 5 in my --
6 MEMBER CATTON: No, I think it's within the 7 structure that they put up that you can do this.
8 MR. PRATT: I'm less uncomfortable ~with this 9 circumstance than other people seem to be, but it seems to 10 me that in the absence of having these numbers and 11 analysis limited or otherwise, these kind of decisions 12 have been going on for a long time. They had been made in 13 a less informed environment and the fact that you didn't
/ \
t
\- 14 know where you were with respect to CDF or large area 15 release frequency and didn't really think quantitatively 16 or qualitatively about whether the risk was going up or l
17 down, seems to me we're all the better off, but now we're l 18 forcing ourselves to think about these things and what 1
19 you're finding out is the first time you do it, you feel a i 20 little uncomfortable because some of the things we're 21 saying says that the decisions we made in the past -- j l
l 22 MR. CARROLL: Weren't very good.
23 MR. PRATT: Well, not necessarily that they 24 weren't very good, but in some cases they're questionable, n
Some of them are close to the line about things
(_) 25 okay?
NEAL R. GROSS COURT REPORTERS AND TRAMSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
324 1 that we would have to say well, I have to be, think about q 2 this one hard and be convinced about it and have some N.) 3 monitoring program to make sure that I didn't make a 4 mistake. So all we're doing is, I think we're making 5 smarter and maybe more cautious decisions than we did 6 before. I don't think it's bad. I think it's good.
7 CHAIRMAN APOSTOLAKIS: Now again though, if I 8 follow the guidelines we have here, you're saying that if 9 you are close to the goal, okay, then you will have the 10 scope, quality and robustness of the analysis, including 11 consideration and quantification of uncertainties would be 12 subject to increased NRC technological management and 13 review.
14 MR. PRATT: Right.
15 CHAIRMAN APOSTOLAKIS: The base of the plan, 16 the cumulative impact blah, blah, blah, blah, so in this 17 particular case of this plan, you would do these things, 18 right?
19 MR. PRATT: If I believe those numbers and the 20 fact is that my reaction those numbers, the staff brought 21 me those numbers for additional management attention, I 22 would send them back to the licensee. I have a hard time 23 believing that the kind of changes we're talking about 24 comport with those numbers. But if they did, I would read n
25 those numbers a little differently from the way that Dr.
()
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. (
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
325 1
1 Powers read them. I would read them to say this is sort 73 2 of marginal.
( !
%J 3 The analysis you did is below the guidance 4 number, but there are things that are left out, so maybe 5 you're at least pretty close and you ought to be careful 6 about what you're doing. And the deltas are -- well, the 7 numbers are positive, but they're awfully small and in 8 fact, something times 10 to the minus 7 is not very 9 distinguishable from o. Okay? And so we ought to --
10 MEMBER POWERS: But you do distinguish in your 11 guidance on that.
12 MR. PRATT: Sorry?
13 MEMBER POWERS: You do make a distinction in i /
14 your guidance. I won't argue with you.
15 MR. PRATT: We made the distinction in our 16 guidance in terms of a guideline, but in terms of a rigid 17 decision criteria. So what that guideline says is I got a 18 flag that says this is a small amount above zero and we 19 have to look at all of the implications of that very 20 cautiously before we make a decision. And the net effect 21 is this could very well turn out to be something that gets 22 approved but gets approved with a more thoughtful analysis 23 and more management attention than something similar would I
l 24 have been done'in the past. I guess that's okay with me.
im
- (,,) 25 CRAIRMAN APOSTOLAKIS
- So then, what I think NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
326 1 you are saying is consistent with the principles, but I g3 2 don't want to say it's inconsistent, but it's not clear l L) 3 it's consistent with the acceptance guidelines. The I
4 acceptance guidelines are stated in Section 2.4.2.1. I 5 Perhaps you can put some extra words to indicae what you I 1
6 just said. But --
7 MEMBER KRESS: Put a figure in?
8 CHAIRMAN APOSTOLAKIS: What?
9 MEMBER KRESS: Put a figure in?
10 CHAIRMAN APOSTOLAKIS: A figure won't help ,
1 11 you. Unless you use a fuzzy line. i l
12 MEMBER CATTON: George, what do you think 13 about that figure they had up there this morning? l
()
'_ 14 CHAIRMAN APOSTOLAKIS: Fuzzy lines, fuzzy 15 lines. I don't remember that figure. That impressed you l
16 a lot.
17 MEMBER CATTON: Yes, it did.
18 CHAIRMAN APOSTOLAKIS: In terms of the 19 principles, I think it's consistent because you're saying 20 proposed changes both individually and cumulatively and I 21 don't have the new version, result in no more than 22 significantly increases. That was the early version.
23 MR. PRATT: It says the increases would be 24 small.
7
(_,) 25 CHAIRMAN APOSTOLAKIS: Yes, but nowhere does NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
327 1 it say that these -- you see, I think most people when g^ 2 they read these, they will interpret it in terms of the i
L 3 numbers that you calculate and you are saying that's not 4 really the whole story. l 5 MR. PRATT: Well, we establish principles and 6 expectations and guidelines and a decision making process.
7 CHAIRMAN APOSTOLAKIS: Yes 8 MR. PRATT: And I think anyone who picks up 9 this document and decides I'm going to quote one part of 10 one sentence is not going to tell the whole story that 11 we've tried to lay out.
12 CHAIRMAN APOSTOLAKIS: Yes.
- 13 MR. PRATT: These are difficult decisions that
(
)
t 14 have got lots of aspects to them and I think the decision ,
1 15 has got to be made looking at the whole process. You i
l 16 can't just stop part way through and say I failed one test .
I 17 or I passed one test and that's going to be the go or no 18 go decision.
19 CHAIRMAN APOSTOLAKIS: I think 2.4.3 really is 20 the key for this on page 11, 2.11, integrated decision 21 making.
22 I hope these -- I don't remember -- these 23 folks expressed up front some place --
24 MEMBER BARTON: Where are you, George?
O
(_/ 25 CHAIRMAN APOSTOLAKIS: Regulatory Guide 1061, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 2000 53701 (202) 234-4433
328 1 page 211. l l
,s 2 MEMBER BARTON: 211, okay.
, ) ,
%.,/ \
3 CHAIRMAN APOSTOLAKIS: Integrated decision 1 4 making.
1 5 MEMBER SHACK: Paragraph 2.4.3.
6 MR. KING: But where those thoughts really 7 show up is in the section before that. It's called 8 Comparison With Acceptance Guidelines and it's a fairly l
9 long section. '
10 CHAIRMAN APOSTOLAKIS: Are they also in the 11 background? The reader who starts reading this, if he i
12 goes to page 11 to see -- this is key to me, it seems to 1
13 me, the full approach. I don't know whether we should l
/ s ,
\' ') 14 have it up front somewhere.
i l
15 MR. ABBOTT: Are you saying this is l l
16 inconsistent with what's just been discussed?
17 CHAIRMAN APOSTOLAKIS: No, this is very 18 consistent, but it's so important that it should not be on 19 page 11. That's what I'm saying. It should be way up 20 front.
21 I think the lines here, the results of the 22 different elements of the engineering analysis discussed 23 in previous sections must be considered in an integrated
! 24 manner, none of the individual analyses is sufficient in Ch In this way, it can be seen that the
, ( ,/ 25 and of itself.
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
329 1 decision is not driven solely by numerical results. I
, . -s 2 mean these are key thoughts and I mean you see from the e
i 3 discussion here, most of us, even though we're talking 4 about risk-informed regulation, we pay a lot of attention 5 to the numbers.
6 MR. JONES: I think, George, we also were very 7 sensitive to this issue from earlier comments of the 8 Committee and specifically, if you go to the start on page 9 2.9 to page 2.10, talking about how you compare it to the 10 guidelines, we say very clearly chat the acceptance 11 guidelines should not be interpreted as being overly 12 prescriptive. They are intended to provide an indication 13 in numerical terms of what is considered acceptable.
l ~
\ )
- 14 Furthermore, the epistemic uncertainties associated with 15 PRA calculations preclude a definitive decision based on I
16 purely numerical results.
17 CHAIRMAN APOSTOLAKIS: Okay, let me ask you 18 this question then. Is this important enough to make it 19 one of the principles?
20 MR. JONES: This follows on to say this is how 21 we demonstrate that principle 4 is met in that same 22 paragraph.
23 CHAIRMAN APOSTOLAKIS: But I'm saying since 24 the principles listed on page 2-1 which is up front, is it l'h
(_) 25 worthwhile to take the thought in this paragraph from page NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433
330 1 2-9 -- where was it? )
73 2 MEMBER BARTON: 2-11.
i 1 N.)
3 CHAIRMAN APOSTOLAKIS: And turn it --
4 MR. JONES: Oh, integrated decision making?
5 CHAIRMAN APOSTOLAKIS: Yes, and turn it into a 6 principle somewhere, especially the results of the 7 different elements of the engineering analysis must be 8 considered in an integrated manner and none of the 9 individual analysis is sufficient in and of itself. It 10 seems to me that's a pretty strong statement that should 11 be way up front.
12 MR. JONES: It is in the sense of it is our 13 very first expectation on page 2.2. All safety impacts --
'/ 14 CHAIRMAN APOSTOLAKIS: Oh, okay.
15 MR. JONES: -- are evaluated in an integrated ;
l 16 manner as part of an overall risk management approach. I 17 CHAIRMAN APOSTOLAKIS: Yes, you are right.
l 18 MR. JONES: And the second one talks about ;
l 19 integrated fashion. That all the principles are met.
20 CHAIRMAN APOSTOLAKIS: If it was not 5 o' clock 21 I would ask you what's the difference between an 22 expectation and the principle.
23 MR. FISCHER: They are great expectations.
24 CHAIRMAN APOSTOLAKIS: It's what?
(__,/ 25 MR. FISCHER: They are great expectations.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l l
331 l l
1 MEMBER SEALE: That's very good. That's very l I
,em 2 good.
)
v 3 CHAIRMAN APOSTOLAKIS: I don't know, does that 1
4 satisfy you, now that we've made it an expectation. It's 1 5 an integrated manner. I think that's going to create a ;
6 lot of problems and I think problems in the sense that it 7 will take people along time to really understand what's 8 going on, but maybe there's no way around it. They have 9 to read it and think about it.
10 MR. GRANTON: Well, I'm just having a problem ,
1 11 with the whole thing because I see you all going through .
1 1
12 all of these pages and all these things and we've never
, _ . 13 seen it. It's hard for us to relate right now. {
[ i
\- 14 CHAIRMAN APOSTOLAKIS: But do you relate to 15 che idea of an integrated approach where none of the 16 individual analysis in and of themselves drive the l l
17 decision?
18 MR. GRANTON: Yes, I think that is the right 19 way to look at things in an integrated fashion. I think 20 you have to look at it that way because you're going to 21 find out that there's not one thing that drives risk.
22 It's the combinations of things and you have to look at 23 them in different lights.
24 MR. CARROLL: And each application has its own
,em Q )\ 25 --
NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
I (202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
332 1 MR. GRANTON: Exactly. ]
2~ MEMBER MILLER: Right up front, a figure.
3 CHAIRMAN APOSTOLAKIS: K is very large. No, l'
l 4 it doesn't. This is really an important detail. So I l
<- i
- 5 don't know whether you want to raise the expectation to a 1
6 principle, but -- the idea that perceived unquantifiable
- 7 benefits will be invoked to effect a change even when the i
8- total CDF is above the goals, I mean I find that new,.even 1
t 9 thouch you can interpret what's in there.in al way that I
\
I i 10 will allow you to do that. I've been reading these things i 11 for a while now.
l ,
12 MEMBER SEALE: That's an awful big target. ]
l 13 CHAIRMAN'APOSTOLAKIS: I guess risk-informed !
O 14 regulation-gives you a lot of latitude.
l l
l 15 MR. CARROLL: Yes, it does. l 1
l 16 MEMBER SEALE: It sure does. i
- 17 CHAIRMAN APOSTOLAKIS
- But inaybe that's the 18- way it should be.
19 MEMBER FONTANA: Well, the plants are licensed
)
20 and they are running. If they have a number that's above ]
21- some line,-you're not going to go shut it down,-are you? .
l
- 22. CHAIRMAN APOSTOLAKIS: :That's the next step. !
'23 MEMBER FONTANA: We11 -- !
1 24 , CHAIRMAN APOSTOLAKIS: This is evaluating the
/ 25 changes. Then we go-to -- it's crazy.
1 NEAL R. GROSS :
COURT REPORTERS AND TRANSCRIBt:P3 1323 RHODE ISLAND AVE., N W.
l WASHINGTON, D.C. 20005-3701 (202) 234-4433 i
(202) 234 4 33
_.._ . - - , _ . _ ._..-._. . . _ _ _ _ _ _ . _ . _ _ . = _ . _ . _ _ - . - . _ - . _ _ - . . _ . . , _ . .
{ 333 1 MEMBER POWERS: Mario, the question is not one .
2 of shutting it down. It's a question of do we want to 3 make changes to request licenses to relieve burden that j 4 make plant even less safe.
5 MEMBER FONTANA: No, that's what we were I
6 saying before. It doesn't matter, not that it doesn't
.7 matter, but it's less important where it starts. I'd have i
8 more confidence in something that tells you what the :
9- direction is and if it makes it.more safe, fine.
10 MEMBER POWERS: But what they're saying is it 11 doesn't matter what you conclude quantitatively and if you 12 can be persuasive and eloquent in saying, in developing 13 ideas of unquantifiable benefits, but then.no matter what O 14 you calculated, this would be approved, no matter where.
i 1
I 15 you stand 16 MEMBER FONTANA: This has to ha one'of the --
17 it has to be one of the. inputs. I don't like your 18 eloquent part.
19 MR. FISCHER: I take a little offense at this 20 statement that it doesn't matter what the bottom line l l
21 number means -- it does matter to the staff. I mean it 22 matters -- ;
23 MEMBER POWERS: No, it can't possibly --
l 24 MR. FISCHER: If it came out 10 to the minus by 25 2, I think-we'd care. !
i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISt.AND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
L .
l 334 l 'l MEMBER POWERS: It cannot possibly matter to l
2 the staff because they don't calculate the number. All we
-t 3 calculate is CDF during operations. So they don't really l
4 care what the bottom line number is. They never f 5 calculated it, so I think we care what the bottom line l
l 6 number is.
l l
L 7 CHAIRMAN APOSTOLAKIS: No, you mean bottom j.-
l- 8- line, delta CDF. I think David means the bottom line, 9 what's the CDF.
10 MR. FISCHER: He wants everything quantified 11 is what I understand.
i 12 MR. PRATT: I think what we're learning from l
13 today's exercise is that what we've done is that we've l- 14 written only about 20 pages, but it's full of such j' 15 fundamental issues with respect'to regulating nuclear 16 power plants, integrated decision making and what the 17 principles are and all these things that, in fact, are so 18 many fundamental things. The next thing you do is you end 1
19 up putting every one of them up in the front of the
(
20 document because none of it makes sense without looking at j- 21 all of it.
22 CH'IRMAN APOSTOLAKIS: No, but integrated 23 decision makiiJ --
24 MR. PRATT: But remember, this morning we were
~OI
(_f. 25 talking about other thing is that were so important that
- NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W (202) 234 4 433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
335 I
1- they needed to be b'rought up in the front and I think l l
2 there's just so many fundamental elements here that I'm i 3 not sure that we.can shorten it more than 20 pages or move i 4 them all up front. g 5 MR. CARROLL: Again though, repeating my PR +
6 theme, this is one that would be desirable to put up front 7' from a PR point of view. ,
i 8 MR. PRATT: I understand and so would the i
9 figure on having alternative approaches, but up front. ;
10 'to a certain extent, look, this is hard stuff. l l
11 This is not easy.
12 MEMBER SEALE: No, it's not.
13 MR. PRATT- And if we convince ourselves just 14 by moving them around a little bit it's' going to get easy, ;
15 iC's not going to get easy.
16 MEMBER CATTON: Some of us are glad we only 17 have to be critical.
18 MR. PRATT: It's hard for me to'finirh l 19 CHAIRMAN APOSTOLAKIS: See, that's the -- it 20 comes back to my question this morning,-can one do useful 21 things with a less than perlict PRA? !
i 22 MR. PRATT: Yes And that's another 1
[
23 fundamental though that's got "o --yes. .
i 24 MEMB M MILLER: You better be able to. j 25 MR. APBOTT: And you can do better things with 1
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N.W.
,(202) 234-4433 WASHINGTON. D.C. 20005 3701 (202) 234 4 33
336 1 a better PRA.
gm 2 MEMBER CATTON: In response to Dana, response
%.)
3 frou the staff --
4 MR. HARDIN: There was a phase of this 5 development, maybe six months or so, maybe even further, 6 where the staff was toying with factors to adjust the base 7 PRA numbers up and down depending ' what was included and 1
8 we backed away from that and I'm not the right one to talk 9 about that, but I think that we decided that that was not 10 the right way to go. I don't know --
11 CHAIRMAN APOSTOLAKIS: But then you had these 12 --
- 13 MR. HARDIN: If somebody would like to comment 14 on that.
15 MR. PRATT: Well, I think it's that way for a )
1 i
16 good reason because I think the last -- it's one thing to )
1 17 be making judgments abe'It what you haven't analyzed. And l l
18 we are recognizing that Dana is right. I mean we'd all l 19 feel more comfortable and we'd make better decisions if we
_0 had all the analysis. But I think what we concluded at 21 the time was it's even worse if you assume what those are 22 and put a numerical value on it and walk away and say I'm 23 basically going to allocate a certain amount of risk to 24 shut down and not think about it any more.
p)
( m, 25 That was the worst approach.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
- . . - . - - ,..,. ~ _ .. . . - . . . . . . ~ .-. - . - ..... - . - ~ . - . . . . . . _ . . -
337 j 1- CHAIRMAN APOSTOLAKIS: No, but isn't that why i 2 you put.all these statements about increased management 3 attention and all that? ;
I 4 MR. PRATT: Yes. l 5 CHAIRMAN APOSTOLAKIS: So here in this i 6 particular case, the numbers we just saw, you would have 7 increased management attention and.--
~
8 MR. PRATT: '
..s. '
v l 9 CHAIRMAN APOSTOLAKIS: And you would look at 10 the scope. j
! 11 MR. PRATT: Yes.
i
! 12 MR. FISCHER: In fact, we're going back. I l
I !
13 put a number of different things we're going to ask the !
l O-i 14 licensees on this vu-graph. One of them is how they can I l 15 change their bottom line numbers or how they might improve l
- l b
l 16 those bottom line numbers. There are a number of other L
17 things we asked them~to do. I mean we are going to give 18 this more management attention. We're not ready to l l \
f:
- l. 19 approve the pilot plan submittals at this point.
l 20 I would like to mention just a couple of these 21 other things that we're-going to ask the pilot licensees i
22 about before my time is-up and then make one final point. l l
I 23 What I'd like you to.come away with from this 4
j 24 vu-graph is that there are a number of very key issues
~
1 i
25 that we need to explore with the pilot licensees yet, the )
l NEAL R. GROSS I
- COURT REPORTERS AND TRANSCRIBERS e 1323 RHODE ISLAND AVE.. N W.
4 (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4 433
338 1 change in risk is the first one, how their proposal 2 affects the licensing basis is another, how they meet the V,r~g 3 five key principles that we've outlined in the general 4 reg. guide and in an IST and SRP, excuse me, reg. guide 5 and SRP.
6 We want more details about the licensee's 7 integrated decision making process. We would like to know 8 how they handle shut down risk if they didn't do it on a 9 quantitative way, how did they do it. One of the pilot 10 licensees said that they used a deliberative process. We 11 would like much more detail than that.
12 CHAIRMAN APOSTOLAKIS: So you are addressing 13 it?
(\' I 14 MR. FISCHER: We want to know how they dealt 15 with shutdown modes of operation, seismic fires and 1
16 floods. We want for them to describe what the process was i
17 that the expert panel used to arrive at decisions. We 18 want it docume.'_Ced in a way that someone can follow it, j 19 The next person can come back and understand what they did 20 and probably come away with similar, if not identical l 21 results, is what we would like.
22 MEMBER CATTON: Or at least know why he 23 disagrees.
24 CHAIRMAN APOSTOLAKIS: Is that an expectation?
/ \
r 1
(_/ 25 MR. FISCHER: It's an expectation in my mind.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND A'/E. N W.
(202) 234-4433 WASHINGTON, D C. 2000!,-3701 (202) 234-4433
l l 339 l 1 We're asking questions along those lines.
i 2 MEMBER CATTON: I think documentation 73 Nl 3 requirements are important.
4 MR. FISCHER: They're very important and Palo 5 Verde is committed to go back and completely redo their 6 expert panel because they failed to document what it was 7 that they had done. And they haven't finished that and 8 come back to the staff, but as a result of our 9 questioning, they said they were going to do that.
10 MEMBER CATTON: Are they going to include the 11 RAIs in the document?
13 MEMBER CATTON: Request for additional
/ T t 4
\2 14 information.
15 CHAIRMAN APOSTOLAKIS: RAIs.
1 16 MR. FISCHER: I know what RAIs are. I'm just 17 -- are they going to -- hopefully they will --
l 18 MEMBER CATTON: Often the documentation 19 doesn't tell you about all the give and take and reasons 20 for positions. If you sort of put in an appendix to the 21 RAIs, then you've got it.
22 MR. CARROLL: And responses.
23 MEMBER CATTON: Of course, the whole package.
24 Westinghouse has committed to do that on their thermal
(~)
(_j 25 hydraulic stuff.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
340 1 MR. FISCHER: I think what we'd like to see is m 2 on a component-specific basis to see the rationale for the i )
<j 3 decision making, whether it be -- including the numerical 4 information associated with the component as well as what 5 the expert panel -- how they dealt with that.
6 MEMBER CATTON: Have you read the 7 recommendations that resulted from Yankee?
8 MR. PRATT: Yes.
9 MEMBER CATTON: In it, it's pretty clear.
10 MR. PRATT: Yes, it's pretty clear that the 11 staff needs to find ways to make licensees' statements and 12 commitments biading in some sense, whether it's getting 13 them in the FSAR or in some other fashion.
,p
- ' ' 14 MR. FISCHER: There is one other thing they 15 didn't provide us with the implementation plan on how they l l
16 were going to implement the programs, how they were going 17 to takj. a step-wise approach to extending test intervals
/ l 18 for low safety significant components. The staff really 19 needs that information in order to do the review.
l
- 20 We want the licensees to consider performance 21 histories and service conditions at pumps and valves and 22 what they're going co see. I would like to just say what 23 the pilot licensees did was they said based on PRA we can
, 24 extend the test interval for these components to 6 years i
\
(/ 25 and the staff has had this one little hyphen here, Oak NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W, (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
l l 341 1 Ridge review event NPRDS data. I'll give an anecdotal l (-) 2 piece of information.
IV 3 They looked at NPRDS data for check valves at 4 the two pilot licensees and in one case like at Comanche 5 Peak they found a check valve, a stop check valve in the 6 component cooling water system which had a number of l
i l 7 repeated failures and that particular check valve in their l 8 IST program was on a refueling outage interval test basis.
9 They had said that it was impractical to test that 10 particular valve for whatever reason and that they were 11 going to do it on a refueling outage.
12 As a result of the NPRDS or the failures that
,_s 13 they had, they stated in the NPRDS input that they decided
/ i i \'") 14 that they needed to test the valves quarterly in order to 15 insure that these stop check valves weren't rusting into 16 their seats, basically.
17 Based on that information, we're going back 18 and saying how are you using performance to determine the 19 test interval. You came in and asked for a 6-year l l
l
- 20 interval on this and yet your own NPRDS input says you 21 need to test them quarterly.
l 22 So what we want is not just a PRA based test l 1
23 interval that they're going to move these components to. l l
24 We want them to explain based on PRA, based on performance
/~^s I e i i
(_.,/ 25 history and service condition why the test interval is l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
l 342 I
l j 1 justified and we don't have that yet.
( .r T 2 So what I'm trying to leave you with there's a )
! V 3 lot of very important stuff we need to get from the !
l 4 licensee before we're able to approve their risk-informed 1
5 in-service test program. They provided us wi.th some !
l 6 bottom line PRA numbers, but it's not enough. We need 7 more to do our integrated decision making and we're asking !
8 them for that.
9 MEMBER CATTON: You need their rationale.
l 10 MR. FISCHER: One last point I want to make 11 here is what we have done, this team has done here is 12 tried to outline an acceptable process for licensees to
- 13 use to implement risk-informed testing programs. We've
! s.
t }
' ~ '
14 asked them to have a quality PRA, maybe not a perfect PRA, 15 but -- and most of the guidance is in the general reg. l l
16 guide on what a quality PRA is.
1 17 I don't want to get hung up there, but what j
! 18 we've asked them to do is to rank components into high and 19 low to try to identify components that are candidates for 20 reduced testing. Okay, just candidates for reduced ]
, 1 l 21 testing.
l i r
22 We've asked them to look at the more safety 1 i
23 significant components and evaluate whether a better test 24 could be done. We've asked them to add components that i
,a a
(_/ 25 are highly risk-significant to their IST program. We've NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234 4433
343 1 asked them to -- we said we'd like to know the basis that
,f- 2 you want to extend test intervals for the low safety 3 significant components. We want you to do it in a step-4 wise manner. We want you to collect data. We want you to 5 have corrective action programs and feedback mechanisms so 6 that you can adjust the ranking in your test programs 7 accordingly. -
8 That's the overall process that we've outlined 9 in the Reg. Guide and the SRP. I don't want you to lose 10 sight of the overall process and get numerical paralysis -
11 -
12 (Laughter.)
13 CHAIRMAN APOSTOLAKIS: That's a new course at O
14 M.I.T.
15 (Laughter.)
16 MR. FISCHER: I think we can take these reg.
17 guides and the SRPs and send them out for public comment 18 and I don't think you know we've committed a major crime.
19 MR. PRATT: And Dave --
20 CHAIRMAN APOSTOLAKIS: It's a civil case, 21 civil, not criminal.
22 MR. PRATT: That whole process that you just 23 outlined with pluses and minuses and all that sort of 24 stuff, sounds like a program that has a net increase or n
( 25
(_)' decrease in safety.
NEOL R. GROSS COURT REPORW.RS AND TRANSCRIBERS 1323 RHODL ISLAND AVE., N W. l 1
(202) 234-4433 WASHINGTOh D C. 20005-3701 (202) 234-4433
344 1 MR. FISCHER: If I had, in the way this -- if 7m 2 properly implemented this program, as described in the
( ')'~
3 reg. guide and the SRP, I believe will improve safety. I 4 think that if we make modest improvements in the 5 unavailability of the more safety significant components, 6 whatever we do to those less safety significant 7 components, if we extend the intervals, it's going to be 8 transparent compared to the benefits you're going to gain.
9 MR. PRATT: Thank you.
10 CHAIRMAN APOSTOLAKIS: And I think what's 11 becoming transparent to me is that most of us really had 12 underestimated the significance of importance of judgment 13 in this whole process, even though we've been talking i
i
'v' 14 about risk-informed regulation, I think when you see i 1
15 numbers you tend to focus on the numbers. This tells us l 16 there's something else.
17 Let's go to 15, just for half a second.
18 MEMBER POWERS: George, I can't resist 19 responding to the items that Gary wanted to get on to the 20 record. I just wondered if this was -- maybe I've been l
l 21 wrong, but instead of core damage frequency based on --
22 power operations based on internal initiator regulations, 23 we're going to a faith-based regulation scheme here?
24 MR. PRATT: No, we're going to a principle --
<x l
(,) 25 MEMBER CATTON: Well documented faith-based --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W-(202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433
345 1 CHAIRMAN APOSTOLAKIS: No, it's a principle --
-y 2 MR. PRATT: It's a principled based l !
~
3 regulation.
. 4 CHAIRMAN APOSTOLAKIS: Just to understand this 5 deliberative process a little better, what if Palo Verde, 6 for example, came to you and said look, this is what we 7 get from our PRA with importance measures, 350 valves are 8 classified of low safety significance, but because we have 9 not done seismic and fire, we think that 11 of these 10 valves really should be left in the high safety category 11 because we haven't done the shutdown thing, PRA.
12 Maybe the next 5 valves should be moved up.
13 So now we're pretty confident that the remaining 314
- ,x
' \ ') 14 valves are indeed of low safety significance. ,
l 15 We moved some up, precisely because our PRA is j l
16 not complete.
17 MR. PRATT: Which is what they said they did.
1 18 CHAIRMAN APOSTOLAKIS: They have already done l
19 it?
20 MR. CHEOCK: They have already done it and 350 21 reflects the expert panel deliberations on modes and --
22 CHAIRMAN APOSTOLAKIS: Ah.
23 MR. CHEOK: And this is not analyzed.
24 MR. FISCHER: But they have it documented very
/'~~N
!.s ) 25 well.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
346 1
1 CHAIRMAN APOSTOLAKIS: You should have told us l l
f~x 2 that because we all thought, we think that the 152 and 350
! )
R./
3 was a result of a PRA that we have.
4 MR. PRATT: But George, you didn't let him get I
5 to the slide.
l 6 MEMBER CATTON: No, we saw this slide.
7 CHAIRMAN APOSTOLAKIS: We saw this slide. l 8 MR. PRATT: You're right. We were distracted 9 by the PRA numbers.
10 CHAIRMAN APOSTOLAKTS: The deliberative part 11 of the process would have benefitted the loss if we had 12 been given this kind of information because that's another
.s 13 way of handling this incompleteness.
i i
~ 14 MEMBER CATTON: I think you have to do that. l l
15 CHAIRMAN APOSTOLAKIS: Yes. ,
1 16 MEMBER CATTON: You can't use risk-based 17 regulation with it. ;
1 18 CHAIRMAN APOSTOLAKIS: I think it's fine.
19 MEMBER BARTON: Risk-informed. I 20 MEMBER CATTON: Risk-informed, sorry. But you 21 also have to be very clear on the required documentation.
! 22 You have to have the documentation because it replaces the 23 numerical calculation. You have to know who did it and 24 why and it needs to be a part. That's why I feel strongly n
'\s ,) 25 about the RAIs. They should be a part also because that's NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
347 1 part of the rationalization process and 6 months or 6
(~S/ 2 years from now when somebody questions your decision, it's r
\._/
3 on the table.
4 MR. FISCHER: I guess my hope wasn't that the 5 RAIS would be in their submittal. I wanted them folded 6 into the description of their program.
7 MEMBER CATTON: But what will happen is 8 they've already written the report and I know that in our 9 dealings with Westinghouse, about the best we're able to 10 do is to get them to cross reference it, the report to the 11 RAIs.
12 CHAIRMAN APOSTOLAKIS: I'm sorry.
13 MEMBER CATTON: That.'s all right.
g
I wonder how many 14 CHAIRMAN APOSTOLAKIS:
15 people who will be reviewers understand this? Is there ]
16 any course, training course at the NRC to enlighten people l
17 about this?
18 MR. JONES: That's the Bob Jones Memorial 1
19 Course that I'll be developing this summer --
l 20 (Laughter.) I i
1 21 -- that Gary talked about repeatedly.
22 CHAIRMAN APOSTOLAKIS: You will be developing 23 a course?
24 MR. JONES: Yes.
i V 25 MEMBER BARTON: And M.I.T. will teach it?
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
348 ,
1 MR. FISCHER: I suspect that after, if we ever
)
2 get these reg. guides out -- l O 3 CHAIRMAN APOSTOLAKIS: In April.
l 4 MR. FISCHER: Okay, when we get these out 5 final, we will develop an inspection manual associated l 6 with this and we'll go out on'the road and teach it to the
.I
'7 residents and probably have workshops for industry to hear i l
8 it. >
9 CHAIRMAN APOSTOLAKIS: I think that's !
10 critical. ,
1 11 MEMBER CATTON: They have found in places ;
12 where'they've tried to put in place risked-informed l 13 regulation. There_were two things that were really ]
14 important. One was'the educational process and in some 15 places they actually delayed implementation for a couple 16 of years until they could bring the people up to speed.
17 CHAIRMAN APOSTOLAKIS: Let's not talk about i
18 additional delays.
i 19 MEMBER CATTON: What's that?
20 CHAIRMAN APOSTOLAKIS: Let's not talk about 21 additional delays. ;
22 MEMBER CATTON: No, I'm not suggesting you '
23 delay it, I'm suggesting you pay a lot of attention to it j 24 because if you don't, it won't work.
25 CHAIRMAN APOSTOLAKIS: Sure. We have a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N W.
(202) 23( 4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
349 1 request here.
2 MR. MARTIN: Lawrence Martin, South Texas fS
'w) 3 Project. I've listened to you all talk about shut down 4 risk and operating risk. I'm not a PRA expert. I carry 5 one with me, actually, he won't let me out of the office 6 unless he's with me.
7 MEMBER CATTON: That's the way you should do 8 it.
9 MR. MARTIN: But one of the things we've done 10 and I'm glad to hear the integrated approach, because had 11 we submitted our diesel tech. spec with more deterministic 12 information in there and less PRA, okay, we might not have 13 gotten it if they didn't have the integrated approach.
A ;
(
'J 14 But one of the things that we're finding out of our PRA 15 information and we're using it different ways every time 16 and we did not have this information available to us when 17 we made our submittal.
18 But one of the things that the extended AOT 19 got for us and we didn't realize that until they turned 20 down our first submittal and we had to go back and put 21 them into the outage, when we went back and put the 22 diesels back into our outage it increased our risk for 23 that outage 41 percent.
24 We've got data that tells us that by doing (n),, 25 them on-line we reduce our overall risk every year and we NEAL R. GROSS COURT REF ORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. '
(202) 234-4433 WASHINGTON, D.t,. 20005-3701 (202) 234-4433 1
f..
.350 1 reduced the outage risk. Had we submitted that with 2 deterministic information that had been turned down, we 4
i 3 would have continued to go through our outages with a 41 i
4 percent increase in risk in.every outage. So I'm glad to 5 see that.
6 MR. ABBOTT: You mean it's more risky to do 7 the diesel. maintenance while you're shut down than when' 8 you're operating?
9 MR. MARTIN: Absolutely, because you had all
-10 the other equipment out at the same time. Now we can do 11 one diesel at a' time, take our time, do the maintenance 12 right and have -- now we're a three train plant, you'need
'i to understand that.
14 Okay?. We have three of them so we always have 15 two available, but the' difference was~a 41 percent.
16- increase in the risk for that outage.
'l 17 MEMBER SEALE: And that's really the i
18 completeness of the PRA that you'd like to have. It's not 19 just the shutdown risk. It's not just the transition J 20 risk. It's also the avoided risk when you do it this way. l
- 27. MR. PRATT: What you want is completeness in 22 your decision making.
.23 MEMBER SEALE: That's right. ,
24 MR. GRANTON: I'd like to add -- I'm Rick 25 Granton. I'd like to add one other point, too, because NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4 133 . WASHINGTON, D C. 20005-3701 (202) 2344433
351 1 you bring up these items about this risk-informed
,r'g 2 approaches here, but I always keep asking myself because I O 3 see the examples that occurred at an operating plant and 4 you rarely hear anybody talk about the risk that we have 5 right now by doing nothing and just staying where we are 6 because you can definitely see instances where risk is 7 increased because of the way regulations are right now.
8 I've seen many examples where that occurred.
9 So the movement towards risk-informed is 10 . adding this other element here where you can optimize risk 11 because right now we can't really cptimize it.
12 MR. ABBOTT: What does what was just said say 13 about limiting conditions for operations for diesel 7_
14 generators for plants that haven't gone through this 15 process because you're committed to take them out of 1
16 service when you're shut down.
l 17 MEMBER CATTON: They're riskier.
l 18 MR. PRATT: They're suboptimizied.
19 MR. MARTIN: I'm not going to make any 20 statement about where the diesels are now at the other 21 plant. All I'm telling you with our plant, specific to 22 our plant it improved.
23 MR. ABBOTT: That question wasn't directed at 24 you. It was just sort of thrown out there for --
,m
! \
'x ,/ 25 MR. CARROLL: No, I think in general the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
352 1 situation is riskier than it needs to be.
f 2 MR. ABBOTT: Right, right.
(]) 3 CHAIRMAN APOSTOLAKIS: I like this integrated 4 approach and all that. One of the downsides, of course, 5 is that does it create incentives for people to actually 6 do these other additional analyses at some point? Because 7 there is benefit there. Maybe it will come with 8 experience.
9 Maybe the licensees as they use these 10 analyses, they will say, gee, if I had a good solid 11 numerical argument here, maybe I would get this. Maybe 12 that's the best way. Then you have a revolution with a
,_s 13 Greek roll, right?
- s 1 /
14 MR. PRATT: Right.
15 MEMBER SEALE: It's the inevitability of 16 gradualness.
i 17 CHAIRMAN APOSTOLAKIS: Pardon? I 18 MEMBER SEALE: It's the inevitability of 19 gradualness.
20 CHAIRMAN APOSTOLAKIS: Yes. It is 5:27. I 21 understand you guys are done?
22 MR. FISCHER: Yes sir.
23 CHAIRMAN APOSTOLAKIS: Do we have any other 24 pressing --
CI 25 MR. CARROLL: I'd like to ask one question.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
353 1 Is it true in this case that the licensees haven't seen p 2 the latest and greatest versions of your --
V CHAIRMAN APOSTOLAKIS: Yes.
3 4 MR. FISCHER: They have not seen a reg. guide 5 or an SRP, They are very anxious.
6 They haven't seen our RAIs because --
7 MR. JONES: The earlier comment, Jay, was 8 across the board on all of the risk-informed SRPs and reg.
9 guides.
10 MR. CARROLL: Because it looks to me like a 11 bunch of your RAIs are based on your current thinking to 12 reflect the --
13 CHAIRMAN APOSTOLAKIS: Oh, they're allowed to
,r)
- s k 14 talk about it. They cannot show the document. They can l l
15 talk about it. l l
16 MR. FISCHER: It was intentional that you see j 17 that the RAIs were a parallel to what as in the reg. guide 18 and SRP. That was totally, that was the way I developed 19 the list of RAIs.
20 CHAIRMAN APOSTOLAKIS: Let me ask Mr Martin or 21 Mr. Granton, is the process becoming clearer to you even 22 though you haven't seen the documents?
23 MR. MARTIN: Yes, but the devil's in the 24 details.
p
( ,/ 25 (Laughter.)
! NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
354 1 CHAIRMAN APOSTOLAKIS: Repeat that?
,x 2 MR. MARTIN: Absolutely, but the devil is in G'
3 the details and I was talking with Bob Gramm today because 4 they made a statement earlier this morning that there's 5 been and I think I detected the word significant, but 6 input from the utilities or the pilots into the SRPs and 7 reg. guides and I needed to check with Bob and I asked 8 Bob, I said Bob, have we had that and I don't recognize 9 it. Bob Gramm works for Suzy Black in rated QA. And 10 they'll be doing a presentation tomorrow. And he informed 11 me yes, we have had significant input into it, I just 12 didn't know it. Okay?
13 I'm very interested and when we started rated
( ')
Y/ 14 QA and we'll talk a little bit about this tomorrow, I 15 recognized that the level of PRAs is significant across 16 the industry, okay? But - and I also know that we're 17 probably driving a Cadillac or close to it. I got a lot 18 of money in it. I hope it's a Cadillac.
19 But I recognize that there is some incremental 20 benefit for somebody that just did, has a level 1, did the 21 IPE. There is still something to be gained and if we 22 don't allow them to get that little gain, then we've lost 23 it, okay?
l 24 So I'm interested in really seeing what the l3
(_) 25 details of those are. First of all, I got a lot invested NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS i
1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
355 1 into it and I'd hate to have to reinvest because I didn't 1
cx \ 2 need it. But more important to me is where the general :
l j
<J l 3 industry is and can they tap into it. :
l 4 CHAIRMAN APOSTOLAKIS: Thank you. Can we i 5 recess at this point? l 1
6 MEMBER CATTON: Right now.
7 CHAIRMAN APOSTOLAKIS: Do we need a vote or 8 not? l l
9 (Whereupon, at 5:34 p.m., the meeting was 10 concluded.) ;
11 12 l
13 I
,r's
~- 14 ,
15 l
1 16 l l
17 18 ,
19 20 21 22 23 24
( ,/ _
25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
4 CERTIFICATE p
V This is to certify that the attached i
proceedings before the United States Nuclear Regulatory Commission in the matter of:
Name of Proceeding: Meeting of the Subcommittee on Probabalistic Ris'i Assessment of the Advisory j Committee on Reactor Safeguards '
Docket Number: not applicable !
Place of Proceeding: Rockville, Maryland were held as,herein appears, and that this is the original transcript thereof for the file of the United States Nuclear i
Regulatory Commission taken by me and, thereafter reduced to j typewriting by me or under the direction of the court I reporting company, and that the transcript is a true and accurate record of the foregoing proceedings.
I Corbett Riner 4 Official Reporter Neal R. Gross and Co., Inc.
l f\h v
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVENUE, N W.
(202) 234-4433 WASHINGTON D.C. 20005 (202) 234 433
}
INTRODUCTORY STATEMENT BY THE CHAIRMAN OF THE i
PROBABILISTIC RISK ASSESSMENT SUBCOMMITTEE 11545 ROCKVILLE PIKE, ROOM T-2B3 ROCKVILLE, MARYLAND FEBRUARY 20-21, 1997 The meeting will now come to order. This is first day of the meeting of the ACRS Subcommittee on Probabilistic Risk Assessment.
I am George Apostulakis, Chairman of the Subcommittee.
ACRS Members in attendance are: John Barton, Ivan Catton, Mario Fontana, Thomas Kress, Dana Powers, Robert Seale, and Bill Shack We also have in attendance Jay Carroll and Ed Abbott who are here as consultants to the Subcommittee.
ACRS Senior Fellow in attendance is: Richard Sherry The purpose of this meeting is to continue our discussion of the NRC staff's approach to codify risk-informed, performance-based regulation through development of Standard Review Plan (SRP) sections and associated regulatory guides. The Subcommittee previously met to discuss these matters on October 31, November 1, 21, and 32, 1996, and January 28, 1997. The Subcommittee will I gather information, analyze relevant issues and facts, and i formulate proposed positions and actions as appropriate, for deliberation by the full Committee. Michael T. Markley is the O Cognizant ACRS Staff Engineer for this meeting.
The rulen for participation in today's meeting have been announced as part of the notice of this meetir.g previously published in the i Federal Register on February 5, 1997.
A transcript of the meeting is being kept and will be made ,
available as stated in the Federal Register Notice. It is '
requested that speakers first identify themselves and speak with l sufficient clarity and volume so that they can be readily heard.
We have received no written comments or requests for time to make oral statements from members of the public.
I would like to inform you that we have one change in the agenda.
Messrs. Biff Bradley of the Nuclear Energy Institute (NEI) and Doug True of ERIN Engineering will not be making a presentation tomorrow regarding large, ear ly release frequency (LERF) as was indicated in the published agenda.
(Chairman's Comments-if any)
We will proceed with the meeting and I call upon Messrs. Gary Holahan, NRR, and Thomas King, RES, to begin.
O
l fL
- .\
s United States O '6*
- } ,j
., Nuclear Regulatory Commission REGELATORY GLIDES AXD STANDARD REVIEW PLANS IN SLPPORT OF RISK INFORWED REGELATION PRESENTATION TO ACRS SEBCOMMITTEE ON PRA FEBRE ARY 20,1997 Gary Holahan, NRR (415-2884)
Thomas King, RES (415-5790) ;
Robert Jones, NRR (415-2198) .
Mark Cunningham, RES (415-6189) g
O O O INTRODUCTION DRAFT REGULATORY GUIDES (RGS) AND STANDARD REVIEW PLANS (SRPS) HAVE BEEN PREPARED TO HELP IMPLEMENT THE COMMISSION'S AUGUST 1995 POLICY ON THE USE OF RISK INFORMATION IN THE REGULATORY PROCESS:-
- FOR REACTOR LICENSEES ;
- FOR PLANT SPECIFIC CHANGES TO A PLANT'S CURRENT LICENSING DASIS (CLB). l OBJECTIVES OF RGS/SRPS:
l
- DESCRIBE OVERALL APPROACH / EXPECTATIONS / PROCESS
- PROVIDE PRINCIPLES AND GUIDANCE FOR DETERMINISTIC j AND PROBABILISTIC ANALYSIS AND INTEGRATED DECISION MAKING
- DESCRIBE PERFORMANCE BASED IMPLEMENTATION STRATEGY
- PROVIDE GUIDANCE ON INFORMATION THAT SHOULD BE SUBMITTED FOR REVIEW. !
USE OF RG/SRPS IS VOLUNTARY FOR LICENSEES 1
i
O O O INTRODUCTION (CONT)-
SCOPE OF RGS/SRPS: /
- GENERAL GUIDANCE (DG-1061 + SRP) ,
i e APPLICATION SPECIFIC GUIDANCE IN-SERVICE TESTING (DG-1062 + SRP) l TECHNICAL SPECIFICATIONS (DG-1065 + SRP)
GRADED QA (RG ONLY) DG-1064) '
i IN-SERVICE INSPECTION (ON A LATER SCHEDULE)
- SUPPORTING DOCUMENT:
NUREG-1602 - SUPPLEMENTAL INFORMATION ON SCOPE AND QUALITY OF PRA ANALYSES i PILOT PROGRAMS UNDERWAY IN ALL APPLICATION SPECIFIC l AREAS:
- PILOTS PROVIDE FEEDBACK ON ISSUES AND GUIDANCE IN RG/SRPS
- DRAFT RG/SRP GUIDANCE USEFUL IN DEVELOPING QUESTIONS ON PILOT APPLICATIONS 2
i
O O O i
INTRODUCTION (CONTl SOME CHANGES IN GENERAL RG/SRP SINCE 2/7/97 ACRS FULL COMMITTEE MEETING:
- PRINCIPLE #4 ON RISK I
{
- REGULATORY PROCESS ;
i PURPOSE OF TODAY'S STAFF PRESENTATION:
i
- PROVIDE OVERVIEW OF APPROACH AND PROCESS I
- REVIEW CONTENT OF GENERAL RG/SRP APPENDICES
- REVIEW CONTENT OF APPLICATION SPECIFIC RGS/SRPS AND DISCUSS RELATED PILOT PROGRAMS
- SOLICIT FEEDBACK FROM ACRS SUBCOMMITTEE MEMBERS i
- LETTER REQUESTED AFTER MARCH 97 FULL COMITTEE MEETING i
3
.__.__.____._.._____m___ _ _ _ _ _ . . _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . -
O O 6 CHANGES TO THE CURRENT LICENSING BASIS STAFF PROPOSES INCREASED REQUIREMENTS - USE 50.109 I
BACKFIT RULE (REG. ANALYSIS GUIDELINES) a i i
" CURRENT LICENSING BASIS" l
LICENSEE REQUESTS CHANGE LONSISTENT LICENSEE MAKES CHANGE ;
WITH APPROVED STAFF POSITION CONSISTENT WITH 50.59 (RULE, RG, SRP, BTP...) PROCESS '
'" NORMAL" STAFF REVIEW" ,
V LICENSEE REQUESTS A CHANGE IN i REQUIREMENTS BEYOND ANY APPROVED STAFF POSITIONS -
10CFR50.90-92 i l
DOES NOT PRESENT DOES PRESENT RISK INFORMATION I RISK INFORMATION
" NORMAL STAFF "USE RISK-INFORMED '
REVIEW" RG/SRP" 4
l
O o 6 i RISK-INFORMED PILOT PLANT PROGRAMS 0
APPLICATION PILOT PLANT INSERVICE TESTING COMANCHE PEAK PALO VERDE ,
GRADED OUALITY ASSURANCE PALO VERDE GRAND GULF SOUTH TEXAS TECHNICAL
- SPECIFICATIONS 10 CE PLANTS j 5
O O O OVERALL APPROACH DEFINE PROPOSED CHANGE '
i DEMONSTRATE THAT CERTAIN FUNDAMENTAL SAFETY PRINCIPLES ARE MET: ,
8 MEET REGULATIONS (OR PROPOSE A CHANGE / EXEMPTION)
- MAINTAIN DEFENSE-IN-DEPTH !
- MAINTAIN SUFFICIENT SAFETY MARGIN
- INCREASES IN RISK AND THEIR CUMULATIVE EFFECT ARE SMALL AND DO NOT CAUSE THE SAFETY GOALS TO BE EXCEEDED
- IMPLEMENT UTILIZING PERFORMANCE BASED MONITORING AND FEEDBACK STRATEGIES EXPECTATIONS ON IMPLEMENTATION:
i
- ASSESS ALL SAFETY IMPACTS I
._.____.-..__._..__._._m___._ - .. -
- _ _ . .-..____._._._-_._________.___..._____m_._-__.__._._-_.-m._.________._._____.._..._..m - - -- _ ._
O oy,,,et3,,$,c,(coNTy O
- SUBSIDIARY OBJECTIVES OF COMMISSION'S SAFETY GOAL POLICY USED TO DEFINE BENCHMARK RISK LEVELS FOR DECISION MAKING: 1 10-4/RY - CORE DAMAGE FREQUENCY (CDF) 10-5/RY - LARGE EARLY RELEASE FREQUENCY (LERF)
!
- PROPOSED CLB CHANGES SHOULD BE MADE IN SMALL !
INCREMENTS (<10% OF BENCHMARK CDF/LERF VALUES) AND WHEN WITHIN A FACTOR OF 10 OF THE BENCHMARK VALUES, MORE ANALYSIS AND MANAGEMENT REVIEW WILL BE NECESSARY.
- SUPPLEMENTAL GUIDANCE IN APPLICATION SPECIFIC
. RG/SRPS :
i
- PERFORM UNCERTAINTY / SENSITIVITY ANALYSIS APPROPRIATE FOR PROPOSED CHANGE
- PERFORM QUALITY ANALYSES AND MAKE IT AVAILABLE FOR !
PUBLIC REVIEW PERFORMANCE MUST BE MONITORED TO HELP VERIFY KEY ASSUMPTIONS AND CHECK AREAS OF LARGE UNCERTAINTY DOCUMENTATION 7
GENERAL RG (DG-1061) APPENDICES e APPENDIX A -
" CATEGORIZATION OF SSCS - SAFETY SIGNIFICANCE" PURPOSE - TO PROVIDE GUIDANCE ON IMPORTANT FACTORS RELATED TO CATEGORIZATION OF SSCS INTO GROUPS WHERE REQUIREMENTS CAN BE RELAXED AND THOSE WHERE NO RELAXATION SHOULD BE MADE CONTENT - FACTORS TO BE CONSIDERED RELATED TO SRP APPENDIX C l
e APPENDIX B -
"AN APPROACH FOR ESTIMATING CONTAINMENT FAILURE AND BYPASS FREQUENCY" PURPOSE IS TO PROVIDE A SIMPLIFIED METHOD FOR USING LEVEL 1 INFORMATION TO ESTIMATE LERF CONTENT: -
GUIDANCE FOR DIFFERENT CONTAINMENT TYPES '
- EVENT TREES BASED UPON NUREG-1150
- DETAILED DISCUSSION BY BNL, INCLUDING TRIAL APPLICATION 8
O O O GENERAL SRP APPENDICES i
l The SRP contains three appendices:
o Appendix A - Guidance for a focused-scope application specific PRA review o Appendix B - Integrated decisionmaking O Appendix C - Categorization of SSCs with respect to safety significance E
~
o o 6 Appendix A: Guidance for a Focused-Scope Review
- PRAs that are used in risk-informed applications are expected to be of adequate quality
- Quality may be accomplished in various ways such as a peer review .
- Staff will perform a focused-scoped review on an application-specific basis ;
i i
)O i
_ . - _ . . - _ . _ . . . _ _ _ - - . - - - _ _ . - _ . - - - _ - _ _ _ _ - - - - - _ _ - _ - _ _ _ . - - - - . _ . - - - - - - - _ _ - - - - ~ - - - _ . . - - _ - - _ - _ - _ - - _ _ - _ - - . _ - - - . _ _ . _ - - -
- O O O Appendix A: Guidance for a Focused-Scope Review Application-specific reviews are expected to focus on:
- Use of appropriate data
- Effects of the application on initiating events
- Effects of mission success criteria on conclusions I
- Modeling of common cause failures and how this affects or is affected by the application
- Modeling of human perforrnance and how this affects or is affected by the application ;
- Effects of truncation limits used 11 t
O O O Appendix B: Integrated Decisionmaking i
i Integrated decisionmaking should:
- consider probabilistic and traditional engineering evaluations, operational experience, and current regulatory requirements t
- be systematic and defensible, and documentation should be available for review
- be based on a technical information basis that is adequate for the scope of the application i
i b
i f
_ _ _ - - . . . - _ _ . _ . _ _ . - _ _ _ . _ _ - . _ _ _ _ - __-____ __ ___ __ -_v' ___ _-__
~
O O O Appendix C: Categorization of SSCs In SSC categorization, the following has to be taken into account:
- Completeness of risk model
- Sensitivity analysis for component data uncertainties, common cause failures, and recovery actions
- Consideration of multiple failure modes
- Multiple component considerations i
- Relationship of importance measures to risk changes I
~_ . - _ _ _ _ _ . _ _ _ _ _ . _ _ _ . _ _ . _ . - _ _ _ _ _ _ _ _ _ , _ . _ _ _ _ .
1A January 22, 1997 MEMORANDUM TO: Hugh L. Thompson,=Jr.
Acting Executive Director for Operations Karen D. Cyr General Counsel FROM: John C. Hoyle, Secretary /s/
SUBJ ECT: STAFF REQUIREMENTS - SECY-96-218 - QUARTERLY j STATUS UPDATE FOR THE PROBABILISTIC RISK ASSESSMENT (PRA) IMPLEMENTATION PLAN, 1 l
INCLUDING A DISCUSSION OF FOUR EMERGING POLICY ISSUES ASSOCIATED WITH RISK-INFORMED PERFORMANCE-BASED REGULATION The Role of Performance Based Reculation in the PRA
! Implementation Plan The Commission has approved Alternative 1 with respect to the 1 l role of performance-based regulation but applications of performance-based approaches should not be limited to risk-informed initiatives. Thus, the Commission also approves l !
elements of Alternative 3 as follows: Performance-based
()
l i
) initiatives that do not explicitly reference criteria derived l l -from PRA insights should not be excluded from consideration.. The l l
i staff should include in the PRA implementation plan, or in a- l separate plan, how these performance-based initiatives will be ,
phased into the overall regulatory improvement and oversight j program. As part of the PRA implementation plan, or its separate plan, the staff should include its plan to solicit input from i industry on (or develop on its own) additional performance-based objectives which are not amenable to probabilistic risk analysis, but could be ranked according to, for example, a relative hazards analysis, and phase in these initiatives.
(EDO) (SECY Suspense: 8/29/97)
The staff should provide the Commission a summary discussion on how performance monitoring is being addressed in current PRA Pilot Applications and, where appropriate, other planned performance-based approaches. The staff should address the technical question concerning how the implementation and monitoring aspects of performance based regulations (Attachment 3, Item IV) are considered in these planned performance-based approaches. For the maintenance-rule implementation activities, address how these issues are considered within the context of the inspection process and inspection program. These items should be addressed in the March 1997 quarterly update and in the next j Commission briefing on the PRA implementation plan.
(SECY Suspense: 3/31/97)
- .(} (EDO)
I t
l The words " intolerable outcome" in the fourth key element are too
?
- vague and require further definition. For example, the words t j could be revised to read " failure to meet a performance criterion will not result in violation of a Safety Limit" or some other specific terminology.
(EDO) (SECY Susper.se: 8/29/97)
- Plant-Specific Anolication of Safety Goals 1 The Commission has tentatively approved Alternative 1 with
! respect to plant-specific application of safety goals and/or subsidiary objectives, but prior to issuance of the final guidance, the staff should explore the legal ramifications of the use of numerical guidelines for plant specific regulatory
, decisions and prepara a legal analysis of the issues for the
- commission. As part of this analysis, the staff should consider '
situations where updates or changes to licensees' PRAs (such as ;
. the underlying assumptions) result in changes to PRA results,
!. which would cause a previously approved action to become
- unacceptable. The analysis should also include a discussion of j the-type of regulatory decisions that might be subject to t litigation, an identification of the problems that such
! litigation might pose for the staff, and an estimate of the level of staff resources and technical support that likely would be
- required to address the issues in such litigation.
j (OGC) (SECY Suspense: 6/30/97)
Risk Neutral vs. Increases in Risk i
The Commission has approved Alternative 1 which would allow for
! small increases in risk under certain conditions, for proposed
! changes to a plant's licensing basis. The legal analysis
- requested above should address the legal ramifications and
- prospects for litigation in making this change. In addition, the
- terms "small" and "under certain conditions" require more precise j- definition. 'The staff should provide a sound rationale for
- judging small increases and provide for explicit consideration of uncertainties. Criteria for judging small increases in risk j should be considered in the context of maintaining reasonable j assurance that.there is no endue risk to public health and safety. The staff should establish procedures to monitor the cumulative changes in risk for a given nuclear facility as the i.
result of license amendments that are conducive to quantitative risk assessments. The staff should develop a methodology for i assessing changes in risk that uses statistical concepts and
- _ gives considerations to uncertainties, j (OGC/EDO) (SECY Suspense
- 8/29/97) i j The staff should, in its development of risk-informed guidance and review of applications regarding risk-informed initiatives i evaluate all safety impacts of proposed changes in an integrated j manner including the use of risk insights to identify areas where i requirements should be increased or improvements could/should be j implemented. In this regard, the staff should encourage i
yg licensees to use risk assessments for purposes of improvement
(_) that may require additional activity or effort on their part, as well as relaxation, in order to realize the full benefit of risk l assessments.
4 The staff should also verify licensee activity in this regard, as j appropriate.
1 i
(EDO) (SECY Suspense: 8/29/97) i Imolementation of Chances to Risk-Informed IST and ISI Recuirements The Commission has approved Alternative 2 allowing the staff to use the acceptable alternative provision of 10 CFR j
- 50. 55a (a) (3) (i) to approve the pilot plants' applications provided appropriate findings can be made. Where the findings
)
necessary to approve the alternative cannot be made, then the use 1
- of exemptions should be considered. The staff should work closely with ASME and with the Code consensus process so as to
! expedite changes to the Code involving ISI and IST.
1 Chairman Jackson
^
cc:
Commissioner Rogers Commissioner Dicus Commissioner Diaz
( Commissioner McGaffigan OGC OCA
! OIG Office Directors, Regions, ACRS, ACNW, ASLBP (via E-Mail) 4
$ l O
O~y O O l
1 A Simplified Approach for Estimating LERF t
Presented by f
I f
W.T. Pratt l
I
, y Presentation to ACRS Subcommittee on PRA February 20,1997 ;
?
~
O O C~
Presentation Outline L
Background
Approach r Example Event Tree Comparison with IPE Results Summary ;
O O O'
Background
- Consequently, not necessary to estimate frequency of late containment failures (or any. failure mode that does not cause early fatalities)
. Based on the NUREG-1150 results, LERF of 104 /ry is a proposed surrogate for the NRCs quantitative individual early fatality health objective
- LERF is defined as the frequency of those accidents and containment failures occurring before or within a few minutes of reactor vessel breach for PWRs and those failures occurring before or within 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> of vessel '
breach for BWRs. The frequency of containment bypass accidents should be included. The definition is consistent with that used in the Commission's Regulatory Analysis Guides.
4
~ '
O
~
O d' ,
Background (Cont'd) t
. The closer the base case estimate of CDF or LERF and the potential impact of the changes on these metrics are to the corresponding acceptance guidelines, more analysis and review will be required.
. A simple approach is proposed to supplement Level 1 PRAs in order to estimate a LERF for use in risk-informed decisionmaking when: .
- an approximate LERF value is acceptable (i.e., not close to benchmark LERF value) ,
- when Level 2 PRA information is not available or complete '
e
O -
O 6
Approach -
. Approach has to distinguish between containment failure modes that I might lead to early fatalities versus those failure modes that will not i
. Failure modes were therefore categorized as
- early containment failure or bypass (potentially leading to large early release, i.e., early fatalities likely) ;
- late containment failure or containment intact (potentially not leading to large early release, i.e., early fatalities unlikely) '
. Aim is to develop simple event trees that will identify those accident sequences that could lead to early containment bypass or result in severe challenges to containment integrity
. Advantageous if failure modes could be identified based on the status of the plant rather than on outcome of uncertain phenomenologicalissues !
t
. . _ - . . _ - - - _ - _ _ - - _ - - _ _ _ _ - _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ - -- -__-_____i
O O Q
?
Coolant No Containment Reactor Coolant Injection Containment Potential for $
Core Damage Isolated or Not Systems Restored Before Failure at No Potential for Large Early .
Frequency Bypassed Depressurized Vessel Failure Vessel Breach Early Fatalities Path Release I 2 3 4 5 6 1 No 2 No ;
3 No 4 Ya 5 No Y
6 No
/\
i 7 No .
V 8 Yes N
9 No ,
10 Ye i Example of Event Tiee (PWR Large Volume Containment) 4 t
Example of Event Tree (PWR Large Volume Containment)
. Most questions quantified by licensee based on status of plant for each CD accident sequence .
. Only a single question concerning the likelihood of containinent failure at vessel breach is prescribed i
. Question 1: The frequency of each accident sequence under consideration is processed through the tree
. Question 2: This question identifies all accidents that cause early loss of containment integrity Containment isolation failure Interfacing system LOCA Steam generator tube rupture
~
O O O Example of Event Tree (Cont'd)
. Question 3: This question is intended to identify accidents in which the reactor coolant system (RCS) remains at high pressure during core meltdown t
Transients and small break LOCAs will be allocated to high pressure category unless licensee takes credit for :
operator action to depressurize RCS Any credit taken for induced hot leg failure as a cause for RCS depressurization should be justified t
i t
't t
o o 6 Example of Event Tree (Cont'd)
?
}
. Question 4: This question identifies those accidents in which coolant j injection is restored and core damage is arrested without reactor vessel breach t
i ?
. ' Question 5: This question identifies those accidents which may result in i early containment failure and is the only question in which !
the split fraction is specified i The simplified tree distinguishes between high and low pressure sequence in terms of the specified likelihood of early containment failure 4
0.1 High pressure sequences 0.01 Low pressure sequences i
- f i ;
i
~'
O .O 6' .
Example of Event Tree (Cont'd) ,
. Question 6: Deterrnines if.those accident sequences that bypass or fail the containment can potentially cause early fatalities. The potential for early fatalities depends on the magnitude and timing of the release relative to two factors:
- 1) the time elapsed from reactor scram to the time at which the release starts (particularly relevant to shutdown accidents).
- 2) the time from the declaration of a gen ~eral emergency to .
the time of the start of the release compared to the time required to effectively warn and evacuate the population in the vicinity of the plant.
i i
l
~ ~
O O d' .
i Comparison with IPE Results t
. Level 1 analyses reported in several IPE submittals were processed through event trees (for different containment designs):
Surry, Sequoyah, Peach Bottom, and Grand Gulf
. LERF was calculated using only the guidance in the event trees i
. LERF was also determined from the IPE submittals .
Frequency of bypass and early failures
- Frequency of source terms with I, Te, Cs release fractions >0.03 frequency of source terms with I, Te, Cs release fractions >0.1
O O O Comparison of Different Estirnates of LERF Surry Sequoyah Peach Grand Gulf (sub- (ice Bottom (Mark lil) atmospheric) condenser) (Mark 1)
Core Damage 7.4E-5 1.7 E-4 5.5E-6 1.6E-5 Frequency Bypass and Early 1.4E-5 1.2E-5 1.6E-6 8.1 E-6 Failure IPE Submittal Large and Early Release 1.3E-5 1.1 E-5 2.6E-7 6.OE-6 1, Te, Cs > 0.03 Large and Early Release 1.1 E 5 8.OE-6 5.8E-8 5.9E-7 1, Te, Cs > O.1 Simplified Event Trees 1.6E-5 7.8E-6 4.2E-6 3.4E-6
O O C~}
- Summary .
i
. Simple event trees have been developed that provide an approach for I l estimating LERF by supplementing Level 1 PRAs ;
i
. A key advantage of this approach is that each accident sequence is allocated to a risk category based largely on the status of the plant i
I
. Event trees for different containment types have been successfully i applied to corresponding Level 1 IPE submittal
- As a result of the applications, staff will consider improvements in Appendix B as part of finalizing general RG after public comment period.
- Some additional clarification and scope related extensions I
- Refinement of split fractions and more detailed justification
?
O O O' Li i
e wv I RISK-INFORMED TECHNICAL SPECIFICATIONS DRAFT REGULATORY GUIDE &
STANDARD REVIEW PLAN ADVISORY COMMITTEE ON REACTOR SAFEGUARDS PRA SUBCOMMITTEE FEBRUARY 20,1997 i
Roy Woods, RES/PRAB Nanette Gilles, NRR/TSB Millard Wohl, NRR/SPSB
Four-Element Approach to Integrated Decision Making Element 1: Define the Proposed Change e Changes to allowed outage times (AOTs) & surveillance test intervals (STis) e Licensee must demonstrate need for request
> Improvement in operational safety
- Consistency of risk basis in requirements
- Reduction of unnecessary burden i
2
~
O O O s
4 i
Element 2: Conduct Engineering Evaluations
- Principle #1
- Proposed Change Meets Current Regulations
- e TS Rule
- 10 CFR 50.36
- License Amendments: 10 CFR 50.90, 50.91, 50.92 l
i 3
o o o '~!
Element 2: Conduct Engineering Evaluations Principle #2: Maintain ~ Defense-in-Depth l,
e Maintain system redundancy, independence, & diversity commensurate with expected frequency & consequences of challenges to system
- Establish appropriate restrictions to preclude simultaneous equipment outages Identify compensatory actions to be taken when entering extended AOT for pre-planned maintenance j
- Do not remove equipment from service when adverse weather conditions are predicted or when plant may be
- subjected to other abnormal conditions 4
l
_ _ _ _ . _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ .7 .
O O O l Element 2: Conduct Engineering Evaluations Principle #2: Maintain Defense-in-Depth e Avoid over-reliance on programmatic activities to compensate for weaknesses in plant design
- Do not rely on programmatic configuration control process to account for large risk increase associated with AOT extension o Maintain defenses against potential common cause failures I f
and human errors Consider whether anticipated operational changes associated with AOT or STI change could introduce new common cause failure modes or human errors i
5 f i
i
O O O I
l Element 2: Conduct Engineering Evaluations i
Principle #3: Maintain Safety Margins l
o AOT or STI change does not adversely affect any t
t assumptions or inputs to safety analysis or, if affected, justification is provided to ensure sufficient safety margin l will continue to exist i
6
O O O' '!
i J
Element 2: Conduct Engineering Evaluations Principle #4: Changes in risk are small and Safety Goals are not exceeded i
e Three-tiered approach to risk-informed technical specification A.OT evaluations i
Tier 1: upper bound guidelines on quantitative risk measures Tier 2: avoidance of risk-significant plant configurations Tier 3: risk-informed plant configuration control i
7
O O O Element 2: Conduct Engineering Evaluations Principle #4: Changes in risk are small and Safety Goals are not exceeded
Element 2: Conduct Engineering Evaluations Principle #4: Changes in risk are small and
' Safety Goals are not exceeded e
e Tier 1 acceptance guidelines
- General RG guidelines for CDF/ACDF & LERF/ALERF i
> Additional guidelines for technical specification AOT changes: !
very small incremental conditional core damage probability (ICCDP) (s 5.0E-7) very small incremental conditional large early release probability (ICLERP) (s 5.0E-8) 4 9
i
o o o ~~!
Element 2: Conduct Engineering Evaluations Principle #4: Changes in risk are small and Safety Goals are not exceeded l e Tier 2 acceptance guideline: Demonstrate appropriate restrictions on dominant risk-significant configurations associated with the change e Tier 3 acceptance guideline: Implemont a risk-informed plant configuration control program & procedures to utilize
& maintain the program 10
Element 3: Develop Implementation & Monitoring Strategies Principle #5: Propose performance-based implementation and monitoring strategies e Three-tiered implementation approach e Monitoring through Maintenance Rule 4
11 i
L;
O O O t
i I
Element 4: Submit Proposed Change i
e Description of & reasons for change e Traditional engineering evaluations performed & results e PRA evaluations performed & results j i
e Description of process to meet three-tiered approach 12 o
. 5 l O O O .
b h
l Pilot Plant Activities l Industry Submittals e CEOG Joint Applications Report for Safety injection Tank i
AOT Extension e CEOG Joint Applications Report for Low Pressure Safety
, injection System AOT Extension l
e CEOG Joint Applications Report for Emergency Diesel Generators AOT Extension e License Amendment Requests from 8 CE plants 13 i
L
Pilot Plant Activities Staff & Industry Interactions e Three requests for additional information e Three pilot plant visits e Numerous meetings with CEOG PSA Working Group e CEOG presentation at ACRS PRA Subcommittee meeting l
i 14
l o o o '~!
i i
i l i Pilot Plant Activities Traditional Engineering issues
- Maintenance of defense in depth
- Some TS restrictions on simultaneous outages
- Three-tiered approach to implementation
- Minimal, if any, effect on safety margins 15
I
' t l Pilot Plant Activities l
Comparison of Lead Pilot Plant Results to Draft Guidelines Safety injection Tanks i
Low Pressure Safety injection System e Almost no discernible change in mean CDF or LERF t
e ICCDP range of 2.3E-8 to 2.9E-7 e Negligible ICLERP (< SE-8) 16
. . . - . . - . - -. - .-. - . _- - - - .-.. ...- -.- - - .. - . - .~. - . . - .- - - - .-
o O O- 9gl
/N United States
,,,,, Nuclear Regulatory Commission t
RISK-INFORMED INSERVICE TESTING i REGULATORY GUIDE (DG-1062)
AND ;
STANDARD REVIEW PLAN (3.9.7)
ACRS PRA SUBCOMMITTEE - February 20,1997 David C. Fischer W. Brad Hardin i I
i t
f 4 !
I
~ ~
O O O i
RISK-INFORMED INSERVICE TESTING REGULATORY GUIDE AND STANDARD REVIEW PLAN DEVELOPMENT TEAM i
t David C. Fischer co-Team Leader NRR/DE/EMEB .
. W. Brad Hardin co-Team Leader RES/ DST /PRAB Michael C. Cheok NRR/DSSA/SPSB ~
Joseph Colaccino NRR/DE/EMEB MANAGEMENT TEAM Richard H. Wessman NRR/DE/EMEB Mark A. Cunningham RES/ DST /PRAB ,
Robert C. Jones NRR/DSSA i
i l
)
i ;
[
- t I
ll
o
~ ~
O O' i
BRIEFING TOPICS I
t i *
= RI-IST team objective i ;
i
= " Classical" IST approach t
i
. RI-IST approach i
. Essential elements of Regulatory Guide (RG) and Standard Review '
Plan (SRP) r
. Status of pilot plant reviews ..
2 !
[
l ii
O O O~
RI-IST TEAM OBJECTIVES .
1 Provide acceptable approach and Guidance to industry for developing and submitting optional RI-IST program Provide guidance to staff for reviewing RI-IST submittals
. Review RI-IST pilot plant program submittals t
i
.f l
l r
t
RI-IST TEAM OBJECTIVES (cont'd) ,
ACCOMPLISHMENTS:
Draft Regulatory Guide DG-1062: "An Approach for Plant Specific, Risk-Informed, Decision Making: Inservice Testing" Draft SRP Section 3.9.7: " Standard Review Plan for the Review of RI-IST Applications" !
. Initial review of pilot applications complete l
t I
FUTURE ACTIONS:
l
. Pilot plant reviews to be completed by 6/97 t
i Continue interaction with ASME on IST code cases I
- -. .. . .. . - . . . = - - . . - - . - - - - . - . - . - . - - . . - . . - _ - . . - - . . - . - - . - - . . - . - . .
O O O' i
" CLASSICAL"IST APPROACH ASME Code referenced in 10 CFR 50.55a i
Program scope includes ASME safety-related code class pumps and ,
i valves i
Like components subjected to prescriptive testing requirements regardless of safety significance
=
Changes in test interval or acceptance criteria reviewed and approved by staff (relief requests) i t
\
=
NRC-endorsed code cases provide clarification of, or alternatives to, ;
code requirements i i
5 I
I i
}
~
O O O ;
RI-IST APPROACH
= Start by ensuring that the plant design and operation is in accordance with the current licensing basis (CLB)
- Ensure that the PRA reflects the actual plant j
=
Establishes RI-IST program using traditional engineering evaluation, PRA insights, and integrated decision making i
i Components categorized as either high or low safety significant (HSSC or LSSC)
RI-IST program scope may include non-code components that are HSSC or LSSC Changes to CLB are identified and justified (defense in depth and safety margins maintained) 6 h
O O O RI-IST APPROACH (cont'd) l
= Test strategy relaxations and possible improvements considered
= Overall effect on plant risk is estimated
. Allows licensee to focus resources commensurate with safety significance
. Step-wise implementation plan
= Performance monitoring and feedback ensures potential problems are promptly detected and corrected t
. Regulatory approval would be an alternative to the Code requirements, as allowed by 10 CFR 50.55a(aX3Xi) 7 i
o o o-ESSENTIAL ASPECTS OF RI-IST REGULATORY GUIDE:
i DG-1962 gives guidance on acceptable methods for utilizing PRA 3 information together with established traditional engineering information in the development ofIST programs that have improved effectiveness regarding the utilization of plant resources while still maintaining acceptable levels of quality and safety.
Consistent with DG-1961 while providing additional application-
! specific guidance.
ESSENTIAL ASPECTS OF RI-IST STANDARD REVIEW PLAN:
SRP Section 3.9.7 provides review procedures and acceptance guidelines for staff reviews. Review procedures are consistent with acceptable methods for implementing a risk-informed inservice testing program. Consistent with SRP Chapter 19 while providing j additional application-specific guidance.
a
O O O ,
ELEMENT 1: DEFINE THE PROPOSED CHANGES TO THE IST PROGRAM
= Identify proposed changes to the current licensing basis (CLB) 1 Identify componc,ats within the scope of the RI-IST program i;
- Identify changes to existing test schedules and methods r
Identify necessary information and analyses to support change i
l
i O O O i I i
ELEMENT 2: GUIDANCE FOR ENGINEERING EVALUATION
- Traditional engineering evaluation (Ensures that the plant is designed and operated in accordance with the CLB)
. Probabilistic risk assessment (PRA reflects actual plant)
. Integrated decision making 1 i
- Key principles addressed:
The proposed change meets the current regulations.
Defense in depth is maintained.
Sufficient safety margins are maintained.
Proposed increases in risk, and their cumulative effect, are small and
. do not cause the NRC Safety Goals to be exceeded.
l =
i
i-s t
ELEMENT 2: GUIDANCE FOR ENGINEERING EVALUATION (cont'd) l
. PRA used to identify pumps and valves that are candidates for relaxation ofIST as well as candidates for enhanced testing Assess change in plant risk from overall change in test frequency or method ,
Expect consideration of common cause failures, results from sensitivity evaluations, human reliability, defense in depth, and safety margin
=
Acceptable plant risk change criteria, uncertainty, and sensitivity analysis in accordance with DG-1061.
-l u
f i
ELEMENT 2: GUIDANCE FOR ENGINEERING EVALUATION ~
(cont'd)
. Consider cumulative effects of RI-IST in relation to other initiatives 1
I
= Licensee expected to review existing approved IST relief requests, maintenance and outage planning, TS, etc.
i l
I I f 12 i t
~
o O O ~! l ELEMENT 3: IMPLEMENTATION, PERFORMANCE MONITORING l AND CORRECTIVE ACTION l
. HSSC typically tested using Code test frequency and method
=
LSSC typically tested at extended interval (up to a maximum limit)
- Expect use of NRC endorsed ASME code cases l
=
Phaseci implementation of RI-IST program ;
l l t
. Monitor component performance to:
- 1) identify degradation (sufficiently early to correct)
- 2) initiate appropriate corrective action
- 3) adjust test frequency based on results
= Periodic reassessment of program
=
Key principle addressed: Performance-based implementation and monitoring strategies 13 l!
~ '
O O O ;
Element 4: DOCUMENTATION i
=
Provide guidance to licensees on minimum documentation necessary to evaluate submittal !
i
=
Must show that proposed CLB change is consistent with the key-principles of risk-informed regulation and NRC staff expectations Provide guidance on PRA records and supporting data to be maintained at the licensee's facility j i
14 I
l s
O O O .;
CURRENT EXPERIENCE WITH RI-IST PILOT PLANTS
. Initial submittals made November 1995
=
Summary of components included in each pump RI-IST Program:
Comanche Peak HSSC LSSC TOTAL pumps 21 12 33 -!
valves 118 516 634 Palo Verde:
valves 152 350 502
= Licensee estimates of change in risk:
Baseline CDF ACDF ALERF Comanche Peak 5.7E-5/RY 7E-6/RY 1.0E-7 with compensatory measures: 8E-7/RY 1.3E-8 ,j Palo Verde 4.7E-5/RY 3.8E-6/RY 3.8E-6 15
o -
O O '
ISSUES TO BE RAISED WITH PILOT PLANTS IN ADDITIONAL RAIS r
=
Methods to improve (or make more realistic) the determination of i baseline CDF, ACDF, and ALERF t 4
t Affect of the proposed RI-IST Program change on the current licensing !
basis Explicitely address the five key safety principles Details of the licensee's integrated decision making process Detailed implementation plans for components (or groups of components)
Step-wise approach Consideration of performance history and service condition ORNL review of NPRDS data Performance monitoring and corrective action plans 16
. _ _ . _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ - _ _ - _ _ _ _ _ _ - _ - _ - _ _ _ _ _ _ _ _ _ _ _ -__ _ _____ _