ML20134P058
| ML20134P058 | |
| Person / Time | |
|---|---|
| Issue date: | 11/22/1996 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-2083, NUDOCS 9611270218 | |
| Download: ML20134P058 (385) | |
Text
.
c .
OfficicI Trcnscript cf Precccdings NUCLEAR REGULATORY COMMISSION O
FCRSD 2623
Title:
Advisory Committee on Reactor Safeguards Probabilistic Risk Assessment Subcommittee f
Docket Number: (not applicable) : TRo4 (Acas) .
RETURN ORIGINAL ,
TO BJWHITE M/S T-2E26 415-7130 Location: Rockville, Maryland l l h Date: Friday, November 22<1996 i
g12 8 961122 T-2083 PDR Work Order No.: NRC-915 Pages 317-619 NEAL R. GROSS AND CO., INC.
n a7007G Court Reporters and Transcribers O;\ -
1323 Rhode Island Avenue, N.W.
' l 43RS? % Sh6E Y:E'
!0r ;r.e UB 0~!';^,8 COTE llIE l
DI5 CLAIMER PUBLIC NOTICE BY THE 1
UNITED STATES NUCLEAR REGUIATORY COMMISSION'S ADVISORY COMMITTEE ON REACTOR SAFEGUARDS i
l NOVEMBER 22, 1996 The contents of this transcript of the ;
proceedings of the United States Nuclear Regulatory Commission's Advisory Committee on Reactor Safeguards on -
NOVEMBER 22, 1996, 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.
i 6 I
i 4
6 4
1 0
317 1 UNITED STATES OF AMERICA
,3
, 2 NUCLEAR REGULATORY COMMISSION
! t 3 +++++
4 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 5 PROBABILISTIC RISK ASSESSMENT SUBCOMMITTEE 6 +++++
7 FRIDAY 8 NOVEMBER 22, 1996 9 +++++
10 ROCKVILLE, MARYLAND 11 +++++
12 The Subcommittee met at the Nuclear Regulatory 13 Commission, Two White Flint North, Room T2B3, 11545 i
G 14 Rockville Pike, at 8:30 a.m., George Apostolakis, 15 Chairman, presiding.
16 17 COMMITTEE MEMBERS:
18 GEORGE E. APOSTOLAKIS Chairman 19 JOHN J. BARTON Member 20 MARIO H. FONTANA Member 21 THOMAS S. KRESS Member 22 DON W. MILLER Member 23 DANA A. POWERS Member 24 ROBERT L. SEALE Member 25 WILLIAM J. SHACK Member NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 2344433
318 1 ACRS STAFF PRESENT:
, m. 2 JOHN T. LARKINS, Executive Director
/
3 MICHAEL T. MARKLEY 4 ACRS FELLOW PRESENT:
5 RICHARD SHERRY 6 ALSO PRESENT:
7 STEVE GOSSELIN 8 JIM CHAPMAN 9 GOVTAM BAGCHI 10 MICHAEL GOLAY 11 BOB JONES 12 GARY HOLAHAN 13 MARK CUNNINGHAM i
\
\
/
L' 14 DICK WESSMAN 15 JOE COACCINO 16 BRAD HARDIN ]
l l
17 MICHEAL CHESK 1 l
18 KEN BALKEY 19 ROBERT GRAMM !
20 ROY WOODS l l
21 STEPHEN DIRSMORE 1 22 MARK RUBIN I l
23 TOM KING 24 GARETH PARRY
(~'s
(_,) 25 JOE MURPHY 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 A-G-E-N-D -A
,- 2 AGENDA ITEM Pace
( /
3 10) Discuss background, goals and objectives, 320 4 expectations for Subcommittee and subsequent 5 full Committee review 6 Industry Presentation 321 7 11) Industry initiatives 321 8 NRC Staff Presentation 9 12) Draft SRP/RG for IST 384 10 13) Draft RG for Graded QA 470 11 14) SRP/RG summary - Questions for public comment 556 12 General Discussion and Adjournment 13 15) General discussion of issues, items for full
,C')
'd 14 ACRS review, PRA Subcommittee recommendations l 15 for ACRS consideration 570 16 17 18 19 20 21 22 23 24 g
() 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND A'K., NY g (202) 234-4433 WASHINGTON, D C. 20005-370) (202) 234-4433
320 1 P-R-O-C-E-E-D-I-N-G-S
,__ 2 (8:34 a.m.)
/sT
\ /
3 CHAIRMAN APOSTOLAKIS: The meeting will now 4 come to order. This is the second day of the meeting of 5 the ACRS Subcommittee on Probablistic Risk Assessment. I 6 am George Apostolakis, Chairman of the Subcommittee.
7 ACRS members in attendance are: Dana Powers, 8 Bill Shack, John Barton, Tom Kress, Bob Seale, Mario 9 Fontana. ACRS senior fellow in attendance is Rick Sherry.
10 The pnrpose of this meeting is to continue our 11 discussion of the NRC staff's approach to codify 12 risk-informed and performance-based regulation through 13 development of standard review plan sections and
,73
(_) 14 associated regulatory guides. The Subcommittee will 15 gather information, analyze relevant issues and facts, and 16 formulate proposed positions and actions as appropriate 17 for deliberation by the full Committee.
18 Michael Markley is the cognizant ACRS staff 19 engineer for this meeting.
20 The rules for participation in today's meeting 21 have been announced as part of the notice of this meeting 22 previously published in the Federal Register on November 23 13, 1996. l 24 A transcript of the meeting is being kept and
/
(h) 25 will be made available as stated in the Federal Register 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 notice. It is requested that speakers first identify
- 2 themselves and speak with sufficient clarity and volume so 1
'~' that they can be readily heard.
3 4 We have received no written comments.
5 However, we have received a request from Mr. Steve 6 Gosselin of the Electric Power Research Institute for time 7 to make a presentation on EPRI initiatives for risk-based 8 in-service inspection, ISI.
9 We will proceed with the meeting. And I call 10 upon Mr. Gosselin to begin.
11 MR. GOSSELIN: Thank you, George.
12 Well, good morning, everyone. It's always 13 nice to be the first one up to bat. Actually, this is my p.,
\_ 14 first presentation at ACRS. I came in. I was looking for 15 the rack. I don't see the rack anywhere here.
16 MEMBER POWERS: There's not a rack. There's 17 not a rack, Mr. Gosselin, but there is a requirement the 18 Committee usually imposes on its first-time speakers to 19 explain why they think they're qualified to come give this 20 talk.
21 (Laughter.)
22 MEMBER POWERS: In all seriousness, sir, we 23 would like to know something about your background before ,
l 24 you begin your presentation. l l
r~
Well, again, thank !
(%) 25 MR. GOSSELIN: Absolutely.
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS <
I 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, O C. 20005-3701 (202) 234-4433 j 1
322 ,
i 1 you very much for the invitation.
s 2 (Slide)
( )
3 MR. GOS3ELIN: My name is Steve Gosselin. I 4 am with the Electric Power Research Institute and the 5 project manager for the risk-based in-service inspection 6 activities that are ongoing at this time.
7 I have approximately 20 years or more 8 experience in commercial nuclear power, 8 years of 9 experience in the naval nuclear program. About the last 10 15 years of that, before I came to EPRI, were with 11 Southern California Edison Corporation at San Onofre 12 Nuclear Plant.
13 Most of my background is in a stress analysis
/'~T i
- 14 and solid mechanics area. I've got a strong background in f 15 systems and design engineering. Most of my work at 16 Southern California Edison was in the design area.
17 My colleague with me is Jim Chapman from 18 Yankee Atomic Electric Company. Jim was a principal !
19 investigator for the work that we did on -- most of the 20 work that I'm going to be presenting to you today with an 21 emphasis in the consequence evaluation and PRA area.
i 22 One other colleague who was going to be here 23 is Ron Gamble from Sartrex Corporation. Ron is an expert 24 in probablistic fracture mechanics and fracture mechanics O)
(, 25 analyses. I'm sure many of you have had dealings with him 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 323 l l
1 before.
I rx 2 What I would like to do this morning -- and I l 3 reali;:e that we do have a rather limited amount of time.
4 And that is to introduce to you the activities that EPRI 5 has ongoing at this time in the area of risk-based or 6 risk-informed in-service inspection.
7 It's our hope that after this morning the 8 Committee might like to have us come back, say after the 9 first of the year, at another time where we can go into 10 much more detail with regard to several of the aspects of 11 this work. We feel that that is going to be important to 12 us to have that opportunity and to get your feedback so we 13 can roll in concerns and comments in the work that we have
,/~T t
\i 14 got going next year. So, with that, I will jump right 15 into it.
16 (Slide) 17 MR. GOSSELIN: As I indicated, this morning I 18 would like to focus really on three areas. I'll give you 19 an overview of the process that we're using to look at 20 risk-informed ISI programs. I'll share with you some of 21 the findings and recommendations that came from the third 22 party review that we had done on this activity and give 23 you a status of the ongoing pilot plant studies that we 24 have in process right now at Arkansas Nuclear One Unit 2 r% \
'w.,)
25 and Fitzpatrick Plant.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234 4433
324 1 In the future, we would like to come back to 2 the ACRS and really get into some detailed technical 7S i
(' / discussions with regard to the risk evaluation, which are 3
4 consequence, damage mechanism assessments, uncertainties 5 comparisons of before and after.
6 But also maybe there might be some time -- and 7 I don't know if this Committee might .oe interested, but a 8 very, very important part of all of this is what we call 9 inspection for cause. After you have determined where you 10 want to look, now we're looking at ISI in a whole 11 different way, developing programs that are really geared 12 to inspect for damage mechanisms and looking at the right i 13 volumes. And if there's an interest, we would like to (D
k_) 14 share that with ACRS as well, the work that we have done 15 in that area.
l 16 (Slide) l 17 MR. GOSSELIN: Just quickly, what were our 18 objectives when we started this work? And we tried to 1
19 bulletize them here for you. First of all, what we're '
20 interested in doing is obviously coming up with a process 21 that we feel is technically sound. But we know today that 1
22 right now this is an alternative for utilities. So it's 23 very important that what we come up with can be 24 implemented in a cost-effective way.
A i ,) 25 Also one of the characteristics we think 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
325 1 very important is that this process be able to have some
,_ 2 level of consistency in its application so that i 1 3 fundamental decisions are made in a similar way from plant 4 to plant and that it be able to be done by plant people.
5 We intend to validate this process through 6 some application studies that are ongoing. And I'll 7 summarize those at the end of this presentation.
8 One key feature that we think is very 9 important is that whatever process you have we believe 10 needs to be what we call process-driven. That is that the 11 process itself will identify the majority, if not all, of 12 the areas in the plant that need to be inspected, that the 13 process provide specific guidance with regard to the s
I \'
k_/ 14 numbers of inspections that need to be performed, and that 15 you provide a selection criteria for personnel that will 16 allow them to select locations to inspection.
17 We think that this is very important because 18 we see the role of an expert panel as really that of an 19 integrative review process. And we believe that their 20 contribution to the scope should be relatively small, more 21 of a refinement, rather than defining the majority of the 22 scope.
23 One other aspect we think is important, as you 24 guys probably know, there are several supplemental types n
() 25 of programs thut are out there that are geared toward NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. i WASHINGTON, D.C. 20005-3701 (202) 234-4433 (202) 234-4433 )
326 1 specific types of degradation mechanisms, like IGSCC and
- 2 Erosion-Corrosion. And these are outside of what's
!' '/ 3 currently included in Section XI. Okay? We think that 4 this is an opportunity to really integrate a lot of these 5 programs into an overall ISI effort.
6 And, consequently and also as a final, we 7 wanted to couple this risk selection method with what we 8 call inspection for cause. Instead of sending somebody 9 out to scrub pipe and simply look for cracks, really we're 10 sending somebody out to look in an area for specific types 11 of damage that are there. So it would be a more focused 12 type of inspection.
1 13 (Slide) )
i
['T
(_) 14 MR. GOSSELIN: We have published an interim !
15 report, this TR. This has been provided to the staff. I i 16 didn't bring any copies with me, but I can provide a copy 17 for Mike that can be -- or copies, how many you would !
18 like, Mike, that could be sent to the Committee to do a 19 review.
20 This is an interim report that we generated ,
l mainly to provide us with some procedures that we could go 21 22 in and do our pilot studies to. And our intention is at 23 the end of these pilot studies to publish a final report, l 24 which would include all the lessons learned and everything r~N
( ) 25 from the last year or so's worth of work.
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 Secondly, there was an independent review that 7s 2 was performed by ERIN Engineering. We have provided this
! \
3 copy to the staff as well and I thought provided a copy 4 for Mike that can be distributed to the members of the 5 Committee for their review as well. We're very interested 6 in any feedback that you might be able to give us in that 7 area.
8 (Slide) 9 MR. GOSSELIN: Okay. Jumping right into this, 10 I want to take just a minute and kind of look at a very 11 big picture here of where we're going with this 12 risk-informed process.
13 In general, the flow of events here, really (3
KI 14 you define some scope. And generally what this process is 15 geared towards is piping and piping components. It 16 includes all the Class 1, 2, and 3 piping and whatever 17 non-code type piping might be in the plant that's 18 important to safety.
19 Now, in this particular process, we look at 20 any number of systems in the plant. For our pilot 21 studies, we're doing a very complete look at anywhere from 22 10 to 15 systems depending on the plant. However, the 23 evaluation is done on a system by system basis.
24 Once you define the scope, we perform a
,r m ,
( ,) 25 failure modes and effects analysis. And it allows us to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHOD 2 ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 I
328 1 break up the piping in a system into segments that have a
- 2 common consequence and common damage mechanism.
3 Once we know that, we're able to arrive at a 4 consequence category and what we call a degradation 5 mechanism category. You can look at these as the 6 consequence of failure and the likelihood of failure, a 7 measure of that.
8 We combine these in a risk matrix that allows 9 us to determine whether the segment falls in a high, 10 medium, or low risk region. This gets an integrated plant 11 review. At one plant, one of our pilot plants is using an 12 expert panel type of an approach. Another pilot plant is l
13 using a QA procedure that's been written. It's l l
l
(
'\ 14 circulated, reviewed in accordance with that procedure, 15 and then comments are resolved.
16 Once this is done, you have those segments 17 within that system that are considered safety-significant.
18 I'll leave this out.
19 (Slide) 20 MR. GOSSELIN: What I'm really going to focus 21 on today in this overview is this activity right in here. l 22 Just for completeness so that you'll see how all of this 23 comes to an end, within a system, once you know what 24 segments are safety-significant, then we define the p
(_) 25 numbers of inspections that are required on those ,
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
329 1 segments.
f3 2 Now, what we were trying to do here was to N] 3 provide a very simple way in which a utility would be able 4 to determine the number of inspections that need to be 5 performed. And we selected initially some percentages 6 that we would use in our pilot studies. And the idea here 7 was we want to try and define a number of inspections that 8 will accomplish three things.
9 Number one, we want to make sure that the 10 number of inspections that we perform give us an adequate 11 representation of the condition that's in the segment that 12 we're inspecting.
13 Secondly, we want to make sure that the number V 14 of inspections that are performed combined with where you 15 look at those locations, the overall effectiveness of the 16 ISI program is improved.
l 17 Finally, the third thing is -- and now it just 18 slipped my mind. Well, it will come back to me. I'm 19 sorry. j l
20 MEMBER SHACK: Steve?
21 MR. GOSSELIN: Yes?
l 22 MEMBER SHACK: What's the time period for j 1
1 23 these inspections?
24 MR. GOSSELIN: Well, currently right now we p)
( 25 have done our work to fit into what's currently being done NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 4
330 1 in Section XI. And for most of these inspections, the
-. *2 time period would be once every ten years.
4
' '/
3 Now, for some degradation mechanisms, like 4 Erosion-Corrosion or MIC, the frequency would be driven by 5 other factors. But the typical inspection on these 6 things, it's once every ten years.
7 MEMBER SHACK: Okay. So this is independent 8 of a degradation-specific mechanism? This you would use 9 __
10 MR. GOSSELIN: This is just a number, how many 11 inspections you would tend to do.
12 MEMBER POWERS: I guess I don't understand.
13 Why is 25 percent an adequate representation of the pipe
\- 14 if it's high risk but 10 percent is an adequate 15 representation if it's a medium risk?
16 MR. GOSSELIN: Okay, i
17 MEMBER POWERS: It seems to me to get an )
18 adequate representation of the pipe, the number of 19 inspections ought to be constant?
20 MR. GOSSELIN: Well, what we have tried to do 21 here -- and this just reminded me of the third thing that 22 I wanted to say. What we tried to do was to weight these 23 inspections more towards the higher-risk areas. In other 24 words, we would have a program where the higher-risk ts l
{ I 25 regions would receive more inspections than the more j uJ NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 23440J WASHINGTON, D C. 20005-3701 (202) 234-4433
331 1 moderate-risk regions. So it would be a weighted fs 2 distribution. And that was the intention here.
( )
~
3 MEMBER POWERS: Well, I can see that in time, 4 but if my objective on any given set of pipe is to have a 5 good representation of what that piping system looks like 6 as a whole that I'm going to select, it seems to me that 7 that selection has something to do with perhaps the piping 8 system and its nature, not its risk significance.
9 MR. GOSSELIN: Yes, it does. You're 10 absolutely correct. And, in fact, once you have completed 11 this page, once you have done here, the risk evaluation is 12 complete. These determinations are made more from the 13 deterministic standpoint on this.
I
\/ 14 MEMBER POWERS: Okay. Suppose I've selected a 15 piping system that I declare to be a medium risk and now 16 I'm going to decide how much of that piping system I have 17 to inspect to understand what its status is. How do I 18 decide ten percent?
19 MR. GOSSELIN: Well, you're saying that all of 20 the segments within the piping system are at medium riak.
21 MEMBER POWERS: Yes.
22 MR. GOSSELIN: Is that how it turns out?
23 MEMBER POWERS: Yes.
24 MR. GOSSELIN: Okay. Well, what we've done is f
() ,
25 we provide in our guidelines this requirement. We've said NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
332 l
1 right now we're going to assume if the segment is at a medium risk, we're going to look at ten percent of the
- p) 2 3 locations in that.
4 Now, through the pilot studies, what we intend 5 to do is to check to see if this is really adequate. In 6 other words, by inspecting this ten percent, will the 7 effectiveness of our ISI Program be as good or better than 8 what we were before or be improved.
9 MEMBER KRESS: How do you do that? By 10 inspecting the other 90 percent to see? How do you decide 11 whether it's been adequate or not?
12 MR. GOSSELIN: Whether it's effective?
13 MEMBER KRESS: Yes.
O i f
'd 14 MR. GOSSELIN: Well, what we're planning to do 15 or what we are doing is once we complete this assessment 16 and we select the locations that we're interested in, then 17 what we do is perform a more quantitative assessment, 18 where we actually look at a comparison of before and 19 after, of the effectiveness of these programs in their 20 ability to address risk. So you can look at the 21 inspections and the locations that they're inspecting 22 before and compare that to what you would do after to see 23 if there is an improvement in effectiveness.
24 These percentages initially were basically
,.cs (j 25 arbitrarily selected on our part as a means of going 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
333 !
1 and doing our pilot studies with. We don't know right now 2 if they will change. We're treating these when we do our f3
( )
t/ So it's a matter of going back 3 pilot studies as minimums.
4 and looking and then determining if we feel that these are 5 adequate.
6 MEMBER POWERS: When you arbitrarily selected 7 these, I suppose that you said when things are at medium 8 risk, it's not as bad to miss something as it is if it's 9 high risk and you made some judgment on the probability 10 of missing an area of localized corrosion, for instance.
11 MR. GOSSELIN: Well, I don't know if our 12 thought process was really that. I think it was much 13 simpler, quite honestly. I mean, we looked, and we said
'/ 14 if you've got a region and a system that has a much higher 15 risk, safety, we felt common sense directed that you would 16 put more attention to those areas and if you had to 17 determine numbers of inspections, that you would tend to 18 weight your inspections towards the higher areas and do 19 less in the medium-risk regions but at the same time make 20 sure that we're at least adequately looking at all of the 21 segments that are important. So it was really not as 22 sophisticated an assessment as you might suggest when we 23 first came up with these.
24 And, again, these were to go in and start
(_,/ 25 looking at some pilot plants and determining whether as a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1J23 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 i
W
334 I 1 starting point -- there we go. !
2 MEMBER POWERS: I would have thought that the 3 frequency of inspections would go up for high-risk 4 sections and lower for the lower-risk sections but the 5 completeness to get an accurate representation would be 6 about the same, pipe.
7 MR. GOSSELIN: Yes. Well, right now I don't 8 believe any of us that are working in this area right now 9 are looking at changing the frequency at which they are 10 inspecting.
11 Now, this question has come up recently by the 12 staff. And our initial efforts were to blend this right 13 in with the current program that's being done. But it's Ik ') 14 clear we're going to have to address also the frequency in ;
15 the sampling selection, including: Do you look at the 16 same location each time or do you look at different l l
17 locations? And that has not been addressed yet.
18 MEMBER SHACK: I guess coming back, how was 19 this connected with your degradation mechanism sort of 20 thing? I mean, somehow I would have assumed that your 21 inspections also should be focused on the place where you, j l
22 in fact, expect to find degradation.
23 MR. GOSSELIN: That's correct. That's 24 correct. What we tried to do here is simply provide some j I
<~x
( ,) 25 guidance as to how many inspections you do. Then we NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C, 20005 3701 (202) 234-4433
335 1 provide very specific guidance as to depending on the type
,_ 2 of degradation mechanism that's present what types of i
' ') 3 areas you should be looking at.
4 MEMBER SHACK: You mean I would expect, then, 5 25 percent, whether I had a piping system, whether, 6 although it was high-risk, I had no mechanisms of 7 degradation that I could really expect to find there and 8 another case that I had a high-risk system that I really 9 expected to have thermal fatigue and IASCC and I'm going 10 to inspect them both 25 percent?
11 MR. GOSSELIN: Yours actually is a good 12 lead-in to my next couple of slides. So let me do that.
13 Okay? And I'll --
,m
/ }
K._/ 14 CHAIRMAN APOSTOLAKIS: Before you go on, for 15 the record, I'd like to note that ACRS Member Don Miller f I
16 joined us a few minutes ago. And also we have Dr. Michael 17 Golay of the Nuclear Safety Research Review Committee in 18 this meeting.
19 (Slide) 20 MR. GOSSELIN: What's really at the heart of 21 what we're doing here is what we refer to as this risk l 22 matrix. And generally the concept behind it is that risk 23 in general is some conditional core damage, core melt 24 potential kinds of potential for that particular damage, i
,x i
( p) 25 so in this case the core melt potential that could be i
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 l
336 0
1 created by a pipe rupture times the potential of having 2 that pipe rupture.
(.,_.s) 3 Now, generally what we are doing for core melt 4 potential, we look at both PSA and deterministic aspects 5 of that evaluation. And for a given particular segment, 6 we classify the impact into one of four impact groups that 7 if that failure would just result in initiating event or 8 would degrade the containment or it would degrade a system 9 or train or if there were some combination that would have 10 to be resolved if you had a rupture in that particular ,
11 segment.
12 Based on the consequence impact group that a 13 rupture in a particular segment might fail, we're able to
,ini Cl 14 rank the consequences to high, medium, and low. And I'll 15 show you how we do that.
16 For the potential for pipe rupture, we focus 17 mainly on the types of damage mechanisms that are present.
18 Once we're able to identify that, knowing what service 19 experience is telling us, we are able to rank the 20 potential for a rupture as either high, medium, or low 21 also. Now --
22 CHAIRMAN APOSTOLAKIS: So the product, then, 1
23 of a high consequence and a low rupture potential is what 1
24 kind of risk?
(g) u-25 MR. GOSSELIN: I'll show you that right now.
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 l I
337 1 Okay?
.2 CHAIRMAN APOSTOLAKIS: Oh, okay.
!g ).
MR. GOSSELIN: Okay. Now --
3 4 (Slide) 5 CHAIRMAN APOSTOLAKIS: There it is.
6 MR. GOSSELIN: Where you have " CONSEQUENCE" up 7 on the top, which is basically a measure of the core melt 8 potential for some limiting break size relative to the 9 potential for a larger break or a rupture and, George, in 10 response to your particular question, if you have a very, 11 very high consequence -- let's say this is a rupture that 12 creates a LOCA and that's a high consequence for that 13 particular plant -- and it's in piping, where there is no
(\
Y' - 14 known degradation mechanism whatsoever, that would have a 15 very small potential for a rupture and a very high ;
16 consequence. And it would put it into what we call a 1 17 medium risk category. In this particular case here, we 18 would look at ten percent of the elements in that 1
1 19 particular segment.
20 If, for example, you had a degradation 21 mechanism, like let's say this was a charging nozzle on a 22 PWR that may be subjected to thermal fatigue, but also if i 23 it breaks you would have a small break LOCA that might be 24 high for a particular plant, in that particular case, the 25 potential with that degradation mechanism, it has a (7_)
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
338 1 potential for a large break, but it's relatively moderate en 2 compared to others.
N 3 So you would be in this medium category and 4 high. And that would put you in a high-risk region. And 5 that high-risk region would mean in that particular 6 segment you would look at least at minimum at 25 percent 7 of the .ons.
8 CHAIRMAN APOSTOLAKIS: Why did you feel you 9 had to develop this new algebra and not stick to 10 probabilities, PSA? Presumably the plant will have an 11 IPE. So now you will have to define, for example, what a 12 medium potential for degradation is and then, you know.
13 For the consequences, perhaps it's easier. But why did Ih i Y- 14 you feel that you had to introduce this and not the IPE 15 itself?
16 MR. GOSSELIN: For our consequence or --
17 CHAIRMAN APOSTOLAKIS: For everything without 18 introducing high, low, and medium, and just say, "If the 19 probability of certain consequences is in this range, 20 inspect 25 percent" or something like that.
21 MR. GOSSELIN: Well, actually, from the 22 consequence standpoint, -- and I'll show you after I get 23 past this slide -- we have very specific ranges of 24 conditional core damage probabilities.
() 25 CHAIRMAN APOSTOLAKIS: Yes. But for the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE , N W.
(202) 2344433 WASHINGTON. D C. 20005-3701 (202) 234-4433
339 1 degradation category, you don't, I don't think. Do you?
2 MR. GOSSELIN: Now, with regard to the 7.
( \
3 degradation mechanism as a failure potential side of the 4 house, we looked at several ways in which that might be 5 done. And we felt that in the traditional -- using 6 traditional structural reliability types of analyses, that 7 the uncertainties -- there were uncertainties associated 8 with that work that we felt would be difficult to -- for a 9 -- difficult to deal with.
10 And also we took a look at our service 11 experience. And our service experience continued to 12 demonstrate to us that the potential for these failures 13 could be derived at by looking at service experience
,rx i
- # 14 itself, that when you account for the uncertainties that 15 are associated with a more rigorous -- not rigorous; 16 that's not the right word -- but a more quantitative type 3 t
17 of TOF assessment, when you consider those uncertainties, I 18 quite honestly, with the variation in the input on the 19 structural reliability analysis, for the most part, you 20 can't do much better than high, medium, or low.
21 CHAIRMAN APOSTOLAKIS: So somebody looks at 22 the evidence, then, that there is some degradation. And 23 that person or persons makes a determination that this is i l
1 24 a medium potential? j r^N l
( ) 25 MR. GOSSELIN: Yes. Well, we have very l l
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS !
1323 RHoDE ISLAND AVE., N W. j (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 ;
340 1 specific guidance. And I'm going to show you some of
,.3 2 that.
/ '4
\ J 3 MR. CHAPMAN: I'm Jim Chapman. Could you put 4 that up, Steve, please, the matrix?
5 MR. GOSSELIN: Yes, Jim.
6 MR. CHAPMAN: We decided on this for two 7 fundamental reasons. One was so that we could use service 8 experience as best we could to characterize the potential 9 for failure, rather than quantitative structural 10 reliability models. That's number one. So that's the 11 vertical column.
12 Secondly, in terms of consequences, as Steven 13 will show you, we wanted our consequence categorization to
,, y t
K~ 't 14 be based on both deterministic principles as well as 15 probablistic principles. So you'll see as Steven goes 16 through this -- and we would love to get more detailed 17 next time -- how we use both deterministic principles and 18 PSA binning categories to get into a low, medium, or high.
19 The other reason we wanted to do it this way 20 really is for display. There is in my opinion significant 21 uncertainty. And the reason that there is not just one 22 high-risk category is to accommodate uncertainties, both 23 in the potential for failure and in the consequences, 24 because I think, as everybody in this room knows, when n
( ,) 25 you're looking at passive failures, there's significant j NEAL R. GROSS COUP.T REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005 3701 (202) 234-4433
341 1 uncertainty, not just in the potential for failure, but in 2 our ability to characterize the consequences. It's a 7
! )
\# 3 significant amount of work. So that's the reason we did 4 it this way.
5 CHAIRMAN APOSTOLAKIS: So you are not trying 6 to avoid the use of probabilities? It's that you have 7 sometimes qualitative information. And it's easier to 8 declare something as being in the high category looking at 9 that information, rather than trying to quantify 10 probability.
11 MR. GOSSELIN: Yes.
12 MEMBER SEALE: Perhaps it was a subliminal 13 wish to avoid getting into arguments about crisp versus A *
(_) 14 fuzzy boundaries.
15 (Laughter.) .
1 16 MR. GOSSELIN: Actually, in all honesty, there l
17 was part of that entered into it. But there was even a 18 more basic desire when we started off. And that was, as I 19 indicated earlier, we wanted to create a process that a 20 utility person at a power plant could use. And when we 21 started looking at doing failure probability calculations, l
22 quite honestly, that is not a typical calculation that's 23 done at a plant.
24 MEMBER SEALE: Yes. We got --
,a l
) 25 MR. GOSSELIN: There are not too many i l
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
342 1 probablistic fracture mechanisms experts at the power zm 2 plant.
(' )
~ ~ ~
3 MEMBER SEALE: Yes. We got into a rather 4 lengthy discussion yesterday about fuzzy boundaries and 5 crisp lines.
6 CHAIRMAN APOSTOLAKIS: Dr. Seale just gave me 7 some ideas.
8 (Laughter.)
9 (Slide) 10 MR. GOSSELIN: Okay. Now, Jim mentioned that 11 we would like to have the opportunity to really get into 12 more detail with this. And, of course, Jim is here to 13 answer a lot of the specific questions with regard to our (v) 14 consequence assessments.
\
l 15 But fundamentally what we did was we looked at I
16 the conditional core damage probabilities that we might 17 want to assign to a high, medium, or low range. And in l
18 doing so, we said, well, first of all, we selected a I 19 conditional core damage probability of less than 10-6 as 20 anything below that, we would consider of relatively low 21 consequence.
22 CHAIRMAN APOSTOLAKIS: Steve, is this --
23 MR. GOSSELIN: I'm sorry?
24 CHAIRMAN APOSTOLAKIS: Go ahead.
, n.
(_) 25 MEMBER POWERS: Could you just remind me.
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 It's conditional on what? I l
2 MR. GOSSELIN: Conditional core damage ems (V) 3 probability given that you have a pipe rupture or what 4 would be the core damage.
5 CHAIRMAN APOSTOLAKIS: The pipe rupture or the 6 --
7 MR. GOSSELIN: It's conditional on having a 8 pipe rupture.
9 CHAIRMAN APOSTOLAKIS: A rupture or the 10 degradation that you have observed?
11 MR. GOSSELIN: Given a rupture.
12 CHAIRMAN APOSTOLAKIS: A rupture, a rupture.
13 Now, why is --
. i
\2 14 MR. GOSSELIN: Or given the limiting break l I
15 size. l l
36 CHAIRMAN APOSTOLAKIS: Yes because you take l 17 care --
18 MR. GOSSELIN: Given the limiting break size.
19 CHAIRMAN APOSTOLAKIS: But is this a real l l
20 probability or is it a frequency?
21 MR. CHAPMAN: It's conditional probability .
I 22 given the pipe has failed. So if you take the piping j
23 failure frequency and multiply it by the conditional core 24 damage probability, you'll get the frequency.
( ,/ 25 CHAIRMAN APOSTOLAKIS: Wait, wait, wait. I'm NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
344 1 confused now. The failure of the pipe itself will serve
, _ , 2 as initiator; right?
(
\-]' 3 MR. CHAPMAN: Yes.
4 CHAIRMAN APOSTOLAKIS: Okay. Okay. Okay.
5 Good.
6 MR. GOSSELIN: Generally we found that these 7 types of consequences are a result of mild initiators and 8 with minimal loss of mitigating capabilities and always 9 maintaining full containment capabilities. So consequence 10 assessments that fall into this area tend to be in this 11 range.
12 CHAIRMAN APOSTOLAKIS: How, if this is a 13 probability, then why did you select these numbers? I s
/i V 14 mean, was there any basis for these numbers?
15 MR. GOSSELIN: Jim, you answer that.
16 MR. CHAPMAN: The basis for low is pretty 1
i 17 straightforward. If one looks at the total frequency of a 18 pipe rupture at a plant, it's on the order of 10-2 per year 1
19 or less. So 10 2 times 10-' we think is low enough as a 20 screen, 10-", per year.
21 CHAIRMAN APOSTOLAKIS: Okay. So you made the i
i 22 connection, then, with standard PRA numbers and --
23 MR. CHAPMAN: Yes, something we thought we 24 could screen.
,/
Q 25 MEMBER SEALE: Actually, it was the age of the 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
345 1 Earth.
2 MR. CHAPMAN: And in the opposite, at the
( )
3 high, on the order of 10 2 times 10-4 we thought made it 4 important, something that should be subject to the higher 5 percentage of inspection.
6 And the medium is, quite frankly, in the 7 middle. And it's what I characterize as a graded 8 approach. it is our safety net in terms of our ability to 9 characterize both the frequency and the consequences.
10 CHAIRMAN APOSTOLAKIS: In your experience, 11 from what you have seen so far, can you tell us where most 12 of the observed incidents fall?
13 MR. GOSSELIN: Well, I can use the reactor es
( i N./ 14 coolant system as an example. For ANO Unit 2, basically 15 about 80 percent of the -- well, let me show you on the 16 matrix here, keeping in mind what these numbers are.
17 We're finding that for PWR RCS, anywhere 18 between 70 to 80 percent of the pipe segments fall into 19 this range.
4 20 CHAIRMAN APOSTOLAKIS: Seventy to 80 percent?
21 MR. GOSSELIN: Excuse me. I've got the wrong.
22 Approximately 20 to 30 percent of the pipe segments fall l 23 into this range, a high consequence with some moderate 24 potential for pipe failure.
- y l 1
(v ) 25 And about 70 to 80 percent of the pipe 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
346 1 segments are falling into this range, which is no active 2 degradation mechanism, very, very low, zero probabilities
\ " l\ 3 of failure with relatively high consequence.
4 MEMBER SEALE: Thinking in terms of work load, 5 which is perhaps mundane at this point, how many segments 6 typically do you wind up with in a PWR RCS?
7 MR. GOSSELIN: Okay. It depends obviously on 8 the number of loops that you have.
9 MEMBER SEALE: Sure, sure.
10 MR. GOSSELIN: At ANO Unit 2, it's a 11 combustion engineering plant. it's got two hot legs and
'l four cold legs. We had approximately 40-plus segments.
13 MEMBER SEALE: Total?
,- m i h V 14 MR. GOSSELIN: Total. Those are continuous 15 runs of pipe that have a common consequence and a common 16 degradation mechanism.
17 MEMBER SEALE: Okay. Okay.
18 MR. GOSSELIN: Okay? Now, we made some ;
l 19 comparisons with the number of segments that were obtained l l
20 by Millstone in the Westinghouse study. And it was 21 somewhat consistent. I believe they had approximately 22 60-some plus segments. And that was four hot legs. So it 23 had an additional two hot legs. so it seemed a pretty 24 reasonable comparison.
,a l
!] 25 CHAIRMAN APOSTOLAKIS: I just noticed 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
347 1 left there you're concerned only about large breaks.
7s 2 MR. GOSSELIN: Yes. We say essentially this I )
\~'/ is a measure of the potential for a large break. Yes.
3 4 CHAIRMAN APOSTOLAKIS: So you don't worry 5 about small LOCAs?
6 MR. GOSSELIN: Well, what we define as a large 7 break is something basically greater than 50 gallons a 8 minute. And that also covers small-break LOCAs. Okay?
9 What we do is we look at a spectrum, really, of break 10 sizes and try and determine the one that's most limiting.
11 Do you want to add to that, Jim?
12 MR. CHAPMAN: When we're doing the consequence 13 analysis, we use the limiting break size, which could be
\l 14 smaller because of the equipment it would challenge.
15 CHAIRMAN APOSTOLAKIS: So it's not necessarily 16 the large break, then?
17 MR. CHAPMAN: It's not necessarily a large 18 break.
19 CRAIRMAN APOSTOLAKIS: So you could drop that 20 word "large" and still represent what you're doing?
21 MR. GOSSELIN: Yes, yes. But I'll tell you in 22 the two plants that we've done thus far, -- and we've ;
j 23 completed all the consequence assessment on Fitzpatrick I l
24 and the majority of the consequence assessment on ANO --
('h
(_) 25 we have not run into a situation yet where a small leak, J NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS j 1323 RHODE ISLAND AVE., N W. l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 i
i 348 1 something less than a large, what we would call a large,
,cy 2 break, ever drove consequence or risk. So we haven't come
( )
3 into that condition where a dripping pipe onto a panel 4 would be of a greater consequence than a very large 5 rupture in the same area.
6 CHAIRMAN APOSTOLAKIS: Well, but there is a 7 spectrum between dripping and large break.
8 MR. GOSSELIN: Yes. So genarally we're 9 finding that anything above the makeup capacity in a plant 10 tends to be the ones that are --
11 MEMBER SEALE: Getting back to my earlier 12 question, what about a boiler? How many segments did you 13 have there?
I
( ) '
\' 14 MR. GOSSELIN: I was hoping you wouldn't ask me 15 that because I don't have that number in the top of my l l
16 head right now. l l
17 MEMBER SEALE: With all those different ways j l
18 to get water into the core, I wouldn't be surprised but 19 what you might have more segments. j l
20 MR. GOSSELIN: Actually, the challenge, I I 21 think, in all honesty, with regard to the boilers is the 22 fact that the code boundaries were defined much 23 differently. And it really necessitated doing a very 24 thorough review of the systems in order to where we're
, /T
\. ,/ 25 going to expand our analysis and how far you were going 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
349 1 expand your analysis to cover non-code-type piping. That 2 obviously may be important. And that was a challenge.
I ,)
~
3 MEMBER SEALE: Yes.
4 MR. CHAPMAN: When Steve responded earlier, it 5 was only on the reactor coolant system piping at ANO.
6 MEMBER SEALE: I underst.and, yes.
7 MR. GOSSELIN: At ANO. And I honestly don't 8 have that number in the top of my head on Fitz.
9 MEMBER SEALE: Okay.
10 (Slide) 11 MR. GOSSELIN: Generally, just to kind of move 12 ahead here a little bit, I skipped the one slide.
13 Generally these impact groups if we look at the
!n!
U/ 14 consequence assessment, I said we look at basically four f
15 impact groups: things that just create initiating events, 26 that that could degrade a system or a train, and something 17 that could bypass containment or jeopardize containment 18 integrity, or some combination of the above. And the --
19 CHAIRMAN APOSTOLAKIS: The " initiating event,"
20 what do you mean? i 21 MR. GOSSELIN: Well, for example --
22 CHAIRMAN APOSTOLAKIS: The break itself is 23 initiating.
24 MR. GOSSELIN: For example, if the result of a r~3
() 25 rupture is simply a LOCA and it doesn't degrade any trains NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
350 1 of safety injection, all the rest of the systems function
,- m 2 normally, then we would place that in a category where the i i
\'J 3 impact of this rupture is basically it causes some 4 initiating event, design basis initiating event.
5 And I've got a table to show you how that is, 6 but it could be anywhere from a LOCA, a small-break LOCA, 7 a steam line break accident, or a reactor trip, loss of 8 load event, loss of flow, the typical design basis 9 initiating events that are in the FSAR.
10 MR. CHAPMAN: Or in the PSA.
11 MR. GOSSELIN: Or in the PSA. I'm sorry.
12 MR. CHAPMAN: Let me add, George, that if you 13 have a pipe rupture, say an ECCS in, say, the suction to a t /
V 14 high-pressure pump, that will not by itself unless there's 15 a spatial issue create an automatic plant trip. It will 16 degrade that standby system. So, even though we treat it 17 as an initiating event, the plant would not be tripped 18 automatically. So basically it's affecting the 19 availability or unavailability of that train.
20 CHAIRMAN APOSTOLAKIS: Okay. So it's not an 21 initiating event in the sense of PSA?
22 MR. GOSSELIN: Yes. That's rlght.
23 MR. CHAPMAN: Correct.
24 MR. GOSSELIN: That's correct.
, ~3
) 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 4 433
I 351 1 MR. GOSSELIN: Now, what we have developed in f~s 2 the reports that I will provide to the Committee, -- these I i 3 tables are in there -- we have tried to create tables *. hat 4 the plant could use in order to make this determination.
5 This is an example of an initiating event 6 impact table. And you can see that basically from this 7 point over, all of this information is in the design basis 8 or PSA of a plant. Okay?
9 And categories high, medium, low, and low for 10 the various initiating events are assigned based on 11 conditional core damage probabilities that exist for these 12 initiating events in the plant PRA.
13 Now, obviously Category I, routine operation, rx i )
\~ ' 14 if your break creates routine operations, obviously its 15 consequences are really none. And it's not really 16 applicable. So the concern, then, really becomes for II, f l
17 III, and IV types of events.
1 18 And, again, these are going to be very 19 plant-specific. And depending on the conditional core ]
l 20 damage probabilities, the appropriate high, medium, or low 21 range would be assigned. So this table would be developed i
1 22 for each particular plant that it looks at. l l
23 Jim? '
24 MR. CHAPMAN: What we wanted to do was to make
(~'\ \
( ,j 25 sure we combine the traditional design basis with the PSA NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
352 1 approach. And this is the traditional characterization 2 for many, many of the plants that are in the fleet right 7S i !
q./
3 now to make sure they looked at both design and PSA and 4 also, quite frankly, to make it better able to communicate 5 it through non-PSA types.
6 MR. GOSSELIN: And, of course, you know, with 7 this type of tables, we're hoping that this adds a great 8 amount of consistency in application from one plant to 9 another, which we thought was important.
10 (Slide) 11 MR. GOSSELIN: This is an example of the 12 system or train impact group table that's in our 13 procedures. And, again, I'll make some general comments 7
, )
' / 14 about this. And, fortunately, Jim is here. We can get 15 into more specifics.
16 But generally you can see we're looking at 17 Category II, Category III, or Category IV type events.
18 Now, if you were in this table, the situation would be 19 aimilar to the example that Jim gave you. Let's say you 20 would have a break. You would assume a break at a 21 location on an ECCS line.
22 That in itself would not create a 23 design-initiating event. It would not create a LOCA.
24 Okay? But obviously the ability to respond to a LOCA has i
n 8
(_,/ 25 been degraded. And maybe you have lost a train out 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
353 1 that ECCS system. So that particular place would put you
<3 2 into this impact group, a particular condition.
( ,!
~
3 And what we have tried to do is structure the 4 consequence categories by taking into account several 5 aspects. One is not only the type of the level of the 6 initiating event that you have to respond to.
7 So if this system was designed to respond to a 8 LOCA, which in a particular plant happened to be a 9 Category IV event, then you would be down at this range.
10 If it was to respond to a reactor trip, if your mitigating 11 capabilities sprayed on some panel and some spatial effect 12 would create a situation where you weren't able to respond 13 properly to a reactor trip, then you would be up into this I i k- 14 table from that end of the spectrum.
15 We try to consider exposure time, how long you 16 might be in this condition before you can take some 17 corrective action to put the plant in a safe condition or 18 correct that situation. And that can range to anywhere 19 from a year or more to either time between testing --
20 maybe this is quarterly testing that might come out; so 21 you would find it when you went there -- either a short or 22 what we call long-term action statements that would be 23 associated with tech specs.
24 And, finally, we consider along here the
/~N
(_), 25 number of available backup trains that one might have 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
354 1 respond. You can see obviously if you have any situation 7~s 2 where you have no backup trains and you've lost complete N ]
3 capability to respond to any of these initiating events, 4 that would always put you in a high consequence area.
5 Also what we're finding that's not directly 6 shown on this table, although when you use it we're 7 finding it's very valuable, is that we're able to really 8 kind of provide a lot of flexibility in how we address 9 varying types of reliability issues, a spatial effect you 10 may ha'.a where you're flooding a room and you may have a 11 motor-operated valve.
12 The question then is: Given I have an 13 exposure time of that amount of time, let's say, how does ,
l r~~S I i
\ -
14 that affect the reliability? What is my response time?
15 And if there is some question, you can either move to the 16 left or up to account for any kind of reliability (
I 17 questions that you might have that may enter it. They're 18 not directly shown on this table.
19 We've got some examples sometime to show how l 20 that works.
l l
21 CHAIRMAN APOSTOLAKIS: Instead of number of l 22 unaffected backup trains, shouldn't you be talking about 23 the number of unavailable backup trains? I mean, they may 24 not be affected by this particular mechanism, but they may 25 be unavailable due to other reasons.
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
355 1 MR. GOSSELIN: No. This is correctly stated.
- g. 2 What we want to know is --
( )
~ But you're right. Behind each 3 MR. CHAPMAN:
4 value here in terms of the number of unaffected trains is 5 a reliability range, an unavailability, unreliability 6 range. So in order to credit one backup train, it has to 7 be in the 10-2 to 10-2 unavailability, unreliability range.
8 Two backup trains are in the 10-' range. So there's a PSA 9 backing to it.
10 In fact, in order to get into a category, 11 high, medium, or low, you have to pass the redundancy 12 test, which Steve has displayed here, as well as the 13 conditional core damage probability test. You have to rm V 14 pass both tests.
15 CHAIRMAN APOSTOLAKIS: Now, these 4 l
16 unavailabilities come from the IPE?
17 MR. CHAPMAN: Yes.
i 18 CHAIRMAN APOSTOLAKIS: Because this is a real 19 case now. You are at a real plant. So you will know 20 whether some of your other trains are down; right? If 21 they're nominally available, you don't know whether they 22 will start. So you can use the PSA numbers.
23 MR. CHAPMAN: Correct.
24 CHAIRMAN APOSTOLAKIS: But if you know they
,n i ) 25 are down for sure -- the PSA, in other words, has to be 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
356 1 adjusted to the current condition of the plant.
, 2 MR. GOSSELIN: I may not understand your i <
3 question when you say "You know that the system is down."
4 CHAIRMAN APOSTOLAKIS: When you do a PSA, you 5 are essentially looking into the future. So you're saying 6 this particular system --
7 MR. GOSSELIN: Oh, I see. Yes.
8 CHAIRMAN APOSTOLAKIS: -- may be down for 9 maintenance or train and so on.
10 MR. GOSSELIN: Yes, yes. I understand.
11 CHAIRMAN APOSTOLAKIS: When you do this, 12 though, you're not doing that. You have discovered the 13 potential problem. And now this is done in real time, is
[
's / 14 it not?
15 MR. CHAPMAN: No.
16 MR. GOSSELIN. No.
17 MR. CHAPMAN: The inspections -- and, Steven, 18 correct me if I'm wrong. Most of the inspections are done 19 during an outage. So what you're trying to do is plan for 20 the inspections on the basis of the importance of those 21 pipe segments to operation into the future.
22 CHAIRMAN APOSTOLAKIS: Until the next outage.
23 MR. CHAPMAN: Until the next outage.
24 MR. GOSSELIN: But we do consider all
,/7
(, ) 25 different modes of operation when we do this assessment.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISI.AND AVE., N W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
357 1 CHAIRMAN APOSTOLAKIS: But you may have some
,y 2 information, though, that goes beyond what went into the
\
\
3 original IPE.
4 MR. CHAPMAN: Yes. And, again, what we're 5 trying to do here and the reason we're trying to do this 6 is to address as many of the uncertainties, both technical o
7 and review, as we can. The first line of defense in ter=c 8 of getting into a high, medium, or low is a deterministic 9 basis.
10 MR. CHAPMAN: As Steve is explaining, the 11 backup to the deterministic basis is the PSA basis. So 12 you hit both. So either one should get you in. And, as 13 Steven said, if there are uncertainties and let's say the
()
i
's Y 14 consequences of an event, the default is to go to a higher 15 category.
16 If you want to spend more effort to better 1
17 characterize it, you can move down if, in fact, onze you (
18 get that information it is a lower consequence category.
19 So it gives flexibility in terms of resources.
20 Some of these events could be so difficult to 21 analyze, you may decide not to implement this program and, 22 therefore, not have the benefit of it. Okay? 1 l
23 CHAIRMAN APOSTOLAKIS: Now, when you declare 24 something as low, according to the first flow chart you A
Iq,,) 25 showed us, there is no requirement for the number 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
1 358 1 inspections, is there?
rx 2 MR. GOSSELIN: You're referring to this?
t i
~
3 CHAIRMAN APOSTOLAKIS: No. Way back.
4 MR. GOSSELIN: I'm sorry.
5 MEMBER SHACK: Your 25 and 10 percent was 6 nothing for low. It was high and medium only.
7' MR. GOSSELIN: Yes. Well, what we were 8 talking about there was volumetric or types of 9 examinations for those segments that fall into this 10 low-risk region.
11 CHAIRMAN APOSTOLAKIS: Yes.
12 MR. GOSSELIN: What we continue to do is 13 normal leak tests and leak tests and system walkdowns and
/ N 14 leakage monitoring systems. We don't specifically call 15 out volumetric inspections.
16 MR. CHAPMAN: Can I enhance that a little bit?
17 MR. GOSSELIN: Sure.
18 MR. CHAPMAN: What we just presented was the 19 consequence category, not the risk category. If you go up 20 to low consequence category, you -- Steven, I need that up 21 there.
22 MR. GOSSELIN: Oh, yes, yes, yes.
23 MR. CHAPMAN: If you go to the low consequence 24 category, if, in fact, the potential for degradation is
,/ m k/ 25 small or medium, you're correct. You fall into 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
1 359 1 low-risk category.
~s 2 But if there's an active degradation mechanism 3 to address precursor events, we want to inspect because we 4 don't want pipe failures. So if there's an active damage 5 mechanism, even if it's low consequence, we will be in a 6 ten percent inspection category. Okay?
7 MR. GOSSELIN: Well, let me enhance that even 8 a little bit more. What we have done is we have said for 9 low consequence types of segments, if the potential for a 10 pipe rupture is very high, we want to make sure that we 11 capture that.
12 MR. CHAPMAN: Right.
13 MR. GOSSELIN: So items like x
! \
k/ 14 Erosion-Corrosion, areas where there is flow-accelerated 15 corrosion, even if the consequence is low, we want to make 16 sure we're inspecting that area.
17 CHAIRMAN APOSTOLAKIS: So what's high, medium, 18 or low takes into account the time until you do something 19 about it. Is that correct?
20 MR. GOSSELIN: It does in a sense for --
21 CHAIRMAN APOSTOLAKIS: If you decide, for 22 example, your next outage is in six months and sey that's 23 low, we're not going to do it, we'll do it at the next 24 one, now you have a longer period of time. You don't --
g3
( ,) 25 MR. BAGCHI: Mr. Chairman, this is Goutam NEJd. R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
360 l l
1 Bagchi from staff.
2 I wanted to point out that these passive 7
i i
\"/
3 segment inspections can be done online, can be done when 4 the plant is in operation.
5 MR. GOSSELIN: That's correct.
6 MR. BAGCHI: So it's not during refueling 7 outages only.
8 MR. GOSSELIN: And, actually, for your 9 particular concern, George, where you have a low 10 consequence and where you have what we believe a 11 relatively high potential for a large break, like 12 flow-accelerated corrosion, a lot of those programs are 13 inspected online.
t
(~ i D/ 14 CHAIRMAN APOSTOLAKIS: But what is high 15 potential? I mean, doesn't that take into account time?
16 MR. GOSSELIN: Yes, it does, actually, in a 17 sense. I know I have been focusing on consequence, but 18 I'm going to get to how we determine this. When you're 19 looking at service experience, indeed it does account for 20 time.
21 MR. CHAPMAN: This is one of the reasons we 22 would like to get with you folks again with more time.
23 MR. GOSSELIN: Yes. I'm quickly just going to 24 give you a sense of the other tables. They're in your
-~
.) 25 handout. Again, we're on the consequence assessment 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
l
~
361 1 now and trying to determine consequence category.
7, 2 (Slide)
( )
3 MR. GOSSELIN: This is the combination impact 4 group table. You can see this is a very straightforward 5 -- or no. This is containment impact group. What we are 6 concerned about here are breaks where you could maybe 7 bypass containment integrity. And we have established 8 some criteria here based on the number of active and 9 passive failures that would be associated with that.
10 MEMBER FONTANA: What would you call a check 11 valve? Passive?
12 MR. GOSSELIN: A check valve.
13 MEMBER FONTANA: Yes.
! / I'm not quite sure how we would is 14 MR. GOSSELIN: ,
j l
15 hancle that.
l 16 MR. CHAPMAN: If it is normally closed, 17 passive.
18 MR. GOSSELIN: Passive, yes, normally closed.
19 (Slide) I I
20 MR. GOSSELIN: For the combination impact i 21 group table, what's significant here that I just wanted to l
22 point out is that when we're looking at these high, i
23 medium, and low categories, one would either pull those i 24 category '- .ls off this table or they would use the
,9 ;
j 25 initiating event table category, whichever is greater. ;
)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS ,
1323 RHODE ISLAND AVE., N W. 1 (202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
362 1 So if this particular table would suggest
, ~s 2 medium but the initiating event was LOCA and you had two
' 1 LJ 3 unaffected trains, a mitigating system, that would suggest 4 a medium consequence. But in the initiating event table, 5 that LOCA may be high. So you would use the high 6 consequence.
7 So, as I indicated --
8 CHAIRMAN APOSTOLAKIS: So you're sure that 9 this is simpler than just looking at the PSA? One would 10 have to train oneself now to work with initiating events 11 and two unaffected trains and so on. If you find it 12 easier, then -- you're the guys doina it.
13 MR. GOSSELIN: Well, actually, what we're
()
kd 14 finding is that this type of terminology is very 15 comfortable at the plant level. Mainly they're used to 16 thinking in these terms.
17 MR. CHAPMAN: You'll also find that in order 18 to use the PSA, you have to know all the impacts in order 19 to, in essence, requantify it, which would tell you -- it 20 already has the information to tell you how many trains 21 are available, given a particular impact. So the 22 information is there.
23 And, again, -- I probably have said it too 24 much -- what we're trying to do is make sure we have a
,m k) 25 solid deterministic basis backed up by a PSA basis. 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
363 1
1 again, the reason for that is the damage mechanism g x, 2 uncertainty in terms of quantifying it and, two, the level ,
x -)
1 3 of review in my mind necessary to fully validate a PSA.
4 So we're trying to have two lines of defense.
5 MEMBER POWERS: It looks to me like a ,
i 6 reasonably risk-informed piping engineer can do this 7 evaluation all at his desk by himself.
8 MR. GOSSELIN: Yes, but you need to be out 9 there, and you need to walk down those systems. You also 10 need --
11 MEMBER POWERS: Well, he has to know.
12 MR. GOSSELIN: -- a systems engineer. You 13 need those sorts of things --
! \
' 14 MEMBER POWERS: He has to know the systems.
15 And he has to be reasonably informed about the results of 16 the PRA. But given those two things, he can line out an 17 inspection program and a justification --
18 MR. GOSSELIN: Yes.
19 MEMBER POWERS: -- just using these tables.
20 MR. GOSSELIN: Yes.
21 MEMBER POWERS: I mean, it looks doable to me.
22 MR. GOSSELIN: Yes.
23 MEMBER POWERS: He might well subsequently 24 give it to another engineer to review, but it's not like
()
(_,) 25 you have to have a panel of people doing this. One guy 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 364 f 1
1 can do it if he's reasonably informed on it. If he's not,
-s 2 then he has to get those resources in.
(
3 MR. GOSSELIN: That's correct.
4 (Slide) 5 MR. GOSSELIN: Well, what we spent some time 6 on here was just really trying to describe how we enter in 7 this portion of this matrix. And I'll just take a few 8 minutes and show you how we are entering this side of this 9 matrix.
10 (Slide) 11 MR. GOSSELIN: Now, this work, just quickly, 12 we know that in general failures are occurring out there 13 because of some act of degradation mechanism that's s
i )
\w/ 14 present. And basically they boil down to fatigue, thermal 15 fatigue types of mechanisms, which could be transient, 16 thermal transients or striping concerns; some 17 corrosion-related activities, which could be corrosion 18 cracking, stress corrosion cracking, some local corrosion, 19 MIC or 0 2 -type stuf f, erosion and cavitation types of 26 events, and flow-accelerated corrosion.
21 And we found that we believe that when you 22 look at the data, that you can make a reasonable 23 determination as to the relative potential for a pipe 24 rupture by understanding what degradation mechanism is
,q
() 25 present.
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
365 i
1 (Slide) 2 MR. GOSSELIN: Now, how do we identify these
,s
- )
'# 3 things? Well, what we have done is we have provided 4 explicit defined attributes, explicitly defined 5 attributes, for these degradation mechanisms. To put it 6 quite simply, -- I've said this before in other 7 presentations -- what we are trying to do here is provide 8 a set of attributes such that the plant engineer, who 9 knows how his plant is designed and knows how it's 10 operated, can make these determinations.
11 These attributes have been determined for each 12 degradation mechanism we're considering. And they're 13 based on industry experience, the causal effects, a lot of
/x
! )
\_/ 14 research that has been done at EPRI over the last years; 15 and many of the NRC requirements that are out there as 16 well.
17 (Slide) !
l 18 MR. GOSSELIN: I've provided this. This was 1
19 just some initial sampling for some of the references that 20 we have used in our report. And those are noted in the I i
21 report that I will be providing to the Committee. ;
i 1
22 (Slide) l 23 MR. GOSSELIN: Now, just to take a look at the 24 service experience, I included the slide to kind of give i
()
25 you some information. We at EPRI are doing a lot of work 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
366 1 in putting together as most extensive database that we can
,_s 2 that's being hopefully validated to the maximum extent I ]
3 possible by the various utilities on piping failures.
4 We've got an ongoing effort at this time where we're 5 working with all the owners' groups to build this database 6 up beyond what was originally done maybe five or six years 7 ago.
8 This is just a snapshot where we've got about 9 1,511 pipe failure events. Basically, these are failures.
10 They could be leaks and breaks, ruptures. These were the 11 failures that occurred and kind of how they broke down 12 damage mechanism-wise.
13 MEMBER FONTANA: None of them failed during an p 4
\/ 14 earthquake, though?
15 MR. GOSSELIN: No, none of them failed during 16 an earthquake. That's correct.
17 MEMBER FONTANA: Okay.
'18 MR. GOSSELIN: Okay? j l
19 (Slide) 20 MR. GOSSELIN: What I provided here is just 21 kind of a snapshot of how some of these broke down with 1
22 regard to some nominal rupture frequency when you look at 23 the data, rupture being any kind of large break.
24 It is pretty interesting results here. These
/~N And you can see here that l
) 25 columns are defined up here.
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
367 1 when you look at vibration fatigue and you look at 2 flow-accelerated corrosion mechanisms, that these rupture
,7 r i
('/ 3 frequencies tended to f all up in the 10-2 area. And, 4 simply stated, this is just looking at the number of 5 ruptures that occur over the number of reactor years of 6 operation that the data included.
7 MEMBER FONTANA: Does that imply to let a pipe 8 vibrate until it breaks?
9 MR. GOSSELIN: Well, we do have failures. And 10 they are vibrating and breaking.
11 MEMBER FONTANA: Okay.
12 MR. GOSSELIN: And it's mainly small bore 13 piping at socket-welded connections where we're having r_3
\v) 14 that problem. It's a very difficult problem because the 15 problem with high cycle vibration is an initiating event, 16 as you probably know, or initiating mechanism. And you 17 could go out and inspect this location today. And 18 tomorrow it could fail. One it initiates, it goes very 19 rapidly to failure. It is a difficult problem to manage.
20 MEMBER BARTON: Would the vibration frequency 21 also include instrumentation lines that vibrate or is it 22 strictly --
23 MR. GOSSELIN: There were failures that 24 included instrumentation lines in the database, but'I
()
25 filtered that out.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS !
1323 RHODE ISLAND AVE., N W. J (202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
368 1 MEMBER BARTON: Okay. So this is strictly --
- 2 MR. GOSSELIN: It's just piping. It didn't 7
N.-]I 3 include swedgelock connections that vibrate loose or 4 anything like that.
5 MEMBER BARTON: I understand.
6 MR. GOSSELIN: Okay. These are pipes. What's 7 also interesting over here is -- take a look down here.
8 This erosion and cavitation, stress corrosion cracking, 9 thermal fatigue, corrosion cracking.
10 In these areas in these mechanisms, there were 11 no large breaks or ruptures in the database. There were 12 actually zero. For the purposes of this, I assumed one, 13 the occurrence of at least one. And we probably could i
,-)
14 have done some Bayesian approach to come up with some more 15 realistic assumption there, but for the purposes of what I 16 was trying to accomplish, I just assumed, well, let's 17 assume I do have one and I've missed it. Okay?
18 And you can kind of see how the breakdown was.
19 These mechanisms, including -- this is local corrosion.
20 These mechanisms, most tend to be around 10. And up in 21 here we're looking about 10-2 And we felt that, hey, the 22 data continues to support that, depending on the mechanism 23 you're looking at. Its relative likelihood of creating a l
24 large rupture or break can be determined. I
,n,
(_,) 25 When you think about it, it makes sense. 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 l l
369 1 mean, think about thermal fatigue. You have this action 2 going on. You get a crack. It grows. It starts to leak.
,-s.
( )
\'~/ And then the plant by tech 3 Eventually it gets picked up.
4 specs has to take some action before it is allowed to grow 5 to anything greater than makeup capacity, let's say.
6 CHAIRMAN APOSTOLAKIS: So this table here, 7 this figure, is derived from the data you just showed us?
8 MR. GOSSELIN: Yes.
9 CHAIRMAN APOSTOLAKIS: But these were actual 10 pipe failures?
11 MR. GOSSELIN: Yes.
12 CHAIRMAN APOSTOLAKIS: But they didn't 13 necessarily lead to a LOCA or a large LOCA --
YM 14 MR. GOSSELIN: That's correct.
15 CHAIRMAN APOSTOLAKIS: -- or a medium LOCA?
16 MR. GOSSELIN: That's correct.
17 CHAIRMAN APOSTOLAKIS: I see.
18 MR. GOSSELIN: Yes, that's correct.
19 CHAIRMAN APOSTOLAKIS: So how did you get the 20 10-27 I mean, that's a judgment or a -- l l
21 MR. GOSSELIN: No. You know, actually, l 22 basically what we did is we took the number of pipe 23 ruptures that occurred over the number of reactor years of j 24 operation.
() 25 CHAIRMAN APOSTOLAKIS: For each one you know NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 23 & 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
370 1 what kind of a LOCA that is or what kind of a leak rate 2 you --
7
('" ) MR. GOSSELIN: Yes. In the database, it's 3
4 characterized into either a leak or a crack or a large 5 break or a rupture or severed pipe. There were several 6 classifications.
7 MR. CHAPMAN: I don't think any of these were 8 reactor coolant system LOCAs. But, as an example, 10-2, 9 roughly 2,000 reactor years experience in this.
10 MR. GOSSELIN: This involved about 2,000 11 reactor years worth of experience.
12 CHAIRMAN APOSTOLAKIS: Doesn't this, though, 13 depend very much on what you called piping failure events?
,s
! \
C/ 14 In other words, what if you had, say, erosion of a piping 15 segment which did not actually lead to a leak? That would 16 not have been included in the numbers that you showed us 17 in the data slide.
18 MR. GOSSELIN: Actually, the number of 19 failures for -- like take this particular case, erosion )
1 20 and cavitation.
21 CHAIRMAN APOSTOLAKIS: Yes.
22 MR. GOSSELIN: The number of failures that 23 occurred in the database were all leaks.
24 CHAIRMAN APOSTOLAKIS: No. But there may be
,m 25 other incidents where you have significant amounts 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
'n 371 1 erosion that did not lead to leaks. Is that correct?
2 MR. GOSSELIN: That's correct. And they may N]
3 have been detected and corrected, yes.
4 CHAIRMAN APOSTOLAKIS: These are not included 5 in the database.
6 MR. GOSSELIN: That's correct. That's 7 correct.
8 CHAIRMAN APOSTOLAKIS: Is it fair, then, to 9 say that we are underestimating a little bit the condition 10 of frequencies here just because you caught them before 11 they led to me leak? Now, that may be a minor effect, 12 but --
13 MR. GOSSELIN: I don't know.
p b
\
V 14 CHAIRMAN APOSTOLAKIS: That probably 15 underestimates the numbers because you had 280 number of 16 failures due to erosion and corrosion. Now, that number 17 may go up depending on what you call a failure; right?
18 MR. GOSSELIN: Well, one of the things that 19 we're also trying to do with this database, George, at 20 this time is we're also getting non-leak identifiers in as 21 well, cracks and non-leaking types of events, and 22 including that. And I think once we get that in, we may 23 be able to factor in some of those concerns.
24 MEMBER SHACK: The number of failures has got (q) v
?5 to be leaks; right? Those aren't ruptures.
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
372 1 MR. GOSSELIN: Depending on the mechanism, 73 2 there are some cases.
}
l'/ 3 MEMBER SHACK: There are a few ruptures in 4 here, but you certainly didn't have 166 stress corrosion 5 ruptures.
6 MR. GOSSELIN: No. In fact, we had none. We 7 only had leaks. There are leaks.
8 MEMBER SHACK: In vibration fatigue, yes, 9 they're probably all failures.
10 MR. GOSSELIN: When we say " rupture" here, 11 we're meaning large breaks on the order of significant 12 gallons a minute. The idea was really to look at that and 13 to see: Hey, this hypothesis that we set out to do, is it 14 doable? And the service experience basically confirmed 15 that for us.
16 So generally what we have done, as you can see 17 in our procedures, is that we set up again a criteria 18 where the measure of a break potential, high, medium, or 19 low, where you have high being large leak conditions, 20 medium leak conditions are none, are classified depending 1
l 21 on the degradation mechanism that you determine is present l
22 in that area.
23 So now remember now that we're dividing our 24 segments into continuous runs of pipe that have not only
,- 1
( ,) 25 the same consequence but have the same degradation NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 l
373 1 mechanism throughout that segment. Okay?
n 2 (Slide)
(N >) 3 MR. GOSSELIN: Just a side point really 4 quickly, and then we'll move on. This comes up all the 5 time. What about water hammer? And we know that water 6 hammer is a significant loading condition and has a 7 potential to create a lot of damage.
8 Just as a perspective, remember what we're 9 trying to do here. We're trying to design an inspection 10 program. So to go ut and inspect your pipe because you 11 think it may have a water hammer is not buying you a lot.
12 So how do we deal with that? Well, what we
,_ 13 have done is this. Generally if you know you're going to I I
- ' 14 have a water hammer, as a designer, I know I have to l
15 design for that in my piping systems. Okay? And I 1
16 generally have some pretty conservative criteria that I 17 use in Section III to account for that. So we're not 18 finding too many cases where if we have a hammer and we 19 have designed for the hammer that it really creates any i l
20 kind of rupture. )
l 21 But let's say I have designed for a water 22 hammer. I know I have the potential for a water hammer. l l
23 But I made the design calculations assuming pretty good 24 piping condition. And now this piping has been in ,
I
/ ~.
k_-) 25 operation. And if there's a presence of a degradation NEAL R. GROSS h COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. l (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 )
l
374 1
1 mechanism that could grade that piping, then I've got to 2 consider the fact that I could have an additional loading 7
i
)
3 condition that's beyond what would be norinally expected.
4 So what we do is this. Let's say you've got a 5 situation where you've got some mechanism in here that 6 would generally have a medium potential for a pipe 7 rupture. But you know in that area whenever the guy goes 8 out and he starts the high-pressure safety injection pumps 9 to do an IST, he gets a huge check valve slam and the 10 pipes move around. Very dynamic. And you may haven't 11 analyzed that in your event. So there is obviously a 12 potential for some hammer loading.
l 13 In that particular case, if we know there's a n
/ 1 V 14 degradation mechanism present and that area could be 15 subjected to hammer, then we move it up into the high i
I 16, region. So that's how we deal with water hammer.
17 (Slide) 18 MR. GOSSELIN: Okay. Now -- j 1
19 MEMBER POWERS: And vibration?
20 MR., GOSSELIN: Well, no. Vibration, again, 21 that's another one of those things that you can't go out 22 and inspect for. And what we have tried to do is provide l
23 some guidance as to where you might look to inspect for l 24 those. Rely on a lot of frequent walkdowns at this time 1
(-) 25 and focusing your atten; an in these areas that have close .
)
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 375 1 to vibration sources with small bore piping and 2 socket-welded connections.
7 1
3 Now I'm going to pick up the pace very 4 rapidly. I just want to touch base on a couple of things.
5 This process in the report that I provide to you, I had a 6 third party review done by ERIN Engineering. The 7 principal reviewers were Karl Fleming, Craig Sellers, and 8 Doug True.
9 CHAIRMAN APOSTOLAKIS: So these are principal, 10 not principle?
11 (Laughter.)
12 MR. GOSSELIN: Oh, yes. Sorry. They were 13 principle stress. I don't know. I must have been ,
fs. l l
V) 14 thinking principle stress. But, anyway, the whole 1 15 purpose, I asked them to provide recommendations to us, 1
16 look at this and provide some recommendations, and 1
l 17 findings which we could build on.
l 18 I've provided a couple of slides in the 19 handout for you that summarize their findings.
20 CHAIRMAN APOSTOLAKIS: So which one of these 21 three reviewers is the one who understands these 22 mechanisms? Mr. Sellers? Of the three reviewers --
23 MR. GOSSELIN: As far as the damage mechanism?
24 CHAIRMAN APOSTOLAKIS: Yes.
(
) 25 MR. GOSSELIN: I had another group do a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
CO2) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
376 1 specific review on that, and that was Structural Integrity
_ 2 Associates. And Pete Ricodell and his staff reviewed !
(( y) 3 that. That report is coming out, and I'll be happy to 4 provide that.
5 CHAIRMAN APOSTOLAKIS: So these three 6 gentlemen really reviewed the methodology and --
7 MR. GOSSELIN: They reviewed the overall 8 methodology and didn't review the validity of the damage 9 mechanism assessment. That work is finishing at this 10 time.
11 (Slide) 12 MR. GOSSELIN: In any case, the findings here 13 are summarized, and you have a copy of the report. I iA)
V 14 won't spend a lot of time on this. Some of the things 15 that were significant were there was a lot of discussion 16 of whether this is a qualitative quantitative approach.
17 And I think that we tended not to get absorbed 18 on what we call this, but it was Karl Fleming's issue. He 19 really kind of viewed this as more of a semi-qualitative 20 kind of approach that tried --
21 CHAIRMAN APOSTOLAKIS: He did not introduce 22 any multiple Greek letter methods here?
23 (Laughter.)
24 MR. GOSSELIN: I'm sorry, George. I didn't n s
) 25 hear 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
377 1 CHAIRMAN APOSTOLAKIS: That's okay. You are
- 2 better off for it.
t
~
3 (Laughter.)
4 MR. GOSSELIN: Is that to bring the rack in?
5 MEMBER SEALE: You won.
6 MR. GOSSELIN: Okay. He did have some 7 recommendations that we have taken some action on. Part 8 of it is some of the things that I showed you here today.
9 And in this area of pipe breaks, leaks, flaws, and 10 associated failure mechanisms, he felt in our report we 11 really needed to beef that discussion up in our interim 12 report, which we didn't include in there when we were 13 trying to put some procedures together. So there's a lot
(\ ') 14 of good recommendations that we're going to be correcting.
15 (Slide) 16 MR. GOSSELIN: All right. In the next couple 17 of minutes here I will just quickly summarize, give you an 18 idea of where we are with some of our pilot studies. At 19 the ANO Unit 2, we have a pilot study ongoing. It's an 20 EPRI-Taylor collaboration project with six CE utilities.
21 At ANO 2 we're looking at ten systems, and 22 I've identified them here. That includes all the Class 1, 23 2, and 3 piping as well as non-code piping. We're a 24 little more than 70 percent complete right now. Actually,
,e
() 25 this is an older slide that I gave about a month and a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 2344433
378 1 half ago. We're looking at final submittal of this to the y 2 NRC right after the first of the year.
(J
)
3 Just to give you an idea of some of the 4 results that we're seeing, in the RCS we're looking at 5 somewhere around about a 65 percent reduction in the UT 6 exams.
7 With regard to PT exams, surface exams, 8 because we're implementing this inspection for cause 9 process, a lot of these degradation mechanisms all come 10 from the inside. And going off and doing surface exams on 11 a weld really doesn't buy you anything. But there are 12 some locations where you have maybe some global bending 13 effects, like due to stratification, where we are looking
.\ '/ 14 at some surface areas where we could have initiation of 15 cracks on the outside surface. But we're finding that for 16 the most part a lot of those exams that are currently 17 doing it Section XI really don't provide much value.
18 (Slide) i 19 MR. GOSSELIN: I've provided a slide in here i 20 to kind of give you an idea of what the systems are that I
21 we're looking at and what their ASME classification is. l l
22 It's not necessary to spend any time on that.
23 (Slide) 24 MR. GOSSELIN: Fitzpatrick -- ]
<s i
) 25 CHAIRMAN APOSTOLAKIS: Are these our pilot 1
1 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
379 1 programs, too? No?
2 MR. JONES: Yes.
[, ,)
'~ CHAIRMAN APOSTOLAKIS: Oh, they are?
3 4 MR. GOSSELIN: Yes.
5 CHAIRMAN APOSTOLAKIS: They are?
6 MR. GOSSELIN: Yes, yes. At the Fitzpatrick 7 plant, it's a GE-BWR/4 plant. We're looking at 14 8 systems, again trying to include non-code piping. And I 9 told you what the challenge is with the BWR systems.
10 That's a little bit tougher nut to crack. We're 11 essentially 90 percent or almost completed with this 12 effort.
13 Again, I expect that the submittal will get
,r 3 i f
'd 14 into the NRC right after the first of the year. But 15 what's interesting to note here -- and this is why I 16 wanted to show this slide. Let's take a look at the RCS 17 on BWR.
18 Most of all, these BWRs, their RCS inspection 19 programs are driven by IGSCC. Okay? And the NRC has got 20 some very specific requirements as to what you're supposed 21 to do.
22 Now, in our process, what we have done is 23 this. I showed you those percentages, but those 24 percentages don't really apply to IGSCC. What our
() 25 requirements say is this: If you have identified 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
380 1 segment that's acceptable to IGSCC and it's part of your
,_s 2 IGSCC program, you continue to implement your IGSCC i \
'~~'
3 program. We want to integrate that right into the Section 4 XI aspect.
5 Similarly, for flow-accelerated corrosion, if 6 this area is susceptible to flow-accelerated corrosion, 7 you implement, you continue to implement your 8 flow-accelerated corrosion process. We didn't want to 9 come into conflict with existing programs that had been 10 reviewed and approved by the staff already. So that's how 11 those are treated. So for --
12 MEMBER SHACK: Section XI is a lousy way to 13 look for flow-assisted corrosion, anyway.
(/ 14 MR. GOSSELIN: Well,Section XI doesn't 15 aduress any of these degradation mechanisms at all except 16 for maybe the fatigue.
l 17 But what's interesting here is: So what can 3S you expect, at least on these pilots that we're doing here 19 on BWRs? Well, there's not going to be much reduction at 20 all in the area in volumetric inspections on the RCS for 21 BWR. And that was the case for Fitz.
l 22 However, the fact that we are doing inspection 23 for cause, there's a big reduction in the number of 24 surface exams that are doing it. Anci the Fitzpatrick
<x !
( , ) 25 people feel this is going to be very helpful because it's NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. j (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433 !
1 381 )
i 1 going to keep the man rem down on the exposure by doing j l
2 these unnecessary exams. So there was a good thing that
( )
3 came out of that for those guys. So that was the point I 4 wanted to make on --
5 MEMBER SHACK: Have you added inspections 6 somewhere? When you looked at this thing sort of throwing 7 away the Section XI classifications, did you end up adding 8 inspections?
9 MR. GOSSELIN: Yes, in two different ways.
10 Number one, there are areas of the plant that are 11 important, non-code piping, that are important that 12 currently weren't selected that we're going to be looking 13 at. Okay? And there was non-code piping that we're
,/~s i
picking up that wasn't expected that we're also 6
'V 14 15 inspecting.
16 MEMBER SHACK: Now, is that true of --
17 MR. GOSSELIN: So you may be reducing the 18 number of inspections, but you may be also inspecting 19 different areas that weren't looked at before.
20 MEMBER SEALE: But yo 're telling us basically 21 that the IGSCC result here is a true result of this 22 methodology and not a reconciliation of this methodology 23 to already develop practice to handle the kinds of 24 problems you run into in recirculation systems and so on?
,a 25 MR. GOSSELIN: Yes. The assessment that's
!.v)
NEAL R. GROSS COURT REPORTERS AND TRANSCf6BERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C 20005-3701 (202) 234-4433
382 1 done looks and says: Do I have IGSCC at this area? If I
,-s 2 have IGSCC at that area, I suspect that that potential is
( )
'^~
3 there. Is this included in my IGSCC program? If it is, 4 then I continue to implement those inspections.
5 MEMBER SEALE: Yes.
6 MR. GOFSELIN: If it isn't, I add it and 7 implement those inspections.
8 MEMBER SEALE: Very good. Very good.
9 MR. GOSSELIN: That was a point I wanted to 10 bring up on the EWR.
11 MEMBER SHACK: Those inspections would be 12 considerably in excess of what you're proposing to do?
13 MR. GOSSELIN: Actually, if it's a high-risk
\
\- 14 area, it's about the same, 25 percent. If it's a low-risk 15 area or a moderate-risk area, it would be greater. But 16 there are some plants that are doing IGSCC inspections 17 that are greater than 25 percent depending on the 18 material. So yes, it would be.
19 MEMBER SHACK: So, in answer to Dr. Seale's 20 question, the IGSCC program is added on top? It didn't 21 fall out of your program?
22 MR. GOSSELIN: No, no. ;
1 23 MEMBER SEALE: But it's confirmed, if you l l
24 will.
1
( >
Well, it's in 0313.
() 25 MEMBER SHACK:
NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS i 1323 RHODE ISLAND AVE., N W. l (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
383 1 MEMBER SEALE: Yes.
,3
, 2 (Slide)
( )
3 MR. GOSSELIN: Okay. So, anyway, again, I 4 have provided a list for your information of the systems 5 and what their classifications are that we're doing. And 6 that really summarizes, ends my presentation with me 7 asking that if the Committee would like, we would be more 8 than happy to come back and would look forward to getting 9 into these various areas in more detail with you and 10 bringing some results.
11 Thank you very much.
12 CHAIRMAN APOSTOLAKIS: Are there any other 13 questions for Steve or the --
? \
'- 14 MEMBER BARTON: No questions. But I think for 15 ISI engineers and system engineers, this is a good 1
16 approach to figuring out the inspection program. It ;
i I
17 really is.
1 18 CHAIRMAN APOSTOLAKIS: Gary?
19 MR. HOLAHAN: Gary Holahan of the staff. I l
20 have one question.
23 I recognize this is a qualitative approach, 22 but would you give me a qualitative judgment about whether l
23 implementing this program is a net safety increase or 24 decrease for the plant?
() 25 MR. GOSSELIN: I think it's a net increase NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
384 1 qualitatively.
2 MR. HOLAHAN: An increase in safety?
(, ,)
' 3 MR. GOSSELIN: Yes.
4 MR. HOLAHAN: Thank you.
5 CHAIRMAN APOSTOLAKIS: An increase.
6 MEMBER SEALE: In safety.
7 CHAIRMAN APOSTOLAKIS: In safety, yes. And 8 that's high, medium, low?
9 (Laughter.)
10 CHAIRMAN APOSTOLAKIS: Any other questions or 11 comments?
12 (No response.)
13 CHAIRMAN APOSTOLAKIS: We appreciate it very
,n t
(.. 14 much.
15 MR. GOSSELIN: Thank you. Thank you very 16 much.
17 CHAIRMAN APOSTOLAKIS: The schedule calls for 18 a break now, and we will do that. We will be back at 5-10 19 minutes after 10:00. j 20 (Whereupon, the foregoing matter went off the i
21 record at 9:55 a.m. and went back on the 22 record at 10:14 a.m.)
1 23 CHAIRMAN APOSTOLAKIS: Now we are talkirg i 24 about the SRP/RG for IST.
r3
!v! 25 MR. WESSMAN: Yes, sir.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 134-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
385 1 CHAIRMAN APOSTOLAKIS: Okay.
2 MR. WESSMAN: Good morning. My name is Dick 7._
(' ' ')
3 Wessman. I am the Chief of the Mechanical Engineering 4 Branch. The Mechanical Engineering Branch in NRR has lead 5 responsibility for the staff's activities regarding review 6 and approval of IST activities.
7 If you will turn to the second page of the 8 viewgraphs, I will introduce the people at the table 9 because a couple of them you may not know. Obviously you 10 have seen a lot of Mark Cunningham on the far end already.
11 We have had a development team that's been 12 involved in this activity. The one member that is not 13 with us today is David Fischer, who was one of the co-team
,.m
! )
(_/ 14 leaders. On my left is Brad Hardin from Research. He has 15 been part of the development team. On the far right is 16 Mike Cheok from NRR, the Probabilistic Branch. You have 17 seen him before. Also with me on wy immediate right is 18 Joe Colaccino, who is part of the Mechanical Engineering 19 Branch. In the far corner is Bob Jones, and you know him.
20 I would point out, of the individuals 21 involved, that several of us have had a fair amount of 22 involvement with the ASME. Of course there's a lot of 23 work by the ASME developing code cases and guidance from 24 the ASME research relating to our activities.
( ) 25 Dave Fischer was the agency's representative NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS f 1323 RHODE ISLAND AVE., N.W. l (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433 4
]
386 1 to the ASME research organization developing the ASME
,s 2 guidelines document that was shown to you a month or so
\
~ '
3 ago when ASME was here. It's a green-covered booklet.
4 Brad Hardin has been the staff's 5 representative to the ASME working group that is 6 developing the general code case.
7 I am a member of the ASME's O&M main 8 conmittee, and represent staff activities to them.
9 Both Joe and Dave Fischer have been principle 10 reviewers on the pilot plant submittals from Commanche 11 Peak and Palo Verde. They have assisted in the plant 12 reviews to a slightly lesser degree by both Brad and Mike 13 Cheok.
,-~ g l i ,
i/ 14 If we'll turn to the next page, I'll summarize l 15 what we want to try and cover with you all today.
16 Obviously indicate what the objectives of the team are. I 17 also want to touch briefly on what the classical IST 18 program is, and then compare and contrast a little bit to 19 the risk informed IST program.
20 Then I want to take you all through the 21 essential elements of the Reg Guide and the Standard 22 Review Plan and point out how from an applications 23 specific standpoint in IST, the staff's principles and 24 guidance are dealt with.
) 25 Finally, I will turn it over to Joe Colaccino N Eld. R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
387 l
1 to summarize on the last four or five viewgraphs, our 2 activities with the pilot plants.
in\
\ /
. Recognize the pilot plants started submittals 4 with us about a year ago, so some of the criteria and 5 activities that they were involved in was evolving almost 6 before some of our work was evolving. But I think we all 7 learned from each other. It has been a valuable 8 interchange working this way. Yet there clearly is more 9 work by these pilots because as our criteria and 10 guidelines are getting clearer, we have to work back with 11 them to assure that they are comporting with what the 12 staff views as the right approach on this.
13 Obviously our team objectives, three major
,,~x i i D' 14 objectives. Develop an approach and guidance for the 15 industry, namely the Regulatory Guide, develop guidance 4
16 for the staff for the Standard Review Plan, and do the i
17 work on the pilot program submittals.
18 Where are we so far? You received the draft 19 Regulatory Guide and the SRP section about a week or so Like the general documents, we've been through a lot 20 ago.
21 of iterations as we have gotten these documents developed.
22 We have done initial review of the pilot i 23 applications and had an initial round of requests for 24 additional information that went to them last spring.
l3 i
() 25 Then we had meetings with both the Palo Verde and 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
388 1 Commanche Peak licensees.
s; 2 Then at that point, we set aside a little of t
( /
3 our work dealing directly with the pilots because of the 4 attention that we have given to the Standard Review Plan 5 and Regulatory Guide development.
6 Our calendar still has the expectation to 7 complete the pilot plant reviews by March of 1997. At 8 that point, we have to provide our recommendations 9 regarding the pilot plan approach to the Commission, and 10 have a Commission understanding of the staff's approval 11 process before we would issue an SER and allow those pilot 12 plants to actually proceed with implementation.
13 Obviously we are continuing on an ongoing
- 14 basis interaction with the ASME on the various code cases.
15 A week or two ago, we received the general code case in 16 their view of a relatively final form for what they call l
1 17 first-round balloting. The staff is reviewing and 18 developing our views on that, which I will respond to the 19 ASME on in early December. j 20 Let me speak briefly as to the classical IST 21 approach. The requirements for in-service testing at the 22 reactors is governed by the regulations in 50.55a. There, l 23 the regulations reference the ASME code. In that sense, l 24 the code becomes an extension of the regulations. 50.55a n
, 25 endorses section 11 of the ASME code and portions of the NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAhD AVE., N W. ]
(202) 234-4433 WASHINGTON. D. ~ 20005-3701 (202) 234-4433
389 1 O&M operations and maintenance code.
, ~3 2 The program scope that the IST program covers
. i N'"' / 3 is the ASME safety related code class pumps and valves.
4 The type of testing that's done though is a rather 5 prescriptive testing and a rather prescriptive frequency 6 of testing.
7 Most tests are done quarterly for active 8 components such as the pumps and valves. Most tests are 9 quite specifically prescripted by portions of the code.
10 There are very definite acceptance criteria, action and 11 alert ranges for a pump when it's tested, and frequencies 12 that the pump would be tested and the requirements for 13 documentation and this sort of thing.
(~
+
ks 1 If a utility seeks to change a test interval 14 15 or change test acceptance criteria, or change the type of 16 test from what is in the ASME code, they have to come in 17 and request approval by the staff. The staff, if it 18 agrees with the licensee's proposal, would approve that as 19 a relief request. Provisions for granting a relief 20 request are provided there in 50.55a.
21 Many of these relief requests are driven by a 22 situation where the plant condition during a normal 23 operating environment of the plant would not allow the 24 conduct of a particular test, and it would require plant 1
/s I
) 25 shutdown to do the test, or that it presents some sort of NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. i (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344 433 l
390 l 1 unusual hardship in order to actually do the tests. So 2 the licensee develops their argument, generally in an
[~,\/
x 3 engineering deterministic sense as to what they would i
4 propose as an alternative, and the staff renders a j 5 decision on it.
6 There are also ASME code cases out there that 7 may provide certain alternative to the actual code 8 requirements.
9 Turning to the IST approach, as discussed on 10 the next two viewgraphs. Let me point out what our vision 11 is of the risk informed IST approach. Approval of such a 12 risk informed program would be done as an alternative to 13 the code requirements, in accordance to 50.55a in the sub-
!")'
(_/ 14 chapter. This was one of the procedural questions that we 15 visited with you on when we had met with you a month or 16 two ago.
17 The approval process essentially is the same 18 as the existing process for a proposed relief request that 19 I mentioned under the classical IST program. The staff 20 has to make a finding that the proposed alternative 21 provides an acceptable level of quality and safety.
22 The IST program, the risk-informed IST prc. gram 23 is established using the same concepts that we have 24 discussed with you before in the general reg guide, the n
( ,) 25 traditional engineering evaluation, application of PRA NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
391 1 insights, and an integrated decision making process, most
-- 2 likely the expert panel concept.
(s'~') 3 Our vision is that components would be 4 categorized in two groups, the high safety significant and 5 the low safety significant components. There is also the 6 expectation that the licensee would look at the components 7 in the plant and identify components that are non code 8 components, in other words, they have not been routinely 9 tested under the existing IST program.
10 If they appear to be risk significant, capture 11 them in the risk informed program. They would be classed 12 in the high safety significant group. As Joe will tell 13 you, the pilots actually identified some instances in this
^
/ 's V 14 area.
15 The view of course by the licensees, and I 16 think held by the staff, is that it allows the licensee to ;
I 17 more closely focus their resources commensurate with the 18 safety significance of components. l 19 In a risk-informed IST program, the 20 expectation is that the testing strategy for the low 21 safety significant components would usually be less 22 prescriptive. Commonly, it would be an extension of the 23 test interval. In other words, test the component less 24 frequently. In some cases, it may be a less prescriptive
! ,) 25 test than the current ASME code test.
NEAL. R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
392 1 Some of the code case work being done by ASME x 2 at this time is in fact developing different levels of
( )
3 tests, a less prescriptive test and a more prescriptive 4 test, for the low safety and high safety significant 5 components.
6 Now the testing strategy for HSSC components 7 may or may not be more rigorous than the current code 8 requirements. Most typically, HSSC components will be 9 continued to be tested at the code required frequency 10 using the ASME code required tests.
11 The licensee would be expected, and again, the 12 pilots have looked at this a little bit, at the HSSC 13 components, to see if there may be some sort of test that
( )
kl 14 may not be currently required by the code, but that it 15 makes good sense to do to more closely understand and 16 predict the performance of the component.
17 For example, in certain situations, a pump 18 lube oil analysis is not currently required by the code, 19 but experience has taught them, and the technology has 20 taught them that taking a lube oil analysis may give them 21 a clue as to bearing where or degradation of this nature.
22 Our vision also would be that once the '
23 licensee has established their test intervals for the LSSC 24 components, that if they chose to change their interval, fM
( j) 25 they would not have to come back to the staff 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 l
393 1 approval, provided the basic process that they are using
,S 2 is not being changed.
('~' )
3' Of course in the current regime, they can't do 4 that. If a licensee under the current IST program wants s
5 to change the interval, they have to come to the staff 6 with a relief request and provide a basis for that.
7 Similarly, test method changes would not 8 require an NRC approval, provided a change in the test 9 method is bounded within the compendium of approved test 10 types, which would be most typically endorsed ASME codes 11 or code cases.
12 Obviously though, if they elected to make a 13 change to their risk-informed IST program that involved a
(' '/ 14 process change, we would view that as a significant enough 15 change that we would want them to come back to the staff 16 for approval.
17 CHAIRMAN APOSTOLAKIS: So in this conter.t l
18 then, is it meaningful to talk about performance?
l 19 MR. WESSMAN: Yes. I think it is. Let me 20 come back to that as we come to the element four about 1
21 performance monitoring and implementation. I think I can 22 give you a couple of observations on that, George.
23 CHAIRMAN APOSTOLAKIS: Okay. The reason why I 24 asked the question is because you will to approve process o
k_) 25 changes, right? You just said that.
NEAL R.-C-ROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. O C. 20005 3701 (202) 234-4433
394 1 MR. WESSMAN: Yes, sir.
2 CHAIRMAN APOSTOLAKIS: That means you can not
- i. !
3 rely only on performance.
4 MR. WESSMAN: No.
5 CHAIRMAN APOSTOLAKIS: Why do you want to get 6 involved in process changes?
7 MR. WESSMAN: No. Maybe I am using the term 8 process differently. Our perception is a process change 9 would be a change to the overall analytical process, 10 ranking process that they have presented and advertised to 11 us. If they elect to use some sort of different 12 methodology for screening components into high and low, 13 for example, that we would view as a process change, and
/'~N O 14 we would want to have a chance to understand that before 15 the licensee implemented it.
16 Okay. The essential aspects of the regulatory 17 guide of course, and I'm not going to read this, is to 18 provide the acceptable methods for using PRA information, l l
19 along with the traditional engineering methods to develop 20 the risk-informed program. The ecsential aspects of the 21 SRP, to provide the procedures and the guidelines for the .
I 22 staff. ;
23 These two documents, as in the case of tech 24 specs or graded QA flow from the general Reg Guide and the 7.
25 SRP, and are very closely linked to them. We want them to
() - ;
1 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 e ;
395 1 comport with these documents as well.
,x 2 Let me turn and take us through the individual
- \
"~
3 elements that we've talked about when we talk about the 4 general Reg Guide and the SRP, and describe briefly how 5 they relate to the consideration of the risk-informed IST 6 program.
7 Of course the first element is to define the 8 proposed changes to the program. In an IST situation, we 9 of course expect the licensee's to define the components 10 that are governed by an IST program that are in the 11 current licensing basis, and also identify those 12 components that may not be in the current licensing basis, 13 but are found to be important to safety. This is likely
(_) 14 to be those components where some risk analysis identifies 15 a component that has not been previously subjected to IST 16 type of testing requirements.
17 We expect them to identify the changes to the 18 existing test schedules and the methods, and obviously )
l 19 then prepare and provide the necessary information and 1
20 analysis that supports the changes that they are going to 21 make.
22 Finally, of course, have an understanding of l l
23 their formal interaction with the NRC and when they may in 24 the future have to come to the staff regarding a relief
,-)
25 request type of activity.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
396 1 CHAIRMAN APOSTOLAKIS: I guess I'm a little 2 bit confused by the second bullet. How can a component be 7 s,
\ )
'^'
3 important to safety and not in the CLB?
4 MR. WESSMAN: I guess our vision there is that 5 it may not be subject to prescribed test requirements, 6 such as an ASME code test, and in that sense, not in the 7 CLB regarding required testing. The risk analysis work 8 identifies that this particular pump or this particular 9 valve, which hadn't previously been bounded by an ASME 10 test, comes up as an important to safety valve.
11 Maybe Joe can give me an example from one of 12 the pilots or something. But our view is, the risk 13 analysis work identifies a component that hasn't been
\~) 14 previously covered by ASME testing. In a risk informed 15 program, the licensee would be expected to offer it up as 16 a candidate for ASME type of testing.
17 MR. HOLAHAN: Yes. I think a good example in 18 this category, I think we may have talked about it before, 19 is a start-up feedwater pump, for example. It's not in 20 the safety analysis. You probably won't even find it 21 described in the FSAR, but it might be an important piece 22 of equipment in the plant. It's in the PRA. If the PRA 23 says it's important, then it looks like it fits in bullet 24 two.
,\
(m- ) 25 MR. COLACCINO: And in fact, for Palo Verde, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
397 1 the start-up feedwater pump discharge check valve is the ley 2 one, and only one component that they do bring in as a NA 3 high safety significant component. It is modeled in their 4 IPE, but it's not in their IST program because it's a non-5 ASME code class component. So it would not be required in 6 the current ASME legulations.
7 MR. WESSMAN: Okay. If I can, I'll --
8 CHAIRMAN APOSTOLAKIS: But other aspects of 9 the component may be in the CLB?
10 MR. HOLAHAN: Probably not.
11 CHAIRMAN APOSTOLAKIS: What is a CLB?
12 MR. HOLAHAN: Current licensing basis.
13 CHAIRMAN APOSTOLAKIS: I know.
m 14 MR. HOLAHAN: It's those things in the plant 15 that are there because of the regulations, orders, license 16 conditions, tech specs. It's those formal -- it's the 17 parts of the plant that have formal regulatory 18 requirements on them, and also licensee commitments and j 19 responses to bulletins. Whatever ends up in the FSAR or 20 on the plant docket as of its having some safety role.
21 CHAIRMAN APOSTOLAKIS: So the start-up l 22 feedwater pump was not recognized as having some safety 23 role before?
24 MR. HOLAHAN: Right. And I can imagine, for
,-~
k _\) 25 example, in some plants the main feedwater pumps, or even NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
398 1 the condensate booster pumps could play a role in some 2 emergency procedure that's not in the normal traditional n) i v
That might make those more 3 engineering design basis.
4 important than they were before, or more important than 5 they were recognized before anyway.
6 MR. WESSMAN: Okay. Turning to the next 7 slide. I want to talk in a little more detail about 8 element two. We actually have, I think there's three 9 pages of your handout that deal with the various aspects 10 of the guidance for the engineering evaluation. There is 11 more meat in the Regulatory Guide in this particular area.
12 This element embraces the traditional 13 engineering evaluation guidelines, the guidelines for use
(_y 14 of the PRA, and the guidelines regarding integrated 15 decision making. It's here where most of our key 16 principles are addressed, and also where our expectations 17 regarding some of the staff expectations are dealt with 18 also.
19 CHAIRMAN APOSTOLAKIS: Excuse me. In the 20 latest version of that diagram, we removed the box 21 integrated decision making, but it's all over the place 22 now.
23 MR. HOLAHAN: Yes. It's so integrated that 24 it's in the engineering box.
,cx
(,) 25 CHAIRMAN APOSTOLAKIS: You can not really 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
399 1 decide anything until you see the *her box, which is
,s 2 proposed monitoring program.
( )
~' MR. WESSMAN: Yes.
3 4 CHAIRMAN APOSTOLAKIS: So this is a part of 5 the integrated decision making or it's preparing the case, 6 so to speak. I'm splitting hairs now, but it's the time 7 to do that.
8 MR. WESSMAN: I think the integrated decision 9 making process does tend to be iterative and go on.
10 But touching briefly on the key principles 11 regarding the regulations are met. Of course we expect 12 the licensee to conform with the ASME code and continue to 13 operate and test the plant in accordance with the 14 requirements.
15 When we think about defense in depth, we 16 assume that the changes that they might propose on an IST 17 program are not going to impact the redundancy or the 18 diversity of the existing systems. All the same 19 components that are in the existing IST program still have 20 to be tested.
21 When we think of the safety margins being 22 maintained, we want to be sure that there is a reasonable 23 effort by the licensee to assure that components don't 24 degrade during a more extended interval, for example. I
() 25 might pursue that slightly with you with the example 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
400 1 l
1 motor operated valves. l
.s 2 Recently, the ASME has published a code case i \
3 dealing with testing motor operated valves, a code case 4 OMN-1. This code case has been endorsed, with a couple of 5 qualifications by the staff in a recent generic letter 6 dealing with periodic verification of motor operated 7 valves.
8 The code case provides guidance to the 9 licensees regarding the expectations regarding monitoring 10 of the valve and forecasting rates of degradation. The 11 expectation is that the next scheduled IST test would be 12 conducted before the rate of degradation reached a point 13 that you found the valve wasn't going to work.
(G
\_/
i 14 The code case allows the selection of ;
15 differing test intervals, and does not force the licensees 16 to be on a quarterly stroke test of the valve, but has 17 this expectation of as an interval is extended, that it's 18 done with care and with the knowledge of the performance 19 of the component, and that you do test it before you 20 discover that it fails.
21 Obviously there may be a situation where a 22 failure occurs or they go to test it, and they find that 23 it is not working as expected, and the expectation is that 24 the interval would be adjusted and a corrective action (p
v
! 25 would be taken.
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
w_
401 1 Changes to the CLB do not impact the 2 individual -- the safety goals and the subsidiary
, ~s
( ")
3 objectives. Of course the expectation is changes to the 4 CLB in a risk-informed IGT program would only result in a 5 small increase in plant risk.
6 CHAIRMAN APOSTOLAKIS: We've changed the 7 wording now, right? This was not before yesterday.
8 MR. WESSMAN: That's true. I didn't put fuzzy 9 words on the page. I didn't know how to do that.
10 CHAIRMAN APOSTOLAKIS: You also didn't put any 11 figures.
12 MR. WESSMAN: There's no figures.
13 CHAIRMAN APOSTOLAKIS: Very clever.
~s V 14 MR. HOLAHAN: But we can fix that too.
15 MR. WESSMAN: Similar to some of our 16 discussions yesterday, the PRA is used to help identify 17 pumps and valves that are candidates for relaxation of IST 18 requirements. It is used to identify candidates for 19 enhanced testing.
20 MEMBER POWERS: When somebody comes to you 21 with a proposal and says I've used my PRA and I have 22 identified these pumps and valves, and you look at the PRA ..
23 and you treat -- and you find that it treats only the
.24 power operations, do you send him home?
,o
( I 25 MR. WESSMM : No. We would not send him home, w/
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
402 1 but we would want to know how he dealt with components in
,~.
2 the shutdown situation. It may be that if the PRA is
'wj' limited, that we would expect more rigorous engineering 3
4 traditional analysis work by the so-called expert panel to 5 help reach a decision as to how components were 6 classified.
7 CHAIRMAN APOSTOLAKIS: Yes. Speaking of that, 8 high and low could change depending on what your measure 9 of risk or safety is. Is that correct?
10 MR. WESSMAN: Yes.
11 CHAIRMAN APOSTOLAKIS: So what do you propose 12 to use, core damage frequency and then the various modes 13 of operation or risk, or a combination?
'- 14 MR. WESSMAN: Well, the acceptance criteria 15 for an overall change in a risk informed IST program is 16 the acceptance criteria that we are still debating and l l
1 17 evolving toward in the general Reg Guide. i 18 CHAIRMAN APOSTOLAKIS: So it might be CDF and 19 LERF?
MR. WESSMAN: Yes, sir. l 20 l
l 21 CHAIRMAN APOSTOLAKIS: Or CDF and something .
l 22 qualitative.
.~ - . .~ .
23 MEMBER KRESS: See how useful it is to have 24 diagrams?
(O) 25 (Laughter.)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON D.C. 20005-3701 (202) 234 4433
403 1 CHAIRMAN APOSTOLAKIS: For the record, the 73 2 diagram did not contribute anything to this.
( )
s_-
3 MEMBER POWERS: It surely would have, had 4 there been one.
5 MR. WESSMAN: Again, as we have discussed 6 yesterday, of course the overall changes in a risk-7 informed test program, there has to be an assessment in 8 the change to plant risk. Again --
9 MEMBER POWERS: But if PRA only treats power 10 operations, how do you assess the change in plant risk?
11 MR. WESSMAN: I may need help here.
12 MR. CUNNINGHAM: If you go back to the diagram
,. 13 that we used yesterday, there was a distinction made, or
(\') 14 you could read the text that we discussed yesterday.
l 1
15 There was a distinction made on the acceptance criteria 16 for plants that have a full scope PRA and which ones which 17 don't have. So the acceptance criteria shifts downward by 18 that half an order of magnitude that showed up on some 19 figure some place.
20 MR. WESSMAN: The vertical line between boxes l
21 two and three, as I recall. l 22 MR. CUNNINGHAM: So if you did not have a j 23 shutdown PRA, but you met the other ones, then you would l l
24 measure against that line rather than the 10 to the minus j
/
() 25 four. 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
404 1 MR. WESSMAN: It sets the tolerance a little 2 higher in that situation.
(m ).
3 MEMBER POWERS: Well, if I remember that 4 wonderful figure that made this very clear to me, and I 5 remember the text which maybe didn't make things too clear 6 to me, that that told me where my starting point and the 7 analysis I had to do were. But it didn't change what 8 magnitude of a change in risk I could have. It just meant 9 I couldn't measure it.
10 MR. CHESK: For the pilot applications, what 11 the licensees have done is resort to the expert panel with 12 a clear set of rules to determine what would not be needed 13 during instances like shutdown.
?
V'/ 14 In one of the pilots, they basically said that 15 if a pump or valve contributed to a success path that's 16 needed during this shutdown load, then I would not do 17 anything with this pump or valve.
18 In another pilot, what they did was they had a 19 different set of rules that basically said if there were 20 other trains available, the redundancy available basically 21 and how often this is needed, we will rank these things 22 either high or low based on that.
23 As a matter of fact, for each pilot, in one 24 pilot for shutdown, 15 valves were added to the high list l 1
-~
! j 25 based just on shutdown. In another one, 22 valves were v
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 405 1 added to the high list just based on shutdown 2 considerations.
(, ,)
\'~) 3 MEMBER POWERS: I think that's terrific 4 information. I appreciate getting that. Now let me ask a 5 question.
6 If an expert panel comes in and they say I can 7 go ahead and change this valve because there are other 8 trains available, doesn't he run afoul of the defense in 9 depth principle?
10 MR. WESSMAN: I think he does. We would 11 expect that as the ranking of a valve or as the changing 12 of a test interval is done, that there is consideration 13 for like components or parallel trains.
i A t
\_/ 14 For example, we would expect that there would 15 be a staggered type of testing. Let me take the example 16 of four essentially identical valves on four paths to the 17 RCS. We would certainly not want to see a licensee say 18 well, okay, these are valves that I might be able to group 19 into a lower risk category and extend the interval for all 20 of them from quarterly to two years, and I won't test any 21 of them now for two years. We wouldn't accept that i
l 22 approach. _
23 We would expect it to be on a staggered basis.
24 So the first one will be the first quarter. Maybe the g_
! ) 25 second valve will be six months out, and this type of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. 1 (202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
406 1 staggering. So that if there is a problem, it's going to n 2 be -- you know, at least there is some chance that it will
( )
3 be intercepted, and we won't wake up two years hence and 4 find out that there was some common cause problem with all 5 four of these valves.
6 MEMBER POWERS: That's not different than 7 what's been done on the other extended testing.
8 MR. WESSMAN: That's true. Part of the 9 condition monitoring concept and check valve testing is 10 built on that concept.
11 MEMBER KRESS: I have a question about the 12 second bullet also.
13 MR. WESSMAN: Yes, sir.
/~)
C'# 14 MEMBER KRESS: Is there a correlation between 15 the test frequency and method and reliability of one of 16 these components? Does such a correlation exist?
17 MEMBER POWERS: It can be.
18 MR. CUNNINGHAM: Well, in the PRA model, you l l
19 would have a model set up that would allow you to show the I I
20 impact of the test duration on the unavailability of the 21 component.
22 MEMBER KRESS: Of course there would be an 1
23 unavailability.
24 MR. CUNNINGHAM: Unavailability. I think Q 25 that's what we're talking about here.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433 l
407 1 MEMBER KRESS: I see. You are not talking 2 about the actual change in the reliability of that 73 N)'~
3 component due to the fact that it may have a different 4 tect interval or a different metnod.
5 MR. CUNNINGHAM: No. It's really how that 6 contributes to the unavailability.
7 MEMBER KRESS: I see. So that's only 8 unavailability.
9 MR. CUNNINGHAM: It's one piece of it, yes.
10 MR. WESSMAN: Okay. If I can nove us along, I 11 think we have sort of touched on the other bullets of this 12 particular viewgraph. We talked about sort of the 13 acceptance criteria, is in the general Reg Guide, so I'll
- e
\/ 14 turn the page on you if I may with just a couple of other 15 bullets relating to element two.
16 As is done in the other application specific 17 initiative, a consideration has to be made of a risk-18 informed IST program in relation to other initiatives. I 19 guess our favorite example is if the graded QA folks say 20 this is a low risk component and we can move it down and 21 not give such good quality to it because we know they are 22 going to keep testing it all the time, and then the test 23 folks come along and say well this is a low risk 24 component, we're not going to test it so often, obviously
,\
( ,)
25 there's been a defeat in the integrated purpose of 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
408 1 whole thing.
2 Then fi.. _1y, of course, under element two the o
( ) licensees are expected to review the existing approved IST 3
4 program, existing relief requests, the tech specs and this 5 sort of thing. If there is a specific change to an 6 existing relief request or a test amendment, we would 7 expect this amendment to come as part of a risk-informed 8 proposal.
9 Turning to element three, and we have a couple 10 of viewgraphs dealing with this that touch on the 11 implementation and'Lonitoring aspects of the thing, as I 12 think I had mentioned earlier, the ESSC component is 13 typically tested using the approved ASME code test a 1 t
frequency and method.
14 LSSC would be typically tested at a 15 more extended interval. We have established a maximum 16 limit.
17 At the moment, what we have specified in our 18 Reg Guide is five years or three outages, whichever is 19 longer. This may be a little bit arbitrary, but it is at 20 least founded on some experience as far as some of the 21 work that's been done on MOV testing. We are getting more 22 work that's been done recently by Oak Ridge regarding 23 degradation method and performance. It appears that this 24 is a reasonable outer limit at this point, but of course (O
(,) 25 even as one steps towards the outer limit, we would expect 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
409 I 1 it to be done in a step-wise fashion so that if an
- 2 unexpected failure mechanism is identified, it can be
/s )
N/ l 3
Intercepted.
4 The expectation on the implemented, that the 5 licensees would use the ASME code cases or existing code 6 tests. Of course we have an expectation that as the ASME 7 develops the risk-informed code cases, that the staff 8 would be able to approve them with limitations if 9 necessary in a process like we have done in the past.
10 We would expect the licensee to do a phased 11 implementation of the program. We kind of touched on it 12 on a component specific basis. We would not want a 13 licensee to put all components out at a four-year interval g
'k -] 14 or something like that when they are like components for 15 example.
16 It may be also necessary that a licensee just 17 from the magnitude of the work involved, has to phase 18 their implementation of the program such that during 19 successi/e plant outages, they gradually work their way 20 into a risk-informed program for pumps and then for MOVS 21 and this sort of thing, so that the program is clearly 22 effectively managed.
23 Of course our expectations from a performance 24 standpoint, and I think we talked on this a little
,n\
ij, 25 yesterday, that there is a monitoring to 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
410 f 1 degradation, initiative corrective action, and adjust the ;
2 test frequency based on results. Much of this concept is l s !
/ s
/
3 prescribed by the ASME code tests that are out there l
i l
4 today.
l 5 Finally, of course there is the expectation to 6 periodically reassess the program overall and make the 7 programmatic adjustments as necessary. I think that this 8 aspect of this element three gets us the essence of the 9 key principle dealing with having a performance-based 10 implementation and monitoring strategies to help deal with 11 potential uncertainties.
12 If I may, I will turn the page to element four 13 and discuss briefly with you the documentation p
N- / 14 expectations. Then I am going to turn it over to Joe and 15 let him talk about the pilots a little bit.
16 Again, there is parallel here between the Reg 17 Guide and the Standard Review Plan and the general 18 documents. Guidelines regarding documentation that we 19 have to have to make a sound evaluation of submittal is 20 provided as part of this portion of the document. The 21 expectatio.1 is that the licensee will show that the CLB 22 changes are consistent with our key principles and our 23 staff expectations.
24 The sort of material that we would expect, and g
( ,) 25 of course that we have seen from the pilots, are things NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE IS!.AND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
411 1 like the summary of the risk assessment work and summaries 2 of the risk -- of the integrated decision making process, l 7s '
3 detailed listings of components and systems involved, and 4 an overall assessment of the CLB change.
5 of course when the pilots made their 6 submittals a year or so ago, they didn't have the benefit 7 of our key principles. But on balance, most of the issues 8 that we saw were important, they saw were important as o
9 well. Where they have not dealt with them, as Joe will 10 point out shortly, we have got additional requests for 11 information from them. We'll have to work our way forward 12 on it.
13 We give the licensees guidance on the records f%
(k /l 14 and supporting data that have to be maintained at the 15 facility. Obviously, a lot of these records are lifetime 16 records and there are Reg Guide requirements there.
17 Guidance about the maintenance of the ASME 18 code test records. Of course this parallels a lot of the 19 requirements that are already in place.
20 These records then are in a situation at the 21 facility where many of them are available for inspection 22 by our inspection staff. The inspectors periodically do 23 look into the conduct of the classical IST programs. We 24 would expect a similar sort of inspection activity in the (n) 25 future of a risk-informed IST program.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 !
412 l
1 Okay. I want to turn you forward to the
,s 2 pilots a little bit if I can, and let Joe take you through 1
)
i \
wi 3 some, I guess some concrete examples and numbers, and some 4 of the things that we encountered if I can.
5 Wait a minute. There's a hand waving. Yes.
6 George? Bob?
7 MR. JONES: I was just going to respond back 8 to Dana's earlier comment on the shutdown stuff. There is 9 a discussion on how we treat shutdown on page 419 of the 10 Reg Guide. Effectively right now, it's written up with a 11 recognition there isn't a lot of PRA related information 12 in the shutdown area. So it basically says those key 13 pumps and valves that you use to maintain safety functions
\/ 14 or to provide defense in depth, or use as part of your 15 operating procedures during shutdown modes, have to be 16 placed in the high category.
17 Given that it's in the high category, and then 18 you're not changing the testing, it will not be a decrease 19 or a risk increase, and so you don't have to evaluate it.
20 But we have addressed it in the Guide. It's the last 21 paragraph on that page.
22 Sorry, Dick.
MR. WESSMAN: ' No , thats fine.' That'give's
~
23 24 Dana a section. I saw him turning pages. I thought if
,m
(,) 25 you want to take a look there for a minute, we'll pause NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
413 1 for a second.
2 MEMBER POWERS: No. That's fine. I'll just rs t-3 comment that taking a very conservative position and 4 restricting things from utilities is an interesting 5 answer. I'm not sure it's a helpful one.
6 CHAIRMAN APOSTOLAKIS: In the context of risk-7 informed regulation --
8 MR. JONES: It's an acceptable way. I mean if 9 they can quantify it and have the full PRA, they could 10 rank it. There's nothing that precludes that in the 11 process, in the categorization. So if they really have 12 looked at those modes of operation, they could categorize
,_ 13 it within the right categories, and then we would just
('-'/ 14 treat it no different than anything else. But given that 15 they can do that, then they ought to also be able to 16 quantify the impact. So you would just pull it onto the 17 whole envelope.
18 MEMBER POWERS: What you have said is if you I
19 can do it, go ahead and do it. If you can't, don't do it.
i 20 MR. JONES: That's correct. I 21 MR. WESSMAN: In that sense, the default mode 22 is to leave components in HSSC category and continue the
^ ~~ ' ~ ~ ~
23 existing test regime. )
24 CHAIRMAN APOSTOLAKIS: Is there any reason why
,s
( ,I 25 you call it RI-IST and not RIPBR-IST?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 2344433 WASHINGTON, D C. 20005 3701 (202) 234-4433
414 1 MR. WESSMAN: No. I don't think I have got 2 any subtle implications there. That's just the way we
,~x i s C/ 3 have I guess chosen to do it. But you are right, there is 4 obviously performance based implicationo in what we're 5 doing. We keep learning and changing the terms as we move 6 forward.
7 MEMBER POWERS: It's verbal drift.
8 CHAIRMAN APOSTOLAKIS: It's a verbal drift, 9 verbal migration.
10 MR. WESSMAN: Okay. Let's talk about 11 Commanche Peak and Palo Verde a little bit then, if we 12 can.
13 MR. COLACCINO: Okay. As Dick said, the
\I 14 initial submittals were made in November, 1995, almost a 15 year ago to the day today. Since then, we have had -- the 16 staff has submitted back to the licensees a lengthy set of 17 RAIs. We did meet with them in April to discus the RAIs.
18 We have gotten responses back from them in June.
19 CHAIRMAN APOSTOLAKIS: What's an RAI? I 20 MR. COLACCINO: Request for additional i
21 information. I might add that the information they j
)
22 submitted in their submittals and in the RAIs was used, a i l
^
goo
d deal of it was used in the~ de've16pment of the Reg
^ '
53' 24 Guide and SRP.
/N l
( ,) 25 I thought it would be helpful to put some !
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 l
415 1 numbers up, to give you a scale of the type of component, i 1
7._
2 numbers of components we're looking at for both Palo Verde k)'~
3 and Commanche Peak. They are two pilot plants.
4 I just wanted to point out, the way I have 5 heard the conversation going on, I thought it might be j 6 worthwhile to find out that for Palo Verde, there are 501 7 valves that are in their IST program. That's the valves 8 that are required by the ASME code. There's one valve 9 that they added through their -- that was modeled in their 10 IPE, and they felt that was important. That was that 11 start-up feedwater, discharge check valve.
12 Of those 502 valves, only 150 some odd valves 13 are actually modeled in their IPE. So about 350 are not
,~.
_ 14 modelled in the IPE but are in their current IST program.
15 So those must be resolved through the integrated decision 16 making process which both these plants decided to use the 17 expert panel.
18 MEMBER FONTANA: Where does the delta CDF come 19 from? Is that like as you said before less time out of 20 service? Delta CDF, that's an improvement. First of all, 21 which direction does it go?
22 MR. WESSMAN: It is an increase in plant risk.
It's'a calculated increase" based upon the changes tha't'
~ ~
2'3 24 they made to their system. I think it becomes a
,m
( ) 25 compromise between components that are out of service NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND A/E., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
416 1 less, but that are now on a longer interval. There is 2 some assumption that they may be not necessarily so 3 reliable.
4 CHAIRMAN APOSTOLAKIS: What is the base case 5 CDF?
6 MR. WESSMAN: I don't know that I have got 7 those numbers.
8 MR. CHESK: The base case for Commanche Peak 9 is 5.7 minus five, and for Palo Verde is 4.7 minus five.
10 These are just internal events without the fires, and does 11 not have any external events or shutdowns.
12 CHAIRMAN APOSTOLAKIS: So we are in Region --
13 MR. CHESK: You are in Region II, basically.
g
) 14 No , wait a minute.
15 MR. WESSMAN: We need the picture again. Then 16 we can see it.
17 MR. CHESK: It seems to me you are in Region 18 III, basically, since you only have a --
19 MEMBER POWERS: No. It seems to me you are in 20 Region IV. If I use the rule of thumb, if the shutdown is 21 worth half an order of magnitude, and I use the rule of 22 thumb that the external events double issue from your
~
'23 psei eveht, I"am in Region IV. Those are the rules of 24 thumb given to me.
25 MR. CHESK: We have the diagram coming.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
417 1 CHAIRMAN APOSTOLAKIS: Say that again, Dana?
2 MEMBER POWERS: Okay. We are at four times 10 73
(~) 3 to the minus fifth. I am told a half an order of 4 magnitude now for shutdown. That gets me to seven times 5 10 to the minus fifth. I am told that external events 6 will roughly double my shutdown. So that adds another 7 four times 10 to the minus fifth. That put me at 1.1 8 times 10 to the minus four, which looks like region IV 9 delta to me.
10 MR. HOLAHAN: No. That's not the way the 11 staff has suggested using the information. It's that the 12 difference between Region II and III is to accommodate the 13 scoce difference.
^
So we would say you don't add it on, C';
\l 14 you use it by reflecting what region you are in.
15 CHAIRMAN APOSTOLAKIS: So you are in Region 16 III according to --
17 MR. HOLAHAN: Well, in fuzzy decision making, 18 you are near the boundary between II and III.
19 CHAIRMAN APOSTOLAKIS: It says 5.7 x 10 to the 20 minus five here. So it's clearly in region III. So they 21 should have done then on the basis of this -- this you 22 said did not have external events, did not have --
23 MR. CHESK: This does not have external 24 events. This is just internal events CDF.
(p,
) 25 CHAIRMAN APOSTOLAKIS: It does not have NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D.C. 20005-3701 (202) 234-4433
418 1 shutdown?
MR. CHESK: It does not have shutdown, q 2 NI 3 correct.
4 CHAIRMAN APOSTOLAKIS: So now I have to 5 remember what region III is, so I go to the text.
6 MR. HOLAHAN: They should be searching for 7 additional compensatory measures because they are in 8 Region III.
9 CHAIRMAN APOSTOLAKIS: Let's see. I am in 10 Region III. So this is applicable for plants with full 11 scope PRAs or plants with internal events PRAs, including 12 fire and internal flooding, supplemented with a 13 demonstration that the proposed changes have no impact on 14 those portions of level I PRA outside the scope of 15 internal events. That is, external events and shutdown l l
16 operations. l I
17 MR. CHESK: That is correct. Both licensees 18 have basically looked at the fires, looked at seismic, and
, 19 looked at shutdown, and showed that the components that l
20 they are going to extend the test frequencies to are not 21 being affected in those regions " Sat the PRA, they have 22 not analyzed.
23 CHAIRMAN APOSTOLAKIS: But internal fire 5 are 24 supposed to be included already.
(,) 25 MR. CHESK: That's right.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344 433 WASHINGTON, D C. 20005-3701 (202) 234-4433
419 1 CHAIRMAN APOSTOLAKIS: They are?
2 MR. CHESK: They are not in this case. I mean
-s
/ \'
\'
3 this was when we had these licensee submittals, this was 4 done before we decided that internal fires was going to be 5 in Region I.
6 CHAIRMAN APOSTOLAKIS: So they have to do it 7 now?
8 MR. CHESK: I am not quite sure which 9 direction we're going to go at this point yet. It's a 10 decision we have to make.
11 MEMBER KRESS: So, Gary, are we supposed to 12 interpret that baseline is the CDF due to internal events 13 only?
7.
14 MR. HOLAHAN: No. I 15 MEMBER KRESS: I thought that's what I j 16 interpreted.
17 MEMBER POWERS: That's what I interpreted you 18 as saying.
19 MEMBER KRESS: Yes.
20 MR. HOLAHAN: No.
21 MR. JONES: The baseline is what has been ,
22 analyzed. If all they have done is analyzed internal 23 events, then the base'line is internal events.
24 MEMBER POWERS: Okay. So my strategy --
,a
) 25 MEMBER KRESS: So my strategy is 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 1
420 1 MEMBER POWERS: Do the simplest possible PRA l l
, 2 you can with the fewest events you can possibly treat.
/ T
\") 3 Then you get a lot more regulatory latitude than if you do 4 a sophisticated one. ;
5 MR. JONES: But if you do that, then you end 6 up in Region III, and then you will have difficulty in 7 getting approval.
l 8 MR. HOLAHAN: Right, because you can't use 9 Region III if you haven't done the scope.
10 MEMBER POWERS: I simply shoot to get into 11 Region II all the time, actually Region I.
12 MR. HOLAHAN: Yes.
13 MEMBER POWERS: See I avoid big changes in ,
3 (3
V 14 risk by doing a bunch of things that go up and a bunch of 15 things going down. So it's very complicated. But my 16 delta is small, so you guys don't get on my case with it.
17 MR. HOLAHAN: Right.
18 CHAIRMAN APOSTOLAKIS: So now they will have ]
19 to do something about internal fires because they are in 20 Region III.
21 MR. WESSMAN: They may have to, or they can
. .22 .take .som.e ,of the gompone.nts that- they may have elected to _,
23 put in an LSSC category, therefore driving the numbers up, 24 and return to an HSSC type of category, helping to bring
,rm l
! ) 25 their overall delta back down. I'm not sure we have made v
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 421 l l
1 a decision on how to deal with fires, but obviously, -- )
l 2 CHAIRMAN APOSTOLAKIS: No, but it's not a
,s
(
~
3 delta that brings them there. It's the CDF values.
4 MR. WESSMAN: That's true, the starting point.
5 CHAIRMAN APOSTOLAKIS: 5.7 x 10 to the minus l
6 five. So independently of the delta, and the Region III 7 already.
8 MR. CHESK: They can still do changes that are 9 below the one minus six delta, so they could still qualify 10 in Region I.
11 CHAIRMAN APOSTOLAKIS: Yes, but that would be 12 a significant change in what they have done. I mean they 13 would have to go down by a factor of seven.
I
,q 1 k
K/ 14 MR. CUNNINGHAM: Obviously in that case, you 15 would have to look and see what's causing it. If it's all 1
16 the valves are doing it, contributing about the same 17 amount, then you are going to have problems doing that.
la If one valve is contributing a large amount, then you 19 maybe just take that valve out and it might dramatically 20 change the delta that you're -- l 21 CHAIRMAN APOSTOLAKIS: So it's really a cost 22 benefit tradeoff here. You don't take full advantage of 23 the risk-informed IST or you do more analysis.
24 MR. CUNNINGHAM: Yes.
,/ 3
( j
! 25 MR. WESSMAN: Or you make a change in 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
422 1
1 plant that changes that baseline.
,a 2 CHAIRMAN APOFTOLAKIS: Yes. Sure. !
l )
3 MR. CUNNINGHAM: So within this, there's lots 4 of different ways of getting an acceptable solution, if 5 you will. One of them is obviously do more analysis. If 6 you think you can save a lot of money by less frequent 7 testing of these valves, it may well pay for a better 8 analysis of fire or of shutdown or something like that.
9 CHAIRMAN APOSTOLAKIS: And then treat 10 everything else qualitatively.
11 MR. WESSMAN: Okay. If we can leave that, I 12 would suggest we --
13 MEMBER MILLER: I had a question on --
i n 4 V 14 MR. WESSMAN: Yes, sir. :
I 15 MEMBER MILLER: I missed this. What criteria ;
16 did they use for categories in your components? How did 17 they do that? l 18 MR. WESSMAN: I'm sorry. I didn't hear you.
19 MEMBER MILLER: How do they categorize their 20 components?
21 MR. WESSMAN: They use two groups of the HSSC 22 and LSSC. l l
23' MEMBER MILLER: How do'they define the l
24 differences?
(-
Q 25 MR. WESSMAN: Okay. They use the Fussel-NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
423 1 Vesely and the risk achievement numbers.
yw 2 Mike, do you recall?
, i
.) 3 MR. CHESK: Basically, what they started off 4 doing was using the PRA model to identify groups of risk-5 significant components. The criteria they ended up using 6 was a Fussel-Vesely greater than .01, and risk achievement 7 was bigger than two, being HSSC, and everything else being 8 LSSC.
9 They also went ahead and I guess with the RAIs 10 we put to them, performed a bunch of sensitivity studies 11 where they varied the recovery factors, common cause 12 factors, truncation values, looked at multiple component p_
13 importance, and came back with a rank list based on all
' 14 these sensitivity studies.
15 These, the inputs to these importance analysis 16 were actually input into an expert panel process. The 17 expert panel then looked at the results of these risk 18 rankings, and then determined based on other factors like 19 plant operational considerations as to what goes in to 20 HSSC and LSSC. So basically, the expert panel is the --
21 MEMBER MILLER: Is the final arbitrator?
22 MR. CHESK: That's correct.
23 MEMBER MILLER: That's the process everybody 24 is going to use until we put some number down in our Reg
/
( ,/ 25 Guide?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
424 1 MR. CHESK: Basically we did not put any
,- 2 criteria for ranking HSSC and LSSC in the Reg Guide. Our
3 criteria basically fell back on this little diagram here, 4 and what your final increase in CDF and LERF is.
5 MR. WESSMAN: We did not want to prescribe a 6 specific raw number, a Fussel-Vesely number, and stick to 7 a minimum number, set of numbers and acceptance criteria, 8 and leave that flexibility to the licensee.
. 9 Now the ASME document, Ken Balkey is standing 10 back here and I think wants to make a comment. But the 11 ASME document does provide guidelines, and I guess I would 12 expect licensees to head down the path of those sort of 13 numbers.
l
/ 'T '
\_ 14 MEMBER MILLER: So if the licensee goes down 15 the path of ASME, that will be a -- there was some i
16 justification that will be approved by the staff?
17 MR. WESSMAN: I think the final approval of a 1
18 licensee's proposed program will turn on what we settle on i 19 in the way of acceptance criteria with our fuzzy lines or ,
l 20 our description or our diagram. ;
i 21 MEMBER MILLER: So in other words --
) At the moment, the sort of I
22 MR. WESSMAN:
23 numbers that are in the ASME document appear quite 24 reasonable to us.
r3
( ) 25 MR. BALKEY: I'm Ken Balkey from Westinghouse.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
425 1 I chaired the ASME research group on risk-based in-service g 2 testing.
( )
3 During your discussion here, one of the 4 options that ASME is trying to make available, if these 5 two plants are indeed in Region III for sake of 6 discussion, you discussed that maybe some of the 7 components for your stretching the interval that's causing 8 the increase, will bring those back into the HSSC or 9 don't' move the interval out, or do further analysis.
10 I think that an option that ASME is trying to 11 provide to the owners is on the high safety significant 12 equipment, to look at better testing itself. One thing 13 about the better testing is that you really are making a (m / 14 change to the risk by that, where by doing more analysis 15 you are really not changing physical risk within the 16 plant.
17 So the ASME is trying to come'up with those 18 code cases that will permit more effective testing on all 19 the different types of components. From the work we did, 20 there are actually some methods that are actually cheaper 21 than what is being done now, such as lube oil analysis, an 22 example in the pump.
23 I think that that's definitely one of the l 24 options that the owners should be able to have. Hopefully r's 1
( ,)
25 ASME working with NRC and dovetailing the code cases and NEAL R. GROSS i
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON. O C. 20005-3701 (202) 234-4433
r 426 1 regulations, that that option will be available. Lecause
, .s 2 that really is an improvement in risk, a true improvement
( )
3 in risk by going to better testing methods.
4 CHAIRMAN APOSTOLAKIS: It just occurred to me 5 that there is no region where you are really asking for a 6 complete PRA, including external events. Right?
7 MR. HOLAHAN: Yes. Region III.
8 CHAIRMAN APOSTOLAKIS: No. It says here in 9 Regions II and III, you have to have a full scope PRA.
10 Oh , or plants with internal events have to consider --
11 have to supplement it with a demonstration that the 12 proposed changes have no impact on those portions of level l 13 I PRA outside its scope, that is, external events and l l
(T
(~) 14 shutdown operations.
15 So you can still get by with a limited level I 16 PRA and just demonstrate that what you are doing does not l 17 affect external events.
18 For some reason, I was under the impression 19 that in Region III, you always had to have a full PRA with 20 external events.
21 MR. WESSMAN: And you see, this may set a 22 smaller bound on the group of components that you could 23 put into a risk-informed program and other components 24 would be left in the " classical" program, which means they
,) 25 are treated like an HSSC and receive the current code test NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
427 1 at the current prescribed interval.
2 CHAIRMAN APOSTOLAKIS: Yes. Well I guess that
.m (s~, )
3 may be enough.
4 MR. HOLAHAN: Let me see if I can help, 5 because I think there's some confusion in the descriptions 6 of Regions II and III, which we mentioned yesterday that 7 we are still working on as part of working the uncertainty 8 discussion in there.
9 The intent, when the words and the pictures 10 are clear, is that Region III would be full scope internal 11 and external and shutdown fires and all those things.
12 CHAIRMAN APOSTOLAKIS: Yes. I understand 13 that. Now given the current state of the art, perhaps
'v) 14 demanding a shutdown PRA may be a little much, but 15 external events are done routinely. So I would make a 16 distinction between shutdown PRA and external events.
17 MR. HOLAHAN. I think in the -- when we talk 18 about this, we keep talking quantitatively. But you 19 remember, the document says that we'll accept qualitative !
20 treatments. So I think some judgement about maybe the 21 quantitative analysis puts you in a certain region. But i 1
22 there are other qualitative aspects to it that you think i
23 either would push you up or down, I think is an integral 24 part of the decision making process.
,rx
()
25 s CHAIRMAN APOSTOLAKIS: Yes. Now regarding 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
428 1 qualitative, in our letter of the 15th of this month, we e3 2 used the word qualitative only in the context of the
( )
v evaluation of inputs or the evaluation of the significance 3
4 of issues related to level 2 PRA.
5 MR. HOLAIAN: Yes.
6 CHAIRMAN APOSTOLAKIS: Not core damage 7 freque cy. So for core damage frequency, I guess we 8 implied that you have to have the core damage frequency, 9 from wherever it comes from. But you are saying that even 10 for the core damage frequency, you are willing to accept 11 qualitative arguments in lieu of actually doing a rigorous 12 analysis of all these activities.
13 MR. HOLAHAN: Yes. So long as there's rN 6 )
14 sufficient basis for the type of decision that's being 15 made. So there's a judgement involved.
16 CHAIRMAN APOSTOLAKIS: So you are extending 17 the use of qualitative evaluations beyond what we had in 18 mind. I guess we have to think about that, because I'm 19 afraid what Dr. Powers said may be a real problem here. I 20 can do a very poor PRA level I analysis, and then on the 21 basis of that, get away with things that I would not have 22 gotten away with.
23 MR. CUNNINGHAM: It becomes a tradeoff again 24 of poor, if you will, or an incomplete analysis there l A i
(' _ / 25 versus the cost of having to justify to us that the impact 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
429 1 is negligible in these other areas. That's another cost
,~
- 2 that you have to bring into it.
i) 3 CHAIRMAN APOSTOLAKIS: In about a year or so 4 though, all the plants will have an IPEEE, won't they?
5 MR. CUNNINGHAM: Yes.
6 CHAIRMAN APOSTOLAKIS: So that issue will --
7 MR. CUNNINGHAM: It will have an IPEEE, but 8 that doesn't necessarily mean they will have a seismic PRA 9 for every plant.
10 CHAIRMAN APOSTOLAKIS: Yes. They might have a 11 margins --
12 MR. CUNNINGHAM: Have a margins.
13 CHAIRMAN APOSTOLAKIS: Which they may decide (O
kl 14 to use for some argument.
15 MR. CUNNINGHAM: And likewise --
16 MR. HOLAHAN: Or they may have a FIVE 17 methodology for fire.
18 MR. CUNNINGHAM: Which it may be cost 19 effective to go on and extend that into, and make it a 20 fire PRA, depending on what each licensee wanted to do, )
21 how serious they wanted to pursue some of these 22 initiatives.
23 MR. JONES: I think, George, one comment about 24 the quality of the PRA that would be submitted with the
,/ s 25 change, and whether you can get away with a less rigorous
()
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 2000 5 3701 (202) 234-4433 l
l
430 1 job on doing it, I think in light of the comments made the 2 last time we discussed this issue, we tried to make it i7_ 1 3 very clear in the Reg Guide that the licensee's 4 responsibility was to identify the necessary scope and 5 quality for the change being proposed. If the change was 6 basically fairly obviously and simple kind of change and 7 would not have a large risk impact, you didn't necessarily 8 have to do a full-blown detailed PRA.
9 On the other hand, if it's a fairly complex 10 change, and I believe I would put this one in that kind of 11 category, I think we would require that the PRA has pretty 12 good quality.
13 So this is along the lines of trying not to be :
('%
k/- 14 too prescriptive in the Reg Guide and allow some judgement 15 to be used depending on the nature of the change being 16 proposed. You are going to have to rely on the industry 17 licensee, number one, to do a good job, and the staff to 18 do its oversight function, and assure that they have done 19 a good job in this area.
20 To define it much more rigorously than that 21 would be -- at that point I think if we had to get much 22 more detailed than that, we would just probably have to 23 punt to thou shalt do a great PRA fcr every change. I 24 mean it would almost come down to that kind of an
-s
( ) 25 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
431 1 It's very difficult when you start to deal in 7
2 the gray areas on defining that to that level.
t t 3 CHAIRMAN APOSTOLAKIS: I understand that. The 4 thing is that we have had several discussions among 5 ourselves regarding the usefulness of having a level III 6 PRA. We finally agreed on what you saw in the letter, the 7 recent letter, that maybe there are ways around it for 8 various reasons. But now you are saying even the level I 9 PRA doesn't have to be complete. I would like my 10 colleagues and I to discuss that at some point, or maybe 11 later this afternoon when we have the free for all 12 discussion.
13 MR. JONES: I think that's a perfectly good q)
( 14 question to put out for comment.
15 MR. CHESK: I was going to point out that as 16 far as the scope is concerned on page 14 of the general 17 SRP, we have three bullets on top there,. that tells the 18 staff how we can get around the fact that the licensee j 1
I 19 might come into a full scope level I type PRA.
I 1
20 MR. SHERRY: But I think it's possible to make 21 judgements, for example, for plants which have a seismic 1 l
22 margins analysis and they are in areas of the country with I
23 a very lcw seismic hazard, that you can probably make a j 1
24 determination fairly easily that there's a low seismic l
,- }
' 25 risk. l m
)
NEAL R. GROSS COURT REPORTERS AND TRANSCP:BERS 1323 RHODE ISLAND AVE., N W. l (202) 234 433 WASHINGTON, D C. 20005-3701 (202) 234-4433
432 1 CHAIRMAN APOSTOLAKIS: But then in my book,
,m 2 you have included it in the PRA. You should do that.
k) 3 That's part of the PRA. That PRA includes seismic events, 4 in my mind, if you do that. See what I'm saying?
5 In other words, you did some analysis and you 6 showed that the base core damage frequency you have 7 doesn't change. You have included it. Including it does 8 not mean going out and doing the whole thing that you find 9 in a typical PRA.
10 MR. CUNNINGHAM: It doesn't necessarily mean 11 that, yes. It can mean that, but it doesn't necessarily 12 mean.
13 CHAIRMAN APOSTOLAKIS: Well, if you managed to n
14 do what Rick just said then yes, it is included. l 15 MR. CHESK: Unless your application affects ,
16 the assumptions you used in doing your seismic margins 17 analysis or your FIVE analysis. So you have to be careful 18 in that sense.
19 CHAIRMAN APOSTOLAKIS: Well, FIVE is really a 20 screening approach. I would make a distinction between 21 HCLPF and FIVE. It's a screening approach. You may still 22 end up with a need of doing a more rigorous analysis, 23 right, for some location.
24 Anyway, we may visit that later today. Now t
( )N 25 one more 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
433 1 MEMBER SHACK: Before we go on, George.
CHAIRMAN APOSTOLAKIS: Yes.
m 2 u
3 MEMBER SHACK: Can I clarify Region III again?
4 I get more confused every time I hear the discussion.
5 From what we read, it sounded as though you didn't have to have the full scope PRA, there were alternatives. Then 6
you said you had to have a full scope PRA, but then you 7
8 said you would accept qualitative arguments.
9 Can you state once more as clearly as 10 possible, what I need in Region III?
MR. HOLAHAN: I can give it a try. Probably 11 12 the best version of this is that we're going to rewrite 13 the words. That's probably going to be better than --
I,,h
\~ 14 MEMbdR POWERS: But you're going to keep the 15 figure, right?
16 MR. HOLAHAN: I'll give it a shot. Okay?
17 MR. JONES: Yes. On the figure.
18 CHAIRMAN APOSTOLAKIS: Because it's clearly 19 irrelevant. It's the words that count.
20 MR. HOLAHAN: I don't know that any words are 21 required since we have the figure.
22 The intent of Region III is for use with full 23 scope PRA, which means including internal, external events 24 at power and shutdown.
n
(_) 25 I think what we have not defined is to 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
434 1 extent pieces of that could be qualitative as opposed to 2 quantitative. I think it's fair to say when the words
,s i
1
( '/
3 were constructed and the figure was drawn, we were 4 thinking of a quantitative analysis. But in the sense 5 that George has expressed it, if you can qualitatively 6 dismiss pieces as being non-contributors, then I think 7 that's -- I would say it would be all right to say that 8 you are in Region III by dismissing some pieces that you 9 made an informed judgement that it was unimportant.
10 I think it becomes a little more difficult if 11 you say there's a piece that I think is maybe an important 12 contributor, but I haven't quantified it. What should I 13 do the.?
(m 14 CHAIRMAN APOSTOLAKIS: Yes. That's right. I 15 think then you will have to think about and give some 16 guidance.
17 MEMBER POWERS: And you would define 18 importance in this context the same way you did importance 19 in the text itself, if it was 10 percent or more of your 20 CDP without the qualitative argument, you would say it was 21 important. If it's less than 10 percent, you would say it 22 was unimportant?
23 MR. HOLAHAN: No. I would say it a little 24 differently. I would say it's important if it would ,
L
() 25 change my view as to where I am in the decision 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
435 1 So if I am already close to one of the boundaries, then a 2 change even smaller than 10 percent might reflect being in I ,)
3 a. different decision arena.
4 MEMBER POWERS: But if you are down around 5 three, then a fairly substantial contribution could move 6 you, could still be unimportant?
7 MR. HOLAHAN: Right.
8 CHAIRMAN APOSTOLAKIS: Enough on this? I have 9 one question.
10 The Commanche Peak delta CDF is 7 x 10 to the 11 minus six. Now let's -- and this comes only from risk-12 informed performance based IST. Let's say for the same 13 plant is also revisiting its tech specs and its graded QA
(./ 14 program or establishing a GQA program. And they get 6x 15 10 to the minus six from the tech specs, right, and a 5 x 16 10 to the minus six from the GQA. Now the total now 17 becomes the 13 plus 518. 1.8 X 10 to the minus five.
18 Now I have an additional problem that I am in 19 the top delta. Right?
20 MR. HOLAHAN: If you were to do them at the 21 same time --
22 CHAIRMAN APOSTOLAKIS: I'm coming to that.
23 Now we said yesterday or you said yesterday that you would 24 like to have some time to digest the significance of
,a l But since -- I mean, they are not asking to change
(/) 25 these. l 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
436 1 the CDF by more than 10 to the minus fifth just by doing 2 changes in the IST. So why would you need serial times to i
t 3 digest the significance of IST changes and tech spec 4 changes when in fact you have different people dealing 5 with these, don't you? So they could be digesting it at 6 the same time, and the utility would be saving much more 7 money.
8 MR. HOLAHAN: But I would say, I think these 9 sort of questions are not now, and will probably never be 10 clarified in the text of any of the guidance documents.
11 Because you are talking about judgmental things that are 12 involved in a decision process.
13 To the extent that you are correct, that you n) v 14 are dealing with two issues which are independent and 15 could be monitored separately so that you were gaining 16 experience in both sort of applications at the same time, 17 I would think it would be acceptable to implement both 18 simultaneously.
19 CHAIRMAN APOSTOLAKIS: There is a way around, 20 and by playing games with the words. You might say each l 1
l 21 one of these is a separate change. i l
22 MR. HOLAHAN: Yes. ;
23 CHAIRMAN APOSTOLAKIS: So then you are below 24 the line?
) 25 MR. HOLAHAN: Yes.
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
437 1 CHAIRMAN APOSTOLAKIS: You're fine.
s 2 MR. HOLAHAN: Yes.
/ i LJ 3 CHAIRMAN APOSTOLAKIS: That's why lines really 4 are meaningless because you can play games like that. But 5 that's a separate issue. It's the same thing with 6 accident sequences. You can break them up, right? Then 7 of course it goes down, and everything is happiness and 8 roses.
9 MR. WESSMAN: But I would take us back to as 10 you make a change in one area, you have to look at see 11 does it impact the other area. I am back to my graded QA 12 example versus IST testing of how to treat valves or 13 components. I think those are very important
( )
'd 14 conciderations across the disciplines.
15 CHAIRMAF APOSTOLAKIS: That's a good point.
16 Yes. ;
17 MR. CUNNINGHAM: It also could come in if you 18 add the delta from one that was going to change your !
19 baseline CDF. You do one first and the second, but 20 baseline CDF changes. So you may be moving horizontally.
21 It might impact you in that respect. It probably won't, )
22 but it could.
l 23 CHAIRMAN APOSTOLAKIS: So that, just if I may l 24 make a suggestion, when you revisit the principles,
) 25 principle number five which used to be changes to CLB I
1 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 I
I
438 1 should be done in a fashion that results in only small
- 2 increases in calculated plant risk. We agreed you were
,3 N) 3 going to change the words yesterday, but maybe you ought 4 to think about this issue now and add a few words to give 5 you the flexibility, to do what you just said, Gary.
6 What exactly is a change. That's really the 7 issue here. What is a change. If you are talking about 8 two entirely different changes, they don't interact, say, 9 then perhaps you would accept both of them and apply this 10 rule to each one separately.
11 MR. HOLAHAN: Yes.
12 CHAIRMAN APOSTOLAKIS: So it's not the total 13 that counts. )
~. ;
/ h 14 MR. HOLAHAN. Yes.
15 MEMBER POWERS: George, I have a question on 16 that from a different point of view. Suppose we discussed 17 at one time packaging.
18 CHAIRMAN APOSTOLAKIS: Yes.
19 MEMBER POWERS: And suppose I come through and 20 I say okay, I've got a big change I'm going to make here 21 and it's going to increase risk, so I am going to counter-22 balance it with another big change such that on the 23 average, on the sum I have a small change, maybe even 10 24 to the minus six small change. Don't I then run afoul of 7S
( ,) 25 principle number six?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
439 1 CHAIRMAN APOSTOLAKIS: In the performance?
2 MEMBER POWERS: Maybe it's five. It's one 7s i \
\'~' /
3 where they --
4 MR. JONES: Five.
5 MEMBER POWERS: I'm sorry. Principle number 6 five.
CHAIRMAN APOSTOLAKIS: In that case, the 7
8 licensee explicitly requests that this be teated as a 9 single change. But in the example I gave, the licensee may request that they be treated separately. Those guys 10 11 have to agree.
12 MR. WESSMAN: Then they would have to develop 13 the argument quite separate.
(
- 14 CHAIRMAN APOSTOLAKIS: It's a different case.
15 MR. WESSMAN: That's correct.
16 MR. JONES: The integrated case, Dana, that 17 you are talking about, is addressed in the Reg Guide now.
18 We are concerned that they come in with large changes 19 above the outside the box, if you wish, and then offset it 20 with another large change to bring you inside the box.
21 That, we're very skeptical of.
22 MEMBER POWERS: You cover it in one of your 23 principles.
24 MR. JONES: Yes. If you are playing the game O 25 within the box, this one is high, 10 to the minus six.
()
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
440 1 This was is 10 to the minus six down, something like that.
,, 2 But when you get outside the box, we're very skeptical at s
)
3 that time, and we want to look at it extremely carefully, 4 if we would allow it at all.
5 We think each individual change then has to be 6 justified on its merit. So the big change up would have 7 to be justified on its merit alone.
8 CHAIRMAN APOSTOLAKIS: You have not decided on 9 it yet. You are thinking about it.
10 MR. JONES: The words are in there now. That 11 is in the guide now. That's on page 210 of the guide, of 12 the general guide. We haven't re-packaged it in the 13 individual, but that is part of the general guide.
p/
'U 14 CHAIRMAN APOSTOLAKIS: We may discuss that a 15 little later. Maybe that's another thing for today. i I
16 But in the example I gave, both resulted in an 17 increase.
18 MR. HOLAHAN: Yes. Can I comment on one 19 additional thing?
20 CHAIRMAN APOSTOLAKIS: Sure.
21 MR. HOLAHAN: A lot of the discussion we've 22 just had shows how easy it is to be caught up in the l
23 numerical analysis as opposed to the integrated decision 24 process that we're suggesting in the guidance documents.
o
) 25 Frankly, my perception of how important in-NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, O C. 20005-3701 (202) 234-4433
441 1 service testing is and the changes in it, I would be very fw 2 surprised to see that it really resulted in a 7 x 10 to
( )
\
' ~ '
/
3 the minus six increase in core damage frequency. I think 4 before I went ahead and plugged that number into a text or 5 figure and ran off and used it in the decision, I would Do I really 6 have to ask where did that number come from.
7 believe that the changes that are being proposed are --
8 that those numbers fully reflect all the changes that are 9 being made.
10 In a lot of cases, I think what happens is you 11 calculate the pieces that are easy to calculate and you 12 don't calculate other changes in the program. The fact 13 that you may be doing smarter tests or tests of stuff that U 14 you didn't do before, or something else that might not 15 happen to be modeled in the PRA, it's not reflected in 16 that number.
17 CHAIRMAN APOSTOLAKIS: But in principle of 18 course, you would always ask these questions.
19 MR. HOLAHAN: Yes.
20 CHAIRMAN APOSTOLAKIS: So what you are saying 21 is that you will be even more skeptical in cases where the 22 result is a bit surprising?
23 MR. HOLAHAN: Yes. Because intuitively, if 24 you look at what's being proposed and if you go through
,m
) 25 the principles, it would say what is being proposed looks NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
442 1
1 fairly sensible. If I take the numerical values, it says
,- 2 what's being proposed has a lot bigger effect than I would
( ')
3 have suspected, which makes me think that this is a good 4 screening tool for saying do I really believe that 5 analysis.
6 CHAIRMAN APOSTOLAKIS: Or maybe you may be led 7 to the conclusion that the PRA level I was really not very 8 good.
9 MR. HOLAHAN: I would have no difficulty in an 10 integrated decision process in which the calculated 11 numerical value for part of how the change affected the 12 plant was 7 x 10 to the minus six, with a qualitative 13 judgement that this is probably really an over-estimate l I
'd 14 because of the things that haven't been accounted for, and 15 what you think the real affect was, and saying that this 16 is as a whole, this looks like a sensible change.
17 CHAIRMAN APOSTOLAKIS: Let me make another --
18 MEMBER POWERS: Does the other possibility 19 exist that the calculated is 6 x 10 to the minus five or 20 six, but your intuition says that that's an under-21 estimate?
22 MR. HOLAHAN Sure. I think so.
23 MEMBER POWERS: Because of the things that are 24 not included?
7m, 4
) 25 MR. HOLAHAN: 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
443 1 MEMBER POWERS: So this is really a crap shoot
,m 2 to come in here. You have got to not only do the PRA
. )
.s s.
Then you have to make a judgement on what 3 calculations.
4 the quantitative evaluation of the things not included in 5 the PRA is going to be drawn by a Holahan.
6 MR. HOLAHAN: I don't think I would describe 7 the process like that. I would say the staff expects --
CHAIRMAN APOSTOLAKIS: Well I mean what they 8
9 mean is that you should have some shaded areas there.
10 (Laughter.)
11 MR. HOLAHAN: It wasn't so bad when he used 12 Murphy's name, but I think things are --
13 CHAIRMAN APOSTOLAKIS: No. Murphy will tell
( !
14 us this afternoon in fact --
15 MEMBER POWERS: It's not only the things that 16 Holahan has thought about, but he has a quantitative 17 judgement on how they might add to this.
18 CHAIRMAN APOSTOLAKIS: Okay. Would someone 19 inform Joe that we expect him to answer?
20 Before we go on though, there is an item that 21 is not in the agenda that I just thought of. At 3:30 22 we'll have to interrupt for a while. So I would suggest 23 maybe half an hour or 45 minutes, that maybe we can speed 24 up the process a little bit. If you feel that some of e
f i C/ 25 this stuff already has been covered yesterday or just skip NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C, 20005-3701 (202) 234-4433
444 1 some of the viewgraphs?
2 MR. WESSMAN: Sure. What I would propose,
[,,)
3 George, is if we can have about 10 more minutes, we'll 4 step you quickly through these other viewgraphs because I t
5 think they show where the pilots were within the bounds of 6 the elements and where we still have additional work to 7 do. I think it's illustrative. I'm glad we had this 8 discussion of the last half an hour, because I think the 9 example of the pilots has stimulated our thinking. That's 10 what we wanted to do.
11 CHAIRMAN APOSTOLAKIS: Just keep the stuff 12 that we have covered already.
i 13 MR. WESSMAN: Okay. Okay. Don't ask any r
ky/
s 14 questions for a couple minutes.
15 CHAIRMAN APOSTOLAKIS: I didn't say that.
16 Just keep the stuff that you have covered already.
17 MR. COLACCINO: I can do this in less than 10 18 minutes. Now we're going to go back through the elements 19 and see what the pilots have actually given to us. Under 20 element one, they have as we said previously, included 1 21 components that were not part of their previous IST 22 program. Palo Verde included one. Commanche Peak 23 included about 26 components. They both identified 24 maximum test intervals. They both set those at six years.
p
( ,) 25 This differs a little bit what we have in our Reg Guide, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
445 1 as we've explained befo 7
,, 2 We do anticipate putting out future requests
( )
\. '
/
4 3 for additional information to supplement the information 4 that we don't have. Things like testing schedules.
5 CHAIRMAN APOSTOLAKIS: By the way, these two 6 plants are doing only these two pilot studies or they are 7 doing others as well?
8 MR. WESSMAN: I believe Palo Verde is involved 9 in working on some of the graded QA work.
10 MR. COLACCINO: Okay. Next slide. When Palo g 11 Verde and Commanche Peak put in their submittal, they 12 stated that they used the guidelines of the PSA 13 applications guides, not something that we have endorsed.
rh
,I \
C/ 14 PRA importance calculations used to identify 15 relative risk contributions of pumps and valves. We 16 talked about that before. They both used the expert panel 17 process.
18 CHAIRMAN APOSTOLAKIS: We're wondering here 19 how the expert panel is chosen. That's maybe not relevant 20 to you, but we should discuss it sometime.
21 Gary?
22 MR. HOLAHAN: Yes.
23 CHAIRMAN APOSTOLAKIS: I think we should 24 discuss the issue of the expert panel and how these guys t' 3 25 are chosen. Maybe later this afternoon?
()
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
446 1 MR. HOLAHAN: Okay.
,m 2 MR. COLACCINO: Both pilots do include
( )
3 criteria, you know, the qualifications of individuals that 4 they want on their expert panel for IST. I believe that 5 it also composes a part of the panel that individuals that 6 will participate in the expert panel for the Maintenance 7 Rule.
8 CHAIRMAN APOSTOLAKIS: Now when you look at 9 the importance results, do you also look at the dominant 10 accident sequences?
11 MR. CHESK: When the staff looks at the 12 results, we do, and I'm not sure if the licensees do or 13 not, but I am going to bet they do look at them.
t k ') 14 CHAIRMAN APOSTOLAKIS: Is it possible to get 15 an importance ranking of pumps and valves that tells me 16 pump 643-1A and so on is at the top, but that pump is not 17 in the dominant accident sequence? And if that is the 18 case, how would that affect your integrated decision l I
19 making?
20 Put it another way. Can I change the 21 importance ranking by including more things in the PRA?
22 MR. CUNNINGHAM: Yes. Yes.
23 CHAIRMAN APOSTOLAKIS: So if I include seismic 24 risk, and strong earthquake makes all the buildings 7_
() 25 collapse, and I get five dominant sequences from that, 4
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
447 1 then the relative significance, importance ranking of the e' 2 pumps is downgraded.
\,s! i 3 MR. CUNNINGHAM: Yes.
4 CHAIRMAN APOSTOLAKIS: To the point where they 5 may become low safety significant?
6 MR. WESSMAN: I would be surprised if a lot of 7 those pumps would make it to that threshold. I think what 8 we have seen so far is most of the pumps, or at least half 9 of them are in the high safety significant. Intuitively, 10 I can not imagine them leaving.
11 CHAIRMAN APOSTOLAKIS: Because they have not 12 done seismic? I'm not saying it's going to happen, but is 13 it conceivable?
i n
! /
N' 14 MR. CUNNINGHAM: Yes. Yes. Sure. You would 15 be having things moving all around in the importance 16 rankings, depending as you bring in seismic, you bring in 17 fire.
18 CHAIRMAN APOSTOLAKIS: So what happens then is 19 that because my plant is very bad from the seismic point 20 of view, I get relief in other areas?
23 MR. CUNNINGHAM: No. But ultimately what 22 happens, I think what you are talking about is an inherent 23 weakness in importance measures. It's only a relative 24 measure. You don't even have to bring other things in.
p) r
(_./ 25 All you have to do is make some equipment less reliable NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
448 1 and the importance of some equipment will be reduced.
,, 2 That is dealt with by the delta CDF, delta
?
3 LERF tests, because if in effect what you are doing is by 4 changing or distorting the importance, you are putting 5 more equipment in the low category, you are in effect 6 making a bigger change. That bigger change is caught in 7 the delta CDF test and in the delta LERF test.
8 CHAIRMAN APOSTOLAKIS: It seems to me that we 9 are using the words importance calculations too much. I 10 mean since this is an integrated decision making process, l
11 maybe we can say importance calculations plus some 12 investigation of the dominant accident sequences. That's 13 where you will take care of issues like that.
'w/' 14 MR. HOLAHAN: In an integrated sense, what 15 happens is the importance measures are screening, because 16 what they do is they give you the first cut at what goes 17 in the high and low category. But then there is an 18 additional test to see whether that cut at high and low 19 has produced acceptable or unacceptable consequences.
20 CHAIRMAN APOSTOLAKIS: And that's where the 21 sequences come in?
22 MR. HOLAHAN: That goes back to what Ken 23 Balkey said, is if you don't like the answers, then you go 24 back and you move things around or you shorten the
,~
( ,) 25 intervals for certain pieces of equipment.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 23M433 WASHINGTON. D.C. 20005-3701 (.J2) 234-4433
449 1 CHAIRMAN APOSTOLAKIS: I would certainly like 2 that kind of thing to be in the bullets to emphasize it.
/, ,)
t'"1 3 Because now we are emphasizing the importance calculations 4 too much I think. What really counts is not importance.
5 What counts is the dominant sequences, isn't it? Or the 6 accident sequence.
7 MR. CUNNINGHAM: What counts is the integrated 8 decision making, including considerations of PRA 9 information, including dominant sequences and importance, 10 and a lot of other stuff.
11 CHAIRMAN APOSTOLAKIS: That's correct.
12 MR. WESSMAN: Okay. I think we've covered 13 most of this on the element two. You have another page of 77 <
- V 14 element two?
15 MR. COLACCINO: Yes. I think we have covered 16 also most of the next page of element two.
17 MR. WESSMAN: Okay. Let's go right to element 18 three then.
19 MR. COLACCINO: As we said, we favor the use 20 of the code cases. The licensees have committed to 21 investigate the use of the ASME code cases when they are 22 approved by the staff. I think this makes total -- you 23 know, it makes perfect sense to us since the code cases 24 are not on the street and they aren't endorsed by the NRC.
f (3 ) 25 Other things that the submittals contained was 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
450 1 that in both of them, in Palo Verde's I know specifically,
-s 2 it said in words that we would do two tests, two quarterly
[ '\
3 tests, and then we would step out the interval for low 4 safety significant components to six years. We were told 5 later in clarification, this wasn't the case. That's when 6 we started asking more questions about scheduling of the 7 tests. We started talking more about staggered testing 8 and things like that. So those are more subjects for 9 future requests for additional information.
10 CHAIRMAN APOSTOLAKIS: But is it conceivable 11 you would accept a six year interval?
12 MR. COLACCINO: I think right now what we're 13 proposing in the Reg Guide is five years or three
,q
)
N/ 14 refueling outages. Certainly that's doable in a step-wise 15 fashion to step out to that length of time.
16 CHAIRMAN APOSTOLAKIS: That's on the basis of 17 what? Is that risk informed or is it we just don't want 18 to have too long periods?
19 MR. COLACCINO: The five years, three 20 refueling outages, is consistent with what's in OMN-1.
21 MR. WESSMAN: But I'm not sure that it's >
22 necessarily risk informed. I think it's primarily based 23 upon the engineering and historical evidence regarding the l 24 performance of certain items of equipment. There has been
(s 25 enough testing in the MOV area, for example, where they
( , )
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
451 1 have been able to demonstrate that there is an expectation n 2 that the valve would not degrade to the point where it
( )
We have seen work in 3 would fall into an inoperable state.
4 this area.
5 Now we haven't seen it in the actual plant 6 environment, as they obviously are in simulated and test 7 environments. So this is why it ought to be a step-wise 8 approach.
9 But six years is three 24-month refuelings.
10 At the moment, that is the outer bound that we have 11 considered.
12 CHAIRMAN APOSTOLAKIS: Actually, since I 13 started the thing yesterday, you can also state that in I, h 14 the PRA space, by saying that after this time, the 15 assumption of the explanation of this is no longer valid.
16 One of the basic assumptions of the PRA is no 17 longer valid.
18 MR. HOLAHAN: Right. I would have entered the 19 question almost entirely the same, except I would have 20 started out saying yes, it is risk informed.
21 CHAIRMAN APOSTOLAKIS: Okay. Well in that 22 sense. In the sense that the basic assumption of the PRA 23 of a constant lambda for that component is no longer 24 valid. Now lambda, you are saying, based on historical ex
) 25 evidence, is beginning to increase.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
452 1 MR. HOLAHAN: But even beyond that, without a rm) 2 numerical -- without a PRA model, I think it is still a
(
v judgement about risk. It's not an arbitrary judgement.
3 4 CHAIRMAN APOSTOLAKIS: My point is not -- yes, 5 it is a judgement about risk. All I am saying is that all 6 these engineering insights can be expressed in terms of 7 the PRA.
8 MR. HOLAHAN: Yes. They have a PRA analog.
9 CHAIRMAN APOSTOLAKIS: Let's not separate the 10 two. In this case this is what happens. Lambda is 11 beginning to increase.
12 Brad?
13 MR. HARDIN: I was just going to add too that
('" )
14 of course the possibility of a plant being allowed to go 15 out to five years is contingent on satisfying all the 16 other risk decisions, including curves and so forth.
17 CHAIRMAN APOSTOLAKIS: Curves?
18 MR. HARDIN: The figures and the review 19 process involving all the other parts.
20 MR. HOLAHAN: And the principles.
21 CHAIRMAN APOSTOLAKIS: We were doing so well 22 up until this point.
23 MR. CUNNINGHAM: It's contingent on the text 24 and/or figures that --
) 25 CHAIRMAN APOSTOLAKIS: 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
453 1 MR. WESSMAN: Obviously the overall assessment
-s 2 that they present to us, and that we are willing to i <
\ )
'"' 3 accept, it all has to come together.
4 I think we have pretty well done this element.
5 We have used our 10 minutes. I think we have one last 6 page we'll mention to you. Then we are ready to conclude.
7 MR. GOLAY: Just to build on this last part, 8 does this mean that if you wanted to go to longer testing 9 intervals, you would require that the equivalent of the 10 bathtub curve be available for the components that would 11 be pushed out, given that they satisfy your safety 12 screening and so on?
13 MR. WESSMAN: I am not willing to leap to the e3
)
~ 14 bathtub curve, but I --
15 MR. GOLAY: No, no. I'm just saying the 16 lambda history.
17 MR. WESSMAN: I certainly think we have to 18 have a performance history that allows us to make that 19 decision. It would have to be stepped out in a step-wise 20 fashion.
21 CHAIRMAN APOSTOLAKIS: But in the spirit of 22 integrated decision making, I can see a licensee come to 23 you with a bathtub curve that they can defend, and based 24 on that, make some argument. Then you guys should have an 7
(,i 25 open mind.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 2LD05-3701 (202) 234-4433
1 1
454 l' MR. WESSMAN: Well that is true. Some of the
,, 2 MOV people have argued that they have data that says it's e i
' But somehow, 10 years sounds so far away, 3 good 10 years.
4 we are not ready to leap to that opportunity yet.
5 Consequently, we felt that the five year or six year time 6 frame we felt we had more information in that area, and 7 until we gradually step our way out and see how the plant 8 experience develops. But we did leave the door open, such 9 that a licensee could propose a longer interval if 10 appropriately justified.
11 MR. HOLAHAN: For example, I could imagine 12 there might be a piece of equipment that's similarly used 13 in the petro chemical industry or something for which
, ~ .
(.J 14 there is experience for long intervals. Maybe that could l l
15 be drawn in. But I agree with Dick's risk-informed )
I 16 judgement that without the performance history and without l 17 understanding what sort of mechanisms and failures and 18 stuff might be involved in that five years, that it's 19 pretty hard to accept such a thing.
20 MR. GOLAY: Well, what prompts the question 21 though is that I think for a lot of components, you are 22 going to find that the data that you are speaking about 23 are not really going to be available. You may end up with 24 a burden of testing a lot of things which are down in the n
{) 25 bug dust from a risk contribution point of view, but 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
455 1 which the library of data simply would be very expensive I
2 to generate. It may be premature to say well they've got !
.s
\
(y 'J l
\
to come in with a lot of additional material if the risk 3
4 significance is very low, j i
5 MR. HOLAHAF. 'Well, one piece of data which j 6 would become available is for the equipment that you 7 extended on for you know, three, fear, five years.
8 Presumably the equipment gets tested after those five ]
l 4
9 years. If you don't see any failures at that point, you 10 have at least that basis for saying okay, now let's j 11 stretch out.
12 But I think that would be he'lpful, even if 13 that's the only information you got.
kl 14 MR. CUNNINGHAM: That's a key piece of the 15 performance monitoring box that we have on our four 16 element, four step process.
17 MR. WESSMAN: It helps develop that confidence 18 level. I think some of this MOV testing is in a different 19 environment, but they may be able to correlate to where it 20 seems reasonable in a nuclear environment.
21 A lot of work that I have had described to me 22 at least in the check valve area says gee, anything beyond 23 five years just does not seem to make sense. They are in 24 that sense, a passive component. The degradation ia
!s ,) 25 mechanisms trat have been seen thus far make it very NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
456 1 difficult to make an extrapolation beyond that, without
~~ s 2 some degree of inspection.
> s
\ l
'~'
3 You may all recall, I think it was the San 4 Onofre event of the mid-1980s or so, where some AFW valves 5 had gone almost five years, check valves, and they found 6 several of them that had failed or weren't even there at 7 that point in time.
8 MR, HOLAHAN: I think the other practical 9 aspect of it, if you think about the costs involved, five 10 years versus quarterly testing is a factor of 20. To go 11 to 10 years only changes you by another factor of two. So 12 I mean most of the benefit is being achieved even just by 13 that first step.
I
/ T
- 14 MR. WESSMAN: So getting from quarterly to two l
15 years is a tremendous incentive. That's what we're here 16 --
17 CHAIRMAN APOSTOLAKIS: On the mechanics of it 18 now, I think the question really was related to the 19 principle of the thing.
20 But even though you may have some aging in the 21 sense that the failure rate starts increasing, if somebody 1
22 can make a convincing case, which is what you are 23 addressing, you may even be willing to live with that and 24 approve an extension.
25 Now Gary gave an example where you do it NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
457 1 gradually. So you are building up the convincing case, in p 2 other words.
> s
'N~J I think that's the only way that 3 MR. WESSMAN:
4 we would be willing to move in that direction.
5 CHAIRMAN APOSTOLAKIS: That's fine.
6 MR. WESSMAN: Let's just very quickly on slide 7 four, some of the documentation material that the pilots 8 gave us.
9 MR. COLACCINO: I just want to point out the 10 second bullet, the documentation on the integrated 11 decision making process, how it's incomplete.
12 When we were going over and looking and asking 13 questions about them, one of the pilots couldn't really (n) v 14 answer our questions very well. As a result of that, they 15 informed us that they were going to reconvene their expert 16 panel to re-do the process. I think it's one of the 17 important aspects of documentation. They weren't even l 18 comfortable with the documentation that they had from 19 their own expert panel deliberations, so that's something 20 we want to see on site and maintained. That's all I have.
21 MR. WESSMAN: Okay. Well, I think we have 22 certainly -- we are learning from these pilots, and 23 certainly we are feeding towards them and they are feeding j 24 toward us.
(~h 25 Similarly, we are learning with our encounters 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
1 458 1 with ASME and have communicated in both directions in this l 7s 2 way. I think in the IST area, it has been a very i
%~)1 3 profitable exchange on both ways.
4 MEMBER SEALE: Have you given any thought to 5 the idea that some of this in-service testing could get to 6 be pretty exotic if you extend the interval out long 7 enough. Particularly in parts of the systems where it's 8 pretty sophisticated process to get to the item to test ;
9 it. Maybe the utilities need to document the in-service 10 tests so that five years from now, they know how to go 11 back and do it again.
12 MR. WESSMAN: Well --
13 MEMBER SEALE: I'm serous. You might want to f%
s-) 14 videotape how you got in to a certain part of the plant, I 15 mean specifically certain components.
16 MR. WESSMAN: Well in certain aspects of 17 maintenance and outage planning, I think some of this is 18 going on right now. Certainly under the present IST 19 regime where licensees have found components where access 20 is particularly difficult or the ALARA considerations are 21 quite onerous, in some cases they have come to the staff 22 and sought either a refueling outage frequency as opposed 23 the code required frequency. That means and if you have 24 parallel train situation and it really means using a
/~N
(,) 25 grouping in a parallel train, they may visit one train one 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 459 .
I 1 year and it could be two years or four years hence until
,_ 2 they revisit that same train it. that same compartment I ,,]
'- 3 again.
4 The type of test is likely to be a prescribed 5 test that isn't likely to change much. So the actual 6 conduct of the test may not be so difficult as the access 7 to instrument the component for this sort of thing.
8 That's all we have on the risk-based IST and 9 Reg Guide. I think some of our examples helped us a 10 little, George. I think that's where we are ready to 11 stop, if you are.
12 CHAIRMAN APOSTOLAKIS: Any other comments or 13 questions? ,
[~'i
\xl 14 MEMBER SHACK: I just wondered if any of the j l
15 pilot plants were here to have any of their comments. i l
16 MR. WESSMAN: No, but Ken was raising his I
17 hand. The pilot folks were not here with us at this time.
18 You may recall when we met a month or so ago and the ASME i 19 made a presentation, the pilots were represented, and Tom 20 Cannon spoke a little bit about the Palo Verde stuff.
21 Ken, perhaps you have a --
22 MR. BALKEY: I just wanted to add a comment, 23 listening to the discussion, that risk-informed movement 24 is actually dovetailing very nicely with movements within t's
( ,) 25 ASME. I believe at the ASME meeting, the ASME code has NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234 4433
460 1 been trying to move towards more performance based 7s 2 standard. The two key pieces that tie in very nicely are
( )
3 the aspect of what Dr. Seale just talked about. You have 4 a difficult component where you might not be able to do a 5 flow test, but the plants are piloting non-intrusive 6 examination methods that can actually characterize the 7 behavior at equipment better than dismantling it.
8 The other piece is that as we moved and did 9 the risk-based research, a number of organizations have 10 now stepped up to start getting the data in order to 11 defend I think the performance histories to help defend 12 extending the intervals.
13 You have the nuclear industry check valve i
! 1
\/~ 14 group that has a data base trying to characterize which 15 components perform best in which environments and relative 16 to particular designs. As Mr. Wessman said, Oak Ridge 17 National Lab is now starting work, very good work on pump 18 behavior, motor operative valve behavior. Then you heard 19 of course Steven Gosselin this morning talking about 20 getting data on piling failures. So the risk-informed 21 movement is actually helping to drive getting other 22 information that will tie in and move towards risk-23 informed performance based process all together.
24 MEMBER FONTANA: Some of the work you have n
( ,) 25 mentioned includes things like motor current signature NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
461 1 analysis?
f3 2 MR. BALKEY: Absolutely.
(
'~'
)
3 MEMBER FONTANA: That is really good stuff.
4 MR. BALKEY: Yes. Like the lube oil analysis 5 work also that Mr. Wessman discussed. There are a number 6 of plants that are piloting and moving forward with these 7 much enhanced methods, because they are much better 8 predictor and actually your benefit in terms of plant 9 availability.
10 MR. WESSMAN: There's pretty intense interest 11 out there, and although we have these two pilots we're 12 looking at, there's eight or 10 others that are eagerly 13 waiting to make submittals to the staff once they see O
\ )
k/ 14 progress by Commanche Peak and Palo Verde. In other 15 arenas, the Vogtle plant is interested in the check valve i l
16 work, and the Shearon Harris plant is interested in pump 17 work. And there's a bunch of people thinking risk 18 informed testing for snubbers and pip supports. So 19 clearly we are hearing a lot of interest from the industry 20 on that, most of it built around economics, I'm sure.
21 CHAIRMAN APOSTOLAKIS: When do you think you 22 will have a decision on this? We heard yesterday that on 23 tech specs it's a matter of a couple of months or so. How 24 about in your case?
['
(_)T 25 MR. WESSMAN: Well, our expectation with Palo 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
462 1 Verde and Commanche Peak is to be to a point that we could
,x
, 2 advise the commission as to the staff's view in March of 3 1997.
4 Now part of the difficulty getting from here 5 to there is getting enough time away from briefings and 6 preparation of papers and things so that we can work on 7 the review. We have some additional questions that we 8 need to get out to the licensees, and of course they have 9 to respond.
10 CHAIRMAN APOSTOLAKIS: Of course you also want 11 to have a quality product at the end.
12 MR. WESSMAN: Yes, sir. Absolutely. If we 13 get to March and we just still don't understand some of
! \
\' # 14 the assertions and aren't ready to make a decision, we are 15 not going to make it.
16 CHAIRMAN APOSTOLAKIS: I think it is really l
l 17 important to speed up the processes as much as we can 18 because of the skepticism that's out there that risk- )
19 informed approach really would not contribute anything, 20 would not give any relief to the licensee.
21 Evidently, there is a large number of 22 licensees that are still extremely skeptical. We heard 23 that yesterday morning and last time when John Garrick was 24 here, and I hear it from other people. So if you guys
[ \
(_,) 25 make some decisions using these things in the next two or NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
463 1 three months, I think that will go a long way towards
,m 2 helping people change their minds.
/ T
('^'/ 3 MR. WESSMAN: I think one part of dealing with 4 the skepticism of course is getting our compendium of 5 documents out in the public arena so the public and the 6 licensees can see where the staff is heading.
7 CHAIRMAN APOSTOLAKIS: Now in your case, you 8 are not releasing anything for public comment, are you?
9 MR. HOLAHAN: Yes.
10 MR. WESSMAN: We are on a uniform schedule so 11 that in about January -- you remember the calendar that 12 Gary showed you yesterday that had us taking the documents 13 to the EDO, and the EDO advising the Commission in mid-
,,y i
(/ 14 December. The expectation is that the Commission will 15 subscribe to the release of the general Reg Guide and its 16 SRP, and several of these application specific reg guides 17 and their SRPs for public comment in January. We still 18 have that countdown with our commitment with the chairman, ]
f that by the end of 1997, we have resolved public comments 19 20 and these documents are in final form. !
21 CHAIRMAN APOSTOLAKIS: Now when we write our i 22 letter we are expected to comment on the individual cases 23 as well, or we choose what we do?
24 MR. HOLAHAN: The staff would appreciate that, p '
l V 25 yes. y i
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 44?3
464 1 CHAIRMAN APOSTOLAKIS: Not just the general?
2 MR. HOLAHAN: Yes.
! t
'- MR. WESSMAN: I think that's the motivation in 3
4 bringing the specific cases to you. If the essence of 5 your comment is bounded by the general, then of course I 6 think we would know how to proceed with the specifics.
7 I think one reason why things like the 8 acceptance criteria in the application specifics are 9 referenced back to the general is we know that's where the 10 hard spot is.
11 I think much of the mechanics of what we 12 described to you in IST is a pretty reasonable approach.
13 MR. HOLAHAN: I think it's important when the
(-
(_/ 14 comment process is going on, that some of the application 15 specific ones are also out there so that people understand 16 how it would apply to issues that they are interested in.
17 MEMBER SEALE: I agree. I think that's very 18 important.
19 CHAIRMAN APOSTOLAKIS: So that means then that 20 Commanche Peak, for example, can not implement what they 21 requested before 1998?
22 MR. HOLAHAN: No. It doesn't mean that. What 23 it means is that it will take a Commission decision to 24 allow that to happen.
im 25 CHAIRMAN APOSTOLAKIS:
(v) But this can happen NEAL R. GROSS COURT REPORTERS AND TPANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
a 465 1 even during the period of public comment? )
,s 2 MR. HOLAHAN: Yes. Yes.
I )
~~
3 CHAlRMAN APOSTOLAKIS: Okay.
4 MR. WESSMAN: From a practical standpoint, 5 when could Commanche Peak test that first valve in their 6 risk informed program, it could be well into 1997. Even 7 if we complete our work with the Commission and they say 8 go ahead in March, and an authorization is given to the 9 licensee, there's a planning horizon and a scheduling 10 horizon. It may be that at the fall outage is when some 11 of the first activity starts, for example.
32 CHAIRMAN APOSTOLAKIS: Yes, but I think the 13 important thing is to give them go ahead from the point of
\/ s 14 view --
15 MR. WESSMAN: When we are ready to give it to 16 him, absolutely.
17 MEMBER POWERS: George, I think they made an 18 important point, that they would want for the public 1
19 comment to have the applications available. We should put i 20 an asterisk by that and include that in our advise to the 21 Commission so that they don't get separated somewhere 22 along the way, at least through inattention on our part.
23 MR. HOLAHAN: It will broaden the scope of 24 people commenting on the applications and the risk-o
( ,) 25 informed things in general, I think, by putting 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
466 1 applications that they are interested in on the table
-~ 2 also. Otherwise, we are in danger of getting comments
( ')
3 from the PRA community. I think if we really are 4 interested in the cultural change, we need to draw other 5 people into the comment process.
6 MR. WESSMAN: And one final point with the 7 pilots that I think it's important to make. That is, we 8 have an expectation and we have conveyed this expectation 9 to them, that should the rules or the criteria change as a 10 result of the public comment process and our learning over 11 the course of 1997, that if they have an approved program, 12 we would ask that -- you know, we would expect them to 13 bring it into comport with the change of the guidelines.
f~'y i
) At the same time, of course, there's all this
\~/ 14 15 activity with the ASME. By late next year, we would hope 16 that ASME has approved code cases, that it's been approved 17 by them and through their Board of Nuclear Codes and 18 Standards, such that the staff can endorse them with 19 whatever qualifications in our final reg guide.
20 MEMBER FONTANA: George, you mentioned that 21 there's a lot of resistance out there on the part of some 22 licensees?
23 CHAIRMAN APOSTOLAKIS: Skepticism, not 24 resistance. That this is useful to them, and that the NRC
() 25 is serious about 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
467 1 MEMBER FONTANA: The skepticism isn't that 73 2 this approach has a potential for significant cost
%.)) 3 savings, but the NRC is just --
4 CHAIRMAN APOSTOLAKIS: That the NRC is willing 5 to do it. I believe that's where the skepticism -- Mike, 6 do you have any --
7 MR. GOLAY: No. I think you have described it 8 accurately.
9 CHAIRMAN APOSTOLAKIS: And we heard that also 10 at the PSA. John Garrick, and also the PSA conference.
11 MEMBER FONTANA: Also the way they handle 12 these teet cases will give a pretty strong signal.
13 CHAIRMAN APOSTOLAKIS: Yes. That's why I am
/ 't U 14 interested in the time here. If we publish something the 15 first quarter of 1997 that shows that we're serious, I i
16 think that will be a very good signal.
17 MR. WESSMAN: It has been speculated to me 18 that some of the licensees see far greater savings in the 19 area of risk informed ISI as opposed to risk informed IST 20 because the difficulties in access and this sort of thing 21 for the inspection and scaffolding and personnel exposures 22 and these sort of things, are much more significant than 23 some of the difficulties in the quarterly test of pumps 24 and valves.
/
V 25 On the other hand, licensees will learn how to 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 I
i
468 1 deal with the process in the IST arena, and then may be
,- 2 able to effectively extend the process to some of these t s
~'
3 other areas.
4 CHAIRMAN APOSTOLAKIS: And as I said, the 5 important thing is really for them to see that you guys 6 are really serious about it, and that you are willing to 7 take full advantage of the risk part, risk informed part.
8 Out of curiosity, how long did it take you to 9 prepare for this presentation?
10 MR. WESSMAN: I don't know. In various ways 11 or another, we have been working on this presentation and 12 other presentations and these documents of course for 13 months. To sit down and make the viewgraphs, we did them n,
$ )
\/ 14 in less than a week because we didn't have time from the 15 last activity.
16 CHAIRMAN APOSTOLAKIS: Wait a minute now.
17 Wait a minute. You have been working for months doing 18 your job, not preparing the presentation. ,
19 MR. WESSMAN: True. It's been a demanding i
20 time for all of the folks that you see here, as it has !
l l
21 been for you all. i 22 MR. HOLAHAN: I wouldn't give too much 23 sympathy to the staff on this subject. These meetings are i i
24 very useful to us, even if it's hard to prepare for. Both I
(~
( ,x) 25 the preparation and the meetings I think serve a useful 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
469 1 role.
, 2 MR. WESSMAN: Absolutely.
(,f
\
)
3 MEMBER SEALE: I'm sure your case is much more 4 convincing as a result of this exercise.
5 CHAIRMAN APOSTOLAKIS: And I take whatever 6 amount of sympathy I gave you, I take it back.
7 (Laughter.)
8 MR. HOLAHAN: Thank you.
9 CHAIRMAN APOSTOLAKIS: This has been very 10 informative. Thank you very much.
11 So we will reconvene at five minutes after 12 1:00, 13 (Whereupon, the foregoing matter went off the p\
'w) 1 f
14 record at 12:09 p.m. and went back on the j l
15 record at 1:05 p.m.)
l 16 17 ,
1 18 19 20 i
21 22 23 24
{} 25
(_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
470 I l
1 A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N 2 (1:34 p.m.)
7-i i
~~
3 CHAIRMAN APOSTOLAKIS: Okay. We are ready to 4 resume our meeting, and the next item is Item 13 on the 5 agenda, the draft regulatory guide for graded QA.
6 Mr. Woods and Mr. Gramm, are they here? Okay.
7 It's yours.
8 MR. GRAMM: Thank you very much, Mr. Chairman.
9 My name is Bob Gramm. I'm in the Quality 10 Assurance Section of NRR. Roy Woods, who is one of the 11 co-team leaders for the development of the regulatory 12 guidance, and Steve Dinsmore represents the PRA Branch.
13 There were three other members on the developmental team g-k/ 14 and one is in our company today as well.
15 I'd like to go over some brief background, if 16 I could, for graded quality assurance. This initiative 17 actually kicked off in the late 1993 timeframe, so we've 18 been going at it now for a number of years. We initially .
l 19 worked with NUMARC, now NEI, in trying to develop a l
20 consensus standard that could be used by the industry to 21 implement this methodology, and we have later participated i
22 with three volunteer plants -- Palo Verde, South Texas, ;
23 and Grand Gulf. I I
24 We have also been in front of you on two prior
()
t8 l 25 occasions, in the fall of '95 and the spring of '96, to 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
i 471 1
1 lay out the methodology. In early 1996, the team was
,\ 2 brought together to articulate in a regulatory guidance i
V; 3 document approaches to graded QA.
4 I'd like to point out that there are a few 5 features which distinguish us, in r..y mind, from the other 6 application activities. First of all, the regulations 7 governing QA enable a licensee to implement quality 8 controls commensurate with an item's importance to safety.
9 That is already recognized in the regulation. So what 10 we're doing here fully comports with 10 CFR 50, 11 Appendix B.
12 Additionally, there are regulations in place 13 which govern how a licensee processes changes to their QA
~- 14 program, and they have a limited amount of flexibility to 15 institute changes, so long as they do not constitute a 16 reduction in commitments. That is also embodied in the 17 regulations.
18 When we look at adjusting quality controls, 19 there is a lack of data at this current point in time j i
20 which would reflect how we reduce quality controls which 21 actually impact or degrade an item's performance. We 22 don't have a lot of quantitative data there. .
1 23 And finally, the PRA model is only a small 24 fraction -- approximately 10 to 20 percent -- of safety-i
(')
(. ,/ 25 related components. And we can go into this later as 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
l 472 1 well, g3 2 CHAIRMAN APOSTOLAKIS: Do you have data on a d 3 similar component that is used under two different 4 regulatory conditions? I mean, in one case you do have QA 5 requirements, the other you don't -- to see whether the 6 failure rates are different? Is there any evidence 7 regarding that, that the component under strict QA has, in 8 fact, a lower failure rate?
9 MR. GRAMM: We have inquired of each licensee, 10 if they could provide those separate databases, to 11 discriminate that. And to date we have not been given any 12 data as -- my understanding is it's basically lumped into 13 one large data set, and they have not attempted to
(,,)
14 distinguish as to how QA may impact the failure rate 15 information.
J 16 MEMBER SEALE: Do you have any experience as 17 to the -- or any idea as to the experience of people in 18 related industries? And, of course, there are obvious 19 examples and then not so obvious examples. We have all 20 heard about the counterfeit bolt problem that aircraft 21 people have had. Sometimes you wonder if the original 22 bolt hadn't been under QA then the bolt wouldn't cost as 23 much. And, therefore, there wouldn't be as much profit in 24 providing the counterfeit bolt.
r~xi l t
(,/ 25 (Laughter.) l NEAL R. GROSS !
I COURT REPORTERS AND TRANSSRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
473 1 And so you may not have the problem if you
,s 2 didn't have the QA program that priced it out of sight.
( )
3 You know, that's a scenario that's as good as any other 4 I've heard.
5 But is there -- I mean, in your discussions 6 with people in other related industries, do you have 7 anything?
8 MR. DINSMORE: I don't think that we've had 9 any discussions with related industries.
10 MEMBER SEALE: Okay.
11 MR. DINSMORE: We probably could investigate 12 this, but it would probably take a couple of years.
13 MEMBER SEALE: Well, okay. But, you know, f)
\/ '
it's something you might be concerned about, because it's 14 15 obvious this is not going to go away. 4 16 CHAIRMAN APOSTOLAKIS: So there are no plans, 17 then, to get any information from the plants to see 18 whether there is a real difference that QA makes?
19 MR. GRAMM: During the latter stages of the 20 discussion, we'll talk about the feedback process that we 21 envision here, where I think over time data will be 22 accumula.ted by the licensees.
23 The four points up here define what we view to 24 be since mid-1994 as the essential ingredients of a graded g)
(, 25 quality assurance program. The first is that it should 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
474 ,
1 provide a process to determine SSC safety significance in
, . -s 2 a reasonable and consistent manner. The second, that the
! \
3 application for QA controls should be based on both safety 4 function and safety significance, and those controls need 5 to provide adequate assurance that the item will still be 6 able to perform its intended safety function.
7 An effective root cause and corrective action 8 program is necessary, as well as an operational feedback 9 program, so it's dynamic in nature and that over time 10 information can be gathered from both the plant and the 11 industry in total to make adjustments in QA control 12 applications.
13 The regulatory guide is structured in the same !
i
/^Ni
\~) 14 format as the master reg. guide. And just as an overview, 15 it does follow the four steps in terms of defining the i
16 change. Steve will speak in detail about the engineering 17 evaluation, qualitative and quantitative assessments. The 18 QA controls are then adjusted, and these embody the 19 corrective action processes and the operational feedback.
20 So these four elements embody the essential ingredients 21 that we have envisioned for graded quality assurance.
1 22 And finally, the documentation that would I 23 either be maintained by the licensee or submitted is !
24 described in the latter section of the reg. guide. j i
("%
! 25 On page 3-1 of the reg. guide we talk about (J
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W. I (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 i l
475 i 1 the process in terms of defining what the QA changes is.
I 7s 2 Before a licensee can embark on this program, they need to
('~#)
3 have an understanding of their commitments, which may be 4 in the FSAR or QA program topical report, various 5 responses to generic communications, or enforcement items.
6 So there is a volume of information which needs to be 7 understood so that as departures are made from that it is 8 done in a controlled process.
9 The candidates for graded quality assurance 10 would include safety-related items. These are items which 11 are necessary to maintain the reactor coolant pressure 12 boundary, those which mitigate accident conditions and 13 provide the ability to shut the plant down and maintain it p
(/ 14 in a shutdown condition, and those items which limit off-15 site exposures in excess of Part 100 guidelines.
16 We have also articulated the non-safety 17 related equipment, which is modeled in the PRA, would be a 18 candidate for grading purposes, and this would be 19 providing augmented quality controls as well as augmented 20 quality controls for non-safety related SSCs which are 21 within the scope of the maintenance rule. These would be 22 items in emergency operating procedures and other items 23 described in that Rule 50.65.
24 The licensee would articulate at the high
(~x.
25 level what changes would be anticipated in the QA program
\
us)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C 20005-3701 (202) 234 4433
476 1 description and then evaluate existing risk studies for 7x 2 appropriate application within this environment.
- )
x_/
3 On page 4-1 of the reg. guide, we move into 4 the engineering evaluation process, and we structured this 5 in a manner where we consider the function performed by 6 the item to then determine either qualitatively or 7 quantitatively its relative safety significance. This is 8 much like the process that's currently employed to 9 determine whether an item is safety-related or not, and 10 then we considered various approaches which licensees have 11 been using at the three volunteer plant sites.
12 In some cases in the document we have avoided 13 providing prescriptive details, and we have articulated a
\- 14 set of questions or issues which would need to be 15 evaluated by the licensee. And this is in recognition 16 there may be more than one pathway, more than one 17 methodology, to achieve a satisfactory end result. So 18 we're providing the opportunity for some flexibility in 19 the implementation approaches, and Steve will go into that 20 further.
21 At this point, I'm going to turn it over to 22 Mr. Dinsmore, and he's going to get into details as to how 23 the engineering evaluation would be carried out.
24 MR. DINSMORE: All right. As Bob said, my
,7
( ,) 25 name is Stephen Dinsmore. I work in NRR in the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ASLAND AVE., N.W.
(202) 2344433 WASHINGTON. D C. 20005-3701 (202) 2344433
477 l 1 Probabilistic Risk Assessment Branch. And I have been
,q 2 reviewing the one submittal and the half submittal that we l
) 3 have on graded quality assurance and working with Bob on 4 these documents with Roy. I guess I'll start off by 5 discussing what we've got up here about risk importance.
6 Just to put things a little bit in j 7 perspective, as Bob said most of the stuff in QA is not in 8 the PRA. We have maybe 20- to 30,000 things called 9 safety-related components, which are those components 10 which are under Appendix B. From the 20- to 30,000, there 11 is -- well, in a PRA there might be 1- or 2,000 components 12 modeled, so we have a rather small subset of components 13 which are modeled in the PRA.
( )
14 As a result of that, we have to both -- or we 15 have to try to figure out how to include stuff --
16 systematically include stuff which is not in the PRA in 17 the whole process. Usually what the utilities are doing 18 is they're having -- they're developing working groups who 19 do all of the detail work and trying to figure out --
20 trying to classify and categorize components, and then 21 they go to something which we term an expert panel. There 22 is different terms for it. But they go to the expert 23 panel and the expert panel would either approve or change 24 what they've come up with.
7
'(j) 25 So this risk importance, the difference 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
478 1 between risk importance and high safety significance
, 2 classification, which is what the IST guys were talking i
~'
3 about this morning, is about five minutes. It's just the 4 difference between what is proposed by the working group 5 and what is accepted by the expert panel.
6 We use it just to keep things straight, since 7 we have expert panel using that judgment to make the final 8 decision. We also have working group members using that 9 judgment to suggest categorization. So that's where risk 10 important comes from.
11 So again, as Bob said, we concentrate on 12 system functions. We want to determine which functions 13 are important, and then we'll determine what components l
[ h is 14 are required to support those functions. So the first 15 step is just to define all operating and accident 1
16 functions performed by the systems. It's unclear how easy l 17 that is. Some people think that's difficult. Some people la don't think it's difficult. One of the utilities said it 19 was real easy to go. I'll give you Just --
l 20 CHAIRMAN APOSTOLAKIS: Which utilities have l 21 these, Steve?
22 MR. DINSMORE: Well, we have South Texas and 23 Grand Gulf.
l 24 MR. GRAMM- And Palo Verde.
,m,
( ,) 25 MR. DINSMORE: Palo Verde we haven't had much NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
479 1 to do with.
2 For example, a couple of system functions for 7..
( )
the chemical and volume control system, which is a nice 3
4 one because there's lots of functions, is to provide high 5 pressure injection from the IRWST to the RCS, provide high 6 pressure recirculation from RHR to RCS, provide high 7 pressure borated water, provide normal RCS makeup flow.
8 So there might -- I guess one of the plants came up with 9 about 150 functions for a specific system.
10 Not many of these are modeled in the PRA, but 11 again, we have the difficulty that most of the stuff we're 12 dealing with isn't modeled in the PRA and this is one of 13 the ways we're trying to deal with all that stuff.
p k_) 14 CHAIRMAN APOSTOLAKIS: The reason, of course, 15 it's not modeled is that it does not contribute to risk.
16 Is that correct?
17 MR. DINSMORE: Well, some of it is not modeled 18 because it's considered inherently reliable. Some of it 19 is not modeled because --
20 CHAIRMAN APOSTOLAKIS: Or there are passive --
21 MR. DINSMORE: Yes. A good example would be a 22 maintenance -- a manual valve in front of a pump, an 23 important pump. During operation that manual valve might 24 be open and locked open. So that's never going to be (CS']) 25 modeled.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1 1323 RHODE ISLAND AVE., N W. l (202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234 4433
480 1 Whether it's high safety significance or not 7x 2 is something we'll work our way down to.
l
\
'~'
)'
3 CHAIRMAN APOSTOLAKIS: Or they are in the 4 minimal cut sets that are screened out perhaps, and you 5 never see those?
6 MR. DINSMORE: Well, that's part of the use of 7 importance measures, to try to determine the importance of 8 stuff which is modeled in the PRA. We developed -- or we 9 discussed a process to deal with the limitations that you 10 must deal with when you're trying to use importance 11 measures, and one of those limitations is the truncation 12 problems.
13 It got very complicated, and that process is (3
+
) That process -- what we
's / 14 not reflected in the reg. guide. ;
i 15 essentially did was laid out the difficulties, and we l l
16 expect the utilities to be able to tell us how they l l
17 addressed each one of these difficulties. l l
1 18 CHAIRMAN APOSTOLAKIS: I think the question is l 19 really, what do we mean by "modeled in the PRA"? If a 20 particular component was not included at all because it !
l 21 was not relevant to what the PRA was trying to achieve, 22 then I think not modeled is an appropriate expression.
I 23 If, on the other hand, it was in one of the 24 minimal cut sets, which was then screened out, then again r x.
( ,) 25 I think it was modeled but perhaps it is not included in NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
481 1 the submittal, because you don't have access to the 2 screened out sequences, do you? But it's not fair, f3 i )
~
3 really, to say it was not modeled.
4 Now, there are three possibilities here. One 5 is it is part of the submittal. The other extreme is it 6 was never modeled. And then in between you have stuff 7 that are not in the -- stuff that is not in the submittal 8 but it has been modeled at some point and it was found to 9 be completely unimportant. I mean, which exactly do you 10 mean? What exactly do you mean?
11 MR. DINSMORE: Right. Okay. We recommended 12 the use of two importance measures -- the Fussel-Vesely 13 and the RAW. The RAW, if you do that calculation based on (3 )
i
\ ~-' 14 the logic models, if you get rid of the truncation 15 problems that importance measure should reflect what is 16 modeled.
17 CHAIRMAN APOSTOLAKIS: Okay. So that includes 18 possible sequences that will be screened out at some 19 point.
20 MR. DINSMORE: Right. One of the reasons that 21 the detailed description of how to deal with these 22 importance measures is not in this reg. guide is because 23 it requires you to do a lot of recalculations of your PRA 24 itself, not just working with the cut sets. So I guess
,/7
(_,) 25 the answer to your question is that we never worried about NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE. N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
482 1 what was not in the cut sets as opposed to what was, or we p 2 never worried about what was screened out, because -- when
\ >
%J 3 we were working with cut sets, because we weren't 4 intending to deal with cut seto.
5 MR. SHERRY: Also, I think that probably a G large class of components are implicitly modeled through 7 their failure being accounted for in initiating event 8 groups.
9 CHAIRMAN APOSTOLAKIS: That's correct.
10 MR. DINSMORE: Yes. That's another one on our 11 list of stuff which we have in the reg. guide.
12 MEMBER POWERS: When you delineate the 13 problems and say, "Here, you licensee, go deal with all of
- m.,
14 these problems," how do you get yourself in a position so 1
15 that you know that indeed he has dealt effectively with l 16 those problems?
17 MR. DINSMORE: Well, Bob was down at a plant l
18 yesterday. I guess we go to the plant and we try to l 19 develop reasonable assurance, which is a phrase which I 20 learned last year, that they are effectively dealing with 21 the problems, l l
22 We don't have specific yes/no criteria. We 23 would have to do this on kind of a case-by-case basis. I l
24 MEMBER POWERS: It seems to me that if you )
A 1 V 25 haven't tried to do it yourself you run the danger of .
l i
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
483 1 being too accepting or being too stringent.
,_ 2 MR. DINSMORE: Well, I think the NEI probably
, i i
r i z/ 3 thinks we're too stringent at this point in time. And too 4 accepting? I guess I'm not quite sure what you mean if 5 you haven't done it yourself. If you haven't done the 6 actual -- that type of work yourself or if you haven't 7 done the study yourself.
8 MEMBER POWERS: If you've never gone through 9 and done something -- for instance, in operational PRAs 10 the staff is in pretty good shape for deciding whether a 11 PRA for operations is of a good PRA. They have access to, 12 have been through, in many cases, part of the NUREG-1150 13 effort. They know every body blow and bell and whistle :
(~)
(/ 14 that had to be confronted and dealt with in those studieu, 1
l 15 and lived with it for a long time, l 16 They know how difficult it is, and so they 17 have some basis to decide whether what they're demanding 18 is just way tco much or if -- they can tell when somebody 19 is snowing them. Just because they listed the topic and 20 set some words under it, they know how to read it and see 21 what he is really communicating there.
22 MR. RUBIN: Dr. Powers, perhaps I could add a 23 little information to give you some feel for some of the 24 process that has been going on, after I put the microphone
,a 25 back together.
(_ )
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 2344 433
484 1 MEMBER POWERS: Do not break the microphones.
,_., 2 (Laughter.)
I )
3 MR. RUBIN: Mark Rubin from the PRA Branch in 4 NRR.
5 We're just working cur way through the graded 6 OA program, so we're still on the fairly steep part of the 7 learning curve. But we've had a number of activities 8 going on to try to get our hands around what would be a 9 good enough process -- I hate to use those words -- and 10 within a number of stages.
11 A lot of the issues Steve started to mention.
12 I know we'll get into it in more detail -- the truncation 13 effects, some vulnerabilities due to common mode, due to
,a
(- 14 areas where you particularly have lower confidence in your 15 knowledge, some of the human performance areas, some of i
16 the common mode effects. We have been investigating those 17 with some of the pilot plants to do some parametric 18 studies, see how some of the various assumptions within 19 the bounds of credibility affect the rank results.
20 We also are trying to get importance 21 information at the system and functional level because it 22 is easier to spot errors or omissions when you're dealing 23 with a higher level function than if you're looking at a 24 list of 8- or 10,000 components. I'll give you an example
,.m 25 from a plant visit.
(u ))
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
485 1 We are down at a plant looking at a program, a s 2 QA program, and they are working their way through the
( )
3 HVAC systems -- and went to the control room HVAC system 4 -- and they ranked it low. And we posed some questions, 5 we wondered why, and their answer was, "Well, it's modeled 6 in the PRA and has no quantitative importance."
7 We pursued it a little further, to what extent 8 it was modeled, and we were told, "Oh, it's just equipment 9 rack cooling, and there's large heat sinks if the cooling 10 is not in operation. We have several hours before the 11 equipment fails and plenty of time to recovery, easy ways 12 to cool the racks, perhaps opening doors." We all know 13 those stories, and so it's not important, i r~'y I l
e N/
)
14 Well, we went a little further. We asked them i 15 what would be the impact on the operators with an abnormal j 16 control room environment? It was something they hadn't 17 thought of, and we asked them if the basic human !
i 18 performance assumptions are impacted at all by, you know, 19 an individual's environment. Would they perform as well?
20 This was a plant in a hot area of the country, especially 21 when we were down there. We wondered if they could 22 perform as per the human performance assumption in the PRA 23 at 110 degrees with a particulate mask on, and they 24 thought perhaps not.
p
( ,) 25 That goes to the types of things we're trying NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
486 1 to get our hands around. We're trying to give tests of 7- 2 credibility to the decisions or the groupings that are t i
~' 3 made. Of course, we have to balance too onerous and not 4 investigating in enough detail.
5 But the analytical studies we're doing looking 6 at some sensitivities of where the rankings are, say, less 7 static, where credible changes in perhaps assumptions or 8 modeling could cause large swings in importances, those 9 are areas we want to look at very carefully. And I think 10 the gentleman up there will go into some more detail on 11 what process structurally we're trying to do.
12 MEMBER POWERS: Well, you've given me some 13 anecdotal accounts of what it takes to get up on the k/ 14 learning curve, and that's exactly what I'm asking about.
15 What does it take to get on this learning curve? Is that 16 the length and the breadth of the learning curve -- HVACs 17 and particulate matter and things like that? Or do you 18 get the length and the breadth of the learning curve from 19 the pilot studies?
20 MR. RUBIN: I think by the time we're through 21 with them we will. I don't believe we have final lists 22 yet. I know they're being --
23 MEMBER POWERS: How do you know that,,the pilot 24 studies give you the length and the breadth of the
( ,) 25 learning curve?
NEAL R. GROSS COURT REPORTERS AND TRANSCRlBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
487 1 MR. RUBIN: They may not. It's a limited
,- 2 number of plants, but the methodologies, trying to i
N._) determine the functional importances, trying to attribute 3
4 those to the level of components that are controlled under 5 the QA program, I think does give us some confidence in 6 the process.
7 MR. DINSMORE: The use of the system functions 8 I think -- I guess what you are trying to ask, is this how 9 we're sure that they've got the high safety significant 10 components selected? How do we review to ensure that they 11 have that selected? Is that what --
12 MEMBER POWERS: What I'm concerned about is 13 when you say, "Here are the problems," and you can't I )
V 14 follow it up by saying, "And here's an approach to those 15 problems that would be acceptable to the staff."
16 MR. DINSMORE: But the list of problems -- we 17 probably do have an approach which would be acceptable.
18 It's just that it would require perhaps more work than a 19 utility --
20 MEMBER POWERS: I mean, that's the problem.
21 MR. DINSMORE: But then if the utility might 22 come back and give us an equipment -- I can't remember the 23 terminology -- an equivalent process, then we would have 24 to determine whether we believe it's equivalent or not.
,m
, \
( ,/ 25 MR. RUBIN: There's likely not -- well, more 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
488 1 than likely, there is not one correct way to complete this 2 process. The maintenance rule used a roughly equivalent T
3 approach to start with using the importance information 4 from the PRA, but then the expert panel filled in some of 5 the holes.
6 This process would fill in more holes, because 7 it goes into somewhat more detail on what areas you may be 8 somewhat vulnerable to for relative rankings. You can use 9 a more rigorous analytical approach to develop system 10 importance and component importance. You could increase 11 the scope of the modeling in the PRA to get more 12 information directly out of it for risk importance -- that 13 would take more work -- or through a more traditional
,w
\-) 14 engineering process, using the so-called experts in system 15 function plant performance, attempt to attribute an 16 equivalent perspective on relative safety.
17 Most likely, there will be a combination of 18 3dditional analytical work but still a strong use of the 19 expert panel. How the expert panel makes that integration 20 process is key, and that's an area where we're trying to 21 develop additional guidance. There is some in that area 22 provided already in the draft reg. guide.
23 MEMBER POWERS: You're making yourself very 24 vulnerable on an expert panel that is not yours.
()
(
25 MR. RUBIN: Well, the expert panel is key to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
489 1 this process if it's used. And to guide them, a way to 7_
2 evaluate them is going to be very important to this
On some of the early visits we saw a disparity 3 process.
4 between the rigor the expert panels used. We saw some 5 that just got together, discussed the issue, and 6 apparently made some decisions on what was more or less 7 important, without not a lot of documentation or criteria 8 on how they did it.
9 We also saw expert panels that created a set 10 of criteria -- traditional engineering deterministic 11 criteria -- on how to attribute high or low importance to 12 things that weren't modeled in their PRA or additional 13 tests beyond the risk input and output from the PRA. And
,rx C h
\_/ 14 so some use a very structured approach with additional 15 criteria, such as something will be high if it supports an 16 important human action from the PRA. That can be either 17 switchgear or a control or instrumentation.
18 So there was a wide spectrum of sophistication 19 or guidance documentation, and we're trying to arrive at 20 what an adequate level of those attributes are. And the 21 first cut of it is in the general reg. guide right now.
22 MR. HOLAHAN: Can I add something to the 23 discussion? I look at the problem from a little bit 24 different point of view, which is I'm not convinced that
,rx
() 25 we're going to come out with the exactly perfect list, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
490 1 which is exactly categorized according to the best
, 2 possible risk insights.
\ /
'~' I go back to a word you used, which is 3
4 vulnerable or vulnerability. And in my mind, this process 5 is acceptable, not because we have so many years of 6 experience and we're going to make exactly the right, you 7 know, perfect decisions about prioritizations and 8 categorizations, is that I think the process is okay so 9 long as you limit the size of the changes you make, and 10 you have a monitoring program in place so that you don't 11 allow your changes to be, you know, much more extreme than 12 you thought you were making.
13 And from what we've seen of the graded QA
(~)
V' 14 proposals, I think they are relatively modest, and I think 15 they have been relatively conservative in the choice of 16 putting equipment in high and low categories. So if you 17 don't have lots of experience -- yes, well, maybe we're 18 not doing it perfectly, and maybe we can't quite tell if 19 the licensees are doing it perfectly. But I don't think 20 there is a big vulnerability because I think there is 21 quite a lot of conservatism in how much is being kept in 22 the current categories.
23 CHAIRMAN APOSTOLAKIS: Okay.
24 MR. DINSMORE: Okay. We do try to put some 7S
() 25 process to work to support our conclusions. It wouldn'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
i 491 1 be just that we would go and say that this is a good list.
I s 2 The first step was just to identify all of the functions, l
\
\' '/
3 and there might be, as I said, a couple of hundred 4 functions for a system. And the second one I guess is 5 what is the real one, which is to categorize the risk 6 importance of a system's functions, either guided by the 7 PRA and the use of importance measure calculations or 8 directly by a working group.
9 I guess I have some more anecdotes about how 10 you can do either of those. I can skip those if you've 11 got enough anecdotes. My own judgment -- I'll give you 12 one. There was -- one of the licensees figured out that 13 operator errors were important. They used the PRA to o
I l
/ 14 determine that this operator action was an important 15 action, had a high impact on the safety of the plant.
16 And then they're working group followed it 17 down. It might seem trivial, this process, but this is 18 what they're supposed to do and this is what we'd like to 19 see them document. The operator error is important, the 20 operator needs instrumentation to know when to do this 21 thing, the instrument needs a vital bus, the vital bus 22 needs the inverter, because their inverters weren't 23 modeled in their PRA. Sometimes they are, sometimes they i
24 aren't, and these guys didn't have it.
( ,/ 25 So that's kind of the stepping down through I
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. 1 WASHINGTON. D C. 20005-3701 I (202) 234-4433 (202) 234 4433
492 1 process that we're hoping will help ensure that we're
,f S 2 picking up everything which is really safety significant.
() 3 I guess I've got -- the third step is to 4 identify the SSCs that support the high risk important 5 functions. We're doing this partly because a lot of the 6 stuff is not modeled. For example, this manual valve 7 again. That wouldn't be modeled, and we'd like them to 8 very clearly lay out specifically what equipment are they 9 depending on for this function to succeed, with no regard 10 to what's in the PRA, exactly what they have to have. And 11 that's another way we're trying to make sure that we're 12 getting appropriate coverage.
13 And then once they've got that list, step 3 is
( I
'- ' 14 pretty easy, probably the easiest step in there. Once 15 they have that list then they have to categorize the 16 actual SSCs themselves because the QA controls are applied 17 mostly to the components themselves, so that there are 18 other controls.
19 And then we have this guiding principle, which 20 I guess we're still under discussion -- which is the SSC 21 operating modes required to support this system function 22 are high risk important, which means that that manual 23 valve -- although I think the utilities would prefer to 24 put that manual valve in a low risk important place
()
(_) 25 because it is a manual valve and it's locked open, and all NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 2344433
493 1 that kind of stuff. We're kind of pursuing the thought
. 2 that at least the remaining open function of that valve i
\
)
/
'"' should be under high QA.
3 4 I don't know if -- Bob is sitting quietly, so 5 he has seen this before. And that's essentially it.
6 That's the end of the process.
7 MEMBER BARTON: Aren't we making this overly 8 complicated? Current systems and components have a QA 9 classification based on importance to safety, safety 10 levels, or whatever, and then you've got to look at the 11 supporting systems needed for safe shutdown, and there is 12 currently some kind of program.
13 The licensees come in and say, "I want to p'
k_/' 14 modify that program," and what they're going to do is come l
15 in and ask for, you know, less regulation, less 1
16 restriction, take stuff out of the program. Why can't we i 1
17 just evaluate what's in the program and the rationale for j 18 eliminating the current requirements that they've got on 19 their systems and components? It seems to me that we're 20 going through an awful lot of complications here to get to 21 a graded QA approach. What am I missing? ,
l 22 MR. DINSMORE: Well, one of the licensees said i
23 that they determine safety significance by taking every )
24 system modeled in the PRA which has a high safety r~s ,
k ,) 25 significant component as indicated by the importance l
NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS ,
1323 RHoDE ISLAND AVE., N W. l (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 .
1 l
494 1 measures, and every component in that system modeled in
,s 2 the PRA is high safety significant. And then they did
. T
/ 3 some deterministic considerations on stuff that is left 4 over.
5 And I guess we looked at that and we weren't 6 completely convinced that they had caught everything that 7 we wanted them to catch, or that we thought might should 8 have been caught. So the answer to your question is kind 9 of that's what we kind of have to do. I mean, they don't 10 have to follow the reg. guide. They can come in and 11 suggest anything or any methodology which they'd like, and 12 we have to evaluate it.
13 But this gives us or at least gives the p
'w/ 14 licensee an indication of what would be easier to convince 15 us to accept. I guess what you just asked was if they 16 could just come in with some type of proposal of their 17 own, and they can still do that, I believe, unless --
18 MR. WOODS: I think what he asked was whether 19 this should be primarily a qualitative process with some 20 insights, whatever you can get from the PRA to add to it, 21 or whether it should be a PRA based or PRA informed 22 process with a little bit of qualitative stuff it's 23 gotten. It's a very good question. We have had many, 24 many discussions throughout the development of this as to rN
! ) 25 which way it would be most appropriate to do it.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 2344433
495 1 And you are largely driven by the fact that g .s 2 PRA is of marginal value because things aren't modeled and i \
'- 3 because, really, largely what you can get your hands on is 4 anecdotal. It's not -- Dr. Seale's question was an 5 anecdote basically about can you get data, and he -- yes, 6 you can get data, but you can't really get data that you 7 can meaningfully use for a large body of the stuff. You 8 can get anecdotes, and you can know, yes, that's important 9 and that's not. But you've got thousands of things and 10 you'll never get through that way.
11 So should you approach it that way or should 12 you approach it more in a qualitative fashion? And that 13 has been the subject of much debate in the staff. I don't A
- 14 have the answer to it. I don't think we have the answer 15 to it, except that, well, pretty much by force you use the 16 PRA only for what you can use it for and the rest of it, 17 if you want to have this kind of program, does have to be 18 qualitative or traditional or whatever.
19 MR. DINSMORE: Okay. Then I guess I'll turn 20 it back over to Bob.
21 MR. GRAMM: The area that I'd like to address 22 now is how the reg. guide deals with two of the safety 23 principles of risk-informed decisionmaking, and that is 24 maintenance of both defense-in-depth and safety margins. 1 fy
) 25 And we've thought through this and articulated on NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
1 496 1 page 4-10 of the document some factors that I think are 73 2 important to remember.
( )
'~
3 We're not changing plant configuration in this 4 process. We're not eliminating any components from the 5 plant boundaries, so we're not degrading the boundaries by 6 changing -- doing design modifications. Those would be 7 controlled under existing design change processes which 8 are outside of graded QA.
9 We're not changing any technical requirements 10 of the components at hand. If a pump has a certain 11 specificity for performance parameters, those remain 12 intact through this entire process. Steve articulated 13 that the factors of consideration, qualitatively, also
,a
'ss'1 14 include new common cause failures, and these factors have 1
15 to be addressed. And we can't identify a formula to do 16 it, but it needs to be thought through by technically l l
17 cognizant personnel of the plant staff.
18 The reduced QA controls on safety-related i
19 items will continue to meet Appendix B. What we're 20 advocating is that there would be a program with reduced I 21 scope and rigor, but it would still meet the 18 criterion 22 and we would be ensured that there is an adequate level of i 23 confidence that the end use items and low safety 24 significant applications can still perform their intended
,n
(_) 25 safety function.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
497 1 So what we have is Appendix B is a very high 2 level description of a quality program. And over the past r3 3 20 years there is a body of information in regulatory 4 guides and ANSI standards -- I could bring a notebook in 5 -- which articulates the staff's perspective on what it 6 takes to meet Appendix B. What we're now saying is if a 7 licensee goes in and studies that documentation in an 8 intelligent manner, there's a way to revise those 9 implementing quality controls in a manner that they still 10 meet Appendix B, yet the prescriptive details in a 11 standard need not be complied with.
12 For example, lead auditor need not have led 13 five audits prior to becoming an auditor to examine low
\
-\ '/ 14 safety significant items. We think that's too 15 prescriptive in this particular case, and we'd be willing 16 to grant some flexibility. So it's the soft controls 17 application that we' re saying is wo- v of change.
18 However, we're going to maintain a minimum floor to ensure 19 that technical requirements are still met by the items 20 before installation in the plant and that they can still 21 perform their design function.
22 We'll have a performance and monitoring 23 program, as Gary alluded to, in place that will assure 24 that over time there aren't catastrophic situations that
/~h
\ ,/ 25 are popping up. The licensee will still study precursor NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
498 1 events to understand apparent causes of those failures. ,
1
-s 2 And if they see a failure occur in a low safety
(
~'
3 significant system that has applicability to a high, 4 they'll have to understand that interrelationship and then 5 take appropriate corrective actions, and to compensate for 6 QA controls.
7 So we see this as a dynamic, a self-adjusting 8 system that QA controls could either be backed off or 9 augmented over time in recognition of operational 10 considerations. So given about six or seven different 11 qualitative assessment criteria, we think generally the 12 program will maintain defense-in-depth and margins of 13 safety and will preserve that through application of this
(-) 14 particular --
15 CHAIRMAN APOSTOLAKIS: Is quality assurance 16 part of defense-in-depth?
17 MEMBER POWERS: It depends on your view of 18 defense-in-depth.
19 CHAIRMAN APOSTOLAKIS: But I'd like to know 20 their view.
21 MR. WOODS: I don't think you can argue that 22 it isn't, but I don't know how much -- it would depend a 23 great deal on what particular component you're talking 24 about, I guess. Pressure vessel, yes, we want to make
, ~\
) 25 sure it's built right and that welds have been X-rayed, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
499 1 and whatever.
7x 2 CHAIRMAN APOSTOLAKIS: So the fact that you i ;
\/ are not changing the plant configuration cannot be the 3
I 4 only argument -- to claim that defense -- I mean, you have 5 satisfied defense-in-depth.
6 MR. DINSMORE: That's right. One other --
7 there is probably dozens of arguments for and against it.
8 But another thing which we could consider is supporting 9 our defense-in-depth is that we work at the function level 10 and not the train level. So we're keeping a higher level 11 type of control or significance determination.
12 MR. GRAMM: What we're also advocating is for 13 those items that are high safety significant, the current
(~
/ 14 QA program will remain intact. And so, therefore, that 15 element of defense-in-depth will remain constant over time 16 for whatever components have been shown to be high safety 17 significant.
18 MR. DINSMORE: And then the last two points I 19 have been elected to do here -- as Bob said earlier, we 20 don't really try to calculate a delta CDF or anything like 21 that. We try to do this classification in a reasonable 22 manner, and then to provide some type of real feedback or 23 real -- some type of useable feedback.
24 We would like to set up -- we were asking the
) 25 utilities to set up and describe to us how they're going 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
500 !
i 1 to quantitatively follow the change or any unexpected j
~ 2 changes which might have come from this QA process. And l (s) 3 we're aware that it's awful difficult to come up with a 4 criteria and a methodology to prove that your 10 4 has 5 stayed 104 6 But on the other hand, it should be possible 7 to see -- to set up a procedure so that if something 8 really starts to go wrong that you might be able to find 9 it. And it's a little bit of the question -- maybe we 10 could determine how much really go wrong is. Does it mean 11 a factor of 10 things are getting worse, or a factor of 12 100? We're not all sure about that yet. That's why those 13 are two very small bullets.
- r EJ 14 MR. WOODS
- I think kind of a summary of why 15 this would be okay is we plan to take very small steps 16 using only equipment that we believe, based on qualitative 17 and quantitative information, whatever we have available, 18 is not terribly safety significant. And in case we are 19 wrong, we have a very strong monitoring and feedback 20 program to see if we have declassified something that we 21 shouldn't have. That's where the strength of the program 22 is, in those steps. It's not in the extensive use of PRA, 23 I think. This is my opinion.
1 24 MR. SHERRY: With what do you compare the g)
. ,) 25 performance of a component which you have relaxed its NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON. D.C. 20005-3701 (202) 234 4433
501 1 quality assurance requirements?
2 MR. WOODS: That's a very good question. You
,S e 4
%~'/ compare it to whatever you have available, which might not 3
4 be all that much, because it is a low safety significant 5 piece of equipment. The utility probably hasn't kept 6 detailed records on that. You just look and see if what 7 you're seeing, which is hopefully nothing, no failures, 8 but if you start to see failures, you want to find out why 9 it's failing, find out if it's related to the QA decrease 10 that you've made, and found out if it has application to 11 something in the plant, maybe the same type of equipment 12 that's in the safety system. You want to make sure that 13 you still have the controls on that.
4
\/~ 14 MR. SHERRY: But even if you have data on a 15 population of components, it may be a population which is 16 mixed. It may be some which have IQA requirements, some 17 which have less.
18 MR. WOODS: It's a very difficult --
19 MR. SHERRY: And you may not know what the mix 20 is and what the --
21 MR. WOODS: It's a very difficult problem.
22 MR. SHERRY: -- contribution from each is.
23 MR. WPODS: Also', you could argue that it 24 would take you hundreds of years to develop statistical n)
(_ 25 significant data. We know that. That's a problem.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
502 1 But you look and see if you think you see a l
ex 2 problem developing based on whatever tools you have I
\
- s 3 available. I don't know how it stands here.
4 MR. DINSMORE: We have faith that the 5 utilities will figure it out and tell us a reasonable 6 method.
7 MR. HOLAHAN: Because of a regulation called 8 maintenance rule, for which they have to do something 9 similar on this equipment anyway.
10 MEMBER POWERS: It sounds like this becomes 11 almost the quintessence of a performance-based system.
12 You identify components that you think have a low safety 13 significance. If they had a high safety significance, e,_'N i !
'/ 14 they would undoubtedly be modeled in a PRA. Maybe or 15 maybe not.
16 And then you allow some reduction in the QA 17 requirements on these things, and then you just look and 18 see if they start failing.
19 MR. WOODS: The added step is you allow some 20 reduction in the QA for a very small number of these 21 things at a time. You take it very incrementally and 22 follow to see if you see any problem. And if you don't 23 see any problem -- you surely don't have a very large 24 problem or you would see it, and you take it in small (g _) 25 steps on equipment that you believe, based on your best NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
503 1 information, is not terribly safety significant.
,- s 2 MEMBER POWERS: So this is more performance
( )
I mean, it's informed in the 3 based than risk informed.
4 sense that things that are in the risk based just don't l 1
5 show up in the relief process.
6 MR. WOODS: It has to be, just by the nature 7 of the problem. There's about that much quantitative 8 information available. This one looks like it might work, 9 actually.
10 (Laughter.)
11 MR. GRAMM: At the end of the engineering 12 evaluation, there's a process we call integrated 13 assessment, and this can go by a broad range of names --
n
)
\/ 14 expert panel processing or whatever. But the purpose here 15 is to integrate information that has become available 16 through both qualitative, quantitative, traditional 17 engineering, probabilistic-based assessments, and to 18 process that information by a group of plant individuals 19 who make sense out of it, and to determine the final 20 safety significance categorization.
21 And here we're making a distinction. What 22 comes into it as risk importance information the panel 23 then processes it and determines if the end use -- the 24 safety categorization, be it high, low, or medium, which
( ,) 25 would then drive the QA control application.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
504 1 The collective judgment would be used of the 2 panel, and then they would consider the assumptions and
(,,, I
- limitations and the various processes that preceded the 3
4 step, and they would consider also maintenance of defense-5 in-depth and safety margin. So this is where the burden 6 would be placed upon those folks to -- all else having 7 failed, to make sure that before they made a decision that 8 it made technically good sense to carry forward in the 9 plant applications.
10 We've had the opportunity, as Mark Rubin 11 indicated, to observe some expert panels in progress, and 12 there are a variety of ways of doing this. And the 13 structure varies from licensee to licensee. But the (3
! )
k/ 14 common factor is it's multi-disciplinary, so we're seeing 15 operations folks, SRO types. The PRA staff is 16 represented, so that they can define what the results mean 17 from a quantitative sense. Plant engineering is there to 18 ensure integrity of design throughout this process. And 19 collectively, these folks then assimilate the information 20 and determine safety significance categorization.
21 Then we move on into the end use application, 22 ultimately would be to adjust the grading of quality 23 programs, and this would be programs for both Appendix B, 24 for safety-related items, as well as for some non-safety
,o
( ,
) 25 items found to be high risk importance or high safety 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
j
505 1 significance, where an augmentation of licensee practices
, -s, 2 would occur in recognition of their higher safety
(
~~'
)
3 significance.
4 The changes, as I indicated before, will still 5 need to conform to Appendix B for their safety-related QA 6 program. It will not be possible to downgrade items to a 7 non-safety related status based merely on PRA. That's an 8 engineering determination. The criteria is articulated in 9 the regulations, and that would be a separate process if 10 they wanted to remove an item completely from the Q list.
11 Quality assurance program changes would be 12 processed and reviewed by the staff if necessary under 13 50. 54 (a) , and other design changes would be processed to o \
t
\-) 14 be under 50.59, or reliefs or exemption requests as 15 necessary. So this is not serving as a method to truncato 16 any other chain process or regulation that is in place.
17 We're fully comporting with those.
18 We have identified a few areas where we think 19 there is opportunity here to adjust QA controls, and the 20 one area of commonality amongst the volunteer plants is 21 procurement. And this would be changing the manner in 22 which they buy items. And rather than going to an 23 Appendix B vendor, there is a move afoot to -- and this 24 is, again, comporting to Part 21, to buy from a commercial o
) 25 grade item supplier and then to dedicate it through 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
506 1 performing appropriate inspections and tests or analysis 2 to demonstrate that it, in fact, is capable of performing 73 )
\ "' -
3 its design function.
4 So the licensee is assuming that burden since 5 the vendor did not have a quality program, but there are 6 technical ways to verify that component will meets its end 7 use application.
8 MR. GOLAY: Can I ask a question on that?
9 MR. GRAMM: Yes.
10 MR. GOLAY: Will it be -- under the way you're 11 approaching it, is it the burden of the licensee to come 12 up with the -- what are effectively revised specs. and to 13 show conformity to those specs., subject to your own i ')
V 14 approval?
i 15 MR. GRAMM: In this case, it would not be )
16 subject to the NRC approval. However, it would be done )
l 17 under the licensee's quality assurance program. They l
18 would define what are called critical characteristics of 19 the item, again based upon the safety function. And then 20 from those critical characteristics would flow various 21 types of verification techniques. It could be hardness 22 testing, it could be dimensional measurements, a wide 23 variety of electrical-type testing. So whatever technique 24 would be appropriate to verify that critical design
,n 1 C/ 25 characteristic would be employed by the licensee.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
507 1 MR. GOLAY: But is it the burden of the p 2 licensee to develop these requirements and then show i i 3 conformance? Or is this something which the NRC is 4 effectively taking some lead in doing?
5 MR. GRAMM: The industry has had developmental 6 activities over the past 10 years to identify typical 7 critical characteristics. So they have industry standards 8 based upon component type, which define typical critical 9 characteristics. Those are then used by the licensee 10 engineering staff and compared to their own site-specific 11 requirements and adjusted accordingly.
12 If we were involved in an inspection activity, 13 those decisions are always subject to the NRC purview, ,
l
! h U 14 particularly if one of these items fails. There would l 1
15 then be a reason for an inspection, to go in and look at i 16 what went wrong in the process. But there is not a burden ,
I 17 on the licensee to submit these for -- all of these types l l
18 of evaluations for staff approval. ,
19 MR. GOLAY: Your answer seems to be that they 20 are actually doing this collaboratively and by conforming !
21 to whatever is the consensus standard that within the 22 industry that is acceptable.
23 MR. GRAMM: They, in fact, can do this whether 24 it's high or low safety significant, so long as they can
! ) 25 verify the critical characteristics. This is fully 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
508 ]
1 available even in today's environment.
1
~,, 2 Under graded QA, what would then happen is the
'~] 3 verification -- a sampling type approach -- might be 4 employed. Rather than verifying each and every item that l 1
5 is delivered, a sampling technique could be devised by the 6 licensee based upon typical -- a MIL standard or an 7 appropriate sampling methodology. And so there would be 8 some flexibility in the degree of verification that would 9 be employed.
10 Other types of changes we envision are 11 elimination of inspection hold points, which are not 12 required by code, and performing a monitoring or a 13 surveillance type overview. The level of detail and t ;
's /
14 records and documentation may be reduced. However, there 15 is some minimum floor that would be necessary to support 16 subsequent corrective action activities and demonstration 17 that qualified people perform the activities.
18 To date, the licensee progress has only I 19 think explored the tip of the iceberg here. For example, 20 we were at Grand Gulf yesterday and they have only 21 processed less than a handful of procurement packages 22 under this new program. So we're not talking about a 23 wholesale change across the site. They're approaching it 24 in a very careful conservative manner, as Roy said
,m i
) 25 earlier.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON. D C. 20005-3701 (202) 234 4433
509 1 And I think they're gaining a common
,_ 2 understanding on their staff and expanding outward from
. 1 3 that. As the staff arrives at confidence and decisions, 4 then the licensee will then explore other areas to employ 5 these types of techniques.
6 Some essential ingredients here at the back 7 end of the process are the operational feedback, and this 8 is already in place so we're just taking advantage of 9 current site practices.
10 MR. MARKLEY: Bob, one quick question. The 11 audits you were talking about relaxing, are those the 12 tech. spec. audits or are those other program audits?
13 MR. GRAMM: What would happen here, Mike, I x/ 14 would envision that the audit scope would be skewed 15 towards high safety significant activity. So that if the 16 audit team had a choice between observing different 17 activities in the field, they would selectively examine 18 design associated with a high safety significant activity 19 rather than a low.
20 So it would be an allocation of the resources 21 in recognition of higher safety significance in some 22 areas. If the tech. spec. audits cannot be eliminated 23 without processing a tech. spec. amendment, that would 24 have to be carried forward. But I think what would be i
/~ s i s
) 25 examined during the course of those audits would be skewed 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
510 1 or biased based upon the safety significance of the
~. 2 activity.
I \
3 For the operational feedback, there would be a 4 consideration of industry and NRC generic communications 5 to make adjustments over time. Plant modifications are 6 made, which may have a bearing on the PRA model or the 7 decisions that were made of a qualitative standpoint, they 8 would have to feed back in the process.
9 As Steve indicated, there would be some 10 performance monitoring techniques employed, so that if 11 some value were exceeded that would trigger an 12 investigation and then a reexamination of quality controls 13 or the safety significance determinations.
(^)
'K / #
14 And corrective actions would include i
15 considerations for cross-system boundaries, common l 16 mode / common cause type failures, and interrelationships !
l 17 between high and low system applications. So we're 18 looking at a process here which will confirm over time the 19 decisions where appropriate, and where necessary make 20 adjustments in quality controls.
21 MEMBER POWERS: The operational feedback 22 you're talking about there are things that you placed in a 23 low safety impact category that were done erroneously. It 24 should be moved up.
( ,) 25 MR. GRAMM: Perhaps either be moved up or an NEAL R. GROSS COURT REPORTERS AND TRANSCRir,ERS 1323 RHODE ISLAND AVE., N V. ,
(202) 234-4433 WASHINGTON, D C. 20005-37's1 (202) 234-4433
511 1 adjustment made in the types of quality control practices 7s 2 that were employed. For example, if the wrong critical i
\
]
3 characteristics were not fully examined, that could then 4 be adjusted for future procurement activity. So we would 5 see either the safety significance could be revised up, 6 down, or down to the upper category, or the QA controls 7 could be either reduced, relaxed, or augmented in 8 recognition of what is happening in --
9 MEMBER POWERS: So there is a feedback that 10 leads to relaxation among those that are in the low safety 11 impact category?
12 MR. DINSMORE: I'd put in a little PRA twist 13 to that, in that if you take all of these low safety n) k/ 14 significant things and fail them at the same time, you're 15 not going to be in very good shape. l l
16 MEMBER POWERS: I think I understand that.
17 MR. DINSMORE: So what we want to make sure is 18 that things don't get along to the point -- not that !
19 they're all failed, of course. That's ridiculous. But we 20 wanted to control any potential large scale degradation. l So if it's -- even though it's low safety significance, we I 21
- 22 don't want to just forget it. There has to be some type 23 of -- or we feel that there should be some type of control 24 to say, "Well, things are really starting to go in an
( ,/ 25 unexpected direction." So we have to do whatever Bob just NEAL R. GROSS COURi REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
i 512 0
1 decided to do.
2 MEMBER POWERS: The operational feedback that 7-3 says those things that I maintained stringent controls on, 4 but shouldn't have in reality, that is really in the 5 bailiwick of the licensee to come in with a new list to 6 you, a new set of arguments.
7 MR. DINSMORE: Well, it would just be an 8 expansion of the classification process which they --
9 MEMBER POWERS: Yes, you'd have to do it 10 again. I mean, it would be up to him.
11 MR. DINSMORE: Yes. But since we don't really 12 consider inherent reliability to be sufficient cause alone
,_s 13 to be put in low safety significance, then I guess I'm not I
14 sure, l
15 MEMBER POWERS: What you're saying is that you !
l 1
I 16 think you do such a good classification at the first part 17 that that's just not a possibility -- a rare --
I 18 MR. DINSMORE: Could you repeat that? I 19 didn't really --
20 MEMBER POWERS: You do such a good 21 classification, such a careful classification, and you are 22 spending an enormous amount of time on this classification 23 here, that operational feedback on that classification is 24 rarely going to cause any changes.
(-- x
) 25 MR. WOODS: Well, that's the hope, of course.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 2344433
I 513 1 But like we said, we want to do this in small steps. He 2 might classify a lot of things in low, where it actually 7s doesn't get around to actually making the change in the QA
\
3 4 program all at once. Certainly, we would like to keep it 5 small. I'm not quite sure what the question was other 6 than that.
7 MEMBER SEALE: He's saying exactly what you --
8 that's what he is referring to. You get a list that no 9 one is so presumptuous as to say is all inclusive. You 10 get two, three, five, whatever years of experience that 11 tells you, " Hey, this works. Now let's make the list 12 longer." Isn't that what you're talking about? I mean, 13 and what is it? Two? Three? Five? Or what --
- 14 MR. WOODS: But what I'm saying is you may j 15 already have a longer list, which he hasn't actually used 16 all of it yet.
17 MEMBER SEALE: I understand.
18 MR. WOODS: That's kind of a -- j 19 MEMBER SEALE: Yes.
20 MR. WOODS: -- nuance. l 1
l 21 MEMBER SEALE: But how do you add to that? I 22 mean, but you're going to approve a certain list.
23 MR. DINSMORE: I don't think so.
24 MR. WOODS: I don't think we have to approve
() 25 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
l 514 1 MR. DINSMORE: We approve a -- my 2 understanding is we approve a process.
( )
3 MEMBER SEALE: You approve a process.
4 MR. DINSMORE: And that they can go ahead in 5 six months and change stuff on the list, which I'm not 6 sure that we're entirely comfortable with. But at least I 7 personally am not. But what we'd approve is how they 8 generate the list.
9 MEMBER SEALE: Okay. The process is what you 10 do.
11 MR. JONES: In fact, Bob, as I understand it, 12 and the experts can correct me like they normally do, in 13 the graded QA area we've actually found ourselves in a
,3
( i V 14 mixed bag as to what we review and don't review. There 15 are licensees who believe they have the authority to do 16 this under their license right now without coming to us.
17 So we would pick them up through inspection.
18 And what they are basically playing around 19 with, if you wish, is the phraseology in Appendix B, which 20 is you do QA commensurate with safety.
21 MEMBER SEALE: Okay.
22 MR. JONES: So then there are other licensees 23 -- and these include the three pilots. We've got the 24 collection and the three pilots we have. There are other
,~
25 licensees who believe we need to actually approve 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
! S15 1 change.
2 So we're kind of in a mixed bag here. What 7~
( !
3 we're trying to do now with the reg. guide and SRP --
4 excuse me -- with the reg. guide -- there is no 5 corresponding SRP because of this issue of the inspection 6 mode -- is to try to define in a sense what is a good 7 process for figuring out what commensurate to safety is, 8 and then what is an acceptable process for reducing QA 9 given that decision.
10 So you see a lot of PRA element in a semi-11 quantitative / qualitative way being used to rank or 12 categorize the components. That tells you which are 13 acceptable to than be reduced commensurate with safety at t
A a
\. / 14 reduced QA. And then, as was discussed by Bob Gramm, then 15 he would define and has defined in the reg. guide some 16 elements of that of where you could reduce the QA process 17 we think and still be meeting the intent -- actually, not 18 the intent -- meeting the regulation, because it would 19 then be consistent or commensurate with safety.
20 And then we will either go out and review 21 somebody's program to it, and then bless the program, 22 which is what we do with QA today, and then monitor it 23 through inspections. And then licensees that currently do l l
24 -- that think they have that authority now, what we --
[^ ~\ l
() 25 where we think we're heading with this is once we get 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
516 1 reg. guide on the street, get comments, continue to work g3 2 with the pilots to refine this, we will then solidify the
! )
3 reg. guide, go final, and in the interim also develop o
4 inspection procedures that we will then go out and test on 5 other licensees. So that's kind of the process we're 6 working with. I hope that's a little helpful.
7 MEMBER SEALE: Yes, it is. It is.
8 MR. GRAMM: I'd like to respond to the 9 question how more items may go from high to low as well.
10 We're finding the expert panels are contemplated to be a 11 standing panel. For a maintenance rule there is a 12 requirement for a periodic assessment, and they envision 13 doing that in graded QA as well, so that one or two years r~s i !
C' 14 down the road, given an accumulation of plant inf orn.o r t on, 15 it would go back through the same type of decisionmaking 16 process that was originally employed and decisions could 17 be made to then take a high safety significant item and 18 put it in low, but given due consideration whether a 19 reliability is the only factor of concern here. So they'd 20 have to go through the same elements of questioning and 21 logic flow that were originally employed.
22 The final chapter of the reg. guide serves to 23 identify some documentation that would either be l
24 maintained by the licensee or submitted, if necessary, to {
/^'; !
(_) 25 support a OA program change. And this would include NEAL R. GROSS l l
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON. O C. 20005-3701 (202) 234-4433
517 1 information that was accumulated on the functions, the 1
, s, 2 justifications that were employed to downgrade items to a i V And so those types of 3 low safety significant category.
4 items are just captured in the last chapter of our reg.
5 guide.
6 That concludes our presentation, so we're at 7 your disposal for any other questions.
8 CHAIRMAN APOSTOLAKIS: All right. Any 9 comments or questions from the table? Public? No?
10 Staff? No?
11 Okay. Thank you very much.
12 MEMBER POWERS: Mr. Chairman?
13 CHAIRMAN APOSTOLAKIS: Yes?
i
- i
'- 14 MEMBER POWERS: Well, maybe just one question.
15 One question about -- how does NEI 96-02 fit into this 16 whole program?
17 MR. GRAMM: We have seen more than one 18 revision of 96-02. We have provided comments to NEI on i
19 that document, where we felt it could be augmented or l 20 changed. At this point, we've certainly considered the 21 content of 96-02 during the development of the reg. guide, 22 and that's the role that it has played in this process.
23 We've considered what is in there, but at this 24 point there is a disparity between the content of the reg. l
(_./ 25 guide and 96-02. So I don't think there is sufficient NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4433
518 1 information in 96-02 that we could endorse it at this
-s 2 point in time.
I b
'"'/ Have you considered just 3 MEMBER POWERS:
4 endorsing it and couldn't do it?
5 MR. GRAMM: That was the original intent of 6 working with NUMARC was to -- for the industry to define 7 such a document where we could have a common approach that 8 would be employed rather than being faced with 70 9 licensees with 70 different methodologies in front of us, 10 and have one document that could be the cornerstone. And, 11 unfortunately, we never reached a point where we could 12 endorse it.
13 MEMBER POWERS: And I can look in the appendix 14 and understand your concerns about 96-02? It has a 15 listing of your correspondence.
16 MR. GRAMM: Yes, sir.
17 MEMBER POWERS: That's an up-to-date listing.
IP MR. GRAMM. The listing in 96-02 would be f
19 dated. I think there has been some evolving of staff l 20 thought over the last six months, particularly with our 21 interactions that have occurred with the pilot. So the l l
22 attachments to 96-02 do not articulate all of the 1
I 23 disparity.
24 MEMBER POWERS: Disparity has gone up, not p
() 25 come down?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
519 1 MR. DINSMORE: Well, as far as use of 7_.
2 importance measures to do ranking, the disparity has
( I 3 probably gone up.
4 MR. WOODS: I was just going to put on the 5 record the fact that we have referred to the three plants 6 as pilots. And because of the fact that Appendix B allows 7 grading of quality assurance commensurate with safety, 8 they don't -- the industry, by and large, doesn't feel 9 that they need our permission to do this. So they haven't 10 -- they really aren't officially pilots who have agreed to 11 provide us with the information that we asked for.
12 They cooperate with us, but they don't feel 13 obligated to answer -- to provide absolutely everything we
,\
i )
k/ 14 ask for. And they don't consider themselves pilots. And 15 if I didn't say that, I'm afraid we'd all be in trouble.
16 So thank you for letting me --
17 (Laughter.)
18 CHAIRMAN APOSTOLAKIS: Mike?
19 MR. GOLAY: Yes. I'm here especially looking 20 at what the Research Branch ought to be doing. And 21 nothing in your presentation really draws attention to the 22 need for support from that part of the agency, yet it 23 seems logical that there would be analyses or tools or 24 something that would be helpful in working through this n
25
(_) problem. So can you tell us what the situation is in 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
520 1 area?
2 MR. WOODS: Go ahead, Mark. I'm not sure what t
,-s)
I'm from Research, but I don't --
3 the question is.
4 MR. GOLAY: Well, let me try the question 5 again. For example, if I were in your shoes, I would be 6 expecting that there would be some analytical tools or 7 data assessment, trends analysis, maybe from AEOD rather 8 than Research, which would help me figure out whether the 9 graded QA approach is working well, especially 10 anticipating experience that might come in the future.
11 Or, for example, examining what should be standard 12 practices for the expert panels.
13 You know, there are a host of questions which l i
\/ '
14 one could address. And I'm sort of looking for what's the 15 agenda that is being served up by this exercise for the 16 Research group, and perhaps for AEOD, and perhaps for 17 others as you may think of them.
l 18 I mean, it sounds like what you're really l
19 going to do is sort of try it out, see what comes back, 20 and then figure out what your research questions are based 21 on whatever feedback you observe.
l 22 MR. WOODS: Well, there may not be any l
23 questions, because we really aren't expectingtany problems 24 with this, given that it's non-safety related equipment
,m
( ,) 25 and they're taking small steps. And we would expect 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
521 1 begin to see problems in small pieces before we see big
<w 2 problems. And it doesn't seem like, to me, you'd want to
( )
'~
3 start some major program. I mean --
4 MR. GOLAY: I'm only asking the question. I'm 5 not trying to say you should be. I only want to 6 understand the state of your thinking and whether you're 7 looking for any support from that group.
8 MR. DINSMORE: I think you should ask our 9 bosses that.
10 MR. CUNNINGHAM: Well, one thing you could --
11 you could pose in the context is -- part of the reason 12 that PRA is not used very much in this process is because 13 you don't have a good understanding of the relationship i<-~)
\/ 14 between quality assurance and reliability of equipment.
15 Conceptually, there is a research question there which 16 says, could you develop some sort of model like that? I'm 17 not suggesting that we're going to volunteer to do that 18 because it may be attractable, I don't know. But that 19 question has kind of come up as we have thought about 20 this, as we have developed these reg. guides.
21 Related to that are the questions on more PRA-22 related questions, like are the importance measures that 23 we're using the right importance measures? And there has 24 been a -- Roy is in the Research office, and he and other
(,) 25 folks have been supporting -- people have been supporting NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 2344 433
522 1 him to try to deal with some of those questions as we've g- 2 gone along.
t /
'~
3 MR. DINSMORE: I guess we've had several --
4 we're working on a schedule. And I guess to be perfectly 5 honest, I personally haven't thought that much about it 6 because we have the -- to present something which we think 7 is workable. And the questions which have arisen during 8 that time I suppose that we'll eventually get around to 9 asking Research.
10 But to actually -- we haven't had any sit-down 11 sessions and tried to decide what we should do over the 12 next two years, because we have been trying to --
13 MR. WOODS: Well, within Research, we have had p_
,~*/
)
14 discussions along the lines of what you're talking about.
15 I know, Mark, you and I both talked to Joe Murphy about 16 the distributions and the tail on the distribution, and 17 how you might -- the QA program really looks to -- for the 18 way out the ends of the tail, and maybe a statistical 19 statistician could help us with this kind of problem.
20 We have looked into the kinds of things that 21 you might ask yourself and do. But as we said here, this 22 thing has got to be published December 31st, 40 days from l 23 now, and that's not the kind of short-term thing you could 24 do. So --
(,,) 25 MR. GOLAY: Well, is there an action item 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
i 523 l I go back and examine this question when you have greater es 2 leisure? j i l !
~
3 MR. WOODS: I would say that there would be 4 such an action item if we start to see a problem.
5 MR. GOLAY: For example, when you were going 6 through your discussion of operational feedback for 7 components which have been ranked low in safety 8 significance, essentially asking might you have gotten 9 things wrong and looking to see if there are any 10 experiences which would tell you that, it sounded to me as 11 you went through it that you're very unlikely to get that 12 signal even if things are going wrong, because you could 13 components that are degrading, for example, and the g
- 14 existence of that degradation would not become manifest to 15 you under most of the operational experience and :
16 inspections, and so on.
17 That, rather, things which would show up 18 importantly, should you have accidents, especially severe )
19 ones, are not going to be revealed by operational i
l 20 feedback. !
21 MR. WOODS: Well, let me take that a bit l
22 further and suggest, then, that what you are alluding to I 23 think is more data collection or requests for more 24 information from the industry in lacking a problem where I l (n) ,
25 could -- that would be a backfit.
1 l
NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
524 1 MR. GOLAY: Well, I think it's probably more 2 subtle than that, actually.
/m\/
\
3 MR. WOODS: Well, whatever it is, it's going 4 to cost somebody some money. And if there's no 5 demonstrable safety benefit for it, then I can't do that 6 either.
7 +
CHAIRMAN APOSTOLAKIS: Yes. But are you 8 getting now into how one would determine that there is a 9 need for information?
10 MR. GOLAY: No. I was actually going beyond 11 that. That is, they have effectively said there is a need 12 for further information, and that's why we had this bullet 13 in their presentation on operational feedback. And so
,/ 8 (V )
14 then I was asking, well, is what they are thinking to do 15 going to provide that kind of feedback in a useful way?
16 My guess is not. But if they really need it, that could 17 be an appropriate topic for Research to pursue with them.
18 And so that's the sort of thing I'm trying to learn more 19 about with these questions.
20 MR. WOODS: I guess my belief is that if there 21 is a significant problem you would see it -- you wouldn't 22 see nothing. A significant problem in this low safety 23 significant equipment is not going to develop and us get 24 no information that it's developing until it's a very
( ,) 25 serious problem. I just don't see how that could happen.
NEAL R. GROSS COURT REFORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
525 1 We do have inspectors out there. We do have residents.
2 You know, it's not like we're talking about 7_
- )
' ' 3 the reactor coolant pump or something here. You would 4 begin to see this before it's a serious problem without a 5 very expensive data gathering, programmatic --
6 MR. JONES: Let me make sure it's clear, and I 7 understand you weren't here yesterday. But the 8 performance monitoring and feedback programs that we're 9 talking about are things that the licensee does. We will 10 look at the program. We will assure the program has the 11 right elements, that it requires certain actions be taken, 12 given certain observed failures, or whatever the 13 monitoring item is, and that's what happens.
,\
t kJ 14 It's not that we will be involved in every 15 failure or item that fails or in the day-to-day 16 performance monitoring of all of these components. We do 17 some monitoring through AEOD on system performance. I 18 mean, and they are trying to develop some new indicators, j l
19 performance indicators, for important safety systems.
20 But a lot of this stuff -- day-to-day 21 operations and failures -- I mean, failures occur all the 22 time in the plants. You know, components will fail, they 23 will look at it, they will determine whether -- what the 24 root cause is. If the root cause appears to be QA, they
()
n 25 will have to take appropriate corrective action and modify 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
526 1 their QA program as appropriate.
f-.s 2 And what we want to capture in this element, I )
3 and the other programs that we're working on, is to assure 4 the mechanisms are in place, that that is what happened, 5 that they look at, is it something about what we changed 6 that may be starting to cause a problem? And if so, then 7 they should feed that back into their normal operations 8 and correct it so it doesn't become a large problem.
9 MEMBER BARTON: I agree. I think most 10 utilities today have got a deviation reporting system with 11 such a low threshold, and they're all analyzing the 12 results of the -- you know, the cause of the deviation. I 13 think the data is out there, and I think it's available to
('~'%
- 14 the inspectors.
15 MR. JONES: Yes. And we would inspect as part 16 of our normal processes, I would anticipate our QA 17 inspectors would assure they are implementing their 18 corrective action programs properly.
19 You can answer that one better than I can, l o 20 Bob. But, I mean, that is part of the role of the QA 21 inspector.
22 CHAIRMAN APOSTOLAKIS: So it seems to me that ,
l 23 the answer you are giving to Mike is that you have no need 24 for additional tools at this time. Is that a correct --
( ,,
) 25 MR. JONES: Well, I 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
527 1 CHAIRMAN APOSTOLAKIS: That's okay.
rx 2 MR. JONES: I would rather answer we haven't k' ) 3 thought about it --
4 CHAIRMAN APOSTOLAKIS: Okay.
5 MR. JONES: -- real critically to do that. I 6 think, though, in the nature of what we're doing at this 7 point in time, in the nature of the kind of relief we're 8 looking at in the QA area, I'm not sure that the 9 development of such activities would be very worthwhile.
10 This is, in my mind --
11 CHAIRMAN APOSTOLAKIS: Okay.
12 MR. JONES: -- relatively small. I think --
,. , 13 CHAIRMAN APOSTOLAKIS: So that's the answer 14 they are giving right now.
15 MR. JONES: But I wanted to make sure that it 16 was put in the context of the overall programs that we 17 have been talking about, including the general program, 18 the role of performance and monitoring and feedback, and 19 what we are trying to accomplish in all of these 20 activities.
21 CHAIRMAN APOSTOLAKIS: And this answer applies 22 only to the graded QA problem.
23 MR. JONES: As far as the need for additional 24 data, yes, rN k._,) 25 CHAIRMAN APOSTOLAKIS: Not a general risk-NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
1 528 1 informed performance-based. Okay.
l 2 MR. JONES: There are going to be some l
, ,)
t' 3 components, and we have the reliability data rule that we 4 have out on the street for comment. And we're working i
5 with industry on that, and whether or not -- to gather 6 some of the more important data and feed that back in.
7 So, I mean, we have something in that area.
8 MEMBER SEALE: It strikes me that this is a 9 very important element, though, of the suite of case 10 studies, if you will, or the pilots -- your disclaimer 11 notwithstanding -- that should be a part of the overall 12 thing that hits the street, because this really tells you 13 what the scope of risk-informed regulation reform can
~s
( )
il 14 include.
15 I mean, you're not asking for a quantitative 16 change or any kind of evaluation of the change in core 17 damage frequency or anything like that. But you are 18 asking for the consideration of the elements of risk in I 19 making decisions.
20 MR. JONES: That's correct.
21 MEMBER SEALE: And I think that's a very 22 powerful message, and this example is a very good way of i
23 sending that message. And so I would not in any way think j l
24 you would want to minimize the importance of this )
,m
/ S j 25 particular element in the grouping of examples or of 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
529 1 supporting pilots.
2 MR. JONES: No. And I don't mean to do that.
v' 3 I -- j l
4 MEMBER SEALE: No , I didn't think you did.
5 But I just wanted -- j 6 MR. JONES: We do want to put this reg. guide l l
7 out with the --
8 MEMBER SEALE: Yes.
9 MR. JONES: -- what the other six --
10 MEMBER SEALE: Yes, okay.
11 MR. WOODS: And I think the most important 12 step is not for us to collect the data, although that 13 would be nice, but to make sure that the licensee himself
$q 1
\s 14 follows it, because he is going to know before we know and 15 there is -- we did make an effort to make it very clear 16 that particularly in graded QA you would need a very 17 strong followup program, a monitoring program and a 18 feedback program, and that's where the data is needed.
19 Ltd we do have words in here to make that clear.
20 MR. MARKLEY: Just one question or 21 clarification. We don't really have QA inspectors per se. l 22 We're still doing that through residents and others, are i
23 we not?
24 MR. GRAMM: But as Bob alluded to, there is
,m i %
( ) 25 one inspection procedure that examines, in particular, 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
530 f 1 corrective actions.
,s 2 MR. MARKLEY: Right.
( )
3 MR. GRAMM: Maybe 40500, and this would 4 certainly be in scope to look at the problems identified 5 by the licensee and then how they were rectified.
l
, 6 MR. MARKLEY: Right.
7 MR. GRAMM: So you are correct. We do not 8 have explicit QA inspectors. However, every NRC inspector 9 should be examining the application of the QA programs in 10 their area when they're on site.
11 CHAIRMAN APOSTOLAKIS: Any other --
12 MR. SHERRY: Just one comment. It seems that 13 it would be interesting for -- as a research program, for n
% 14 a limited number of components to reexamine the database i 15 and to see if you could separate and distinguish different 16 failure rates for different --
l 17 CHAIRMAN APOSTOLAKIS: I remember vaguely 18 somebody from Livermore years ago mentioning that they did 19 that. Is there smiling over there?
l 20 MR. CUNNINGHAM: I have the same vague memory, )
21 but that's --
l 22 CHAIRMAN APOSTOLAKIS: Yes. )
i 23 MR. CUNNINGHAM: -- and it kind of continues 24 that they weren't able to differentiate very much.
77
() 25 CHAIRMAN APOSTOLAKIS: And they were not able 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
531 1 to differentiate within components, so they --
- 2 MR. CUNNINGHAM: Very, very vague memory.
3 MR. SHERRY: Doesn't that --
4 CHAIRMAN APOSTOLAKIS: I know, but --
5 MR. SHERRY: That information itself is --
6 MR. CUNNINGHAM: You're better than I --
7 MR. SHERRY: -- valuable, I would think.
8 MR. CUNNINGHAM: You're better off than I. I 9 don't recall --
o 10 CHAIRMAN APOSTOLAKIS: Yes. I remember who 11 did it. I don't remember the details.
12 AUDIENCE MEMBER: Do you remember who?
13 CHAIRMAN APOSTOLAKIS: Yes.
(~~N, 5
'NY- 14 AUDIENCE MEMBER: I'd be interested in reading 15 it.
16 CHAIRMAN APOSTOLAKIS: Sure. Sergio Guerro.
17 MR. CUNNINGHAM: Sergio?
18 CHAIRMAN APOSTOLAKIS: Yes. He was working at j 1
I 19 Livermore at the time, and he said they did a study like !
20 that.
21 MR. WOODS: You might find it's like Dr.
22 Seale's aircraft bolts. Remember the story he told I 23 earlier that QA really -- I mean, it's going to be real 24 hard to tell any effect.
(.
(,) 25 CHAIRMAN APOSTOLAKIS: Yes. You have to look NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
532 1 at --
2 MR. WOODS: In some cases, in the other
(
/
3 direction --
4 CHAIRMAN APOSTOLAKIS: You have to look at 5 what they did. I'm not saying that what they did is 6 valid. All I'm saying is that somebody tried to do that 7 some time.
8 MR. WOODS: Somebody tried, yes.
9 CHAIRMAN APOSTOLAKIS: And as Rick s it 10 might be worthwhile to revisit the issue and se , first of -
11 all, what was done, and second, whether you want to try 12 again.
13 MEMBER SEALE: I want to deny any ownership of I )
\M ~
14 defective aircraft bolts.
15 (Laughter.)
16 MEMBER MILLER: Mr. Chairman, I have a 17 question, 18 If I understand it, the resident inspectors 19 are going to have to carry -- kind of carry this 20 particular reg. guic'.e and be the ones who do this, not 21 headquarters or anybody else?
22 MR. GRAMM: We haven't yet developed the 23 inspection procedure, the companion inspection procedure, 24 and I think a decision would have to be made at that point I
im \
() 25 in time who the appropriate staff would be to get NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
533 1 involved.
,- 2 For example, looking at the volunteer plant (3) 3 activities, the NRC headquarters office has been examining ,
4 those in lieu of the resident staff, because they simply 5 don't have the knowledge of everything that is occurring 6 in this environment.
7 MEMBER MILLER: That's why I'm asking the 8 question.
9 MR. GRAMM: So that will be a factor in terms 10 of when the inspection procedure is written who would be 11 assigned to actually --
12 MEMBER MILLER: Well, the next question -- I 13 would assume you then want to develop a training program
! )
() 14 for whoever is going to carry it out. I mean, if it's 15 writing inspection procedures, I don't think it's going to 16 be enough.
17 MR. GRAMM: Each time an inspection procedure 18 is written that's a mandatory element of consideration, 19 whether training is needed and who is going to do it, and 20 the content -- developing the content of the training 21 course that would be accounted for you. Yes.
22 MEMBER MILLER: So this -- you know, a 23 resident inspector has a lot of things to do and --
24 MR. GRAMM: Absolutely.
() 25 MEMBER MILLER: -- getting up to speed on 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
534 1 one might not be easy.
1 2 MEMBER BARTON: They always did a good job of ]
\ )
- beating me up enough on the QA program.
3 4 MEMBER MILLER: Well, they like to beat you 5 up.
6 (Laughter.)
7 CHAIRMAN APOSTOLAKIS: Anything else on this 8 subject? Thank you very much, gentlemen..
9 What I think we should do, since we have to 10 interrupt this in about 20 minutes, is identify the issues 11 that we would like to discuss the remaining time, maybe 12 create a list so that we will have some idea how many of 13 those we have. So we'll try to manage the time.
,. s
! 1
\_/ 14 I have here five that may be parts of the list 15 and, of course, you are welcome to suggest others. I 16 think one that we identified today that must be discussed 17 either today or some other time is whether parts of the 18 CDF or the contributors to CDF could be handled 19 qualitatively.
20 We are on record that that should be a 21 quantitative evaluation. So do you think we should 22 discuss this today with the members? Well, let me give 23 you my whole list and then we'll decide.
24 The second one had to do with the package p
( ,) 25 changes again. I think Bob brought that up. And I 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
535 1 know whether we want to revisit that today in light of
,-.,.s 2 what we said about separate independent increases being
( l
handled individually.
3 4 I think we have not paid enough attention to 5 this issue of quantified versus unquantified 6 uncertainties, and in the reg. guides some places there is 7 an explicit mention of the unquantified uncertainties. In 8 others, not so much. And we suggested yesterday that 1061 9 should have some sort of a statement or introduction for 10 this subject and then adjust everything else accordingly.
11 So --
12 MR. CUNNINGHAM: And that we were doing that.
13 We're doing that now.
/~'s
_ 14 CHAIRMAN APOSTOLAKIS: Okay. Now, the j 15 question is whether we want to have some discussion on the 16 subject today, so that we will help you when you do that.
17 The fourth one on my list is a broader 18 question that occurred to me as we were doing this 19 yesterday and today, that one of the major attributes or l
20 features of a performance-based regulatory system is to l J
21 give flexibility to the licensees. And I would like to i
22 have some discussion independently of an individual reg.
23 guide to talk about it.
24 I mean, where does the s't'aff feel that such f3 (j 25 flexibility is afforded now? And are we doing enough on 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
536 1 that front?
,e 3 2 And the final question here was prompted by
(')'
3 Mike Golay's question -- and, again, I don't know whether 4 it's for today or some other time -- what kinds of 5 research needs does this whole effort raise? And I know 6 that Mark is planning to -- he threatened us to come back 7 with a program at some time in the future. So maybe --
8 I'm sure you will be thinking about it. And I don't know 9 whether it's for today or some other day that we 10 discussed, but that's certainly an interesting question.
11 So that exhausts my list. Dana, do you want 12 to add anything? Let's go around the table.
13 MEMBER POWERS: I was trying to stay
,e~
~ 14 orthogonal to your list. I find your list coincident with l 15 a lot of my topics.
16 I'd like to understand a little better about 17 the time-dependent CDF values.
l 18 CHAIRMAN APOSTOLAKIS: Okay. Yes.
19 MEMBER POWERS: And in that context, I'd like 20 to understand a little better this criterion that was 21 described --
22 CHAIRMAN APOSTOLAKIS: Yes.
23 MEMBER POWERS: -- to us on the AOT, because I 24 don't understand why it's consistent with the rest of the (9) 25 discussion.
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
537 1 CHAIRMAN APOSTOLAKIS: Okay.
n 2 MEMBER POWERS: I'd like to know -- understand
( i C/ 3 better some of this integrated decision process and why an 4 expert panel is an appropriate way to doing that, and how 5 this dependence on a licensee's expert panels and a 6 licensee's peer review of its PRA is going to play out.
7 It seems to me that that puts some interesting obligations 8 on how you specify both experts and peer reviewers.
9 I think that's -- oh, there's another one that 10 shows up in the written language. It seemed to place an 11 obligation on the licensees, in looking at probabilistic 12 risk assessment, to identify areas where increased 13 regulations are required. And I would like to know how p i V 14 that plays vis-a-vis Part 20. I think maybe it's Part 21.
15 One of them.
16 MR. HOLAHAN: Dose requirements for personnel?
17 MEMBER POWERS: No, it's the requirements to 18 notify the NRC when you see significant hazards.
19 MR. HOLAHAN: 21. Part 21.
20 MEMBER POWERS: 21, then.
21 The question is: what happens if they do a 22 PRA that someone else looking at it would say, "This 23 clearly indicates the need for regulations"? But they 24 just didn't see it. Are they subject to enforcement 7-y) l 25 action because they have the PRA in their hands?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 2344433
538 1 MR. HOLAHAN: I think that one is easy.
- 2 MEMBER POWERS: That's easy?
\
That's easy.
~
3 CHAIRMAN APOSTOLAKIS:
4 (Laughter.)
5 MEMBER POWERS: Those are the ones I have that 6 were --
7 CHAIRMAN APOSTOLAKIS: And we can always add 8 later.
9 Bill?
10 MEMBER SHACK: No, it was covered.
11 CHAIRMAN APOSTOLAKIS: John?
12 MEMBER BARTON: I don't have anything else.
13 CHAIRMAN APOSTOLAKIS: Tom?
/ \
! )
kJ 14 MEMBER KRESS: Two additions. The definition 15 of LERF and how that relates to actually meeting the QHOs.
16 The other one is I'd like to hear more about their views i 17 on what scope of PRA is going to be required to enter on 18 the abscissa of their chart, assuming they have such a 19 chart, and/or whether or not the acceptance criterias are 20 fixed on that thing, or do they move depending on the 21 scope of the PRA that determines the value of the CDF.
22 That's the only two additional ones I had over 23 the ones you guys mentioned.
24 CHAIRMAN APOSTOLAKIS: Bob? l l
q) 25 MEMBER SEALE: I don't have anything to add 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
539 1 right now.
7- 2 CHAIRMAN APOSTOLAKIS: Don?
(' !
3 MEMBER MILLER: I guess I'd like to see a more 4 defined -- I realize that's not the right thing to do --
5 more defined criteria on the difference between -- or 6 defining high safety significance and low safety 7 significance. They allude to using Fussel-Vesely and RAW, 8 but they don't really give you any numbers except ASME 9 gives you numbers.
10 CHAIRMAN APOSTOLAKIS: Okay.
11 MEMBER MILLER: One thing -- another one just 12 hit me. It really has nothing to do with it directly, but 13 indirectly is the reg. -- when all of this hits the street i
f')/
(- 14 &nd it becomes the thing to do, is there a program to 1
15 bring everybody up to speed through training? See, that's )
16 not related directly to the reg. guides.
17 But if one year from now all of these reg.
18 guides are in place -- this is a framework for PRA -- is ,
1 19 anybody going to be able to actually go out and do it? If 20 something gets started now, I would think --
1 l
21 CHAIRMAN APOSTOLAKIS: Mike? Nothing?
22 Mario?
23 MEMBER FONTANA: I have a couple of thoughts, 24 not too well thought out.
/^\
() 25 One is that these plants come in with NEAL R. GROSS COURT REPORTERS Af4D TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D.C. 200053701 (202) 234-4433
540 1 different levels of PRAs, and you're going through all 7
2 kinds of gyrations to try to accommodate them, depending (s '~'
) 3 on whether they have an internal or an external --
4 external full scope for external events, or whatever. And 5 in case they should do this, or in case they do that, it 6 seems like you're bending over backwards. Why don't you 7 just tell them, "Just all go out and do a full PRA. Go do 8 it." You know, and no bellyaching, all that kind of 9 stuff.
10 MEMBER KRESS: Is your name Tom Kress?
11 CHAIRMAN APOSTOLAKIS: Are you talking about i
12 the full scope Level 3 PPA including uncertainties? j 13 Organizational factors? l I
t' % l
-( )
'w/ 14 MEMBER KRESS: Where were you when I needed 15 you?
16 (Laughter.)
17 CHAIRMAN APOSTOLAKIS: He was here, and you 18 were in Vienna.
19 (Laughter.)
20 MEMBER POWERS: And he sandbagged you good, l 21 Tom.
22 (Laughter.)
23 CHAIRMAN APOSTOLAKIS: I think we should go 24 back and read the paragraph we have in our second page of
(
) 25 our letter on plant-specific applications of safety goals.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
541 1 Why --
2 MEMBER POWERS: He railed about that for days 7q Y
3 after --
4 CHAIRMAN APOSTOLAKIS: And he wants to do it 5 again.
6 MEMBER POWERS: And he's going to come right 7 back again.
8 CHAIRMAN APOSTOLAKIS: Yes.
9 MEMBER FONTANA: The other thing on -- the 10 thing that I have some potential for research -- I have 11 been in situations where equipment, experimental equipment 12 built under tight QA rules have cost something like 10 13 times as much, very similar equipment built under standard
( )
U 14 commercial rules. Is there in the data somewhere a way to 15 mine failure rates of equipment that has been put into 16 nuclear plants and done under QA, compared to similar 17 equipment that is under commercial use built under good i l
18 commercial standards -- practice?
19 And the next step is to figure what the cost 20 difference is. And then the next step is that if you put 21 it in the commercial and it's compared to nuclear, what is 22 the delta CDF? What is the delta CDF?
23 CHAIRMAN APOSTOLAKIS: Well, then, use that 24 chart.
,a i i
() 25 MR. GOLAY: I don't think, actually, 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
542 1 the problem. Once these guys finally pay attention to it, rw 2 they'11 have no problem filling up the agenda on research.
(q) )
3 MEMBER FONTANA: No. It seems that in a lot 4 of cases there has been an absolutely humongous amount of 5 money spent on OA. A lot of it probably really wasn't 6 necessary. The trouble is you have no way to put a handle 7 on it.
8 CHAIRMAN APOSTOLAKIS: Rick?
9 MR. SHERRY: Yes. Just --
10 CHAIRMAN APOSTOLAKIS: Oh, I'm sorry. Are you 11 done?
12 MEMBER FONTANA: Well, I should be done.
13 (Laughter.)
14 MR. HOLAHAN: Was there anything left after !
1 l
15 that --
l 16 (Laughter.) l l
17 MR. SHERRY: Just two thoughts. One is that 18 there wasn't much discussion on the aspects of Level 2 19 analysis and the requirements, and I think that touches on 20 the definition of LERF and requirements for the LERF. And 21 second is it's -- the way I understand it is a small 22 increase in risk is allowable under a specific regulation, 23 50.91 and 50.92. And there are some words in there which 24 define a significant hazard.
(~._\
i
(_/ 25 I'm interested in how that relates to a small NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
543 1 increase, and when is a small increase sufficiently large
,3 2 that it would trigger different actions under that
( I 3 regulation.
4 CHAIRMAN APOSTOLAKIS: Okay.
5 MEMBER KRESS: George, I was looking back 6 through my notes and I have a couple more items to add to 7 the list.
8 (Laughter.)
9 CHAIRMAN APOSTOLAKIS: All right.
10 MEMBER KRESS: One of them was -- it seems 11 that a lot of this involves allowing changes in the test 12 frequency -- testing equipment or inspection frequency.
13 And I don't see any strong correlations between that and g^3 U 14 reliability, and I'd like to know if there are plans to 15 look at that or if you have some way to relate that test 16 frequency to reliability, or how you incorporate that into l 17 determining delta CDF right now -- what you're doing.
18 I didn't hear enough on that.
19 CHAIRMAN APOSTOLAKIS: Okay.
20 MEMBER KRESS: The other one has to do with --
l l
21 I gueso I've been concerned for some time that we're not I
22 quite using the right importance measures to get the 23 ranking of equipment for stuff, and I'd like a little more 24 discussion on the importance.
,w 25 CHAIRMAN APOSTOLAKIS: Okay. That's it?
' .s)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 2344433
544 1 MEMBER KRESS: That's the other two I had 2 noted.
i t
. ,)
'~' All right, Golay?
3 CHAIRMAN APOSTOLAKIS:
4 MR. GOLAY: I only have one. I think both 5 this committee and my own would benefit from having a good 6 idea of the vulnerabilities of the probablistic -- I mean 7 the performance based regulatory approach in the sense 8 that there may be things which will fall between the 9 cracks in the moving from the current to whatever is the 10 future approach.
11 Conceivably, safety might be degraded, and 12 there are lots of ways that things could go wrong in a 13 practical way. And it seems to me that would start with 77 j i
(/
1 14 some statement of that assessment by the staff members who l i
15 have been involved. And at least today, I didn't hear it.
16 Perhaps it was done yesterday.
l 17 But ultimately, I think the committees have to 18 form that understanding and make recommendations.
19 CHAIRMAN APOSTOLAKIS: Okay, I have here at 20 least -- I don't know, 15 or so, which obviously we cannot 21 discuss today. So is there any way that a group of 11, 12 22 people can -- oh, I'm sorry; maybe you guys want to add to 23 the list.
24 Gary, you want to add anything?
n
) 25 MR. HOLAHAN: Are we finished speaking 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
545 1 figures and fuzziness?
-s
, 2 CHAIRMA'J APOSTOLAKIS: The figures?
i )
3 MEMBER POWERS: Oh , definitely we're done with 4 that.
5 MR. HOLAHAN: We're done with that?
6 CHAIRMAN APOSTOLAKIS: I -- let me tell you 7 what I was trying to do. Since everybody seems to think 8 that the figures are useful, I was going to try to put a 9 few lines somewhere in the letter indicating that the 10 figures are not worthwhile.
11 (Laughter.)
12 Now, clearly, that would not make it in the 13 letter itself.
p i
14 MR. HOLAHAN: A minority opinion.
15 MEMBER POWERS: Well, George, understand that
. 16 I'm extremely sympathetic with regard to horizontal lines 17 in the figure. It's the vertical ones that I find useful.
18 CHAIRMAN APOSTOLAKIS: Well, maybe we should 19 discuss it again and see -- maybe it's not a matter of yes 20 or no with regard to the figures. Maybe part of the -- in 21 fact, I like the suggestion yesterday by Jim Fulford of 22 dropping the lines inside. I mean, the more -- the 23 fuzzier you make them, the happier I am.
24 So, do you want to discuss it for another ten
) 25 minutes sometime?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
546 1 MEMBER FONTANA: Those inside lines are --
,-s 2 yeah, but the inside lines are the ones that -- the
' )
\~/ 3 difference between if you had an internal or an external -
4 - and all that kind of stuff.
5 CHAIRMAN APOSTOLAKIS: Yes, yes. I thought 6 that was a good suggestion. But, but, but -- my
' experience with this committee is that you can dissent one 8 day, and then you can find people embracing your views 9 three days later after they have time to think about it.
10 MEMBER SHACK: Don't hold your breath.
11 (Laughter.)
12 MEMBER MILLER: After they went to Vienna, you 13 got a lot of concurrence.
O k/ }
I 14 CHAIRMAN APOSTOLAKIS: So, shall we add that j l
15 to the list and --
16 MR. HOLAHAN: I think we're going to revise l l
17 the figure and the text anyway. If you'd like to have it i 18 on the overhead while we mark it around, I think that's 19 perfectly all right.
l l
20 CHAIRMAN APOSTOLAKIS: Well, the idea is also 1 21 to give you feedback regarding your thoughts; so, if you 22 think that would be useful, --
23 MR. HOLAHAN: Okay.
24 CHAIRMAN APOSTOLAKIS: Okay.
/S i
)
25 MR. HOLAHAN: For a few minutes. Maybe we 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
547 1 should do it last.
f3 2 MEMBER FONTANA: I hadn't really thought this i \
C/ 3 out all the way through, but a guidance that -- it's 4 something like no matter where you are -- because you 5 don't want to go up to the limit in one big step. That's 6 one of your basic ground rules. Something like you can go 7 -- no matter where you are, you can go halfway.
8 In other words, the further away you are, the 9 bigger jump you can take. And by definition, you'.ll never 10 get there because you don't want to get there anyway.
11 MEMBER POWERS: Wrong.
12 MEMBER MILLER: That one goes the way of the
,, 13 fuzzy diagram.
14 MEMBER FONTANA: But is there some 15 simplification like that?
16 MEMBER SEALE: Yes, but the 10% in one bin 17 takes care of that.
18 MEMBER FONTANA: I don't see --
19 MEMBER SEALE: You can't go halfway if you can 20 only change it by 10%.
21 MEMBER FONTANA: I thought we were going to 22 let them change it more than that.
23 CHAIRMAN APOSTOLAKIS: Any other issues?
24 Nobody?
(q ,,/ 25 MR. HOLAHAN: We do want to talk about Federal 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
548 1 Register notice questions, but that's already on the --
2 CHAIRMAN APOSTOLAKIS: Federal Register 7 .x i~'} 3 notice, meaning?
4 MR. HOLAHAN: Meaning the questions that we 5 would put out for public --
6 CHAIRMAN APOSTOLAKIS: Oh, yes, yes.
7 MR. KING: I think we phrased our issues in 8 the form of questions that are going to go in the Federal 9 Register notice. And we're prepared to put those on the 10 screen and how them to you.
11 MR. HOLAHAN. And it may be that some of these 12 issues are things that we can't resolve in the next few 13 weeks, but we'll submit the questions.
f) <
4 x/- 14 MR. KING: Some of the things you just talked 15 about here are embodied in those questions.
16 CHAIRMAN APOSTOLAKIS: Which means that you 17 expect the public comment period to resolve them, is that 18 what you're saying?
19 MR. KING: No.
20 MR. HOLAHAN: We expect to resolve them with 21 input following the public comment period.
22 CHAIRNAN APOSTOLAKIS: Okay, now we have an 23 even longer list. I don't know, is there any way we could 24 prioritize here?
()
O 25 MR. JONES: Well, I would suggest, George, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 2000 5 3701 f.202) 2344433
549 l l
1 that why don't we quickly put up our Federal Register
,_s 2 questions if they're covered under that; and if the
(' ~ ' ')
3 committee thinks that that's good enough for treatment as 4 opposed to further discussion on these items, we can just 5 leave them at that point, and that might throw some off 6 the list right away, and then prioritize what's left -- as 7 a suggestion.
8 CHAIRMAN APOSTOLAKIS: So your suggestion is 9 if it's in the list -- on the list -- the Federal Register 10 notice list, then we chould not discuss it?
11 MR. JONES: No, I don't want to go -- I want 12 to say recognize that it's on the list. Then, if the 13 committee judges that, for now, is a good enough
' _/l
( 14 treatment, those we don't discuss. If there's particular 15 items that you would like to discuss further, that's fine; 16 we can discuss them. But we could eliminate a few that 17 way.
18 CHAIRMAN APOSTOLAKIS: Maybe one of the 19 guiding principles here for selecting what to discuss 20 first and then second is whether any of these questions 21 would affect our ability or the contents -- our ability to 22 write the letter that we owe the Commission regarding 23 this.
24 For example, one of the issues that was raised
( ,) 25 was training. That may be important, but I don't think 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
550 l' affects our letter. Other questions do affect our letter.
,S
, 2 MEMBER FONTANA: Yes.
! )
~' CHAIRMAN APOSTOLAKIS: And maybe that should 3
4 be a screening criteria and at least to prioritize for the 5 short run.
6 MR. HOLAHAN: Well, I just went through the 7 items quickly myself to think of which items were 8 fundamental to the way --
9 CHAIRMAN APOSTOLAKIS: Yes.
10 MR. HOLAHAN: -- the thing is structured. And 11 I think there are three related ones that are pretty 12 fundamental. One is -- George, your first one, which is -
13 --
U 14 CHAIRMAN APOSTOLAKIS: The CDF's.
15 MR. HOLAHAN: -- the CDF's. Can you really 16 deal with CDF in a qualitative way or how do you deal with 17 that? And I think the LERF, how that's --
18 CHAIRMAN APOSTOLAKIS: Yes.
19 MR. HOLAHAN: -- dealt with has turned up in a 20 couple of the questions, and I think it's fundamentally 21 related to the QHO's, and I think we need to understand 22 that. And then I think we need to dispose of the Fontana 23 question because it -- for one, that says why not require 24 everyone to do full scope, full PRA's -- because that (3
25 would be a fundamental change.
()
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, O C. 20005-3701 (202) 234-4433
551 1 CHAIRMAN APOSTOLAKIS: We have started
, 2 disposing of in our letter that was at least on the 18th.
( )
3 We can't go against a letter that's only four days old.
4 (Laughter.)
o 5 Give it six months, for heaven's sake.
6 MEMBER FONTANA: Look at the last sentence in 7 the letter.
8 CHAIRMAN APOSTOLAKIS: The last sentence? "We 9 believe that the use of Level 3 PRA's in the future should 10 be encouraged." I mean, the only thing that's missing 11 here is perhaps.
12 (Laughter.)
13 MR. HOLAHAN: I raise it not because I endorse fm, 14 it, but because it is a fundamental question.
15 CHAIRMAN APOSTOLAKIS: I would disagree with 16 that last item. I don't think that's a fundamental 17 question for the time being.
18 MEMBER FONTANA: It's not saying that we would 19 -- it's not anything that --
20 MR. HOLAHAN: The question is fundamental.
21 CHAIRMAN APOSTOLAKIS: If the question is 22 whether -- given this letter, I think we can wait on that 23 a little bit.
24 MR. MARKLEY: But they're going to have to
,o, 25 deal with partial PRA's. That's the reality of it.
()
NEAL R. GROSS COURT REPCRTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
552 0
1 CHAIRMAN APOSTOLAKIS: But the question -- I
,s 2 think the more fundamental part of Mario's question is Q_,)
3 whether the Level 1 PRA can be a partial PRA, which is 4 related to the first issue.
5 MEMBER FONTANA: I had real problems with 6 trying --
7 CHAIRMAN APOSTOLAKIS: And I'm having a 8 problem with that too. I mean, after the IPE exercise, it 9 seems to me that we are really going out of our way to 10' accommodate incomplete efforts.
11 Yes, I'm sorry.
12 MEMBER FONTANA: And he's already raised it --
13 he's already raised the question of -- I have real
/ _\
l )
'~2 14 problems of trying to develop a LERF built only on a Level 15 1 PRA, 16 CHAIRMAN APOSTOLAKIS: But we have the 17 question of -- definition of LERF, so all that --
18 MEMBER FONTANA: That's already in there.
19 CHAIRMAN APOSTOLAKIS: Yes.
20 MEMBER FONTANA: Okay, 21 CHAIRMAN APOSTOLAKIS: I really think we ought 22 to discuss --
23 MEMBER SEALE: Do you have anything else on --
24 CHAIRMAN APOSTOLAKIS: Oh, yes; I'm sorry.
(\ 4
(/ 25 You have --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C 20005-3701 (202) 234-4433
1 553 1 MR. HOLAHAN: Well, I think I agree with Bob,
, 2 and I think it turns up at least once or maybe twice in I i 3 the list, which is the issue of expert panel and 4 integrated decision --
5 CHAIRMAN APOSTOLAKIS: Yes, that question has 6 come up a few times. That was Dana's question regarding 7 better integrated decision making. If we're talking about 8 today, I'm not sure we can cover more than this -- one, 9 two, three -- three so far. Maybe we can.
10 MEMBER SEALE: Four.
11 CHAIRMAN APOSTOLAKIS: Which one is the 12 fourth?
13 MEMBER SEALE: The first one, --
p I
14 CHAIRMAN APOSTOLAKIS: Yes.
15 MEMBER SEALE: -- and then the integrated 16 decision process, --
17 CHAIRMAN APOSTOLAKIS: Yes.
1 18 MEMBER SEALE: -- LERF, -- l 19 CHAIRMAN APOSTOLAKIS: Yes.
20 MEMBER SEALE: -- PRA.
21 CHAIRMAN APOSTOLAKIS: That's part of one.
22 MEMBER SEALE: Okay, I'm sorry.
23 CHAIRMAN APOSTOLAKIS: Level 1 stuff.
24 MEMBER KRESS: It seems to me like what goes 25 into the definition of abscissa on CDF -- what sort of PRA NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
554 1 is required to determine that entry --
, ,, 2 MR. JONES: I would agree with that.
( I 3' CHAIRMAN APOSTOLAKIS: Say that again. Which 4 one is this?
5 MR. JONES: Basically I think it has to deal 6 with the application and use of the decision criteria and 7 what the role -- the base line is, and what does base line 8 mean and are in those figures. It's that whole thing of 9 how do you decide the acceptance guidelines.
10 CHAIRMAN APOSTOLAKIS: Can we discuss it with 11 the integrated --
12 MR. JONES: Oh, yes, yes. I would just say 13 that --
rs C 14 CHAIRMAN APOSTOLAKIS: Bob, I would like to j i
15 see some discussion on quantified and unquantified l l
16 uncertainties. You are already doing something about it, i 17 so -- J 18 MR. JONES: Yes.
19 CHAIRMAN APOSTOLAKIS: -- maybe that will be 20 quick discussion.
21 MR. PARRY: George, I think that's also 22 related to the same issue though to some extent.
23 CHAIRMAN APOSTOLAKIS: Okay, so let's --
24 MR. PARRY: I think a lot of these issues are
,r y
!, ) 25 related.
x.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
555 l 1 CHAIRMAN APOSTOLAKIS: In the decision making I l
7~
2 process, yes. Maybe -- so you guys don't want to discuss
~~'
3 flexibility. Flexibility if there is time? Maybe you can 4 just give us an answer so we don't debate it. Not right 5 now though. I have to go. So maybe we can -- 4:00?
6 Okay, and I hope the list of issues will not 7 change.
8 (Whereupon, the foregoing matter went off the 9 record at 3:24 p.m. and went back on the 10 record at 4:17 p.m.)
11 CHAIRMAN APOSTOLAKIS: I met with the Chairman 12 and I advised her that the committee feels that it would 13 be a good idea to be given the month of January to read x_/ 14 the documents -- the final documents; no revisions any 15 more -- and have a subcommittee meeting with the staff, 16 which we have tentatively scheduled, subject to the 17 Chairman's decision to grant this request, for January --
18 MEMBER SEALE: Twenty-eight and twenty-nine.
19 CHAIRMAN APOSTOLAKIS: -- twenty-eight and 20 twenty-nine, Tuesday and Wednesday. And that we write our 21 letter at the February meeting rather than the December l
22 meeting, which we feel would be too soon. I also advised 23 her -- although there was no need for that -- that your 24 deadline not change.
fm
) 25 (Laughter.)
NEAL R. GROSS i I
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. 1 (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l l
556 1 I can assure you my advise on that subject --
-~
s 2 MEMBER POWERS: That was good not to try to
'~'
)
3 disrupt their lives just because of us. He's a very 4 thoughtful individual, i
5 CRAIRMAN APOSTOLAKIS: So the Chairman said 6 that she will think about it. She will talk to you, the 7 staff; and she will make a decision by Monday or Tuesday.
8 So this is where we stand.
9 MR. HOLAHAN: Okay.
10 CHAIRMAN APOSTOLAKIS: Okay. So you guys had 11 your break.
12 MEMBER SEALE: You got yours too.
13 CHAIRMAN APOSTOLAKIS: Okay, so shall we start
(~N
\ t K/ 14 discussing these issues?
15 MEMBER SEALE: No, they've got their last 16 presentation to make.
17 CHAIRMAN APOSTOLAKIS: Presentation?
18 MR. KING: We have an item 14 on the agenda 19 which is called summary of what we've done.
20 CHAIRMAN APOSTOLAKIS: Oh , the questions --
i 21 yeah, I thought there was --
22 MR. KING: The questions are in here. What I 23 also wanted to try and do was to try and point out where 24 we were going to -- the things we thought we could point n
( ,) 25 out to the Commission, the important points in the SECY 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
557 1 paper that transmits all this stuff.
,-y
- 2 I had also put together from yesterday's
. 1
~
3 discussion what I thought were the major issues raised in 4 yesterday's discussion. You've added to that today. And 5 then I think the last bullet is important before we break 6 up today, if we are going to discuss it at the December 7 full committee meeting, what do you need from us.
8 And in the interest of time, why don't we just 9 go to page three, and we can go through the questions that 10 we've got sketched out so far for the Federal Register 11 notice.
12 And I've broken those into basically four 13 categories: some general questions that apply to the I )
14 whole process, ones that deal with the engineering i 1
15 evaluation, some that deal with the performance monitoring 16 part, and then the fourth one will be with the application 17 specific reg. guides.
18 And I'm not going to read all of these, but 19 the ones with the asterisks are ones that deal with the 20 four policy issues that we've gone to the Commission on.
21 We recommended that we solicit public comment on those 22 issues, and that the Commission not make any final 23 decision until that -- public comments are received and 24 reviewed.
()
( 25 CHAIRMAN APOSTOLAKIS: So that means another NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
558 1 year then? l l
,, 2 MR. KING: That means --
i \
i /
3 CHAIRMAN APOSTOLAKIS: What is --
4 MR. KING: Public comment periods are 5 generally 90 days.
6 CHAIRMAN APOSTOLAKIS: I'm sorry?
7 MR. KING: Ninety days.
8 CHAIRMAN APOSTOLAKIS: Oh, okay.
9 MR. KING: So if it's published in January, in 10 April it would be over.
11 You can read these, but let me just point out 12 I think number E on this page -- E, F, and G deal with 13 ACRS issues that were raised over the past couple of days.
.O
(_)
[ 4 14 Question on E is guidance on the treatment of 15 uncertainties: is it clear and sufficient; or is 16 additional guidance desired.
37 This would deal with uncertainties -- revising 18 the uncertainty section to talk about the uncertainties --
19 modeling uncertainties, data uncertainties, and the 20 uncertainty due to the fact that there's some things that 21 are unanalyzed; and also, how do you deal with that in 22 terms of how close can you come to the limits.
23 So I think all of that would be embodied in 24 this question. F, the issue of do we want guidance on the gm,
) 25 treatment of temporary changes in the CLB. Right now, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
559 1 there is none in there. We talked about this yesterday.
l
,_ 2 We're talking about the general reg. guide now. There is i 1
'# 3 guidance like that in the specific one.
4 So we wanted to solicit a comment there. And 5 then G deals specifically with the area of shut down. We 6 just have one set of guidance now that deals with risk, 7 shut down, or operation. We were really focusing on the 8 operation when we wrote that.
9 We're asking the specific question: is that 10 good enough for shut down, or should we get something else 11 for the shut down condition? Like, for example, LERF may 12 not mean much when the containment's open.
13 MEMBER POWERS: It may mean a whole lot.
q
(- 14 MR. KING: Page --
15 MEMBER POWERS: Could you explain H to me a 16 little bit?
17 MR. KING: Right now, we're approaching -- the 18 policy statement is not a rule. We're writing the Federal 19 Register notice and these reg. guides that these things 20 are voluntary for licensees to apply. But we thought 21 about putting this question in, and we framed it as H 22 here. Should licensees be required to submit this 23 information, which would make it -- certainly raise back 24 fit issues and so forth.
,m
( ,) 25 But we'd like to get some feedback as to, 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
560 1 know, should this be something more than voluntary. So
,, 2 that's the reason H is in there.
i
-,' ') Page four are issues dealing with the 3
4 engineering evaluation, the principles, the risk criteria.
5 Item C --
6 MEMBER POWERS: Well, on A, you asked if the 7 principles are clear and sufficient. You don't ask the 8 question if they're all necessary.
9 MR. KING: No, that's a good point. We could 10 put in are they all -- add that sub-element to the 11 question as well.
1 12 Item C dealt with the -- some of the issues 13 raised here in that we're talking about defense in depth ;
[d 14 and safety margins. I think George brought up the issue l I
15 of should we treat things differently that are in the PRA l 16 versus things that aren't analyzed in the PRA. And the 17 first bullet under C is an attempt to raise that question.
18 And then also, we talked about should we try 19 and put some quantitative guidelines in that would help 20 take PRA information and make some decisions on defense in !
21 depth and safety margins. And, you know, the number 10%
22 was thrown out. You know, the word outliers was thrown 23 out and so forth. This would be an attempt to solicit 24 some feedback on that.
p
() 25 We're going to be thinking about that as 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
561 1 try and finalize this reg. guide as well. We may put some o 2 things in the final version, but we did want to solicit
( )
3 comment on it. D, dealing with the probablistic analysis.
4 I think the first bullet addresses an ACRS issue which has 5 to do with the qualitative nature of the risk evaluations.
6 Is there sufficient guidance in the document 7 now? If not, what else should be added? Page five 8 continues the list of PRA questions. I think most of 9 these we've talked about.
10 The second bullet on page five was put in 11 specifically as a result of yesterday's discussion dealing 12 with how useful is the figure that illustrates CDF and 13 LERF and would they want something more or less than 14 what's in there now.
15 The third bullet was put in because of Dr.
16 Kress' concern, which I think is a good question on LERF !
l l
17 for PWR's versus BWR's. The fourth bullet, we had had it I
18 in, but I think it also addresses Dr. Fontana's question )
19 about the appendix that deals with how do you estimate 1
20 LERF if you only have a Level 1 PRA -- is that sufficient.
21 And then the last two bullets and about the 22 top half of the next page deal with PRA scope issues. And 23 particularly they deal with the regions, you know, we have 24 on the figure now, the various regions which are really --
(\ i i s ,/ 25 to some extent, define the scope of the PRA that has 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
562 1 -- that you're willing to accept to -- if you're in that
,g 2 particular region.
() 3 So we've asked some questions in that regard 4 regarding PRA scope. I think that carries you over to the 5 middle of page six.
6 MEMBER POWERS: In our discussions of 7 uncertainty, and in your question here, you raised the 8 issue on uncertainty only for things that increase either 9 CDF or LERF. But because we're dealing oftentimes with 10 point estimates and not with means, don't you have to look 11 at uncertainty even if it's going -- a proposed change 12 based on a point estimate appears to decrease CDF or LERF?
13 MR. KING: I'm not sure I follow your (3
t Y- 14 question. Right now we've get -- for decreases in risk or 15 small increases in risk, we're saying point estimates are 16 good enough. You don't really need to do any quantitative 17 uncertainty analysis.
18 MEMBER POWERS: That's right; that's what 19 you've got now. And you're very comfortable with that --
20 point estimate and the mean are not the same things?
21 MR. KING: No, no; that's right. We're 22 convinced that it's a decrease, you know, so what -- and 23 if we're convinced it's just a small increase, what if 24 you're wrong. Even if you're off by a factor of ten, it
(._) 25 isn't going to make much difference.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W, (202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
563 1 So that's sort of the logic behind -- be 7 .s 2 willing to accept point estimates in those cases.
( )
3 MEMBER POWERS: If you're at 10-6, then you're 4 off by a factor of ten, then you've hit your other 5 boundary. I mean, your boundaries are very narrow 6 relative to some of the uncertainty bounds that are not 7 uncommon to find in these things.
8 And so why is it acceptable to use a point 9 estimate to show decrease when in fact your uncertainties 10 may be such that the mean value is actually an increase?
11 The ranges on these things typically are not small 12 compared to a factor of ten.
13 MR. HOLAHAN. But the ranges on the change is
/^%a r
i
\/ 14 not nearly as large as the range on the baseline PRA. I 15 think we can distinguish an increase from a decrease.
16 MEMBER POWERS: Based on a point estimate.
17 MR. HOLAHAN: I think you feel confident in 18 saying a decrease in the point estimate doesn't correspond 19 to a large increase in the mean. Maybe there's enough 20 fuzz in it that there could be maybe a small increase.
21 In most cases, I would expect the change in a 22 point estimate to be pretty close to a change in the mean.
23 But even in unusual cases, I think -- I can't think of how 24 you would get to a big change -- big, meaning bigger than 7
(,) 25 10 '* . So it would change your decision by basing it 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
564 1 point estimate.
,, 2 MEMBER POWERS: If we take similar plants and
- 2 3 the IPE's as an ensemble so that the independent estimates 4 for similar plants constitutes different trials as a point 5 estimate, do we arrive at that same confidence?
6 MR. HOLAHAN: I think it's a different 7 question. Because then I think you're testing how well do 8 you know the mean, not how well do you know the difference 9 between a before and after change that you're trying to 10 review.
11 I think if you took two of those plants and 12 said oh, I was going to change the same thing in both 13 cases, it's very likely to have the same sign and to be on
/o )
D' 14 the same order.
15 MEMBER KRESS: Where all you do is go into the 16 same PRA and change that one thing.
17 MR. HOLAHAN: Yes.
18 MR. KING: The same change.
19 MR. HOLAHAN: You could get different answers.
20 I mean, the sensitivities would be different based on how 21 it's modeled and things, but I --
22 MEMBER POWERS: Well, you're talking about the 1
23 change. But the change is not independent of the absolute 24 position in this world. And the formalism you've set up,
,r )
) 25 I've got to know both axes.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
565 1 MR. HOLAHAN: Not for decreases.
l s 2 MEMBER POWERS: Not for decreases.
I 1
ii
'~' MR. HOLAHAN: Which is well shown by a figure.
3 4 MEMBER POWERS: Explained well by a figure.
5 MR. JONES: And supporting text. i l
6 MEMBER POWERS: And explained somewhat roughly 7 by a supporting text.
8 MR. MURPHY: Let me offer a -- this is Joe 9 Murphy from --
10 MEMBER SEALE: Oh, Joe finally showed up.
11 MEMBER KRESS: You've got your list with you 12 that --
13 MR. MURPHY: No, I forgot the list. I'm f~)
k/# 14 waiting for your input from your contributions to the 15 list, and I'll add them to mine, and we'll produce it as 16 soon as possible. I 17 What I'd like to refer back to is the first 18 draft of NUREG-1150 where we tried to look at the impact 19 of some changes in design and how they affected risk.
1 20 Throughout -- particularly in terms of core 21 damage frequency, we found sort of pragmatically, as we i
22 look at the results that have been done with point i 1
23 estimates as a scoping study and then gone back and done l 24 the full uncertainty analysis, now it's the t'%
(_,) 25 illustratorious uncertainty only. I had to get that word NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005 3701 (202) 234-4433
566 1 in.
-y 2 The difference between the point estimate and
\
)
3 the mean is usually within a factor of three, and it's 4 often less than a factor of two. It's not a big 5 difference between -- at least for Level 1. At Level 2, 6 that number increases a little bit, but not by a lot.
7 This is just based on a -- it's not a 8 mathematical theorem by any means, but it's the 9 observation of a whole bunch of analyses. That generally 10 comes -- the way they come out. The second thing is, 11 though, if you look at 1150, we looked at changes.
12 There were sometimes where we found that the 13 point estimate was up, and there was a portion of the x- 14 distribution was down. In other words, we showed that the j 15 point estimate would indicate a risk increase, but the j l
16 distribution would indicate that some fraction of that ;
17 distribution was in fact a risk decrease. )
18 And of course, the opposite was true too. But 19 never did we find a case where the change was large. I 20 mean, if the point estimate change was small, then the 21 changes in distribution were small. And probably the 22 right way to look at it is, when you're down in the 23 numbers in the 10 4 and 10" range, what we're really trying 24 to define is zero, or something very close to it.
q
,) 25 We shouldn't really think that 10" is much 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
567 1 smaller than 104 They both mean small. It doesn't mean
~~
2 much more than that. And I think that's probably the
)
V We're down in the fuzz.
3 right way to look at that.
4 MR. KING: Okay, just to finish this up, the 5 two remaining questions on policy issues are marked with 6 the asterisks there. And then at the bottom of the page, 7 I think there will be some specific questions for the 8 application specific reg. guides which we haven't 9 developed yet. As those get more final, we'll do that.
10 Just quickly, you know, when we send all this 11 stuff up to the Commission, there's going to be a SECY 12 paper that transmits this package of material. We thought !
13 it important in that SECY paper to highlight a few things !
(~\
t
\
l i
\ '/ 14 to the Commission. I don't expect them to wade through 15 the whole package and pick out what's important.
16 The things at this point that we'd recommend 17 be highlighted in the paper that -- the fact that this 18 information or the use of this process is voluntary and 19 the submittal of risk information voluntary. You know, 20 licensee could say well, I'm not going to do it. But 21 what's that mean?
22 And I think from our standpoint what it means 23 is if they don't want to use risk and follow the guidance 24 or the encouragement of the policy statement, they won't r~N t
( ,/
25 be given as high a priority as people that do in terms 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
568 1 staff resources.
2 We clearly want to describe to them the 73 \
s /
'~'
3 overall process, approach, and the principles; and the 4 fact that, you know, they allow flexibility and there's 5 guidance as to what each of those means. But those are 6 the things that are really important, meaning those 7 principles.
8 We clearly want to describe the basis for the 9 CDF and delta CDF and the LERF and delta LERF guidance 10 that we've developed, you know, and it's consistency with 11 the regulatory analysis guidelines and the previous SRM's 12 the Commission has issued on safety goals.
13 I think it's important to point out to them ,
/~'N l
\) - 14 that our -- what we call Region IV on the graph is really 15 a practice. Maybe the Commission will consider it a 16 policy that if people exceed the CDF and LERF values, 17 we're -- and want to make changes, those changes need to 18 be in the direction of reducing risk. And I think it's 19 important to point that out to them.
20 And also, how that relates to the IPE follow 21 up activities that Gary mentioned previously that we have 22 risk information. We know some plants have reported CDF's 23 greater than 10-4 and LERF's greater than 10-5 That 24 doesn't mean we're going to sit back and do nothing with
,r y
() 25 those. There is going to be some follow up activity to 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
569 1 look at those plants and see if anything should be done.
2 And then I know the Chairman is very 7-( )
~
3 interested in the treatment of uncertainties, so I think 4 we want to summarize and point out how we're dealing with 5 that in this package of materials. So, there may be other 6 things we want to add to that as well, but that -- those 7 ought to be pointed out.
8 And the other thing I'd quickly mention is we 9 do have some open items -- clearly some open items from 10 the past couple of days. But in addition to that, we 11 mentioned we are expanding the discussion on 12 uncertainties. We're moving some guidance from the SRP 13 and NUREG-1602 into the reg. guide.
(_ 14 Major items in that regard are the discussion 15 on peer review process and the SSC categorization 16 discussion. We're still having discussions on the LERF 17 guidelines internally. That's not settled yet. We're 18 updating NUREG-1602 to be more readable, as well as get 19 any guidance out of it and put it into the reg. guide.
20 We've gotten CRGR comments as -- I think those 21 have been -- affected the SRP more than the reg. guide, 22 but we're dealing with those. And then anything that 23 comes out of the past couple of days that we want to go in 24 and make some changes, there's a number of those that ry
(_,) 25 we're going to have to update before this package goes to 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 570 l l
1 the Commission. ;
I gs 2 The last page, I had made an attempt to l 3 summarize a few comments from yesterday, but I think we 4 have a better list from before the break, so we don't need 5 to go through the last page. !
6 With that, we can go talk about the important 7 issues and see if we can settle on some of those.
8 CHAIRMAN APOSTOLAKIS: Okay, so which one 9 shall we start with? Let's start with the quantified and 10 unquantified uncertainties, which may be a short 11 discussion. It's also important.
12 MR. HOLAHAN: Mr. Jones has offered to lead 13 such a discussion.
(~3 sl 14 MR. JONES: I offered to give you the slide.
15 Is this like my offer to have another draft next week?
16 MR. HOLAHAN- No, this is your slide from 17 yesterday's presentation.
18 MR. JONES: Oh, Mark's slide.
19 MR. HOLAHAN: Is it Mark's slide? Oh, okay.
20 MR. JONES: But -- you're going to take it, 21 Mark? Go ahead.
22 MR. CUNNINGHAM: Let me read the slide again 23 for a second. Okay, --
24 CHAIRMAN APOSTOLAKIS: Did you use that
,m,
( ,) 25 yesterday?
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
571 1 MR. CUNNINGHAM: We skipped over it.
r~x 2 MR. HOLAHAN: It was in the package yesterday, f )
'O 3 but it was not used.
4 CHAIRMAN APOSTOLAKIS: Did I see all this in 5 the -- in 1061?
6 MR. JONES: No, no; this information is in one 7 of the drafts. Actually, it's not in the SRP either.
8 This information is being developed and expanded for 9 putting in the reg. guide.
10 MR. HOLAHAN: Right.
11 MR. JONES: And we have an internal draft of 12 that with some internal comments we've started going 13 through. So what you've got here is -- yeah, we have fs t
N- ) 14 identified now the sources -- potential sources of 15 uncertainty and discussed how we would treat those. And 16 we've got words along those lines about dealing with what 17 you can calculate, model uncertainty on the key 18 assumptions, and then completeness uncertainty.
19 On the completeness uncertainty issue, in all 20 honesty, the way we're dealing with the completeness 21 uncertainty issue is basically, if you go back to the CDF 22 slide, it's just the two regions.
23 CHAIRMAN APOSTOLAKIS: Yeah, but --
24 MR. JONES: It's the Regions 2 and 3 part.
(7 If what Gary said
()_
25 CHAIRMAN APOSTOLAKIS: l l
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W. i WASHINGTON, D C. 20005-3701 I (202) 234 4433 (202) 2344433 I
l
572 1 yesterday is correct that the vertical line really was r x, 2 placed there by the commissioners having in mind what the N.-] state of the art can do, then we don't need that, do we?
3 4 MR. JONES: Well, what we were looking at in 5 completeness is scope, primarily internal, external; and 6 that segregation we're talking there.
7 CHAIRMAN APO3TOLAKIS: I proposed yesterday to 8 include the margin, and you guys said no.
9 MR. JONES: No, what I'm saying -- that's what 10 we were planning on in this other write up.
11 CHAIRMAN APOSTOLAKIS: You're planning to say 12 no or you're planning to think about it?
13 MR. JONES: We're saying no on that part But
(' ') 14 the write up -- I was saying the write up we had primarily 15 was dealing with the completeness as far as the scope and 16 dealing with the Region 2 and 3, and the figures and text 17 -- and accompanying text.
18 The last part, the sources of uncertainty, or I
19 --
20 MR. HOLAHAN. Do we need to say it again?
21 CHAIRMAN APOSTOLAKIS: Yes.
22 MR. HOLAHAN. Okay. Margin for what you know i
23 you don't know, incompleteness in scope. But no I i
1 24 additional margin for things that --
Ch U 25 CHAIRMAN APOSTOLAKIS: Oh, oh, oh .
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433 .
573 1 MR. HOLAHAN: -- that Mr. Murphy hasn't 2 thought of yet.
7_3
! )
3 CHAIRMAN APOSTOLAKIS: Okay.
4 MR. CUNNINGHAM: Within that though, what you 5 know you don't know in terms of scope of the analysis.
6 But we know we don't know much about organizational 7 factors. That's not built into that.
8 MR. HOLAHAN: No.
9 MR. CUNNINGHAM: So that's more than --
10 CHAIRMAN APOSTOLAKIS: So what you know --
11 could have been known, but it is not known.
12 MR. HOLAHAN: But I am not convinced we really 13 know what we don't know about organizational factors.
/~N s
'\~d 14 CHAIRMAN APOSTOLAKIS: Well, I think what you 15 mean is things that could have been quantified with 16 present state of the art --
17 MR. HOLAHAN- Scope.
18 CHAIRMAN APOSTOLAKIS: -- and they have not.
19 MR. CUNNINGHAM: Yes.
20 MR. HOLAHAN: For the scope issue.
21 MR. CUNNINGHAM: Yes, it comes down to --
22 actually, it's really a large scope issue, not the details 23 of the scope.
24 MR. PARRY: Scope in terms of initiating 7m
) 25 points and operational modes, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N W.
(202) 2344433 WASHINGTON. D C. 20005-3701 (202) 234-4433
574 1 MR. CUNNINGHAM: Yes.
,3 2 MEMBER POWERS: But the direction from the 3 Commission seems to me, at least without remembering the 4 words exactly, seem to me to say go to the state of the i
5 art.
6 MR. KING: Best capability of current 7 evaluation techniques.
8 MEMBER POWERS: Which -- I mean, it's not my 9 problem. It's going to be your problem. You're going to 10 have to answer her question of gee, how dare you come 11 before me with something that does not do what I told you 12 to do.
13 MR. KING: Well, I think -- the way I y'~%
- I V 14 interpret that is you do the shut down, the full power, 15 the external events. You don't do things where you don't 16 have techniques to analyze like --
17 MEMBER POWERS: That's the way I interpret it 18 as well.
19 MR. KING: And you judge that against this 20 line.
21 MR. HOLAHAN: Right.
22 MEMBER POWERS: And that's how I would 23 interpret it. But you're saying oh, you can -- if you're j 24 up here close, you can do this, what you've been told to n
25 do; but if you're not close, then you don't have to do
(]
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
575 1 what you've been told to do.
2 MR. HOLAHAN: But you have a restriction for 7s
, \
'~
3 doing that. You're restricted.
4 MR. KING: The limit is different. Full power 5 only, the limit is different.
6 MR. PARRY: Yes, you put a negative margin on 7 what you're allowed on CDF effectively.
8 MEMBER POWERS: I think I understand your 9 position. I would not want to be in your shoes going to 10 my boss to explain that after he had told me to go to 11 state of the art. I mean, it's not my pr.iblem; it's your 12 problem.
13 MR. KING: But the alternative is this becomes Ch
-) 14 a process to make everybody do full scope PRA's, and I 15 don't think that was the intent. ,
I 16 MEMBER POWERS: And I think --
I 17 MR. KING: They may want to do it voluntarily.
18 That's okay.
19 MEMBER POWERS: And I think we --
)
20 MR. HOLAHAN: But it's clear that --
21 MEMBER POWERS: We've written an interesting 22 letter that says for Level 1 scope, the implication is 23 pretty strong it should be a full scope.
24 CHAIRMAN APOSTOLAKIS: That's why I raised the 7
[ } 25 issue.
w/
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
576 1 MEMBER POWERS: And I think we feel that's the
,- . 2 state of the art, and at ler.st a vocal --
l )
\ /
3 MEMBER KRESS: But in order to know where to back off of it -- from the margin, they have to know what 5 the full scope results aro likely to give them.
6 Otherwise, you can't do that.
7 MEMBER POWERS: That's the next step, having 8 said okay, it's okay to back off on what you were told to 9 do , --
10 MEMBER KRESS: How do you know how far to back 11 off.
12 MEMBER POWERS: -- how far do you have to back 13 off, and we'll get to that discussion in a second. Right IV h now, what I --
14 15 MEMBER KRESS: But that's the natural follow 16 on.
17 MEMBER POWERS: Yes, absolutely. And what I'm 18 wondering, is this something we need to discuss among 1
19 ourselves? You've got an SRM that says go do this to the l
)
1 20 extent that the state of the art will allow. Don't we 21 have to advise the Commission on the idea that we would do I l
22 less than what the -- I mean, maybe it's the thing that we l
23 support.
24 And in fact, I think -- I'm clearly very p 1 Q 25 sympathetic that because of just what we said, we 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
577 1 want to make this transition to risk based regulation so !
2 revolutionary and so hard to do that it deters people.
( ~)
Q) 3 Because we inherently believe that it leads to safer 4 plants and better protection of the public that we're 5 sympathetic to this idea --
6 MEMBER KRESS: It goes to this question of 7 whether we would have qualitative evaluation as opposed to 8 full quantitative.
9 CHAIRMAN APOSTOLAKIS: What would you leave 10 out that would require a qualitative evaluation?
11 MEMBER POWERS: External events.
12 MEMBER KRESS: I think external events would
,_ 13 be one that --
i )
Y' 14 CHAIRMAN APOSTOLAKIS: Yeah, I would say 15 external events.
16 MEMBER KRESS: And I would have to think about j l
17 it, but I even suspect at this point in time that the shut f 1
18 down risk.
19 MEMBER POWERS: I think there are many, many 20 things that come up that I could make qualitative 21 arguments on to say these particular changes I'm proposing 22 here will really have no impact plus or minus on my shut 23 dbwn operations.
24 MEMBER KRESS: Yes.
k s) m 25 MEMBER POWERS: I'm quite certain that 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
578 j 1 the case. And I'm very sympathetic with this, but I think
,s 2 it's something that you better have a defense for.
( ')
3 MEMBER KRESS: Technical defense for.
4 MEMBER POWERS: And why you want to do it.
5 And if we're -- I mean, we could leave the question to 6 you, because we've already written --
7 MEMBER KRESS: We know what we've already 8 said.
9 MEMBER POWERS: We've already said do a full 10 scope Level 1 that's within the state of the art.
11 MEMBER KRESS: But we really didn't define 12 what we meant.
13 MEMBER POWERS: No, we did not. We're not
.rm l'
's- 14 terribly clear. But we can always -- out ambiguity is 15 steadied here maybe. If we're going to be supportive of 16 the staff in this position that they've taken, --
17 CHAIRMAN APOSTOLAKIS: We are kind of silent, 18 Dana; but you can use the same arguments we have here to 19 argue that shut down and external events --
20 MEMBER POWERS: Our vagueness is steadied 21 here.
22 CHAIRMAN APOSTOLAKIS: Yes, yes.
23 MEMBER POWERS: But sooner or later, you have 24 to fish or cut bait in this debate.
(p) 25 CHAIRMAN APOSTOLAKIS: Right; oh, 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
579 1 MEMBER POWERS: And you have to say yes, we 7._x 2 like this idea that the staff has taken to back off and
)
- ~# 3 quibble with how well they know the back off position.
4 MEMBER SEALE: Well, you could --
5 MEMBER POWERS: And if we do that, we have to 6 have a reason for that. I mean, I expect to get -- I 7 would expect to get interrogated on this. And I think I 8 do have a reason, but I would sure like to get that set 9 down very clearly.
10 CHAIRMAN APOSTOLAKIS: Listen to what -- many 11 licensees do not have access to the technologies for 12 facile conduct of full scope Level 2 or Level 3 PRA's, the 13 three power operations, low power / shut down operations, as s
t )
\/ 14 well as accidents initiated by external events.
15 MEMBER KRESS: A Level 1. You're talking
, 16 about Level 1.
17 CHAIRMAN APOSTOLAKIS: Level 1. This is Level l 18 1. It says Level 2 or Level 3, shut down, as well as l l
j 19 accidents initiated by external events.
20' MEMBER POWERS: The intention of the author, I 21 will assure you, was exactly as he interpreted it.
22 CHAIRMAN APOSTOLAKIS: Yes.
23 MEMBER POWERS: That it's Level 2, 3 for power 24 operations, and it's anything for shut down, and anything
(-
( j 25 for accidents. And I firmly believe that, and I believe l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
580 1 the introductory sentence in that paragraph that says you
, 2 want this to be an evolutionary process and not so
,3 i )
3 revolutionary that it deters people.
4 CHAIRMAN APOSTOLAKIS: So it seems to me when 5 we were writing this letter and we're saying CDF, we had 6 in mind internal events and we didn't say it. Certainly 7 this argument here on page two will justify this new 8 position, if it is indeed new.
9 MEMBER POWERS: But we have to make that 10 argument.
11 CHAIRMAN APOSTOLAKIS: Yes, make it 12 explicitly.
13 MEMBER POWERS: I mean, I think -- and there
\/ 14 are several of these things --
15 MEMBER SEALE: And I could say that perhaps 16 the way to -- one way to make that argument is that the 17 interpretation of the commissioners' letter is the state 18 of the art as it resides within the applicant's 19 capability.
20 MEMBER POWERS: Absolutely, Bob. And I think 21 we've got a -- there's an excellent precedent that I bear 22 in mind all the time about this. The Rand Corporation in 23 the late '60s did an interesting study of first of a kind 24 engineering and how it was successful and not successful. l
.m There are lots (J) 25 And they came up to lots of conclusions.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIDERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 )
l
581 1 of problems with first of the kind engineering.
2 But what sticks in my mind most is their
,. S I/
i 3 definition of first of the kind. It was not you were the 4 first person on the face of the earth to do it. It was 5 the first time your particular organization had done 6 something.
7 MEMBER SEALE: Yes, yes, yes.
8 MEMBER POWERS: And they did not find big 9 differences between people doing something for the first 10 time on the face of the earth versus an organization doing 11 it for the first time. And I think that's the idea you 12 have to bear in mind, that the applicant and licensees 13 have this problem.
( \
\~/ 14 This is a first of the kind engineering job 15 for them, --
16 MEMBER SEALE: Exactly.
17 MEMBER POWERS: -- and you have to go slow. I 18 mean, it's the reason that it takes so long to translate I
19 laboratory science into industrial practice.
i 20 MEMBER SEALE: You have to calibrate your I
21 expectation --
22 MEMBER POWERS: That's right. ;
23 MEMBER SEALE: -- against their existing !
24 capabilities.
,e-(,)3 25 MEMBER POWERS: That's 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
582 1 CHAIRMAN APOSTOLAKIS: We saw an example
,s 2 yesterday. They were ecstatic with the risk monitor, and i i
\ ,/
3 all it does is internal events -- Southern California 4 Edison. All it does is internal events. And they thought 5 -- in fact, they made some good arguments, you know, that 6 now the culture at the plant is changing.
7 Now I wouldn't want to miss that opportunity 8 just because I want to see them doing external events.
9 So, I think that's great. So all we need to do is state 10 it explicitly. Okay, I think the ground work is here.
11 MEMBER POWERS: Well, I think we have to state 12 it explicitly, but I also think that I would be 13 uncomfortable sitting in front of the boss without a pre-p
\* 'l 14 articulated and technically defensible defense -- I mean 15 technically sound defense on that.
16 MEMBER KRESS: That involves the margins.
17 MEMBER POWERS: I think it involves cultural 18 considerations. I mean, --
19 CHAIRMAN APOSTOLAKIS: Yes.
20 MEMBER POWERS: The mechanics is good. I 21 mean, this is a rational thing. But why did you do it 22 this way rather than going all the way? It has to do with 23 some of these cultural and first of a kind engineering 1
24 things. l
/~'N
( ,/ 25 CHAIRMAN APOSTOLAKIS: Are we all convinced 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
583 1 that the 10-* is intended only for things we can quantify?
,- 2 If you say yes, I'll go along, just as I did yesterday.
~' 3 MR. HOLAHAN: I think the questions are a 4 little different. I'm not -- when you say intended, I 5 think we're free to establish a new intent.
6 The Commission's policy statement was directed 7 towards the whole of the regulations for all of the 8 plants. Okay, and that's the context in which you should, 9 you know, analyze everything to the state of the art in 10 making judgements about, you know, the current and future 11 regulations.
12 Now we have, you know, a slightly different 13 question, which is how do you make decisions on individual ry w 14 issues on individual plants. And I think what we've 15 chosen to do is to use the same number in a similar way.
16 But I don't think it flows exactly logically from the l
17 policy statement in such a way that we're constrained to 18 do it with the same number or with the same definition. I 19 MEMBER KRESS: It's a new policy.
20 MR. HOLAHAN: I think it's a new policy and 21 it's been forwarded to the Commission.
22 CHAIRMAN APOSTOLAKIS: I'm having the problem 23 I stated yesterday that if we interpret it that way and 24 then -- we manage to quantify something that is not
(-
! j 25 quantifiable right now, should we move the limit? Because NEAL R. GFN3SS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
584 1 now we can do more.
2 MEMBER FONTANA: As I remember the history of
( \
\"' /
3 the thing, originally, on a core damage frequency, they 4 were talking about 10'3 5 MR. KING: I don't remember that.
6 MEMBER FONTANA: And then they moved it -I 7 think one reason they moved it to 10 is to take into 8 consideration some of the -- I might be wrong, but I seem 9 to remember the first numbers that they were talking about 10 was 10-3 I don't know if that means anything, but --
11 CHAIRMAN APOSTOLAKIS: Well, maybe they were 12 doing the similar calculations for those that Rick has 13 done. They came up with 10-3 They didn't like it, and f)
I <
(_/ 14 they went down.
15 MR. HOLAHAN: Yes. If you back calculate from l l
16 the QHO's, I'm not surprised you'll get a 10'3 Then if 17 you think about the implications of living with the 10'3, i
18 you don't stay there very long.
I 19 CHAIRMAN APOSTOLAKIS: Maybe that's also -- i 1
20 MEMBER SEALE: And built into it is the fact 21 that you know that gives you some margin.
l 22 CHAIRMAN APOSTOLAKIS: Okay, so what is the 23 conclusion here? That we seem to agree that what is 24 absolutely required is a Level 1 internal events PRA, A Now,
( ) 25 right? Not necessarily with uncertainty analysis.
NEA!. R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
585 1 that's pushing it too far.
n 2 MR. KING: We've also added fires -- internal
/
I 3 fires.
4 CHAIRMAN APOSTOLAKIS: Yes, yes; fires are 5 part of the internal development. It's about time we --
6 MEMBER FONTANA: What's pushing too far?
7 CHAIRIGN APOSTOLAKIS: Well, let's see what 8 Gary says.
9 MR. HOLAHAN: First, remember that there's 10 flexibility in the approach that says you don't always 11 have to do a quantitative analysis. We're saying for 12 those cases in which you are doing a quantitative analysis 13 and making a judgement based on it, what kind of analysis
,/
V' 14 should it be?
15 CHAIRMAN APOSTOLAKIS: Well, if you are to use l
16 these new reg. guides, you have to do a Level 1 internal 17 events PRA, don't you?
18 MR. HOLAHAN: No.
19 MR. KING: Clearly reductions in risk could be 20 all qualitative.
21 MR. HOLAHAN: It could be based on risk 22 insights from general studies or from the IPE insights 23 reports, as these things are of a low significance; and if 24 it's -- I think for some decisions, a qualitative approach
,, o 25 would be acceptable. And then you'd never to get to 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
586 1 part of the reg. guide.
fs 2 But if you are using a quantitative approach,
( )
\J then the question is, what kind of analysis.
3 4 CHAIRMIJ4 APOSTOLAKIS: If you are to do that, e 5 Level 1 internal events including internal fires.
6 MR. HOLAHAN: Let me say a little bit -- if 7 you want to use this decision approach in which you would 8 allow changes as large as -- knowingly as large as 9 somewhere near 104, you would need an internal events with 10 fire analysis. I think that's what --
11 CHAIRMAN APOSTOLAKIS: You mean you could go 12 below -- if you're close to 10^6 , you may not need fires.
13 That's a fundamental problem.
/ \
( l
'k / 14 MR. HOLAHAN: I think -- well, the way it's 15 structured is Region II says internal and fires --
16 including fires. But it says -- j 1
17 MR. JONES: And uncertainty. l l
l 18 MR. HOLAHAN: With uncertainty. But Region I, l 19 I think, is internal with fires, but allows for a point 20 est inia t e .
21 MR. KING: Right.
l l
22 MR. JONES: And without a detailed uncertainty l
23 analysis. l 24 MR. HOLAHAN: Okay. Indicating that you have
,/ 3
) 25 to use the figure in combination with the text.
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
587 1 CHAIRMAN APOSTOLAKIS: What I like is your fm 2 SRP. The SRP states it the way it's supposed to be
( )
v 3 stated.
4 MEMBER FONTANA: Actually, what I did is I 5 wrote the text into the figure.
6 CHAIRMAN APOSTOLAKIS: Which is --
7 MEMBER KRESS: A real compromise. How would 8 you like that?
9 MR. JONES: We had one of those at one time.
10 CHAIRMAN APOSTOLAKIS: I have to find it 11 again. "Where applicable, the modeling and the 12 quantification of the effects of the change should also be 13 reviewed to ensure that the models are appropriate, and t
p_ )
'd 14 that the results are supported by plant and/or industry 15 data."
16 And where spending now an inordinate amount of 17 time trying to predict in advance for what case, what 18 models are appropriate, what kind of analysis -- I mean, 19 why can't we say that?
20 MR. JONES: We did. We said it in the reg.
21 guide too. The reg. guide says --
22 CHAIRMAN APOSTOLAKIS: We spent all this time 23 doing this.
24 MR. JONES: No, the reg. guide starts off with (O
Q / 25 saying that appropriate consideration -- justified NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
588 1 proposal be appropriate for the nature and the scope of
,_.s 2 the change. The scope and quality of the engineering
/ ;
i' /
3 analysis will be appropriate for the nature and scope of 4 the change. Appropriate consideration given uncertainty, 5 used their judgement.
6 CHAIRMAN APOSTOLAKIS: If they see all that 7 stuff later -- I mean, this appropriate stuff disappears.
8 They're going to go with the figures and the text.
9 MEMBER KRESS: If they don't say that, 10 somebody's going to ask them what in the heck they mean by 11 appropriate.
12 CHAIRMAN APOSTOLAKIS: Appropriate means that 13 when you give me the 10", you have a convincing argument l 14 that it is indeed 104 So if somebody who knows their 15 plant comes and says excuse me, but by changing this, it 16 may affect your fire response.
17 MEMBER KRESS: And what argument will convince 18 you?
19 CHAIRMAN APOSTOLAKIS: Well, what is the --
20 MEMBER KRESS: That's what they're calling 21 for.
22 CHAIRMAN APOSTOLAKIS: No, no; I don't think 23 that's convincing at all. I mean, if somebody tells you 24 that by changing this, I need this piece of equipment to n
( ,) 25 respond to a fire in this room, that's a convincing NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
589 1 argument. I mean, you can't go beyond that unless n 2 somebody doubts or disputes it.
I (v ) 1 mean, how do we convince people, right? We l 3 1 4 don't convince them with figures. We convince them by -- l 5 and I think we are spending a lot of time arguing, you l 6 know, what should you do in two or three, where should you 7 draw the line.
8 I like the idea -- by the way, your idea 9 yesterday of eliminating those lines. Keep the two and 10 three and one, but don't draw any lines.
11 MEMBER KRESS: How will we know where they are 12 then?
13 CHAIRMAN APOSTOLAKIS: You don't have to know.
O \
i C/ 14 MR. MARKLEY: Somebody's going to draw them if 15 they're not there, and they may not put them in the right 16 place.
17 CHAIRMAN APOSTOLAKIS: But that's my whole 18 point, that there isn't a place where they belong. They 19 don't belong anywhere. All you are saying is look, as you 20 move away from that vertical line, the requirements of 21 doing certain things become weaker and weaker. That's all 22 you're saying.
23 Now, where it is weak is the traditional, 24 prescriptive approach that we're all used to and will
(- 25 love, and we want to bring it over here too. The best way
( !
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344 433
590 1 is page 13 of the SRP: where applicable, do the right 2 job. But do the appropriate job. j
! )
3 MEMBER FONTANA: In practice, what's going to 4 happen is the first case will set the precedent, and 5 that's where --
l 6 CHAIRMAN APOSTOLAKIS: Well, --
7 MEMBER FONTANA: They've got to regulate to 8 something..
9 CHAIRMAN APOSTOLAKIS: Nobody in the pilot 10 program says use this figure, is that correct? Because 11 they seem to be doing fine. They didn't need anybody to 12 tell them the line is there and not a little bit to the 13 left.
/ 'N l \
'v 14 Yes, sir?
15 Joe Murphy.
16 MR. MURPHY: I have to beg to differ a little 17 bit. If we take the lines off the figure, we have to 18 recognize -- and I think Dana put it well earlier. We 19 have people that are essentially doing first of a kind 20 engineering.
21 And we're putting them in a position where 22 we've giving them a guidance that basically says if you 23 were expert for a number of years in this area and you had 24 an expert reviewer with the same kind of experience, I'm
,a (j 25 sure you would know what those numbers mean.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
591 1 In other words, one, two, and three, and four, 2 if I took the lines off. But for the average utility and 7s N~~
3 for the average reviewer, I think what you're saying is 4 bring me a rock. You do it, and I'll tell you if you did 5 it well enough; but I'm not willing to give you any 6 guidance as to what well enough means.
7 I think if we leave the lines on and 8 characterize them as I would characterize them, that these 9 are our guidelines. We are willing to depart from those 10 guidelines from a good reason. The degree to which we are 11 willing to depart from the guidelines depends on how close 12 you are to the guidelines when you get done.
13 Certainly we're not going to look at 1 x 10"
/ N
\
i -) 14 as different than 1.1 x 10". We're not that crazy. But 15 we're sure going to look at 5 x 10 4 as different than 1 x 16 10". And you're going to need a much better argument to 17 get -- to change it by a factor of five than you will by 18 one or two percent.
19 I think as long as we put the lines --
20 interpret the lines as guidelines that are flexible 21 guidelines and there's an expression of the degree of i 1
l 22 where you put this, how far you get away from that line, I 23 think, is a much better prospect than taking the lines off i
24 and leaving in the roman numerals.
~'
N I iq./ ) 25 CHAIRMAN APOSTOLAKIS: Okay, now let me show NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-5701 (202) 234-4433
)
592 I to you why what you just said is self-inconsistent. On j l
3 2 the one hand, you and many others -- not just you -- l t i
\
'~'/
3 including the EPRI presenter this morning -- take a 4 paternalistic approach and say you people really don't 5 know much, so we'll give you guidelines.
6 I don't expect you to be sophisticated. I 7 mean, the users of this. So I'll tell you where these 8 things are. And at the same time, you expect them to be 9 sophisticated enough to understand that 3 x 10'5 could be 4 10 x 10 4 in some instances -- could be less than that in 11 other instances, which takes a lot of experience, 12 And you have that experience, Joe, because you 13 have been in the middle of PRA's from the days of Adam and I \
14 Eve. So this, to me, sounds contradictory. And it was 15 the same thing this morning.
16 Why are you doing this new algebra, high, 17 medium, low? Oh, this is for utility people.
18 MR. JONES: Let me --
19 CHAIRMAN APOSTOLAKIS: What does that mean?
20 MR. JONES: Let me give you a different 21 perspective. I'm the Branch Chief of the PRA Branch. And 22 I'll -- now. I don't know for how long, maybe forever.
23 But my experience in PRA is minimal. Okay, and I've been 24 doing traditional engineering, deterministic reviews for
(_,/ 25 quite a while.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. O C. 20005-3701 (202) 234-4433
593 1 MEMBER KRESS: Thermal hydraulics.
,~s 2 MR. JONES: Thermal hydraulics.
I )
'- 3 MEMBER POWERS: Since Adam and Eve?
4 MR. JONES: Since Adam and Eve.
5 MR. HOLAHAN: Before, I think.
6 MR. JONES: Thank you. But, you know, from a 7 reviewer's perspective, from my simplistic perspective in 8 trying to apply this, it would be extremely helpful to me 9 to have some general guidance on when do I really have to 10 worry real hard about uncertainties and how good it has to 11 be.
12 And I think these kind of regions are useful 13 to me in doing that. And I think that if I totally throw p,
\- ~ 14 it in words and say throughout this range, you just have 15 this very, very flexible approach and you do a little bit 1
16 more here, a little bit less there, I think it -- I don't l 1
)
17 think it's going to be helpful to a reviewer. i 18 CHAIRMAN APOSTOLAKIS: Bob, what I'm saying is 19 we're becoming awfully prescriptive. I think you could 20 get the same kind of guideline either by dropping the line 21 completely or putting some shade there.
22 But the reason why you would not make a big 23 mistake here is because you've been in the middle of it 24 for a while now, and you have to produce something.
,q
( ) 25 MR. JONES: Oh , yes; oh, 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
594 1 CHAIRMAN APOSTOLAKIS: Other people haven't
,_ 2 spent all this time --
/ \
(' 'l 3 MEMBER KRESS: But the way to put that shading 4 in is just what Joe said.
5 CHAIRMAN APOSTOLAKIS: Which is?
6 MEMBER KRESS: Just say in words that the 7 closer you are to this line, -- I mean, that's --
8 CHAIRMAN APOSTOLAKIS: If these people don't 9 know enough to realize themselves what this document says, 10 that these are the models that are appropriate for this 11 situation, why would they be sophisticated and 12 knowledgeable enough to realize that the line is not a 13 line?
sa b 14 Now we are making tremendous progress.
15 MR. JONES: I was going to say let me --
'. 6 MR. HOLAHAN. Those are not cross outs; those L7 are --
~8 2 CHAIRMAN APOSTOLAKIS: I realize thet.
It MR. HOLAHAN. -- those are shades.
20 MR. JONES: But let me make another comment 21 just -- I think that's germane to this. We went to CRGR 22 with this package two weeks ago. I don't know. I get my 1
1 23 meetings mixed up -- to go through and explain this same 24 kind of decision criteria to them.
,, m.
(%j ) 25 We probably spent a half an hour discussing 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
i 595 l
1 the bright line at 1E-4 minus four, and why are we
,,, 2 allowing a change of 1E-5 when I'm at 1E-4, which puts me
/s \
/
I say I can't have change and exceeds
3 into the region.
4 the safety goals. And we went round and around, and we i 1
5 tried to explain that we wouldn't do this. l
\
6 It didn't come off very well in that. But we 7 are going to be adding words in there to deal with some of 8 the precision issues for exactly that reason. So I think 9 there is a sensitivity to the lines. I mean, and I would 10 agree that an unsophisticated review will have difficulty 11 with bright lines.
12 And I think that's what we saw when we went to 13 CRGR. I think an unsophisticated -- what is that?
O, kj 14 (Laughter.)
15 I was going to say it would also have 1
16 difficulty the other way without them. So I think we're j 17 going to have to discuss precision as part of what's here.
18 CHAIRMAN APOSTOLAKIS: For years now, I've 19 seen this attitude that you have to simplify these things i
20 for the utility people. And John here can give us is 21 perspective. For some reason, make it as simple as I 22 possible. Now, I would remind you of what Einstein said.
23 "Make it as simple as possible, but not 24 simpler."
,7 t ) 25 There's a lot of depth in that. I remember NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1223 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 2344433 J
596 1 when we finished one with the very first industry response
, 2 on PRA's. John Garrick and Caplan were talking about the i) risk curves, and the guy -- our contact with the utility o 3 4 hit the roof. He said, "You're going to present this to 5 my vice president?"
6 I like when they say "my". I don't know why 7 they say "my". But he said my vice president. They will 8 never understand this. The vice presidents, I can assure 9 you, understood it within ten minutes. They liked it, and 10 now everybody's using those. So, why -- and again, this 11 morning we saw that.
12 High, medium, low, and a new algebra now. l 13 Maybe we'll have axioms pretty soon. Why? Because the ,
/, \ l, t
Nj 14 utility people have to see this very simple thing. They ;
l l
15 don't understand anything else. Although they had some 16 other arguments too. So I don't understand that.
17 I think -- first of all, would you go to a guy l l
18 who doesn't know thermal hydraulics and ask him to do a 19 thermal hydraulic calculation?
20 Why would you go to a guy who doesn't know 21 anything about PRA and ask him to do a PRA calculation? I 22 don't understand that. Why?
23 MEMBER POWERS: George, you have to understand 24 that the many, many situations arise where people are f'3
( ,) 25 asked to do things at plants that are -- they're not NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
597 1 really specialists. They're generalists.
2 CHAIRMAN APOSTOLAKIS: I have never heard at
,f-~3 a
3 any of our meetings anybody coming here, presenting 4 something that's related to thermal hydraulics, and say 5 now I'm going to simplify it because plant people are 6 going to use it. I never heard that about the steam 7 generator rule, nor anything else.
8 But when it comes to PRA, we have to sit here 9 and try to anticipate the situation. The guy won't 10 understand what uncertainty is. The guy won't understand 11 -- well, I mean, he has no business getting into this then 12 if he doesn't understand --
13 MEMBER KRESS: I think we need to let Joe
,n
(' ') '
14 finish his comments. He needs to say something.
15 CHAIRMAN APOSTOLAKIS: How about my comments?
16 (Laughter.)
17 MR. MURPHY: I'd like to defend my honor a 18 little bit since I was told my position was illogical. I l 19 think the premise you're operating under is incorrect, to 20 be honest with you. If I take the lines off the figure )
21 and I just leave the roman numerals, the question is not 22 whether somebody in the utility can understand this and l
23 what that means in terms of a general region.
24 Certainly they can. But what they can't
,~.
( ,) 25 understand is what I mean by it as the regulator. I'm 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
598 1 guy that's going to review their submittal, and all I've 2 got is a number on a graph -- a point on a graph. And
)
(
3 what do I mean by it? And is that the same that Bob means 4 by it, and is that the same that Gary and Tom mean by it?
5 And what we need instead -- as I say, if we do 6 it the other way and we put a structure of a line and --
7 but then identify that the line itself - it's a fuzzy 8 line. And it would be nice if we could draw it in some 9 manner that it was heavy in the middle and it sort of 10 dwindled off to oblivion.
11 CHAIRMAN APOSTOLAKIS: Why do you say it would 12 be nice? We can do that.
13 MR. HOLAHAN: We can do that, f
V 14 CHAIRMAN APOSTOLAKIS: We can do it, and 15 that's what I've been advocating.
16 MR. MURPHY: I had not heard you advocate 17 that.
18 CHAIRMAN APOSTOLAKIS: Well, you were not here 19 yesterday. I 20 MR. MURPHY: I wasn't here.
21 CHAIRMAN APOSTOLAKIS: My point, Joe, is that 22 there are two extreme positions here. One extreme, which 23 I like personally is what's stated on page 13 of the SRP.
l 24 Where applicable, do the right thing.
n,
) 25 MR. MURPHY: Amen. But what does that mean?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234 4433
T 599 1 CHAIRMAN APOSTOLAKIS: Okay, that's one
,_ 2 extreme. Then you give me these arguments, and I say
/ ;
(/ 3 well, those guys have a point. The other extreme, in my 4 opinion, is this -- especially before Gary worked with it.
5 Now you put the approximately 10"3, you drew a few lines.
6 Then it seems to me we're beginning to converge.
7 I could live with a figure like that, in other 8 words.
9 MEMBER BARTON: George, I think -- utilities 10 don't want over prescription. But I think you're going to 11 find that if this helps -- everybody in the utility's not 12 a PRA expert. You've got to remember that. People that 13 are developing changes to submit to the Commission don't l
(~'N
(_ 14 necessarily have a lot of PRA experience.
15 CHAIRMAN APOSTOLAKIS: But shouldn't they have l
16 access to it, John?
17 MEMBER BARTON: They may have access to it, 18 but they, themselves, may not be sophisticated in the 19 process. A chart like this, I think, would help them in 20 developing the rational for the change and the riske 21 associated with submittal of change. I think it will also !
22 help the reviewer on the NRC's part.
23 I think there's -- this allows the 24 communication channel, I think, between the regulator and A
( ) 25 the licensee in a particular change submittal.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W-(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
600 1 CHAIRMAN APOSTOLAKIS: I can live with this.
2 MR. PARRY: George?
i 7 ,l
'/
3 CHAIRMAN APOSTOLAKIS: I can live with this.
4 But -- yeah, I'll come to you, Gareth, in a 5 moment.
6 I have yet to talk to a utility person who 7 started talking about PRA or didn't understand PRA. I 8 don't know where people get that impression.
9 MEMBER POWERS: I've talked to a lot of folks 10 that don't.
11 CHAIRMAN APOSTOLAKIS: Well, you were talking 12 to them about what?
13 MEMBER POWERS: PRA.
p b) 14 MEMBER SHACK: When you came to digital 15 instrumentation, do the right thing wasn't nearly good 16 enough for you.
17 (Laughter.)
18 CHAIRMAN APOSTOLAKIS: You don't talk much, 19 Bill; but when you do, --
20 (Laughter.)
21 -- you are a nuisance.
22 MR. KING: There's another aspect of this that 23 has nothing to do with whether they understand it or not 24 that went into coming up with something like this, and it fm, 25 has to do with -- you know, one of the main purposes 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
i 601 1 this whole risk informed exercise is burden reduction.
,-y 2 You know, analysis costs money.
0 3 The IPE's cost money. For them to submit an 4 application for a CLB change, it costs money. And we 5 tried to come up with a process that would say how could 6 we build upon what's out there today, the IPE's which have 7 a limited scope -- you know, what could that be used for?
8 Point estimate -- is that, you know, something 9 where you don't have to do the full blown uncertainty 10 analysis. And people do that today. They submit 11 applications for things based on point estimate analysis.
12 Where can we use that?
13 That was a major consideration in trying to
.o
( )
V 14 come up and define those regit 15 CHAIRMAN APOSTOLAKIS: And I fully appreciate 16 that.
17 MR. KING: And it's not so much understanding 18 -- you know, do they understand it, do we understand it.
19 CHAIRMAN APOSTOLAKIS: Okay, but I fully 20 appreciate that motivation and I agree with it. I agree 21 with it. But it seems to me you can handle that 22 differently. But, I repeat, if you do something like 23 that, I will --
24 MEMBER KRESS: Put those little crossbars --
/ \
j 25 CHAIRMAN APOSTOLAKIS: Or something like 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
602 1 shade or something.
2 MEMBER SEALE: Why not use dashed lines
(
r3
)
%_/'
3 instead of solid lines? .
4 CHAIRMAN APOSTOLAKIS: Oh, I'm sorry, Gareth.
5 MEMBER KRESS: If you put a shade in there, 6 I'm going to argue with the gradation -- the rate of 7 gradation.
8 CHAIRMAN APOSTOLAKIS: Well, that's because 9 you want to get back at me.
10 (Laughter.)
11 MEMBER KRESS: I mean, it's the same thing.
12 MR. HOLAHAN: We'll say not to scale.
13 CHAIRMAN APOSTOLAKIS: He will say that too.
I \
14 MR. PARRY: There's something on this figure, 15 I think, is that it's too complicated. We try to put too 16 much on one thing.
17 MR. HOLAHAN: I think we need two figures. I
\
18 MR. PARRY: We need two figures. And the 19 first figure has the single line along the top and a 20 single line down the right-hand side. That defines what i
21 the -- if you like, what the goals are for CDF and delta )
l 22 CDF.
23 And what the internal lines are is really 24 there to help the implementation -- help the decision
~3
(_) 25 maker to understand what he needs to do to convince the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
603 1 NRC that the change is acceptable.
2 That's the only reason for having --
'~'
3 CHAIRMAN APOSTOLAKIS: There's many things 4 that you can do.
5 MR. PARRY: The internal -- that's what 6 they're there for. So in a sense, sticking an internal 7 line up 3 x 10 4 is probably too -- I agree, is probably 8 too prescriptive. Because, for some plants, the stuff 9 that we've left out muy be very small; for other plants, 10 it may be large.
11 So that one at least ought to be a little 12 moveable. But I think the external lines, the thick ones, 13 really are the goals, and those are the things that really
,s i i
's l 14 do need to be communicated. j 15 CHAIRMAN APOSTOLAKIS: Except for the --
16 MR. PARRY: Well, I think even that needs to 17 be communicated. But the --
18 MR. HOLAHAN: It's thinner.
19 MR. PARRY: But it's a definition of small. ;
20 And so the real issue becomes that we've got to do a 21 better job of prescribing how, with a limited scope, the l 4
l 22 PRA -- you can convince the NRC that you have an 1 23 acceptable change.
24 And that includes then -- then we'll have to 25 show how to address uncertainties, how to address 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
604 1 missing scope issues, and things like that.
-s 2 CHAIRMAN APOSTOLAKIS: Okay, okay; I think we t x Q~'.) 3 discussed it enough.
4 Any other comments on this issue that have not 5 been raised?
G Steven?
7 MR. DINSMORE: Yes, Stephen Dinsmore from NRR.
8 I'm going to agree with Dr. Apostolakis for a little 9 different reason. It seems to me if you draw a nice clean 10 line at 1 x 10'6, that -- if the utility has a 1.1 x 10-'
11 PRA, the first thing they're going to do is tweak it down.
12 CHAIRMAN APOSTOLAKIS: Exactly.
13 MR. DINSMORE: So, if you leave it gray, they p\
Cl 14 come in as 1.1, they won't worry about it. I l
l 15 CHAIRMAN APOSTOLAKIS: That's the reality. l 16 MR. DINSMORE: Nice line. We're going to i
1 17 spend a lot of time arguing with them about do you just l l
18 change your number, why did you do that.
19 CHAIRMAN APOSTOLAKIS: Well spoken.
20 MR. GOLAY: It strikes me that there's an 21 analysis that's still needed here before you can answer l 22 the questions that are being discussed, which is 23 effectively the arguments for fuzziness have to do with 24 some way to treat -- recognize factors of uncertainty.
f))
( 25 There are other factors of uncertainty that show up NEAL R. GROSS COllRT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W. ,
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
605 1 outside of a PRA scope.
,s 2 For example, let's say the SALP scores as a
( )
'# surrogate for lots of other things. I think until 3
4 someone's worked through how you would take into account 5 all of the factors which would affect your bottom line 6 assessment of the safety of a plant and put it in the 7 space of this figure, you don't really know whether this 8 figure is workable.
9 And I would suggest as an exercise figure out 10 whether you can actually use this figure, taking into 11 account, let us say, the SALP scores, just as one trial 12 balloon. I think you can't, and that you're much better 13 off actually trying to use the two coordinates here as f~s V) 14 input into a subjective evaluation and leaving it 15 subjective, rather than trying to make it definitive in 16 this way.
17 Because fundamentally, you end up always i 18 having some arbitrariness in the ultimate answer if you I
19 try to be quantitative about it.
i l
20 And secondarily, if you try to use this, 21 someone's going to eventually have you in court having to 1
22 defend the bright lines. j 23 MR. HOLAHAN: Well, I think an important 24 element of what you said came up this morning also, which n 25 is this is not the one and only decision element; that 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
606 1 have to take other things into consideration. Now, g~g 2 whether you convert the SAI P scores into delta CDF and try 4
%) 3 to put them on the chart, c2 just say I've taken that into 4 consideration also, --
5 MR. GOLAY: The latter aLrikes me as the 6 reality.
7 MR. HOLAHAN: Well, it is the reality because 8 I can tell you, among other things, some of the Millstone 9 plants were the pilot applications, and they're not there 10 anymore because we've made a judgement that this isn't a 11 good time to be reducing requirements on the plants, that 12 we have some concerns about how it's being managed.
13 So that goes into the decision process. But
/% '
i L
1 14 you didn't convert that to the -- you didn't put it on the j 15 chart to do that.
16 CHAIRMAN APOSTOLAKIS: You can ask the same 17 question Mike asked in a different way. Suppose you get 18 two applications from two plants. One is -- what, SALP 19 has three categories?
20 MEMBER BARTON: Yes.
21 CHAIRMAN APOSTOLAKIS: One -- and the other 22 one is in number three. And their numbers are more or 23 less the same.
24 MR. HOLAHAN: Right.
,a
) 25 CHAIRMAN APOSTOLAKIS: Your decision 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
607 1 probably be different.
2 MR. HOLAHAN: It could be, depending upon the (3
(v) __
4 CHAIRMAN APOSTOLAKIS: Depending upon the 5 issue.
6 MR. HOLAHAN: Yes.
7 MEMBER KRESS: I think all you're saying is 8 there are additional screens besides this. And if it 9 passes this screen, then you can enter into this.
10 CHAIRMAN APOSTOLAKIS: Yes.
11 MEMBER KRESS: Which is perfectly reasonable.
12 MR. HOLAHAN: That's the integrated decision 13 making process.
- x 1
14 CHAIRMAN APOSTOLAKIS: And to bring up another 15 issue, could it be possible to go to the IPE's or the 16 people who did the IPE's and maybe can give you a quick 17 answer to get some sort of understanding as to what a 18 change of 10"" means? I mean, why did we decide that 19 small?
20 For example, I think people work with orders 21 of magnitude as if they were, you know, nothing. And the 22 reason is that we are so used to working with 10-2,3 As,6 __
23 they say ch, no, my judgement is this is not 10'5, it's 24 10~3 Okay, if it's your judgement, I'll use it. I mean,
/~
'__! 25 two orders of magnitude, just like 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 l
608 1 So, is there any way we can go to this ex 2 spectrum of PRA's that we have and try to understand and
< )
t 3 calibrate ourselves? If I remove a whole train from a 4 system, what kind of change do I expect in a PRA -- in a 5 CDF?
6 If I remove a whole system, what do I expect 7 to see? Or, if I start playing now with tech specs and so 8 on. Gareth, do you have any idea about that?
9 MR. HOLAHAN: There are a number of studies 10 like that. I don't think it's in the IPE's. The 1150 11 follow up studies have those kind of sensitivities in 12 them, taking equipment out of service and things like 13 that.
14 CHAIRMAN APOSTOLAKIS: Well, that would be a 15 useful input as well. But I think the IPE's being, you 16 know, the latest --
17 MEMBER KRESS: Isn't that the definition of 18 some of the importance factors?
19 MEMBER POWERS: George -- I'd just point out 20 to you, George, that there's been a tradition within the 21 nuclear business that I can trace back to at least 1954 to 22 take 10% as a small change.
23 CHAIRMAN APOSTOLAKIS: Yeah, but that's out of 24 the blue. I mean, --
/ \
) 25 MEMBER POWERS: I mean, it's been traditional NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
609 1 for a long, long time to say that a 10% change in a number !
i g 2 was negligibly small.
+ l 3 CHAIRMAN APOSTOLAKIS: But, for example, do I 4 have a reasonable assurance that if someone is proposing a 5 change that is -- leads to a delta CDF of 10-5 that that 6 person has not removed a train from a safety system?
7 MR. HOLAHAN: No, you don't. That's why we 8 also have key principles to address those sort of 9 situations.
10 CHAIRMAN APOSTOLAKIS: It seems to me that 11 moving a train would be pretty drastic.
12 MEMBER KRESS: One of them is maintaining --
13 CHAIRMAN APOSTOLAKIS: No, I understand that.
C' 14 But can I tell in advance that if they did that, the delta 15 CDF would be certainly higher than 10'S -- that kind of 16 calibration?
17 MR. HOLAHAN: No, I don't think so.
18 CHAIRMAN APOSTOLAKIS: But on the other hand, 19 we do also know, and we've cited it a couple of times 20 coming around the table here in the last couple of months 21 that the idea that a plant that has its risk scattered 22 among -- well, approximately ten -- of more or less equal 23 size is something we think of as being a balanced plant.
24 CHAIRMAN APOSTOLAKIS: That's right.
g3 Q 25 MEMBER SEALE: And what we're saying here 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
610 1 don't upset the balance.
2 CHAIRMAN APOSTOLAKIS: Well, you know, you may
/
3 have a point. And my point is that we haven't really 4 talked about these things in that context. I mean, trying 5 to calibrate ourselves --
6 MR. HOLAHAN. There is a lot of --
7 CHAIRMAN APOSTOLAKIS: You just gave me an 8 argument that would help me in calibrating myself.
9 MR. HOLAHAN. I think there's quite a lot of 10 anaafsis that's been done on these sort of things. Some 11 of the utilities have tables which show what it means to 12 take pieces of equipment out of service and even have l
13 combinations of them and you can see what the effects are.
Q U 14 CHAIRMAN APOSTOLAKIS: Do you have access to l
15 those? )
16 MR. HOLAHAN- Yes. You know, I think we got l 17 some of them from the pilot applications.
18 CHAIRMAN APOSTOLAKIS: I'd like to see --
19 MR. HOLAHAN: I think we can --
20 CHAIRMAN APOSTOLAKIS: Is that available?
21 MR. KING: The IPE's have some of that too.
22 They have their risk distributed among the systems as well 23 as among the initiating events and so forth.
24 MR. HOLAHAN: Yeah, I think what you'11 find Q(y 35 is you can't talk about a train. You have to talk about 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
611 1 train of something.
2 CHAIRMAN APOSTOLAKIS: Of course.
eq i
~
3 MR. HOLAHAN: Because a train of low pressure 4 injection is not worth an order of magnitude, but a train 5 of auxiliary feedwater is.
6 CHAIRMAN APOSTOLAKIS: Well, you're already 7 beginning to calibrate then.
8 MR. JONES: Station blackout improvements is 9 worth about 2 x 10-5 10 MEMBER SEALE: Yeah, that's a double whammy.
11 CHAIRMAN APOSTOLAKIS: That's interesting.
12 MR. JONES: And that was considered a fairly 13 significant action. So these relaxations are -- and k- 14 again, the standards on the reg. analysis is a 1 x 10-5 15 change basically is considered to be substantial.
16 MEMBER KRESS: So what you're saying is these 17 delta numbers are -- could mean real changes and could --
18 CHAIRMAN APOSTOLAKIS: That's what I'm trying 19 to understand. What exactly can they mean?
20 MEMBER KRESS: Yeah, they're not just some 21 trivial change.
22 MR. HOLAHAN: Yes. That was probably why I --
23 MEMBER SEALE: They recently put in gas 24 turbines.
/ i
)
,. 25 MR. HOLAHAN: Yes.
NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234 4433
612 1 MEMBER SEALE: What did they think they got 2 when they did it? We know they spent a lot of money to do (p) 3 it.
4 MR. KING: Some of the IPE's have quantified 5 that. I'd have to -- I don't remember off hand what the -
6 -
7 MR. HOLAHAN: I think it's on the order of 10" 5
8 9 MEMBER SEALE: I think -- that's my 10 recollection.
11 MR. HOLAHAN: I think there's been a study on 12 that subject. Were the numbers pulled out of the IPE's?
13 Well, we can collect information like that and put it
/ 't
\- 14 together.
15 CHAIRMAN APOSTOLAKIS: Do you think it would 16 be useful as part of this effort to include some 17 discussion of that type, or would anybody else find it 18 useful?
19 MEMBER KRESS: I would.
20 CHAIRMAN APOSTOLAKIS: I mean, you guys se em 21 to be on top of it, so it shouldn't be much of a problem.
22 It doesn't have to be an exhaustive discussion, but at 23 least --
24 MR. RUBIN: There was very little data q _) 25 available. I believe most of the IPE's did not quantify NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
613 1 before and after. There were, I think, three data points
,m 2 from a preliminary study done before the summary IPE
( !
3 insights document was done.
4 MR. HOLAHAN: You're talking about the station 5 blackout.
6 MR. RUBIN: Yes, the blackout.
7 MR. HOLAHAN: But just various examples of 8 what it means to take equipment out of service or to 9 change the programs.
10 CHAIRMAN APOSTOLAKIS: A number of examples 11 would be --
12 MR. HOLAHAN: All you have to do is look at --
13 when the staff does a regulatory analysis to have a new
, ~y,
<.k/ '
14 requirement, we do such an analysis; take a few examples.
15 MEMBER SEALE: -- decided to spend Anmc money.
16 I mean, that -- the gas turbine thing, seems to me that's 17 kind of -- you know, you pay your money and you get your 18 benefit. Somebody made that decision.
1 19 MR. KING: It wouldn't be too much trouble to l 1
l 20 pull together some available information. l 21 CHAIRMAN APOSTOLAKIS: I would appreciate 22 that. I would find it very informative.
23 MR. KING: For -- you're talking about for the 24 full committee meeting in a couple of weeks?
n
( ) 25 CHAIRMAN APOSTOLAKIS: Oh, no, no. Well, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISt.AND AVE., N W.
(202) 234-4433 WASHINGTON, O C. 20005-3701 (202) 234-4433
614 1 let's talk about things like that now because I think in 2 about ten, 15 minutes, we'll probably have to end this 7_
\
\
'~'/ because the committee has to discuss a few other things.
3 4 And now the question is, what do we do -- I 5 mean, I don't think we can decide what we can do on the 6 5th of December before we know the decision of the 7 Chairman.
8 MR. HOLAHAN: I agree.
9 MEMBER POWERS: Are we anticipating writing a 10 new letter to anyone on the 5th -- on December meeting if 11 the Chairman decides that we can have the extra couple of 12 months?
13 CHAIRMAN APOSTOLAKIS: We will not write a
/~h 4 l
\_ ' 14 letter if we will have the extra couple of months. l 15 MEMBER POWERS: Is it possible --
16 MEMBER KRESS: Unless she decides the other 17 way, we'll have to write a --
l 18 MEMBER POWERS: I understand that. )
l 19 MR. LARKINS: It won't be a couple of months.
20 MEMBER POWERS: It seems to me that there have 21 been enough of us here and there's been enough feedback 22 given that -- wouldn't it be better to enjoin these people 23 to go off and do their fixing rather than preparing 24 presentations for the full committee?
f~s t
i
) 25 MEMBER KRESS: That's my opinion.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
615 1 CHAIRMAN APOSTOLAKIS: It depends critically es 2 on the Chairman's decision. If she wants a letter by
/ \
!' /
~'
3 December 7th, then they better come back and make a 4 presentation. If she does not, then we can make drastic 5 changes in the current schedule. In other words, they may 6 not appear at all before the committee.
7 MEMBER POWERS: But we could, at this point --
8 since we don't have the Chairman's decision until Tuesday, 9 we could decide that if she decides in the affirmative, 10 it's not necessary for these gentlemen to come back to the 11 committee.
12 CHAIRMAN APOSTOLAKIS: I believe that's a 13 correct --
/~N iK. )i 14 MEMBER POWERS: Unless you guys just had a 15 desperate --
16 CHAIRMAN APOSTOLAKIS: We know that these l
I 17 three gentlemen want to come back, so --
18 MEMBER POWERS: Well, they may have a sort of 19 quota of viewgraphs that they have to produce each 20 quarter. And you may not be quite there.
21 MR. HOLAHAN: Who are you speaking to?
22 CHAIRMAN APOSTOLAKIS: Oh, yes; if the 23 Chairman says that we can write a letter in February, we 24 will -- let me put it this way, Dana, and see if you guys
,,3
! ) 25 agree. We don't need a presentation like the presentation w./
net.L R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
616 1 that we would have if we have to write the letter.
2 On the other hand, we may take advantage -- we
~s b 3 may want to take advantage of this time where the whole 4 committee will be together for you to come back and maybe 5 discuss one or two or three of these issues and maybe 6 brief us as to what changes have been made so we don't put 7 every -- postpone everything until the January 8 subcommittee meeting.
9 In other words, part of the ACRS full 10 committee meeting will be turned into a subcommittee 11 meeting. Well, this subcommittee meeting is really a full 12 committee meeting.
13 MEMBER SEALE: Could I -- would you remind me
()
'-- 14 again when we're going to have the final version of the 15 documents available so that we can --
MR. KING: Well, our schedule -- we're l 16 17 planning now for the general reg. guide and SRP to have 18 all the changes incorporated including CRGR's December 4th 19 and issue a new version of the CRGR. And if you want one, 20 you can have one. !
21 CHAIRMAN APOSTOLAKIS: We do.
22 MEMBER SEALE: We do.
23 CHAIRMAN APOSTOLAKIS: We do.
1 24 MR. KING: Now the application specific items,
/7 i
- s. ,) 25 we're meeting with -- we're meeting with the office j 1
1 NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 2344433
617 1 directors on Monday. We're meeting with CRGR on Tuesday
,e 3 2 Depending on what comes out of that, I can't really give i )
'~'
3 you a schedule. It's not going to be much beyond December 4 4th. We're going to ham to have those --
5 MEMBER SEALE: So we can walk away from the 6 next full committee meeting in any event having the 7 material which will be the basis --
8 CHAIRMAN APOSTOLAKIS: Yes.
9 MEMBER SEALE: -- for our middle of February 10 letter.
11 MR. HOLAHAN: Except, we're serving a lot of 12 different people in this process. Among -- in addition to 13 the CRGR and ACRS, we all have office management. And so, i a 4
V 14 between the 4th and, let's say, early December when this 15 is delivered to the EDO, we have to go through office 16 concurrences.
17 Now, so it could be that the 4th is not the 18 ultimate draft, that there's another -- in taking it 19 through management review, there may be a mid December 20 draft also.
21 CHAIRMAN APOSTOLAKIS: I hope -- again, to 22 repeat it -- the Chairman's decision will not change your 23 deadline.
24 MR. HOLAHAN: We would still expect to give a
/ \
() 25 draft for comment to the EDO's office to be delivered 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
618 1 the Commission by --
2 CHAIRMAN APOSTOLAKIS: Okay.
7,3
( )
~
3 MR. HOLAHAN: -- December.
4 CHAIRMAN APOSTOLAKIS: Her decision will be 5 whether the ACRS will be allowed to write a letter in 6 February or not. Your deadlines are fixed.
7 MR. HOLAHAN: What I think it would affect in 8 the long run is it would affect the point at which it 9 would go out for public comment, because we presumably --
10 if the Chairman -- the Commission asks for the ACRS views, 11 what we would want is a more refined version or another 12 draft, however you want to explain it, before public 13 comment.
f C'N k/
4 14 CHAIRMAN APOSTOLAKIS: These go together. My 15 request was for us to write a letter in February and the 16 public period comment to be postponed -- public comment 17 period to be postponed. Okay?
18 MR. HOLAHAN: Yes.
19 CHAIRMAN APOSTOLAKIS: Is there anything else?
20 I don't think we can start another issue now.
21 MR. KING: We should plan to be here for the 22 December meeting? ,
23 CHAIRMAN APOSTOLAKIS: We will contact you I
24 through Mike after we get a decision from the Chairman, ]
q
(
) 25 which may come on Monday. l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N W. l WASHINGTON, D.C 20005-3701 i (202) 234-4433 (202) 234-4433 I
l 619 1 MR. MARKLEY: Right now it's scheduled for the 7
2 first presentation on the 5th, 8:35, 8:45, something like l I
4 )
- 3 that. j 4 MR. HOLAHAN: How long? !
l 5 MR. MARKLEY: Two hours. It's 8:45 to 10:45 !
6 on the 5th.
7 MEMBER POWERS: But it does strike me that the 8 committee, without these gentlemen there, has plenty to 9 discuss. You know, if we had those two hours to discuss, 10 we've got things to talk about that these guys don't need 11 to be there if they've got other things to do. Not that 12 they do, but --
13 CHAIRMAN APOSTOLAKIS: Okay, thank you very a \
-/ 14 much. This is the end of the subcommittee meeting.
15 (Whereupon, the proceedings were adjourned at I
16 5:35 p.m.)
17 18 19 1 20 21 I l
l 22 23 l
24 t
m,/
! 25 I
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
o (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
O CERTIFICATE This is to certify that the attached j proceedings before the United States Nuclear j Regulatory Commission in the matter of:
Name of Proceeding: ACRS SUBCOMMITTEE ON PROBABILISTIC RISK ASSESSMENT Docket Number: N/A 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 Regulatory Commission taken by me and, thereafter reduced to typewriting by me or under the direction of the court
, reporting company, and that the transcript is a true and accurate record of the foregoing proceedings.
- 1 1ML
, TORBETT RINER Official Reporter Neal R. Gross and Co., Inc.
s a
0
t ,
p/ :
L INTRODUCTORY STATEMENT BY THE CHAIRMAN OF THE 1-PROBABILISTIC RISK ASSESSMENT SUBCOMMITTEE 11545 ROCKVILLE PIKE,-ROOM T-2B3 -
ROCKVILLE, MARYLAND
- NOVEMBER
- 21-22, 1996 l
' The 1 meeting - will now come to order. This is second day of the !
meeting of the ACRS Subcommittee on Probabilistic Risk Assessment. !
i I am George Apostolakis, Chairman of the Subcommittee.
1 ;
{ ACRS Members in attendance are: John Barton, Mario Fontana, Thomas l j Kress, Don' Miller, Dana' Powers, Robert Seale, and Bill Shack l- ACRS Senior Fellow in attendance is: Rick Sherry l ,
.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) t sections and associated regulatory guides. The Subcommittee will .
gather .information, analyze relevant issues and facts, and :
formulate proposed positions and actions as appropriate, for ,
j deliberation by the full Committee.. .
j Michael T. Markley is the Cognizant ACRS Staff Engineer for this s j meeting, a
[p t'
The rules for participation in today's meeting have been announced i as part of the no.tice of this meeting previously published in the i Federal Register on November 13, 1996. l
- j. A transcript of the meeting is being kept and will be made l available - as stated in the Federal Register Notice. It is I requested that speakers first identify themselves and speak with i
p sufficient clarity and volume so that they can be readily heard.
We have received no written comments. However, we have received a request from Mr. Steve Gosselin of the Electric Power Research 4 Institute (EPRI) for time to make a presentation on EPRI j initiatives for risk-based inservice inspection (ISI) .
l (Chairman's Comments-if any) 4 1
We will proceed with the meeting and I call upon Mr. Steve Gosselin
[ to begin, i
A
) .. .
C :
l l
1
O O O P.
- EPRI '
- Adisory Committee on Reactor Safegaurds Meeting of
the Probabilistic Risk Assessment Subcommittee EPRI RISK-INFORMED INSERVICE INSPECTION .
EVALUATION PROCEDURE OVERVIEW '
presentedby:
S. R. Gosselin, P.E.
Electric PowerResearchInstitute Jim Chapman Yankee Atomic Electric .
Ron Gamble i Sartrex Corporation November 22,1996 Washington, DC System & Component Integrity Technology -
O O
~
O .
- EPRI -
Purpose I
Introduce EPRI Risk-Informed ISI Activities Overview of Process Third Party Review Findings and Recomendations
- Status Ongoing Pilot Plant Studies Future Meetings with ACRS i
- Detail 2d Technical Discussions Risk Evalution (Conseque'nce, Damage Mechanism, etc.)
Inspection for Cause Methods .
i soosesse.per- 2 System & Component Integrity Technology -
l w ____ _ m -
___ _m -+
O O O
- EPRI '
Risk-informed ISI Program Objectives l4 Technically sound, cost effective approach that can be ,
consistently applied by plant personnel Validate process through pilot applications studies Process Driven Identify areas of the plant to be inspected ,
i -
Define the number ofinspections required i Provide selection criteria that plant personnel can apply to select inspection locations Integrate other supplemental integrity management programs such as IGSCC, Erosion-Corrosion, etc.
l Couple risk selection with an inspection for cause process System & Component integrity Technology seer soooosserPT- 3
l_- O O O
-- EPRI -
l Risk-informed ISI Evaluation Procedure Interim Report
! EPRI Report TR-106706, " Risk-Informed Inservice Inspection Evaluation Procedure",
June 1996 ERIN Engineering, " Independent Review EPRI Risk-Informed Inservice inspection Procedure, Final Report", September 1996.
System & Component integrity Technology -_
SG090896.PFT - 4
l
(
lg - EPRI-f EPRI Risk-Informed ISI SYSTEM SCOPE All Class 1,2, & 3 Piping Non-Code Piping Important to Safety i
,r FMEA
,e SEGMENT Conseques:ce + DM's ,
l ir ir Degradation CONSEQUENCE h" CATEGORY TEG RY
> RISK MATRIX :
j
' )
HIGH, MEDIUM, LOW l
,P Integrated Plant ;
Review l
,r SAFETY !
SIGNIFICANT SEGMENTS System & Component Integrity Technology -,
I O - EPRI .
SAFETY SIGNIFICANT SEGMENTS ir
{
Defines Number of ;
inspections Required l 25% - High Risk Region 10% - Medium Risk Region
,r SELECT INSPECTION O LOCATIONS l EXAMINATION l REQUIREMENTS l l
i
,P DOCUMENT REVISED ISI l i
PROGRAM .
l
\
1 O
- System & Component integrity Technology - )
i
EPRI-RISK MATRIX CONCEPT 1
I
) RISK = (Core Melt Potential l Pipe Rupture) x (Potential for Pipe Rupture)
Core Melt Potential Potential for Pipe Ruoture -
i PSA & Deterministic
- Degradation Mechanisms t
- Impact Groups -
Service Expenence Initiating Event
- Rupture Potential Ranking
- Degraded Containment -
HIGH
- Degraded System /frain -
MEDIUM Combination -
LOW i Consequence Ranking HIGH
- MEDIUM LOW System & Component Integrity Technology amin.
SG090896. PPT - 5
- EPRI ~
i
. RISK MATRIX CONCEPT I
- i
- i: CON. SEQUENCE CATEGORYi ..
- Core Melt Potential for Umiting Break Sizel RISK REGIONS . ,
HIGH MEDIUM LOW NONE LQR MEDIUM jilfdi
= j T ie
.r
','b n;&g m -
... ;: ::$.! ,' s F:t '
d '^
-, WeltRMK ' , -3}(jiiff$%;$$
?a,i HIGH LOW RISK - MEDIUM RISK j!: M - t':^ ': a 190H_ fESK[s.
a@ ~ R;
~
jm ,;":
~
s M;w;.gN r' ;
x y ~._ im
?k f? ! le jMany i fio;.ms
! MEDIUM LOW RISK w
,gp NNIROillgKh
.. 1.w LOW RISK IEEDIUlf RISK I
t
{:p,:]! yj^'tydwgg- g:{d3 ,
- 4:
- O i!
- n jg;ggeggq k$ i
- Q - :: ,
!_.!Q ;..e } SMALL LOW RISK LOW RISK LOW RISK . ISEDlutt RISK
- .' O ;;..
O. . .
j System & Component Integrity Technology -
SG090896. PPT - 6 i
O O O
- EPRI -
Consequence Ranking Criteria I
- HIGH CCDP >104 Severe Initiating Events Severe Loss of Mitigation -
- High Risk of Containment Bypass
- MEDIUM 104 < CCDP < 104 Moderate Initiators Moderate Loss of Mitigation
- Moderate Risk of Containment Bypass
- LOW CCDP < 104 MildInitiators Minimal Loss of Mitigation Full Containment SG090896. PPT - 7 System & Component Integrity Technology -l
O O O
~
- EPRI -
Consequence Evaluation Break Size (large or " worst case size") ;
lsolability of Break
- Automatic / Operator Action
- Detection and Exposure Time Spatial Effects
- Flooding
- Impact on equipment in the vicinity of the break Initiating Events -
i
- PRA/IPE, design basis documents i I
Consider direct and indirect effects -
4 Loss of System / Recovery Loss of Train / Recovery l
i l
l 1
i System & Component Integrity Technology ,
SG0908M. PPT - 8
O O O
~
- EPRI - ~
impact Groups I
CONSEQUENCE IMPACT GROUP IMPACT Initiating Event initiating Event System Loss of System (S)
Loss of Multiple Systems (MS)
Degradation of System Degradation of Multiple Systems Train Loss of Train .
Loss of Multiple Trains Degradation of Train Degradation of Multiple Trains Containment Loss of Containment Integrity Combinations Any Combination of impacts System & Component Integrity Technology sumer-SG090096. PPT - 9
O O O EPRI ~
Initiating Event impact Group Table l
Design Basis initiating Event Recommended initiating Event initiating Event Frequency (1/yr.) Initiating Event Consequence Category Type Examples Category Startup, Routine operation >1 shutdown, I standby, N/A refueling, etc.
21 Anticipated (events that might Reactor trip, Low 11 operational occur during a turbine trip, occurrence calendaryearin a partialloss of particular plant) MFW Loss of primary Low -
flow 4
10 /104 (events Excessive Medium ll! Infrequent events that might occur feedwater/ steam during the lifetime removal of a particular plant)
LOSP High
<104 SLOCA Limiting faults or (events not MLOCA/LLOCA Mediurn/High IV accidents expected to occur SLB during the plant's ISLOCA lifetime)
SyS:em & Component Integrity Technology -
SG090896. PPT - 10
_ _ _ _ _ _ _ _ - _ - - _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ - - _ _ - _ - - _ _ . _ _ - _ _ . - - - - - _ _ _ _ . _ _ _ _ _ _ _ _ _ _ __s
_. .. .. . . - . . . . _ - - - . . . . . - . - - - . = . . - . - . . . . . - - _ - - . . - - . .
O O O
~
EPRI -
System / Train impact Group Table l
Affected Systems Number of Unaffected Backup Trains Frequency of Exposure Time o 1 2 =3 Challenge to Challenge All year jM[* fem ypi
}l .M L Anticipated Between tests (1-3 month) j%g L !$f"$ 'M. L (DB Cat II) Long AOT(1-2 week) MJ L L M$}
Short AOT (s 3 day) - Id
, M .- L L All year H' gH ,
M- L
. .,r Infrequent Between tests (1-3 month) H- M L L (DB Cat til) Long AOT (1-2 week) , : H, M L L Short AOT (s 3 day) / H ':;' L L L All year /H M L L Unexpected Between tests (1-3 month) [H M L L (DB Cat IV) Long AOT (1-2 week) 'lH L L L Short AOT (s 3 day) IH) L L L System & Component Integrity Technology -
SGO90896. PPT - 11
O O O
~
EPRI Containment impact Group Table Remaining Means of Protection Against Containment Bypass Consequence Category 1 Active "High" 1 Passive "High" 2 Active " Medium" 1 Active,1 Passive " Low" 2 Passive " Low" More than 2 "None" 1
l l ,
System & Component Integrity Technology sess_
SG090996. PPT - 12
_ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ . _ _ _ _ _ _ - ._.___ _ _ _ ._ _ _ _ _.__ _. _ _.___ _ _ ______ _ __ _ _ _ _ ___ i
O O O
~
- EPRI Combination impact Group Table Event Consequence Category initiating event and one unaffected train of High mitigating systems N is available initiating event and two unaffected trains of Medium mitigating systems are available (or IE category from Table 3.2)
Initiating event and more than two Low unaffected trains of mitigating systems (or IE category from Table 3.2) available initiating event and no mitigating system IE category from Table 3.2 affected Note 1 - Mitigating systems always correspond to the analyzed limiting event.
System & Component Integrity Technology name:
SG090096. PPT - 13
~
- ~ ' ' ' - - ' ' '
RISK MATRIX CONCEPT "
t l
i I , . . _ .
! . ..sCONSEQUENCE_ CATEGORY .. ._..
l E Core Melt Potential for Limiting (Break. Size ! t RISK REGIONS i
. r HIGH '
MEDIUM LOW NONE LQE MEDIUM HIGH :
r l ,, ; g W '
>s' *
~
m..., s s, s ,
- s:e s ,,
- s. ..
~ ~
.h [: :*d,:1 HIGH LOW RSK MEDIUM RISK . l~ HIGH AIOK , '
- a ' ' ',HIGH
's ftBKl'
' *a,'- .. , .
s s< e <
! 2 (E4 < 's
- ,s i 3
i U 2E -
~>- J s
< s>,' -
- K so -
digd:24: ' '^ ' ' '
EU: .8D.ii.
ij2:$
s
'[c's '
,r i MEDIUM LOW RISK LOW RISK MEDIUM RISK HIGHRIOKvf 1:em .
i di- d: .
s xl
's ya x-
- o : ; ' '
- 4,g,i L
- 0 g:
["og8 SMALL LOW RISK LOW RISK LOW RISK - MEDIUM RISK
- oi
': IL::
a System & Component Integrity Technology -
SG090896. PPT - 14 i
6
0 O O
~
EPRI ~
Degradation Mechanism Evaluation Degradation Mechanism Description & Attributes Damage mechanisms are identified by comparing component design and fabrication characteristics; operating conditions; and service experience with degradation attributes. The following damage mechanisms are considered:
+ Thermal Fatigue (Thermal Transient & TASCS)
Corrosion Cracking Stress Corrosion Cracking (IGSCC, TGSCC, PWSCC, etc.)
Local Corrosion Attsck (MIC,02Pitting, etc.)
- Erosion / Cavitation i Erosion Corrosion (i.e., Flow Accellerated Corrosion) j
- Degradation Mechanism Categories Damage potential categories are assigned to each component according to the degradation method present and itspotentialto create apipe rupture.
System & Component Integrity Technology -
SG090896. PPT - 15 L------ - - - - _ _ --- - - - ---_-_- - - - - - - _ - - - _ - _ _ _ - - - - - - - - , - - - - _ _ _-- -
- EPRI
~
Identifying Degradation Mechcanisms Degradation mechanisms are idendtified by enlicit. well defined atitihutes, which provide the means for a uniform. consistent evaluation Generic attributes have been determined for each degradation mechanism based on industry experience, basic causal effects, EPRI studies, and NRC requirements Generic effects are used on a pipe system /segement specific basis to identify the presence of potentially active degradation mechanisms and and their locations System & Component Integrity Technology SG090896. PPT - 16
- EPRI
~
Degradation Mechanism Evaluation Supporting References EPRI Report TR-104534, Volumes 1-4, " Fatigue Management Handbook" EPRI Report TR-103696, "PWSCC of Alloy 600 Materials in PWR Primary System Penetrations"
- ' Generic Letter 88-01, "NRC Position on IGSCC in BWR Austenitic Stainless l Steel Piping"
- NUREG-0313, Revision 2 .
l EPRI TR-103403, " Service Water System Corrosion and Deposition Source Book" EPRI Report NP-6815, " Detection and Control of Microbiologically influenced Corrosion"
- EPRI Report TR-103198, Volumes T1 and T2, "A Method to Predict Cavitation and the Extent of Damage in Power Plant Piping"
- EPRI Report NSAC/201 L, " Recommendations for an Effective Flow Accelerated Corrosion Program" System & Component integrity Technology -
SG090896. PPT - 17 l
O O O
~
. - EPRI ~
l Service Failure Data l
l SKI Report 96-20 " Piping Failures in United States Nuclear Power Plants: 1961-1995", January 1996 1511 Piping Failure Events 4
Failure Mechanism Number of Failures Corrosion Fatigue / Cracking 14 l Construction Defects 184 Design Dynamic Loading 8 Water Hammer 35 Vibration Fatigue 364 Thermal Fatigue 38 Erosion / Corrosion 280 l Stress Corrosion Cracking 166 Erosion-Cavitation 15 !
Corrosion 72 Other 43 Unknown 292 System & Component Integrity Technology em ;
SOO90896. PPT - 18 '
9
- fm
~ .
l U oukq. ) - -
? .
Damage Mechanism Nominal Frequency iE-01
- CC: Corrosion fatigue,
. TF: Thermal fatigue,
- SC: Stress corrosion cracking,
- E-C: Erosen-cavitation.
- LC: Local Corrosion,
~ WH: WaterHammer, p
- FAC: Flow assisted corrosion.
g VF: Vibratory fatigue h 1E-02 .
E :
l a
1 -
c g .
N -
E
.E 1E-03 .
z -
l t
a !
1 E-04
'* - + -
~#
- -i- -t-
O O O
~
- EPRI -
Degradation Mechanism Rupture Failure Potential Categories Large Pipe Leak Conditions Degradation Mechanism Break Potential HIGH Large Erosion Corrosion (FAC)
Thermal Fatigue Corrosion Fatigue / Cracking -
MEDIUM Small Stress Corrosion Cracking (IGSCC, TGSCC,PWSCC, ECSCC)
Local Corrosion Attack (MIC, O2 .
Pitting)
Erosion / Cavitation LOW None No Degradation Mechanisms For piping segments with the potential for catastrophic failure, the consequence analysis should be conducted consistent with the plant's design basis. The design basis should determine whether a large leak or catastrophic failure is appropriate. Absent this information, alternatives include use of the IPE/PRA or other engineering basis for establishing break characteristics.
i System & Component Integrity Technology men i SG090896. PPT - 20 ;
O O EPRI ~
RISK MATRIX a
' E CON. SEQUENC. E CA... TE. , GORY?. . . . .~.
E Core Melt Potential for Limiting Break Sizei: _
HIGH MEDIUM LOW NONE LQE MEDIUM Mifdi f.hh ,' ,
. ^-
- . . . . e . f. , x; ' , _ . ,;py.3% qg'
- ghijg ug..wyy. .
- jy ggg.g;g y
~, q-
.a ;
,:,,..z-..,,~
M
.s..<.
- :. ': LOW RISK . IIEDIURI RISK .: '#m,..x d:'ttWl8WK'g'. ..r.s .. .,. . :-
? IMOHtWKAi g: ; ;7~ -jf4 ,gg.gg '~ "';
- ::3 7 5; m: g g 'i ygj' ag 7
y
- ,'I
- s, e
s..g x>
- .s
., :, ^ 4 &
g>( g@gy x
g[~"4', +
T,
?: i$f '
v ^;d73 3 , y
- m
- ; e - - V
?: ,g ... .
, .: "s 's e s ,.
jf:;:p}r
MEDIUM LOW RISK LOW RISK . IEEDIUti RISK '. 411000flI0K' \
- 3 ;;-
~. g ; ;
.. a ,': v. 3- - , ~
- Js
_ a.
E _e F:
SMALL LOW RISK LOW RISK LOW RISK 10EDIUtd RISK':
E.
-: n.::
System & Component integrity Technology mem_
SG090896. PPT - 21
Performed by ERIN Engineering and Research, Inc.
Principle Reviewers
- Mr. KarlFleming
- Mr. Craig Sellers
- Mr. Doug True Purpose .
Provide overall review of methodology ;
, Identifyanypotential weaknesses Recommend improvements to better refine the process l
i System & Component Integrity Technology name SG090696. PPT - 22
O O O
~
- EPRI ~
Independent Review of EPRI RBISI Process Findings l
Technically sound and robust approach l
- Current interim report, which documents methods to be used in ongoing pilot applications, will need to be expanded to fully L document the process technical basis l
- Process is " semi-qualitative" PSA application and is consistent '
with the guiding principles in the EPRI PSA Application l
Guideline Process is based on current state-of-the-art.
Process is based in part on an understanding piping degradation mechanisms and ability to account for piping failure mechanism experience.
System & Component Integrity Technology suss SG090096. PPT - 23 i
l - EPRI -
! Independent Review of EPRI RBISI Process Recommendations Fully develop the technical basis in the final report.
- Quantification PSA basis of the approach should be more visible The opportunities for risk reduction as well as cost savings by
! application of the EPRI RBISI process in two pilot plant studies
! should be documented and explained Operating experience with pipe breaks, leaks, flaws and
, associated failure mechanisms and how the resulting insights have been applied need to be fully documented System & Component Integrity Technology -
SG090896. PPT - 24
- EPRI .
WO-4230 - CE-PWR Pilot Application Study EPRI T/C Project with 6 CE Utilities Pilot Plant - ANO Unit 2 Scope
- All Class 1,2, & 3 piping and non-code piping important to safety
- Approximately 70% complete Schedule ,
Final Submittal to NRC for all systems first quarter 1997 Results !
I 65% reduction of UTinspections the RCS
- 85% reduction of UTinspections for HPSI 95% of surface PT exams are eliminated System & Component Integrity Technology -
sansee96. PPT - 25 ,
s
ANO-2 RISI Evaluation Scope l
S Count SYSTEMTITLE C S 1 Reactor Coolant 1 !
2 Chemical & Volume Control 1, 2 3 High Pressure SafetyInjection 1, 2 4 Low Pressure SafetyInjection 1, 2 1
5 Shutdown Cooling 1, 2 6 "'"i"**"t Sgr y 2 O
7 Main Steam 2 8 Main Feedwater 2, NNS 9 Emergency Feedwater 2, 3, NNS 10 Service Water 2, 3, NNS
!O
,e
O O O EPRI -
WO-4231 - BWR Pilot Application Study EPRI T/C Project with NYPA i 4
Pilot Plant: James A. Fitzpatrick (GE-BWR/4)
Scope !
14 systems All Class 1, 2, & 3 piping and non-code piping important to safety Approximately 90% complete Schedule - -
Final Submittal to NRC for all systems first quarter 1997 Results 40% reduction in UT inspections currently being performed on the Core i Spray System from all inspection programs (Section XI, IGSCC, and FAC)
No significant change in UT scope for the Recire System (current IGSCC ,
inspection scope will be incorporated into ISI program)
~100% of surface PT exams are eliminated in both systems System & Component Integrity Technology -
SG090896. PPT - 26
_ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ - 2
1 5
JAF RISI Evaluation Scope 1
Count SYSTEM TTTLE
[ g[N 1 Reactor Water Recirculation (02-2) 1 2 Main Steam (29) 1 3 Feedwater(34) 1 4 Core Spray (14) 1, 2 5 Reactor Water Cleanup (12) 1 6 Control Rod Drive System (CRD; 03) 2 j
O 7 High Pressure Coolant Injection (23) 1, 2, NNS 1
3 Residual Heat Removal (10) 1, 2 l
- 9 Reactor Core Isolation Cooling (RCIC; 13) 1, 2, 3,
- NNS 10 Nuclear Boiler Vessel Instrumentation (02-3) 1 ,
i 11 Standby Liquid Control (11) 1, 'l l l
12 FuelPool Cooling (19) 3
( 13 Service Water (46) & RHRSW(10) 3, NNS 14 Emergency Service Water (46 & 15) 3 m
. . _ . _ _ . _ . - _ _ .- - - - - -- -- ~
- s,
! United Statea
\,,,,)s Nuclear Regulatory Commission RISK-INFORMED INSERVICE TESTING OVERVIEW OF REGULATORY GUIDE DG-1062 AND SRP CHAPTER 3.9.7 Richard H. Wessman, NRR Mark Cunningham, RES and RI-IST Development Team Presentation to the ACRS Subcommittee on PRA November 22,1996
RISK-INFORMED REGULATORY GUIDE AND STANDARD REVIEW PLAN DEVELOPMENT TEAM 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 1
O 9 9 .
'l
BRIEFING TOPICS G RI-IST team objective O " Classical" IST approach O RI-IST approach
. O Essential elements of Regulatory Guide CRG) and Standard Review Plan (SRP)
O Current experience with RI-IST pilot plants i
2 l
i
RI-IST TEAM OBJECTIVES G Provide acceptable approach and guidance to industry for developing and submitting optional RI-IST program O Provide guidance to staff for reviewing RI-IST submittals '
i G Review RI-IST pilot plant program submittals t
3
O O O .
RI-IST TEAM OBJECTIVES (cont'd)
ACCOMPLISHMENTS:
O Draft Regulatory Guide DG-1062: "An Approach for Plant Specific, Risk-informed, Decision Making:
4 Inservice Testing 9 Draft SRP Section 3.9.7: " Standard Review Plan -
for the Review of RI-IST Applications" 9 initial review of pilot applications complete l FUTURE ACTIONS:
O Pilot plant reviews to be completed by 3/97 9 Continue interaction with ASME on IST code cases 4
i
" CLASSICAL" IST APPROACH O ASME Code referenced in 10 CFR 50.55a O Program scope includes ASME safety-related code class pumps and valves O Like components subjected to prescriptive testing requirements regardless of safety significance 9 Changes in test interval or acceptance criteria reviewed and approved by staff (relief request) i G NRC-endorsed code cases provide alternatives to code requirements 5
4 O O O .i RI-IST APPROACH O Regulatory approval would be an alternative to the Code requirements, as allowed by 10 CFR 50.55a(a)(3D(i)
G Establishes IST program using traditional engineering evaluation, PRA insights, and integrated decision making j
G Components categorized as either HSSC or LSSC 9 Program scope expected to include non-code components that are HSSC i
G Allows licensee to focus resources commensurate with safety significance 6
RI-IST APPROACH (cont'd)
G Testing strategy for LSSCs usually less prescriptive G Testing strategy for selected HSSCs may be more rigorous than current code requirements G Test interval changes for LSSCs will not require NRC approval provided the approved process is not changed i
l O Test method changes will not require NRC approval provided the test is an approved test (endorsed ASME Code or Code Case)
O Changes to RI-IST program that involve process changes require NRC approval 1
7 O O O l
ESSENTIAL ASPECTS OF RI-IST REGULATORY GUIDE:
~
DG-1062 gives guidance on acceptable methods for utilizing PRA information together with established .
traditional engineering information in the development of IST programs that have improved effectiveness regarding the utilization of plant resources while still maintaining acceptable levels of quality and safety.
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.
8
i ELEMENT 1: DEFINE THE PROPOSED CHANGES TO THE IST PROGRAM G Identify components in current licensing basis (CLB) i G Identify Components not in CLB but important to safety G Identify changes to existing test schedules and methods G Identify necessary information and analyses to support change G Determine required formal interaction with NRC 9
O O O .
ELEMENT 2: GUIDANCE FOR ENGINEERING EVALUATION 9 Traditional engineering evaluation a
- 9 Probabilistic risk assessment 9 Integrated decision making 9 Key principles addressed
-Current regulations are met
-Defense in depth maintained
-Safety margins maintained
-Changes to CLB do not result in individual plant risk exceeding safety goals and subsidiary objectives '
-CLB changes result in only small increases in plant risk 10
ELEMENT 2: GUIDANCE FOR ENGINEERING '
EVALUATION-(cont'd) l 9 PRA used to identify pumps and valves that are l candidates for relaxation of IST as well as candidates for enhanced testing 9 Assess change in plant risk from overall change in test frequency or method l 9 Expect consideration of common cause failures, I
results from sensitivity evaluations, human reliability, defense in depth, and safety margin O Acceptable plant risk change criteria, uncertainty, and sensitivity analysis in accordance with DG-1061 11
ELEMENT 2: GUIDANCE FOR ENGINEERING EVALUATION (cont'd) i G Consider cumulative effects of RI-IST in relation to other initiatives ,
G Licensee expected to review existing approved IST '
relief requests, maintenance and outage planning, TS, etc.
i 12
ELEMENT 3: IMPLEMENTATION, PERFORMANCE MONITORING AND CORRECTIVE ACTION O HSSC typically tested using Code test frequency l and method G LSSC typically tested at extended interval Cup to a maximum limit)
G Expect use of ASME. code cases
- O Phased implementation of RI-IST program G Monitor component performance to
- 1) identify degradation
- 2) initiate appropriate corrective action 3D adjust test frequency based on results O O O
O O O .
l ELEMENT 3: IMPLEMENTATION, PERFORMANCE
^
MONITORING AND CORRECTIVE ACTION (cont'dD O Periodic reassessment of program 9 Key Principle Addressed: Performance-based l implementation and monitoring strategies consider uncertainties 14
t Element 4: DOCUMENTATION O Provide guidance to licensees on minimum '
documentation necessary to evaluate submittal
, 9 Must show that proposed CLB change is consistent with the key principles of risk-informed regulation and NRC staff expectations 9 Provide guidance on PRA records and supporting data maintained at licensee's facility 9 Provide guidance on maintenance of ASME Code test records and component test interval schedules to support RI-IST .
15 9 9 9 .
LO O O ,
CURRENT EXPERIENCE WITH RI-IST PILOT PLANTS S Initial submittals made November 1995 9 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 S ACDF estimated by licensees as:
Comanche Peak 7 X 10-6/RY Palo Verde 4 X 10-6/RY 16
ASSESSMENT OF RI-IST PILOT PLANT'S ADHERENCE TO ELEMENTS ELEMENT 1: DEFINE THE PROPOSED CHANGES TO THE IST PROGRAM O Both pilots include in their RI-IST program proposal components that are in their current IST program and components determined to be important to safety (components in CLB and not in CLB)
O identified maximum test interval. Details of schedule subject of future RAl.
j O Information to support change contained in license submittal and response to RAI incomplete. Future RAI needed.
.i G 9 9 j
ASSESSMENT OF RI-IST PILOT PLANT'S ADHERENCE TO ELEMENTS (cont'd)
ELEMENT 2: GUIDANCE FOR ENGINEERING EVALUATION i
e Guidelines of EPRI "PSA Applications Guide" used by pilots 9 PRA importance calculations to identify relative risk contributions of pumps and valves j
O Expert panel process integrates traditional engineering information (defense in depth, historical performance data, test methodsD with risk insights j G Pilots did not discuss the impact of previously approved IST relief requests or current TS 18
- i ASSESSMENT OF RI-IST PILOT PLANT'S ADHERENCE TO ELEMENTS (cont'd)
ELEMENT 2: GUIDANCE FOR ENGINEERING EVALUATION (cont'd)
O Expert panel also used to compensate for PRA scope (external events and shutdown), and to integrate results from sensitivity analyses 9 Licensee's estimate of change in CDF show risk impact less than DE-1061 criteria O Pilots did not consider cumulative effects of RI-IST in relation to other initiatives i
G Pilots did not explicitly address defense-in-depth except to state that PRA implicitly treats issue G 9' S
O O O ASSESSMENT OF RI-IST PILOT PLANT'S ADHERENCE TO ELEMENTS (cont'dD ELEMENT 3: IMPLEMENTATION, PERFORMANCE MONITORING AND CORRECTIVE ACTION G Commitment to investiaate use of ASME Code Cases in RI-IST programs, when they are approved by staff O Specific testing intervals for low safety significant components not included in submittals (overall maximum interval of 6 years is proposedD i
l 9 Assessing performance monitoring, feedback
! mechanism, and corrective action plan O Details of step-wise implementation scheme will be requested in future RAI 20 j
ASSESSMENT OF RI-IST PILOT PLANT'S ADHERENCE TO ELEMENTS (cont'd)
ELEMENT 4: DOCUMENTATION l
O Assessing completeness of information to support acceptability of the proposed CLB change 9 Documentation of integrated decision-making process incomplete !
O Additional RAI required to obtain documentation identified in DG-1062 l e
D' S .
o o O x GRADED QUALITY ASSURANCE
- REGULATORY GUIDE 1 1
6 PRESENTATION i 3 THE ADVISORY COMMITTEE ON
! REACTOR SAFEGUARDS NOVEMBER 22,1996 Robert Gramm, NRR 415-1910 Roy Woods, RES 415-6622 Stephen Dinsmore, NRR 415-8482 i i
O .
O O -
GRADED QUALITY ASSURANCE (GQA) e Development of Regulatory Guide e Prior ACRS Briefings e Background of GQA - SECY 95-059 ,
- Process to identify SSC safety significance
- Application of QA controls based on safety function and significance
- Effective root cause and corrective action program
- Operational feedback to assess QA controls and safety significance 1
l
O O O -
~
GRADED QA REGULATORY GUIDE METHODOLOGY
- USE OF RISK INFORMED PROCESS 1
- 1. DEFINE THE QA CHANGES
- 2. PERFORM ENGINEERING EVALUATION ,
- 3. IMPLEMENT GQA AND PERFORM MONITORING
- 4. GQA DOCUMENTATION i
2
ELEMENT 1: DEMNE THE QA CHANGES
o Response to generic communications
- 2. IDENTIFY CANDIDATES FOR GRADED QA
- Subject to current Appendix'B requirements
- SSCs modelled in the risk studies e
SSCs within the scope of the Maintenance Rule
- 3. IDENTIFY EXPECTED REVISIONS TO QA PROGRAM
- 4. EVALUATE RISK STUDY APPLICABILITY 3
ELEMENT 2: ENGIlNEERING EVALUATION i
I
- CATEGORIZE SSC RISK IMPORTANCE 6 .
e CONFORMANCE WITH SAFETY PRINCIPLES
- INTEGRATED ASSESSMENT DETERMINING SSC SAFETY-
- SIGNIFICANCE 1
l 4
o .
o 0 -
, ELEMENT 2: ENGBEERING EVALUATION CATEGORIZE SSC RISK IMPORTANCE
- 1. IDENTIFY SYSTEM FUNCTIONS e Define all operating and accident functions performed by systems
- 2. CATEGORIZE RISK IMPORTANCE OF SYSTEM FUNCTIONS e Risk importance classification, guided by PRA importance measures e Risk Importance classification, guded by qualitative considerations
- 3. IDENTIFY SSCs THAT SUPPORT HIGH-RISK-IMPORTANT FUNCTIONS e Complete list of all system components whose successful operation (passive and/or active) is needed for performance of system level function
- 4. CATEGORIZE RISK-IMPORTANCE OF SSCs e Guiding principle: SSC operating modes (open, remain open, etc) required to support high-risk-important system functions are also high-risk-important 5
ELEMENT 2: ENGINEERING EVALUATION CONFORMANCE WITH SAFETY PRINCIPLES e DEFENSE-IN-DEPTH IS MAINTAINED e SAFETY MARGINS ARE MAINTAINED e INDIVIDUAL PLANT RISK DOES NOT EXCEED SAFETY GOAL
- ONLY SMALL INCREASES IN PLANT RISK Instead of risk change calculations, develop Quantitative performance criteria Qualitative performance criteria i
6 i i
i ELEMENT 2: ENGBEERING EVALUATION -
i INTEGRATED ASSESSMENT e INTEGRATION OF QUALITATIVE AND QUANTITATIVE INFORMATION ,
Review and acceptance of resolution of quantitative importance !
measure limitations i Review and acceptance of qualitative importance determination Review and acceptance of performance criteria derived from control of change in risk level
]
o DETERMINATION OF SAFETY SIGNIFICANCE AND ADJUSTMENTS IN QA PROGRAM Review, accept, or modify high-risk-importance as high-safety-significant 7
~
O O O -
ELEMENT 3: DEVELOP IMPLEMENTATION
- AND MOMTORING STRATEGIES
!
- GRADING OF QA ACTIVITIES
- The proposed change conforms with regulations o PotentialImplementation Areas
- Procurement
- Frequency of inspections
- Audits ;
- Level of document approval
- Records and documentation 8
i f
vf
o .
o 0 -
ELEMENT 3: DEVELOP IMPLEMENTATION AND MOMTORING STRATEGIES
- OPERATIONAL FEEDBACK !
l - Operating Experience -
- - Plant modifications
- Component Degradation Monitoring i
- Update FRA to reflect plant modifications o CORRECTIVE ACTIONS l
9 i
_ . _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ . _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ____ _ _m .. ._ _ . . - . -. _ _ _ _ . - . . - . .
ELEMElNT 4: DOCEMENTATION e SCOPE OF SYSTEMS FOR GQA e FINAL CHARACTERIZATION OF SSC SAFETY SIGNIFICANCE e PLANT OPERATING EXPERIENCE e ENGINEERING EVALUATION o DEVELOPMENT OF GQA PROGRAM 10
O O O &!
w 2
SUMMARY
REGULATORY GUIDES AND STANDARD REVIEW PLANS FOR RISK INFORMED DECISION MAKING ,
ON PLANT SPECIFIC CLB CHANGES t
GARY M. HOLAHAN, NRR THOMAS L. KING, RES i i
PRESENTATION TO THE ACRS SUBCOMMITTEE ON PRA .
- NOVEMBER 22, 1996 l
I l
O O O Y
.i i
ITEMS FOR DISCUSSION i
e OUTLINE OF~ PROPOSED FEDERAL REGISTER NOTICE e QUESTIONS FOR PUBLIC COW 4ENT e COMMISSION PAPER e OPEN ITEMS I
e PLAN FOR ACRS DECEMBER FULL COMMITTEE MEETING I
l t
t I
i 1
b I
PROPOSED FEDERAL REGISTER NOTICE TITLE: USE OF PRA In PLANT-SPECIFIC REACTOR REGULATORY APPLICATIONS (10CFR50)
- ACTION: PROPOSED REGULATORY GUIDES AND STANDARD REVIEW PLANS I
e SUPPLEMENTARY INFORMATION:
BACKGROUND DESCRIPTION OF REGULATORY GUIDES AND STANDARD REVIEW PLANS e OBJECTIVES l e DESCRIPTION OF DOCUMENTS e STRUCTURE OF DOCUMENTS e RATIONALE FOR -
PROCESS 1
PRINCIPLES GUIDELINES e POLICY ISSUES QUESTIONS FOR PuBLIC COMMENT e CONTACTS, TIMING, BACKFIT CONSIDERATIONS, OMB CLEARANCE, ETC.
2
FEDERAL REGISTER NOTICE QUESTIONS -
GENERAL:
A)* IS IT APPROPRIATE TO APPLY THE COM4ISSION'S SAFETY GOALS AND THEIR SUBSIDIARY OBJECTIVES ON A PLANT SPECIFIC BASIS? IF NOT, EXPLAIN WHY.
B)* IS'IT APPROPRIATE TO ALLOW, UNDER CERTAIN CONDITIONS, CHANGES TO A PLANT'S CLB THAT INCREASE CDF AND/OR LERF? IF NOT, EXPLAIN WHY.
IS THE LEVEL OF DETAIL IN THE GUIDANCE CONTAINED IN THE PROPOSED C)
REGULATORY GUIDES AND SRPS CLEAR AND SUFFICIENT, OR IS MORE DETAILED GUIDANCE DESIRED? EXPLAIN WHAT LEVEL OF DETAIL IS DESIRED.
D) ARE THE ELEMENTS OF THE REVIEW PROCESS DESCRIBED N THE REG GUIDES SUFFICIENT? IF NOT, EXPLAIN WHY.
E) IS THE GUIDA.MCE ON THE TREATMENT OF UNCERTAINTIES CLEAR AND SUFFICIENT, OR IS ADDITIONAL GUIDANCE DESIRED?
F) IS GUIDANCE ON THE ACCEPTABILITY AND TREATMENT OF TEMPORARY CHANGES IN THE CLB NEEDED? IF SO, WHAT GUIDANCE AND ACCEPTANCE GUIDELINES WOULD BE SUFFICIENT?
l G) IS THE GUIDANCE APPROPRIATE FOR SHUTDOWN CONDITIONS? IF NOT, WHAT GUIDANCE SHOULD BE PROVIDED?
H) SHOULD LICENSEES BE REQUIRED TO SUBMIT RISK INFORMATION IN SUPPORT OF PROPOSED CHANGES TO THEIR CLB ?
- POLICY ISSUE 3
O O O -
FRN QUESTIONS (CONT) ,j I ENGINEERING EVALUATION:
A) ARE THE PROPOSED SAFETY PRINCIPLES CLEAR AND SUFFICIENT? IF NOT, WHAT SHOULD BE CLARIFIED AND/OR ADDED?
B) IS SUFFICIENT GUIDANCE PROVIDED REGARDING THE INTENT AND LEVEL OF DETAIL REQUESTED IN THE SUBMITTAL WITH RESPECT TO THE EVALUATION OF THE SAFETY PRINCIPLES? IF NOT, WHAT SHOULD BE ADDED?
C) IS THE GUIDANCE ASSOCIATED WITH DID AND SAFETY MARGIN SUFFICIENT?
SPECIFICALLY"
- e SHOULD THERE BE DIFFERENT GUIDANCE FOR THOSE ITEMS ANALYZED IN .
THE PRA VERSUS THOSE NOT ANALYZED? IF SO, WHAT SHOULD THE DIFFERENCES BE?
e SHOULD THERE BE QUANTITATIVE GUIDELINES FOR DETERMINING THE SUFFICIENCY OF DID AND SAFETY MARGINS? IF SO, WHAT WOULD BE APPROPRIATE GUIDELINES?
D) IS THE GUIDANCE ASSOCIATED WITH THE PROBABILISTIC ANALYSIS SUFFICIENT? SPECIFICALLY:
1 e IS ADDITIONAL GUIDANCE ON THE USE OF QUALITATIVE RISK EVALUATIONS NECESSARY? IF SO, WHAT GUIDANCE WOULD BE APPROPRIATE?
4
O O O -
- FRN QUESTIONS (CONT)
_l e ARE THE PROPOSED ACCEPTANCE GUIDELINES FOR CDF AND LERF APPROPRIATE? IF NOT, WHAT WOULD BE APPROPRIATE? WHY?
e ARE THE FIGURES IN DG-1061 lLLUSTRATING CDF AND LERF GUIDELINES USEFUL? IS MORE SPECIFIC OR LESS DETAILED GUIDANCE DESIRED?
WHY? j e SHOULD THERE BE SEPARATE LERF GUIDELINES FOR PWRS AND BWRS? IF SO, WHAT SHOULD THEY BE?
e IS THE GUIDANCE IN APPENDIX A OF DG-1061 FOR ESTIMATING LERF SUFFICIENT? IF NOT, WHAT IS NEEDED?
e SHOULD THERE BE ACCEPTANCE GUIDELINES FOR THE JSE OF PRA LEVEL 3 INFORMATION? IF SO, WHAT WOULD BE APPROPRIATE?
e IS IT APPROPRIATE TO REQUEST THAT INTERNAL FIRES BE INCLUDED IN THE MINIMUM SCOPE OF PRA ANALYSIS FOR THE USE OF QUANTITATIVE
( PRA INFORMATION? IF NOT, EXPLAIN WHY.
e IS IT APPROPRIATE TO REQUEST THAT FULL SCOPE PRA ANALYSIS (FULL POWER, LOW POWER, SHUT DOWN, EXTERNAL EVENTS) BE USED FOR LANT'S CDF ND/OR LERF ARE ASSESSING NEAR THE SUBSIDIARY CHANGES TO THE CLB OBJECTIVES OF 10- WHERE A p/RY AND 10 p/RY, RESPECTIVELY.
5
FRN QUESTIONS (CONT) ,
e IS THE USE OF POINT ESTIMATE PRA ANALYSIS APPROPRIATE FOR ANALYZING SMALL INCREASES OR ANY DECREASES ON CDF AND LERF? IF NOT, EXPLAIN WHY.
e IS IT APPROPRIATE TO REQUEST UNCERTAINTY AND SENSITIVITY ANALYSIS BE PERFORMED FOR CERTAIN CHANGES TO THE CLB THAT INCREASE CDF AND/OR LERF? IF NOT, EXPLAIN WHY.
e IS THE INFORMATION ON PRA QUALITY IN DRAFT NUREG-1602 SUFFICIENT? SHOULD IT SPECIFY ACCEPTABLE PRA METHODS?
t PERFORMANCE MONITORING AND FEEDBACK:
A)* SHOULD THE USE OF PERFORMANCE MONITORING BE MORE LIDELY APPLIED IN REGULATION AND REGULATORY PRACTICE, OR IS IT SUFF;CIENT TO IMPLEMENT IT THROUGH THE ELEMENTS DESCRIBED IN PROPOSED REGULATORY GUIDE DG-1061?
B) IS IT APPROPRIATE TO REQUEST THAT PERFORMANCE MONITORING AND i
! FEEDBACK BE INCLUDED IN THE SCOPE OF THE EVALUATION AND PROPOSED i CHANGES TO THE CLB?
l APPLICATION SPECIFIC R.G.S/SRPS:
A)* IS IT APPROPRIATE TO UTILIZE THE "OTHER ACCEPTABLE METHODS" PROVISION OF 10CFR50.55A TO IMPLEMENT CHANGES TO THE CLB FOR ISl?
B) ADD ADDITIONAL QUESTIONS AS APPROPRIATE FOR THE APPLICATION SPECIFIC i R.G.S/SRPS 6
COmISSION PAPER ISSUES TO BE HIGHLIGHTED:
APPLICATIONS FOR CHANGES IN CLB UTILIZING RISK INFORMATION WILL BE GIVEN HIGHER PRIORITY OVERALL PROCESS / APPROACH / PRINCIPLES USE OF 10-4/RY CDF AND *CDF AS CORE DAMAGE PREVENTION GOAL USEgRYF 10-5/RY LERF AND *LERF AS LARGE RELEASE GOAL (VERSUS 10- S'JGGESTED FOR STUDY IN SAFETY GOAL POL _CY STATEMENT)
PROCESS SETS A PRACTICE OF ONLY APPROVING RISK NEUTRAL OR RISV REDUCTIOrt WHEN CDF/LERF EXCEEDED (ALSO RELATE TO IPE FOLLOWUP ACTIVITIES)
TREATMENT OF UNCERTAINTIES l
7
._ . - . . . . -. .. . _...... . - ~ . -. - . . . - ~ - . _ . . - . . . . . - . . . . . . -._-__-. ...._ ....--_ .- - - . -
O O O -
~
GENERAL R.G./SRP - OPEN ITEMS 1
e EXPAND DISCUSSION ON TREATMENT OF UNCERTAINTIES
- MOVE LICENSEE GUIDANCE IN SRP AND NUREG-1602 TO R.G.
PEER REVIEW-PROCESS SSC CATEGORIZATION DUE TO SAFETY SIGNIFICANCE e DECISION GUIDELINES FOR LERF STILL UNDER DISCUSSION e UPDATE NUREG-1602 e CRGR COMENTS e ACRS COMENTS 8
. - - - v- - _ _ _ _ - _ _ _ _ _ m-.__ _ _ _ . ______ _ __.___m.__ _ _ _ _ _ _ _ _ _ . _ _ . _ . _ _ _ _ _
MAJOR ACRS COMMENTS e CLARIFY PROCESS / ANALYSIS ONLY APPLIES TO PROPOSED CHANGE.
e ALLOW FLEXIBILITY IN APPLYING PRINCIPLES AND PROBABILISTIC GUIDELINES:
QUALITATIVE ANALYSIS ACDF/ALERF e ADD CAUTION ON APPROACHING CDF/LERF LIMITS l e IS PROPOSED LERF GUIDELINE CORRECT? SHOULD IT BE OIFFERENT FOR PWRS vS. BWRS?
s ARE GUIDELINES IN THE FORM OF ANNUALIZED CDF/LERF SUFFICIENT?
e IS APPD A TO DG-1061 ON LERF CONSERVATIVE?
1 e REWORD PRINCIPLE ON IMPLEMENTING INCREASES IN RISK IN SMALL STEPS.
e HOW MUCH OF THE PERFORMANCE MONITORING PLAN SHOULD TNE STAFF REVIEW AND APPROVE VS. JUST REVIEWING PERFORMANCE?
9
_ _ _ _ _ _ _ _ _ _ - - _ - - _ _ _ - _ _ _ - - _ -__ - __-_ - - - _ _ - - - - - _ _ _ _ _ - _ _ - - ..