ML18004A157
| ML18004A157 | |
| Person / Time | |
|---|---|
| Issue date: | 10/18/2017 |
| From: | Hossein Nourbakhsh Advisory Committee on Reactor Safeguards |
| To: | |
| Nourbakhsh H | |
| References | |
| NRC-3327 | |
| Download: ML18004A157 (198) | |
Text
Official Transcript of Proceedings NUCLEAR REGULATORY COMMISSION
Title:
Advisory Committee on Reactor Safeguards Regulatory Policies and Practices Docket Number: (n/a)
Location: Rockville, Maryland Date: Wednesday, October 18, 2017 Work Order No.: NRC-3327 Pages 1-168 NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers 1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005 (202) 234-4433
1 1
2 3
4 DISCLAIMER 5
6 7 UNITED STATES NUCLEAR REGULATORY COMMISSIONS 8 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 9
10 11 The contents of this transcript of the 12 proceeding of the United States Nuclear Regulatory 13 Commission Advisory Committee on Reactor Safeguards, 14 as reported herein, is a record of the discussions 15 recorded at the meeting.
16 17 This transcript has not been reviewed, 18 corrected, and edited, and it may contain 19 inaccuracies.
20 21 22 23 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 www.nealrgross.com
1 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION
+ + + + +
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS (ACRS)
+ + + + +
REGULATORY POLICIES AND PRACTICES SUBCOMMITTEE
+ + + + +
WEDNESDAY OCTOBER 18, 2017
+ + + + +
ROCKVILLE, MARYLAND
+ + + + +
The Subcommittee met at the Nuclear Regulatory Commission, Two White Flint North, Room T2B3, 11545 Rockville Pike, at 8:30 a.m., John W.
Stetkar, Chairman, presiding.
COMMITTEE MEMBERS:
JOHN W. STETKAR, Chairman RONALD G. BALLINGER, Member DENNIS C. BLEY, Member CHARLES H. BROWN, JR. Member MICHAEL L. CORRADINI, Member JOSE MARCH-LEUBA, Member DANA A. POWERS, Member NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
2 JOY L. REMPE, Member GORDON R. SKILLMAN, Member MATTHEW SUNSERI, Member DESIGNATED FEDERAL OFFICIAL:
HOSSEIN NOURBAKHSH ALSO PRESENT:
ANDREA D. VEIL, ACRS Executive Director NATE BIXLER, Sandia National Laboratories KEITH COMPTON, RES HOSSEIN ESMAILI, RES ED FULLER, RES TINA GHOSH, RES TREY HATHAWAY, RES SALMAN HAQ, RES DONALD HELTON, RES WILLIAM ORDERS, NRR EDWARD ROACH, NSIR PATRICIA SANTIAGO, RES AMY SHARP, RES TODD SMITH, NSIR CASEY WAGNER, Dycoda LLC KIMBERLY WEBBER, RES NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
3 C-O-N-T-E-N-T-S Introductory Remarks...............................6 Patricia Santiago, RES Focused Pressurizer Safety Valve Study (New Appendix I of Draft Report)..................26 Trey Hathaway, RES Summary of Report Updates in Response to ACRS Member Comments from Prior Subcommittee Meetings.............................89 Tina Ghosh, RES Casey Wagner, Dycoda LLC Nathan Bixler, SNL Discussion.......................................143 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 1 P R O C E E D I N G S 2 8:31 a.m.
3 CHAIR STETKAR: The meeting will now come 4 to order. This is a meeting of the Advisory Committee 5 on Reactor Safeguards, Regulatory Policies and 6 Practices Subcommittee.
7 I'm John Stetkar, Chairman of the 8 Subcommittee Meeting. And you're not.
9 Members in attendance today are Ron 10 Ballinger, Matt Sunseri, Dick Stillman, Dana Powers, 11 Mike Corradini, Jose March-Leuba, Dennis Bley, I 12 forgot you, Charlie Brown, and Joy Rempe.
13 Hossein Nourbakhsh is the Designated 14 Federal Official for this meeting.
15 The purpose of today's meeting is to 16 discuss the State-of-the-Art Reactor Consequence 17 Analyses Project for the Sequoyah Integrated 18 Deterministic and Uncertainty Analyses.
19 Today we have Members of the NRC Staff and 20 Sandia National Laboratories to brief the 21 Subcommittee.
22 The ACRS was established by statute and is 23 governed by the Federal Advisory Committee Act. That 24 means that the Committee can only speak through its 25 published letter reports.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
5 1 We hold meetings to gather information to 2 support our deliberations. Interested parties who 3 wish to provide comments can contact our office 4 requesting time after the meeting announcement is 5 published in the Federal Register.
6 That said, we set aside ten minutes for 7 spur-of-the moment comments from members of the public 8 attending or listening to our meetings. Written 9 comments are also welcome.
10 The ACRS section of the NRC public website 11 provides our charter bylaws, letter reports and full 12 transcripts of all full and Subcommittee meetings, 13 including slides presented there.
14 The rules for participation in today's 15 meeting were announced in the Federal Register on 16 October 10, 2017. The meeting was announced as an 17 open meeting.
18 No written statement or request for making 19 an oral statement to the Subcommittee has been 20 received from the public concerning this meeting.
21 A transcript of the meeting is being kept 22 and will be made available as stated in the Federal 23 Register Notice.
24 Therefore, we will request that all 25 participants in this meeting use the microphones NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
6 1 located throughout the meeting room when addressing 2 the Subcommittee.
3 The participants should first identify 4 themselves and speak with sufficient clarity and 5 volume so that they can be readily heard.
6 We have a bridge line established for the 7 public to listen into the meeting.
8 To minimize disturbance, the public line 9 will be kept in a listen-in-only mode, and I'll open 10 it at the end of the meeting for public comments.
11 To avoid disturbance, I request that all 12 attendees and everyone else in the meeting room put 13 your electronic devices like cell phones, beepy things 14 and whatever, in the off or noise-free mode.
15 Now, I'll proceed with the meeting, and 16 I'll call upon Pat Santiago of the NRC Office of 17 Nuclear Regulatory Research to begin today's 18 presentations. Pat?
19 MS. SANTIAGO: Thank you, good morning.
20 We appreciate all the feedback that we got 21 from the Subcommittee Members in the May of 2016 22 Subcommittee Meeting, as well as the June meeting of 23 this year.
24 And we worked the last few months to 25 address all the comments that we received.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
7 1 On this second slide, I want to talk a 2 little bit about our schedule.
3 Our Sequoyah Analyses is going to be 4 published as a NUREG/CFR report, and it's due to the 5 Commission November 30th. And we'll be sending it to 6 the NRC Publications Branch at the same time.
7 We're relying on the ACRS Members to 8 service our peer reviewers and we're requesting a 9 letter documenting your review. To facilitate that 10 process, we're briefing the full Committee in November 11 of this year.
12 And this works well with our schedule 13 since the NUREG is due to the Commission November 14 30th. So, we appreciate your support for that.
15 We hope to be able to address any final 16 comments from today's Subcommittee meeting in the next 17 few weeks, but as you can see from our schedule, any 18 additional analysis would not be documented in this 19 report.
20 Rather, we could possibility add 21 additional clarifications on what was done within the 22 scope of this study. And time to include it within 23 the NUREG will be submitted for the Commission, and 24 for publication.
25 Also, separately, much of the same team 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
8 1 in the process of updating the Surry Uncertainty 2 Analysis. And currently, that NUREG report is due to 3 the Commission June of 2018.
4 Previously, the Subcommittee had indicated 5 that you may not need to review the updated 6 Uncertainty Analysis.
7 However, we can discuss whether you'd like 8 the team to come back with the updated Uncertainty 9 Analysis next year.
10 MEMBER BLEY: Pat?
11 MS. SANTIAGO: In addition to that --
12 MEMBER BLEY: Pat? Excuse me. Can you 13 say something about the extent of that update?
14 And is it to make things at least 15 methodologically consistent with what's been done 16 here?
17 MS. SANTIAGO: Okay, so the Sequoyah 18 Update or the Surry --
19 MEMBER BLEY: The Surry Uncertainty.
20 MS. SANTIAGO: We'll talk a little bit 21 about that --
22 MEMBER BLEY: Is that in the presentation?
23 MS. SANTIAGO: Okay, yes.
24 MS. GHOSH: So, there are a couple of main 25 things.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
9 1 CHAIR STETKAR: Move a little closer, 2 please.
3 MS. GHOSH: Is this better?
4 You're aware of some of the major updates 5 we made to the Sequoyah Analysis between last year and 6 this year. Some of them are very relevant to Surry as 7 well.
8 All of the safety valve parameters are 9 important for Surry too, because of the induced steam 10 generator tube rupture in that case.
11 And so for a different reason but still 12 very important. So, we definitely wanted to make that 13 update.
14 So, basically, all of the relevant updates 15 in the Sequoyah Analysis in the last year, we are 16 implementing in Surry, and that will change the 17 results somewhat.
18 The other thing is at the Subcommittee on 19 Surry, we got some comments from Bill Shack and others 20 on our steam-generator tube rupture modeling. So, 21 right now, we are currently in the throes of updating 22 that modeling.
23 So, we've gone and gathered more 24 information, talk to more experts, and we're in the 25 process of updating that modeling. So, those are the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
10 1 two main areas.
2 MEMBER BLEY: Well, since you brought up 3 that last one, are you going to talk about the reason 4 why you don't have any steam-generator tube rupture in 5 the Sequoyah Analysis?
6 Is that part of your presentation?
7 MS. GHOSH: No, we had tried -- okay, so 8 we did not address induced steam-generator tube 9 rupture.
10 MEMBER BLEY: So, you don't have to update 11 it?
12 MS. GHOSH: For Sequoyah. No, that was 13 not within the scope of the study, and we tried to 14 make that clear in the introduction that we did not do 15 that.
16 MEMBER BLEY: You made it very clear.
17 What wasn't clear was why the scopes are different, 18 why you didn't do it here and you've done it there.
19 And now you're going back and redoing it 20 over on the other plant.
21 MS. GHOSH: That's what we say in the 22 Sequoyah study, is look at the Surry study for 23 insights on tube rupture because --
24 MEMBER BLEY: Yes, you do.
25 MS. GHOSH: -- because those should 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
11 1 consistent. The scope of the Sequoyah study was really 2 focused on containment issues.
3 MEMBER BLEY: You say that kind of at the 4 end, that's where the focus was, but you don't say 5 anything about why.
6 Is there a simple why, or it was just the 7 way the scope was set up?
8 MS. GHOSH: You know, at the time --
9 MEMBER BLEY: This was added in to look at 10 an ice condenser, I guess.
11 MS. GHOSH: Yes, when we went to the 12 Commission and said we should finish this third pilot 13 study for these reasons.
14 Some of the big motivators at the time 15 were still doing the post-Fukushima regulatory 16 actions, and we wanted to make sure we weren't missing 17 anything in terms of hydrogen challenges for the ice 18 condenser containment.
19 So, that was a major focus of this third 20 plant. It was really focused on the unique aspects of 21 the containment.
22 MEMBER BLEY: This kind of makes sense 23 but, at least to this reader, it didn't jump off the 24 page at me.
25 In fact, we're focused on the containment NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
12 1 as kind of an afterthought at the end. It's like we 2 didn't do these things but that's because we're 3 focused on the containment.
4 So, you didn't really introduce it with --
5 MS. GHOSH: Ah, okay.
6 MEMBER BLEY: -- why you're doing this.
7 MS. SANTIAGO: I would re-look at the 8 Executive Summary. I think that's an important point, 9 to make sure that it's clear why we did what we did.
10 MEMBER BLEY: Maybe it is, but it wasn't 11 clear to me.
12 MEMBER REMPE: Let's go a bit further.
13 If you had looked at it -- and since you 14 are doing it at Surry, maybe you have some insights on 15 what would have happened if you had looked at 16 Sequoyah, especially with the Level 3.
17 It's a Westinghouse plant, right? Is 18 there something you think you might be able to say?
19 For example, you did -- most of the time, 20 as I recall, and Don's in the audience and he can 21 correct me, but when they were doing the Level 3, they 22 decided, well, because it's a Westinghouse plant, it 23 wouldn't have been a big deal, except for if you had a 24 loop seal that was about 480 gallons per minute.
25 And do you think you might have the same NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
13 1 kind of conclusions that they had in the Level 3? And 2 then, there's only that one sensitivity study you did, 3 where you did have that high of a loop seal leak.
4 And I'm just wondering if you can kind of 5 make some comments in your Executive Summary, if 6 you're updating it, that could be based on engineering 7 judgment from all the ongoing evaluations you've done.
8 And the only place I think you might want 9 to think about acknowledging it makes a difference is 10 how what happens with this 480 gallon-per-minute leak.
11 Because instead of talking about the 12 containment failure, you might have had a bypass. And 13 it looks like that someone is agreeing with me by --
14 up and down.
15 But anyway, it seems like there's some 16 things you might want to go a bit further and say that 17 it's probably -- again, you guys should say it, not 18 me. But maybe it won't be that important except for 19 that one case.
20 MS. GHOSH: That's a good comment. We can 21 modify it. We can work on modifying it.
22 CHAIR STETKAR: I have to be cognizant of 23 time. We have a 12:00 p.m. hard stop today.
24 We shouldn't try to speculate about 25 differences between Surry and Sequoyah because 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
14 1 are difference plants. Period. They're different 2 plants.
3 So, trying to speculate about lessons 4 learned at Sequoyah based on what we know for Surry is 5 against what we should be doing.
6 These are plant-specific studies and a 7 conditional probability of a high-dry-low situation.
8 And Sequoyah could be substantially different than 9 high-dry-low.
10 At Surry, we don't know that, and that's 11 the important insight from SOARCA, is that you must do 12 plant-specific and site-specific, integrated, 13 uncertainty analysis.
14 MEMBER REMPE: But then I would argue with 15 you, sir, that the Executive Summary is really coming 16 up with conclusions that may not be valid for a large 17 lead rate.
18 CHAIR STETKAR: That's correct.
19 MEMBER REMPE: And again, the Vogtle 20 analysis did something based on generic knowledge 21 about --
22 CHAIR STETKAR: We're not talking about 23 Vogtle here.
24 MEMBER REMPE: That's true, but again, I 25 think you may want to -- without trying to extrapolate NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
15 1 things from other plants, you may have some 2 conclusions that are not valid for this particular 3 plant --
4 CHAIR STETKAR: That's right.
5 MEMBER REMPE: Because you did not 6 consider the consequential steam-generator tube 7 rupture.
8 So, something needs to be acknowledged in 9 that Executive Summary on the downsides of not 10 considering it.
11 CHAIR STETKAR: It's fair game to talk 12 about the Executive Summary for this study.
13 And the Executive Summary should, in my 14 opinion, clearly delineate what was done, why it was 15 done, what was not done, and why it was not done.
16 MEMBER REMPE: And what are the downfalls 17 of not doing it? I would go a bit further on 18 potential downfalls of not doing it, because some of 19 the conclusions may not be valid because they had a 20 limited scope.
21 And they should have some acknowledgment.
22 And without extrapolating, they could say, hey, we've 23 done a bunch of analyses and this may be important, 24 you may have to do it plant-specific to determine it.
25 But I think you need to have some caveats NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
16 1 in your conclusions. Okay?
2 CHAIR STETKAR: Okay, good. Pat, we kind 3 of interrupted you midstream. Pick up the ball.
4 MS. SANTIAGO: One other thing that may 5 help some of this discussion, actually, is we plan on 6 doing -- basically, after the Surry RAY's done, we 7 want to work on a compendium of important insights 8 that we've gained from the three Uncertainty Analyses.
9 And we're going to publish that as a summary NUREG as 10 well.
11 But I do what to point out that in each of 12 the SOARCA studies, each bottom study during Sequoyah, 13 we do say that this is a site-specific study, and you 14 may have to look at the specific site and other 15 designs of other plants.
16 CHAIR STETKAR: And I think that whenever 17 you come to that sort of general insight, NUREG or 18 whatever, we would be very interested in seeing that.
19 Because I'm very concerned about trying to 20 make the entire pressurized-water nuclear power 21 industry look like Surry, and the entire boiling-water 22 industry look like Peach Bottom. And the entire ice-23 condenser plants look like Sequoyah.
24 So, trying to develop broad generic 25 insights from these very site-specific, plant-specific NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
17 1 studies is dangerous and misguided. And that's on the 2 record.
3 So, that's why I'd be really interested in 4 strength seeing that Overall Insights study.
5 MS. SANTIAGO: Okay, and maybe when we do 6 the Surry, we can give you an outline of what our 7 plans are for that compendium.
8 CHAIR STETKAR: Yes, that would be good.
9 MS. GHOSH: Can I just make one quick 10 clarification?
11 I don't think the point is to try to draw 12 some broad-brush -- you know what, the main issue is 13 just to make something that's more practical than 14 having -- I mean, at this point, we have like 1900 15 pages of Uncertainty Analyses.
16 And people go in and they'll pluck out 17 what they need, but we get a lot of complaints about 18 just how much material you have to go through to pick 19 out that nugget.
20 So, we're trying to make something that's 21 a little bit easier to digest in one place, rather 22 than having to go to the library and dig through all 23 that material.
24 CHAIR STETKAR: Pat? Since you still have 25 the ball to carry.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
18 1 (Laughter.)
2 MS. SANTIAGO: Next slide, I did want to 3 compliment the team members that worked on this 4 because they worked very hard and diligently on this 5 particular project, at the same time that they're 6 working on the Surry Uncertainty Analyses and other 7 projects, as you heard me talk about on October 3rd, 8 when the Division of Systems Analysis briefed.
9 So, I did want to thank them all. Dr.
10 Tina Ghosh and Doug Osborn were the co-leads. Doug 11 Osborn's from Sandia National Labs.
12 We have a handful of folks in the audience 13 to help us answer any questions as we brief, as well 14 as those on the phone line.
15 On the last slide, I just wanted to talk a 16 little bit about the meeting today, and the focus is 17 going to be on the Sequoyah draft report.
18 Since our last June 6th Subcommittee 19 Meeting, there's two significant additions that we've 20 added to this draft report.
21 The first change is the new introductory 22 material in Section 4 on accident progression, which 23 was presented in June by Hossein Esmaili. And at that 24 time, it wasn't included in the report.
25 We don't plan to brief on that material NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
19 1 since we did brief in June, but I just wanted to 2 mention that.
3 The second addition is a new Appendix I 4 in the draft report, which documents the supplemental 5 Uncertainty Analyses that was focused on a range of 6 safety valve parameter values where an early 7 containment failure is possible.
8 Dr. Trey Hathaway will present this work 9 today.
10 After's Trey's presentation, the team will 11 walk through Member comments and discuss how they were 12 investigated and addressed in this new draft report 13 that we provided you last month.
14 And lastly, we'll seek the Subcommittee 15 Members' feedback on how we should focus our 16 presentation to the full Committee in November, since 17 we'll only have 90 minutes at that particular meeting.
18 And I'll now turn it over to Trey and my 19 Staff to present.
20 CHAIR STETKAR: Pat, thanks. A couple 21 things.
22 First of all, in your introductory 23 remarks, you characterized the ACRS's role in this 24 project as providing a peer review. I personally take 25 issue with 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
20 1 The ACRS has not performed a peer-review 2 service. We do not delve into the depth of technical 3 details that is implied by a full-scope technical peer 4 review.
5 We focus on integration, we focus on 6 consistency, we focus on high-level issues related to 7 technical scope content, things like that.
8 But by no means do we perform a peer 9 review service in the way that those words are 10 typically understood throughout the industry, and 11 throughout the Staff.
12 So, I would appreciate it if you don't 13 characterize what we're doing as a peer review.
14 MS. SANTIAGO: Okay, so for Surry and 15 Peach Bottom, we did have a full, complex, external 16 peer review group.
17 And so for Sequoyah, based on knowing what 18 the comments were for that and addressing all of that, 19 we have guidance that says we can use or ask the ACRS 20 to support review of the next document, which was 21 Sequoyah.
22 So, I apologize for suggesting it's a peer 23 review.
24 It's an internal review, I guess, that we 25 would like to have a letter from the Committee, so NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
21 1 that we can accompany or send it to the Commission 2 about the same time as the report is going to the 3 Commission.
4 So, thank you.
5 MEMBER BLEY: I have to add a little bit 6 to what John said. We're giving the some kind of 7 oversight in this study as we did to the others.
8 To imply that we're providing any sort of 9 detailed technical review of the sort you'd had from 10 your peer reviews is just wrong, and we don't do that.
11 And you ought to not tell people that that's what ACRS 12 has done for you.
13 CHAIR STETKAR: Also, please don't 14 characterize this as an internal body. We are not 15 internal to the NRC, we are an independent advisory 16 committee.
17 MS. SANTIAGO: I forgot the right words.
18 CHAIR STETKAR: Are you going to discuss 19 the Executive Summary?
20 I didn't see your slides until this 21 morning, so I haven't had a chance to look through 22 them.
23 Are you going to discuss the Executive 24 Summary at all?
25 MS. SANTIAGO: We weren't planning on 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
22 1 CHAIR STETKAR: Okay, let me ask one 2 general question and just get it out on the table, and 3 kind of think about it.
4 As I read the Executive Summary and a few 5 of the Chapters that have the new material in the 6 current version of the report, I found kind of a 7 different tenor from the previous version of the 8 study.
9 And the difference seemed to be more 10 emphasis on comparisons between the current version of 11 the study, and what I'll characterize as the April 12 2016 version of the study that we saw a year and a 13 half ago.
14 There's a lot of comparative stuff saying, 15 well, we had this in the April 2016 version, and look, 16 the same trends are here, but the results are 17 different.
18 That obviously had to be a conscious 19 decision because it's pervasive throughout the new 20 material.
21 Why do you feel it's necessary to compare 22 the results, insights, conclusions, models, and so 23 forth from this current version of the study, to an 24 outdated, non-technically-supported interim version of 25 the study?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
23 1 I mean, when I've done projects in the 2 past and have made mistakes in the middle, I don't go 3 back and say, hey, I made a mistake but, look, you 4 know, this other stuff is all the same.
5 I just don't see the benefit of it. You 6 changed models, you changed uncertainty distributions; 7 those changes made a difference.
8 Why go back and say, well -- I mean, there 9 are statements that say, well, the uncertainty 10 distribution moved the results around but, look, the 11 general trends would have been the same.
12 That's obvious. You can present that same 13 level of insights and conclusions, just given the 14 results of the current study.
15 It's obvious to a reader that if a 16 pressurizer safety valve sticks open, big, early, it's 17 not a good day.
18 You don't need to compare that to the fact 19 that it used to stick open more, bigger, in a previous 20 version of the study that was found to have some 21 technical questions.
22 So, I'm not sure, because it was a 23 conscious decision. The only reason I bring it up is 24 why was that decision made to compare and contrast to 25 the previous version of the study?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
24 1 Because to me, it was not only from a 2 philosophical perspective that I tried to mention 3 here, it was actually distracting.
4 MS. GHOSH: Okay, so I think we understand 5 your comment.
6 I mean, as to answering the why, I think 7 in part, we've put out this big study, it's publicly 8 available. It's going to be forever in Adams and --
9 CHAIR STETKAR: When you say this big 10 study, you mean the 2016 version of the study?
11 MS. GHOSH: Yes, because it was out there.
12 It was the first time we did this with the ice 13 condensers. A lot of people had looked at it.
14 So, maybe it was in recognition that 15 people might have been familiar with last year's 16 study. But I understand your comment and I think we 17 can take a look at --
18 MEMBER CORRADINI: I have a slightly 19 different opinion than my colleagues. It's your 20 study, you can do whatever you want. Then you'll get 21 criticized for it. So, my view of it is 22 that John's saying if you were reading it today, this 23 would be his reaction. And another way to do it 24 is to write the Executive Summary in such a manner 25 that nobody had seen anything before, and say, oh, by NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
25 1 the way, we had a draft version.
2 Go look at Appendix blah-blah and you'll 3 be able to compare and contrast in blah-blah, but 4 don't do it in the main body.
5 One, it detracts from what I assume you 6 think is better. And two, it would confuse the reader 7 that never took the time or the energy to read the one 8 before.
9 But that's your style, it's your report.
10 MEMBER BLEY: And it sounds very 11 defensive.
12 We would have certainly asked and we would 13 have expected in this presentation you would do those 14 sorts of things and show us what changed.
15 But I agree with John and, I think, with 16 Mike.
17 CHAIR STETKAR: Certainly, as I read it, 18 quite honestly, because I read the Executive Summary 19 first, it's the first thing you come to, and quite 20 honestly, it's what most folks will pay most attention 21 to.
22 So, I wanted to kind of see what flavor of 23 information was in there.
24 I found it very, very distracting in the 25 Executive Summary. Once I got past that, it 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
26 1 infused in parts of Chapter 4 and in Chapter 7 in the 2 results.
3 There, because those are -- well, Chapter 4 7 is kind of summary of results, so that's a companion 5 to the Executive Summary. But Chapter 4 is more of a 6 technical detail thing and I was less confused by it 7 there.
8 But I think Mike's idea of pointing to an 9 Appendix that compares and contrasts it -- I 10 understand the desire to capture the differences and 11 draw insights from the differences.
12 But, anyway, you've heard our comments.
13 MS. SANTIAGO: That's a good comment, 14 thank you.
15 CHAIR STETKAR: And with that, I don't 16 know who's -- Tina?
17 MS. GHOSH: Trey.
18 CHAIR STETKAR: Trey, you're up.
19 DR. HATHAWAY: Okay, what I'm going to 20 present today is the third results of this sort of 21 Focus Study that we performed kind of after the full 22 UA --
23 CHAIR STETKAR: Is your mic on?
24 DR. HATHAWAY: I've got it here.
25 CHAIR STETKAR: You're just a very soft-NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
27 1 spoken person.
2 DR. HATHAWAY: Sorry about that.
3 CHAIR STETKAR: Scream at the mic.
4 DR. HATHAWAY: So, the point of this was 5 to really try to reinforce the conclusions and the 6 study.
7 We knew we only had a very few number of 8 early containment failures so we wanted to kind of 9 focus in on this region of ventures to see if it 10 reinforced conclusions of the full UA.
11 So, that's what I'm going to present to 12 you today. Also, kind of see if it provided any 13 additional insights, focusing in on this region.
14 So, the Figure on the right is the results 15 of the 2016 UA, and we just kind of -- we were drawing 16 conclusions -- not drawing conclusions, we were trying 17 draw insights from these previous analyses, all the 18 analyses that were performed to really hone in on what 19 region of interest led to early containment failure.
20 And the reason we chose this is because we 21 did have -- well, I'm presenting this because we had 22 fewer early containment failures than the current 23 study.
24 So, what you see here, the Figure on the 25 right, plots the number, the total 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
28 1 pressurizer safety valve cycles plus the open area 2 fraction.
3 The points in blue represent the -- excuse 4 me, all the points represent the various individual 5 realizations, the Monte Carlo realizations. And the 6 points in red represent cases that had an early 7 containment failure.
8 And what you see is those sort of cases 9 that begets early containment failure kind of housed 10 in this region of cycles greater than 0.3 and less 11 than 65, is what was chosen.
12 Really, my particular interest was less 13 than hot leg failure, try to have a dent before hot 14 leg failure.
15 So, yes, again, we were trying to use the 16 previous studies to inform what we were going to look 17 at for this sort of focused look.
18 So, we took the current model and just 19 sort of used the identical distributions as before, 20 except for two.
21 We bounded the distribution for the open 22 area fraction to 30 percent, and then we also 23 constructed a distribution based off the sample data 24 from the previous study to try to focus in on this 25 range of 165 cycles.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
29 1 For this focused look, originally, it was 2 for whether or not we had early containment failures.
3 So, we limited the calculation to 15 hours1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br />, and that 4 was just really to -- again, this was informed by 5 previous studies.
6 Previous studies indicated that the early 7 containment failures were before 15 hours1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br />. The model 8 was identical to the previous study.
9 The only difference is there was this 10 minor error in the study having to do with the fabric 11 seal pressure.
12 That was corrected; it was a formatting 13 error on a text file that was written so that was 14 corrected for this study.
15 MEMBER CORRADINI: So, I'm going to ask a 16 general question and you can address it whichever way 17 you want. But I've been looking through the new 18 version, and again, I could be missing it.
19 I'm just looking for data, test data, that 20 helps decide if this is realistic, optimistic, or 21 conservative.
22 I'm assuming you were trying to buy us at 23 conservative, but I'm looking for safety-valve test 24 data.
25 DR. HATHAWAY: While this particular study NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
30 1 wasn't looking at --
2 MEMBER CORRADINI: No, I know you weren't.
3 I'm just looking for a reference that tells me to go 4 look over here.
5 So, your distribution of fraction of valve 6 open after how many cycles bounds it, or represents 7 it, or whatever.
8 MS. GHOSH: So, from the perspective of 9 having the potential for early containment failure, 10 we're looking at a more conservative range for this 11 two-dimensional sample space.
12 Because that black box is basically what 13 we're sampling, and that only represents about four 14 percent, four or five percent, of the entire sample 15 distribution.
16 But it's in that four or five percent 17 where the early containment failure is possible.
18 So, we're artificially constraining it to 19 explore this area where you can get early containment 20 failure. But it's only a small percentage of the 21 total sample distribution, which is --
22 DR. HATHAWAY: Okay, yes.
23 Again, this was to really try to focus in 24 on the full study and just look at that region where 25 early containment was most probable to really see if NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
31 1 there were any additional insights.
2 CHAIR STETKAR: Trey, your slides, I read 3 through this, I see what you're trying to do and what 4 you did. But I got confused a bit.
5 Your slides don't have the Figures that 6 I'm going to talk about, and the Figures I'm going to 7 talk about right now are Figures I-2 and I-3 of the 8 report, that develop the values and the range of 9 values that you sampled from for this study.
10 I got confused by those because they do 11 not, to me, anyway, seem consistent with the numerical 12 values that are used in the main body of the study.
13 So, I'd like somebody to explain to me where I'm 14 wrong.
15 In particular, I look at your fifth-order 16 polynomial fit to the valve failure rate for the valve 17 data in Figure I-2.
18 And I noticed the first term is 3.26 times 19 10 to the -2, which, to me, ought to be the mean value 20 for the conditional probability that a valve sticks 21 open on the first demand.
22 If I'm wrong about that, please tell me 23 because that's kind of fundamental.
24 Because in the main body of the report, we 25 don't use that value. We use a value of 2.65e to the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
32 1 -2, not 3.26. So, I'm curious why the first-order 2 term is 3.26 versus 2.65.
3 The next-order term for failure is per 4 demand. If I use just a linear model, it's 1.83e to 5 the -3, and in the main body of the report, we use a 6 value of 2.23e to the -3 for the demand.
7 So, the shape is -- the intercept at zero 8 is slightly different and the shape is somewhat 9 different.
10 So, somebody fit a fifth-order polynomial 11 to stuff that ostensibly uses data from the main body 12 of the report, and I can't draw that link.
13 Now, why is that important? Well, 14 overall, I'm not sure, but it's certainly going to 15 affect the number of samples in different regions.
16 The second -- let me get through the 17 second part of it.
18 The second part of it is when you do the 19 differentiation of that curve to get the density 20 function in Figure I-3, that density function does 21 indeed start out at 1.83e to the -3, at a value of 1.
22 Now, that's actually at a value of one-23 plus because if you differentiate the delta function 24 on the first demand, you get infinite. So, you can't 25 do 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
33 1 So, I want to make sure that the actual 2 sampling used the correct probability for failure on 3 the first demand, that it didn't use 1.83e to the -3, 4 that it used something between 2 and 3e to the -2 in 5 the actual runs that you made.
6 Why is that important? Because that's 7 going to force more earlier failures.
8 Certainly, a lot more earlier failures 9 than might be implied by the density function here in 10 I-3.
11 So, if somebody could confirm to me where 12 the numbers came from, and in particular, when you did 13 your sampling?
14 Give me assurance that you used something 15 on the order of two or three times ten to the -2 for 16 the probability that it fails on the first demand?
17 The first demand, number one, not 1.83e to 18 the -3.
19 I'd be a lot happier.
20 DR. HATHAWAY: I'll try to address your 21 comment, I'll take a look at the comments, I can look 22 at them.
23 But the distribution extended well past 24 what I fit, because I was interested in just this 25 range. I think that's what I 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
34 1 I was interested in the range of 0 to 250.
2 So, I was really trying to get a fit to the data in 3 that range.
4 The problem is, as you extend past that 5 range, the shape of the distribution is different. It 6 starts to curve.
7 CHAIR STETKAR: You're starting to talk 8 about somebody who really likes to do curve-fitting.
9 I'm asking you a more fundamental question. I'm 10 asking a question that your first term says 3.26e to 11 the -2.
12 In the main body of the report, there's a 13 table that says the mean value for failure on the 14 first demand is, and I've lost it again, --
15 MEMBER BLEY: It's about 3 --
16 CHAIR STETKAR: It's about 2.65. Now, 17 it's about 3, but why is the value different? And the 18 first term is also different.
19 So, why didn't you use the values that are 20 tabulated, and then you can do all of your fifth-order 21 polynomial curve-fitting, whatever you want to do, and 22 extrapolate out?
23 It's the same mathematical problem, it's 24 just a different curve, and I don't know why it has to 25 be different. It's the same lambda.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
35 1 MEMBER BLEY: You make a big, and I think 2 reasonable, case that it's a different failure rate 3 for the first actuation than for the rest?
4 CHAIR STETKAR: Yes.
5 MEMBER BLEY: So, you have --
6 CHAIR STETKAR: And that's what's used in 7 the study. That's what's used in the real study.
8 MEMBER BLEY: Yes.
9 CHAIR STETKAR: This is supposed to be 10 something that focuses on the first, nominally, 65 or 11 so demands, and tries to force samples within that 12 regime, that are consistent with the first study, 13 given the fact that the first study extends -- uses 14 all three valves.
15 I mean, the first study, the baseline 16 study if you want to call it, that is very elegant in 17 the way that it characterizes the valves and multiple 18 failures, and things like that.
19 But the intent of this shouldn't be to 20 develop different failure rates.
21 MS. GHOSH: So, I think we're going to 22 take your comment and re-check the numbers after this 23 meeting, because we're not going to be able to resolve 24 it right here.
25 CHAIR STETKAR: 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
36 1 MS. GHOSH: So, we understand the comment.
2 The one thing I will say, I don't think we 3 should attribute too much precision to all of this, 4 because our main focus was to make sure we were within 5 this range of sampling where we could get early 6 containment failure.
7 So, in terms of the high-level insights 8 that we can draw from this study, we feel that we 9 accurately explored the two-dimensional sample space.
10 There may be some discrepancies between 11 the exact frequency and in parts of the sample space.
12 We'll have to go back and check that.
13 CHAIR STETKAR: Tina?
14 MS. GHOSH: Yes?
15 CHAIR STETKAR: I get it.
16 But you spend a lot of time in the 17 Appendix, like a couple of paragraphs, and this plot, 18 saying you fit a fifth-order polynomial because, well, 19 if you tried to fit a fourth-order polynomial, you had 20 a little bit of different glitches when you did the 21 differentiation.
22 So, you had to fit a fifth order. If you 23 wanted to use a simple model, why don't you use the 24 linear model? Failure on demand at time zero plus 25 lambda X. It would have been a straight line.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
37 1 The differentiation is pretty good; it's 2 got the same value at every demand. That's a real 3 simple model.
4 It's not as elegant, it doesn't have this 5 nice curvy shape to it.
6 MS. GHOSH: I understand that.
7 CHAIR STETKAR: But it's pretty simple in 8 the sense of we're just trying to get an approximation 9 and focus on this range of what might be important in 10 that range.
11 MR. FULLER: Ed Fuller, from the Office of 12 Research.
13 I want to make a comment on what John was 14 saying because it was an issue that we encountered 15 years ago, when I worked at Pole Star, doing a 16 steam-generator tube integrity risk assessment for a 17 variety of plants.
18 And we looked at operational data and test 19 data at the time, to come up with some notion of 20 whether or not there was a distinct difference between 21 what would happen on the first demand versus all the 22 rest.
23 And we concluded that at the time, anyway, 24 most of the time when the data was pertaining to one 25 demand right away failure, it was due to some NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
38 1 maintenance errors.
2 And so I think, my own view is, take what 3 John said and refit, starting with Demand 2; exclude 4 the first demand from your Focus Study.
5 CHAIR STETKAR: The main report does that, 6 and it does it in -- I tend to be not the most 7 pleasant person in the world -- but in a very elegant 8 way, the main body of the report.
9 The way it treats three individual demands 10 -- three individual valves, sequential demands, 11 sampling of the number of cycles to failure, is a very 12 elegant model.
13 The only questions that I had on this 14 Focus Study was the values that are used to plot that 15 curve.
16 And then as a separate, non-related error, 17 a question to make sure that you did use that initial 18 demand failure for sampling for Demand 1 in the Focus 19 Study.
20 DR. HATHAWAY: I don't know if this will 21 clear this up or not --
22 CHAIR STETKAR: It's a factor of ten 23 higher.
24 DR. HATHAWAY: In the UA, it was two 25 distributions. It has one on the first sample 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
39 1 then it has one on the second.
2 What I'm trying to do by using the data is 3 fit to that so it convolves those two distributions 4 together.
5 CHAIR STETKAR: But your curve I-2 shows 6 the concept of what's in the baseline study.
7 It shows a delta function at one, which 8 corresponds to the conditional probability that it 9 fails on the first time.
10 And then it shows some sort of shape in a 11 accumulative form as you accumulate demands.
12 Now, the shape of that curve kind of 13 depends on the model that you use, and the slope of 14 that curve depends on the failure rate that you use 15 per successive demand.
16 The same failure rate is used for each 17 successive after the first, in the baseline model.
18 And depending on -- it's not a linear 19 model if you account for the accumulative effects from 20 failures.
21 But in a simple term, you could use a 22 linear model, especially over the first 50, 60, 70 23 demands or so. This didn't.
24 And then when you differentiate this 25 curve, you'd better account for that delta function at NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
40 1 one.
2 And if you're not, you've got a real 3 problem because you're ignoring a factor of ten or so, 4 or more than a factor of ten, conditional probability 5 that it fails on the first demand, which is, indeed, 6 much more closer to the regime that you're trying to 7 test in this Focus Study.
8 You're trying to see how sensitive the 9 results are to earlier failures with a large open 10 area.
11 MEMBER BLEY: Just maybe Figure I-3, which 12 is the density function, was drawn without showing the 13 delta function either.
14 CHAIR STETKAR: That's the reason why I 15 ask it. It could be.
16 MEMBER BLEY: Maybe it's just the way you 17 drew the Figure, and you didn't show that three minus 18 two, right, at one.
19 That's kind of what John was hoping you 20 would say.
21 CHAIR STETKAR: Yes, I was really hoping 22 you would say, well, of course we used the three minus 23 two for the first demand. Or two and a half minus 24 two, or whatever it should have been.
25 But I didn't hear 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
41 1 MS. GHOSH: If you go to Figure I-4.
2 CHAIR STETKAR: I-7?
3 MEMBER BLEY: I-4?
4 MS. GHOSH: Yes, if you have that.
5 It has the actual sample then successful -
6 - we were showing this for a different reason, to show 7 that there wasn't a huge dependence in terms of the 8 completed runs.
9 But at least to the first order of 10 magnitude, we did sample that first demand higher.
11 CHAIR STETKAR: You did? I'm trying to 12 pull it up right now.
13 MS. GHOSH: But I think in terms of the 14 larger question of the exact numbers, we'll have to 15 get back to you on that and we'll double-check.
16 MEMBER BLEY: Somehow, I'm not seeing it.
17 CHAIR STETKAR: Yes, I mean, if I had seen 18 it in that Figure, I would have probably mentioned it.
19 Yes, I'm not seeing it here.
20 MS. GHOSH: So, maybe we can talk during a 21 break. We should probably move on and we can get back 22 to you on that.
23 MEMBER BLEY: Are you sure you meant I-4?
24 Okay.
25 BB: I know there's a bunch 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
42 1 CHAIR STETKAR: Turn your mic back on, 2 Tina. Don't --
3 MS. GHOSH: Sorry, the bottom right of 4 that six-figure -- there's a bunch of charts in that 5 Figure.
6 Yes, I apologize, I meant to release all 7 of them. Maybe we'll catch up later, and we can move 8 on. Yes? Yes, sorry.
9 MEMBER BLEY: Thanks.
10 DR. HATHAWAY: So, we ran the models, and 11 what we really looked at is when we constrained our 12 distributions to this small portion of the sample 13 space, where early containment failure was the most 14 probable, we got about that 17 percent of our 15 realizations resulted in early containment failure.
16 And looking at it, about 15 percent of the 17 realizations, the BOC core conditions had an early 18 containment failure of 16 percent of the MOC and 19 19 percent of the EOC.
20 So, the table on the right shows just the 21 statistics for the various realizations. The mean of 22 the BOC was about 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />, the mean of the MOC was 23 about 6.6, and the EOC is about 6.6.
24 And what you also see is the BOC's kind of 25 constrained closer to the mean, where the MOC and the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
43 1 EOC are spread out further.
2 And we'll see this a little later on when 3 we look at the cesium iodine release rates as a 4 function of time.
5 CHAIR STETKAR: Trey, I don't know where 6 to ask this so I'll ask it now, and tell me if I 7 should ask it later. Let me get my notes here.
8 See, you see this number 17 percent? 17 9 percent of the realizations in the Focus Study 10 resulted in early containment failure.
11 What you really mean is 17 percent of the 12 361 successful realizations had that behavior.
13 Out of 600 realizations, there were 239 14 that did not run to completion, which is 40 percent.
15 40, 4-0, percent did not run to completion when you 16 tried to force this.
17 It's curious, and I went back and I did 18 the numbers, that in the baseline study, 40 percent of 19 the runs in the area that had a large early open 20 fraction did not run to completion. Same percentage.
21 So, something about MELCOR in this region 22 is not very good. 40 percent of the runs are not 23 running to completion in this particular region.
24 A much higher percentage run to completion 25 everywhere else.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
44 1 MEMBER CORRADINI: But I thought we had 2 asked that.
3 CHAIR STETKAR: Yes, and we never got an 4 answer.
5 MEMBER BLEY: We talked about it a lot.
6 CHAIR STETKAR: We never got an answer.
7 MEMBER BLEY: The way it is is kind of --
8 CHAIR STETKAR: Yes, that's the way it is.
9 Yes, we looked at it and it doesn't finish. Now --
10 MEMBER CORRADINI: But I thought -- if I 11 just want to interrupt you, John -- I thought you guys 12 promised, I thought, that you were going to go back 13 and try to unravel the reasons for failure in that 14 area?
15 CHAIR STETKAR: That's why I brought it up 16 now, because I don't think they're planning to discuss 17 it.
18 The reason I bring it up now is that in 19 the results from this Focused Safety Study, they were 20 reflected back through the overall insights and 21 conclusions.
22 There is an inference that the same 17 23 percent early containment failure and the same 24 conditional consequences apply for that 40 percent of 25 the runs that did not finish.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
45 1 And I do not know why that must be true. I 2 have no evidence --
3 MEMBER BLEY: Or at least one wonders --
4 CHAIR STETKAR: I don't know if it is 5 true.
6 MEMBER BLEY: -- if there's something 7 going on where there could be a lot more of them.
8 CHAIR STETKAR: I would have been really 9 happy if like 95 percent of these runs ran to 10 completion and you saw that same fraction, but they 11 didn't.
12 MEMBER BALLINGER: To turn it around, if 13 you have 40 percent, I think I made this comment 14 earlier.
15 If you have 40 percent that don't run to 16 completion, what's to say that the numbers that did 17 run to completion are actually good numbers?
18 I mean, if there's something fundamentally 19 going on here, the fact that you've got runs that went 20 to completion and you get numbers kind of makes you 21 feel good, but what if there's something fundamentally 22 wrong?
23 So, that's why I think we ask that you 24 folks go back and find out exactly why the runs that 25 didn't go to completion, what caused them to do 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
46 1 Was it some convergence issue? Or whether 2 it's some flat-out mistake in a model or something.
3 Because it does reflect back on these numbers.
4 MS. GHOSH: So, I'm going to just note 5 something that I --
6 MEMBER BALLINGER: Remember, I'm 7 metallurgist so...
8 MS. GHOSH: Yes, I'm not going to give a 9 satisfactory answer, but one of the primary 10 motivations to do this Focus Study, to do a bunch more 11 realizations in this area was that we knew from the 12 overall UA that we had 20-some, I think, realizations 13 that we had tried to sample in this area.
14 And we had a 30-some percentage failure 15 rate, which is much higher than the overall UA, which 16 is pretty small.
17 So, we knew that there was something going 18 on in this sample space, that we were getting higher 19 incompletion rates.
20 So, that was one of the major motivations 21 to do this study, is just to explore the response 22 surface and the smaller area much more.
23 So, we have 361 additional points in this 24 area, so we have more information but it doesn't 25 answer the second part of your 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
47 1 MEMBER BLEY: But you didn't put any 2 diagnostics in the code or anything to see where it 3 was failing, what was going on?
4 MS. GHOSH: So, I'll let Casey, and if 5 Hossein wants to speak, speak next. We did have our 6 Code Guides.
7 We look at the reason for the failure in a 8 bunch of these cases, and in many of them, it was 9 frustrating.
10 I think we had documented some, or maybe 11 we didn't end up putting it in the Appendix --
12 CHAIR STETKAR: It is not documented 13 anywhere in the report, I can guarantee that.
14 MS. GHOSH: Okay, so that's probably 15 something we could add because at least we know what 16 the error meant. We know what the error message was, 17 and in some cases it's easily explainable.
18 I think some cases, things are just 19 happening too fast, the time steps are too small and 20 you get hung up.
21 But in a bunch of cases, it was this 22 executive abort signal detected, which is really 23 frustrating, because it's very hard to untangle.
24 But let me turn it over --
25 MEMBER BALLINGER: I'll say the same thing NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
48 1 as I did last time, and that is sometimes, exploring 2 why these things didn't go to completion provides you 3 with what I would call surprise, which actually means 4 something.
5 And so I may be a metallurgist, but if a 6 student comes to me with results from a model he's 7 written -- he or she, excuse me, has written -- and 8 says that 40 percent of the time, the thing didn't 9 converge, then I would reject the results out of hand.
10 I would say there's no reason for me to 11 believe that the ones that did run to completion were 12 not fortuitous.
13 MEMBER CORRADINI: So, you guys don't have 14 to try to explain more. My interpretation of what 15 you're saying is you don't know, even though you tried 16 to look.
17 So, if there is a documentation of however 18 many it is and the failure message, it would be, to me 19 -- and this is not a criticism because these are 20 highly non-linear calculations that go off the 21 reservation, potentially more easily, and the fact 22 that you had the small percentage of the bigger 23 population is a good thing.
24 But I do think you've got to do some sort 25 of postmortem as to why.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
49 1 MEMBER BLEY: I'd like to hear what Casey 2 has got to say.
3 MEMBER BALLINGER: He's a thermo-4 hydrologist guy so he gets away with more.
5 MR. WAGNER: Casey Wagner, Dycoda.
6 There's a couple of things that I have to point out.
7 This was already done and it used the same 8 code as the last calculations.
9 So, whatever problems we were running into 10 on the last set of calculations, we were going to run 11 into it on this set of calculations.
12 MEMBER BALLINGER: See, I rest my case 13 then.
14 MR. WAGNER: Well, more calculations were 15 done there.
16 It went to a small time step and 17 eventually, if you have enough time steps that are 18 very small, it would give up and say that it was not 19 able to converge. And so that was the stoppage.
20 The convergence criteria is there, I don't 21 know if this will satisfy you, to assure that you do 22 get convergence, and if you don't have convergence, it 23 keeps on cutting the time step, trying to get there.
24 If it doesn't get there, those 25 calculations are stopped, and we are considered NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
50 1 failed.
2 The other ones, presumably, did satisfy 3 that convergence criteria, and so we do accept them as 4 successful and reasonable calculations.
5 MEMBER BLEY: It just seems there's no --
6 since you don't know quite what's going on, I don't 7 see any basis for assuming that the same percentage of 8 containment failures are in this group that died as in 9 the other calcs.
10 There might be something that's happening 11 on the way to containment failure that's blowing up 12 the code.
13 MEMBER MARCH-LEUBA: Casey, can I ask you 14 a couple of questions?
15 You mentioned you're using the same code.
16 Do you mean the same version of the code, or the same 17 number?
18 MR. WAGNER: It is the exact same version.
19 MEMBER MARCH-LEUBA: And all these 20 variables didn't happen on initialization? It 21 happened in the middle.
22 It did not happen during initialization, 23 it happened in the middle of the transfer?
24 That typically happens when one of the 25 nodes goes into other reservations that we mentioned NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
51 1 before. Another's supposed to have water, it doesn't.
2 A little one goes through somewhere, and 3 that happens all the time in other codes. But if it's 4 water, somebody dies.
5 CHAIR STETKAR: When you say all the time, 6 it's, to me, knowing nothing about the codes, it's 7 just curious that in the baseline study, at least 8 according to what I understand from the report, they 9 only had 40 chances to get into this regime.
10 And out of those 40 chances, 17 didn't 11 run. Here, they forced it to get 600 chances to get 12 into this regime, and the same fraction of times 13 didn't run.
14 MEMBER MARCH-LEUBA: You give these codes 15 too much credit.
16 CHAIR STETKAR: But, okay, you folks are 17 familiar with it, but we are in fact relying on this 18 code to draw conclusions and insights about kind of a 19 fundamental behavior of this machine.
20 MEMBER MARCH-LEUBA: The experience of 21 people that run this code will tell you that they run 22 10 runs when it crashes for every run when it 23 survives. And you converge on the model and you fix 24 everything.
25 CHAIR STETKAR: But that wasn't the case.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
52 1 In more of the well-behaved regime of this whole 2 study, they got something, I don't know, Casey, 90 3 percent success rate or better?
4 MEMBER MARCH-LEUBA: Eventually.
5 CHAIR STETKAR: No, not eventually.
6 MEMBER MARCH-LEUBA: No, when you're 7 trying to develop the input there, you're trying to 8 describe your plant, you try once.
9 You run it, it crashes. Then you fix it.
10 You run it, it crashes. And eventually, you fix it --
11 MEMBER BLEY: We kind of get that, but 12 what John's talking about is there's this whole space 13 they did calculations, 90 percent of the time, the 14 code runs great.
15 There's this little vulnerable area for 16 early containment failure. When they run in that 17 regime, 40 percent of the time it crashes.
18 So, there's something in there that's 19 causing trouble.
20 MEMBER MARCH-LEUBA: My point that I 21 obviously didn't do correctly is that's worrisome 22 because it happens in the middle of the run.
23 Something is happening.
24 MR. WAGNER: Yes, I can explain what's 25 happening to get it to make sense 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
53 1 For those cases where the valve sticks 2 open, we're moving, we're generating more hydrogen, 3 we're going into early containment failures, we're 4 getting more core damage.
5 And then the accumulator dumps, and so 6 we're dumping an accumulator into a much more degraded 7 core than the cases that are high-pressure, that we 8 get the hotleg failure, and the core's relatively 9 intact, and it quenches it pretty quickly.
10 And then we move onto a boil-off severe 11 accident. This is much more challenging because the 12 core has formed a debris bed and then we did a hotleg 13 failure, or it seeps into the accumulator dump.
14 MEMBER CORRADINI: So, let's just backup.
15 So, I don't do this anymore because I probably 16 couldn't do it right. I couldn't even get the model 17 to run at times.
18 So, if I were to do a detective job on 19 this, the first question I ask is of the 40 percent 20 that failed, did they all fail at approximately the 21 same time in terms of event sequence?
22 You are hinting that the event of an 23 accumulator, essentially, hotleg creep rupture, and 24 then an accumulator dump, is a common time-to-receive 25 failure.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
54 1 Is that the case?
2 CHAIR STETKAR: This is a step up in valve 3 but I just think --
4 MR. WAGNER: Yes, we also get the hotleg 5 failure, even if the primary system pressure has come 6 down.
7 The one thing that I don't think Trey 8 shared with you is that we did forensics because we 9 had so many more samples here on our uncertain 10 parameters.
11 And was there in space in them where we 12 didn't have representative successful samples?
13 And Trey did a very thorough job of going 14 through all the different parameters, and they're 15 interspersed. They're like this, failures and 16 successes. And so there wasn't a common theme there.
17 On our sample parameters, we were missing 18 a space.
19 We unfortunately got the 40 percent, which 20 is very high, but the interspersed end of 21 representation of what we were trying to look at for 22 uncertain parameters was very good.
23 And we judged that as success in a lot 24 more cases.
25 MEMBER CORRADINI: Okay, I get your point.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
55 1 But what I'm asking, from the standpoint 2 of just rudimentary failure times, is there an event 3 that MELCOR records that at, or right after, that 4 event, everything goes to hell in a handbasket 5 regardless of the uncertain parameters or the variable 6 parameters.
7 MR. ESMAILI: This is Hossein Esmaili, can 8 I just say something? I think we don't know right now 9 what caused this error.
10 We can going back and look at it, and when 11 Larry was here on April 18, we went through the whole 12 thing. We said that this is our quench model, this is 13 how we do this, this is our comparisons with the 14 experiments, et cetera.
15 When you apply those models to 16 sophisticated whole-plan models, you can run into 17 whether it's on the core side or whether it's the 18 input model side.
19 We cannot answer this until we go to the 20 actual source code to try to find out what caused this 21 problem. But in most cases, we can resolve it.
22 It was not important, I guess as the panel 23 here discussed. It was not important because we had 24 an overall high success rate, and most of calculations 25 that we did was over 99 percent success rate.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
56 1 It's just that we got into this issue 2 here. And we are going to find out exactly what 3 happened in these calculations, and maybe we can let 4 you know later.
5 Right now, I don't think we can tell you 6 with certainty what happened in these calculations.
7 It's often one thing, one thing in the code, that once 8 you get over it, the calculations would go forward.
9 But I don't know until I talk to my 10 colleagues about it.
11 MEMBER REMPE: And so your diagnostics 12 don't tell you it's in --
13 MR. ESMAILI: No.
14 MEMBER REMPE: It doesn't tell you which 15 nodes? It doesn't give you any clues? It just --
16 MR. ESMAILI: It gives you some idea but 17 it doesn't tell you everything about it.
18 So, you really have to go run the code in 19 a de-bug mode, find out what happened there, and find 20 out exactly what was going on.
21 And in some cases, it could be a small 22 error in how things are done, and --
23 MEMBER REMPE: And in the diagnostics that 24 you've had, you've not gone through and the 17 --
25 MR. ESMAILI: It doesn'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
57 1 (Simultaneous Speaking.)
2 MEMBER REMPE: And the 17 or however many 3 hundred cases, you've not ever tried to say, oh, it's 4 always in that same place?
5 You've not had the opportunity to --
6 MR. ESMAILI: I don't know if we have done 7 that. I don't know if we have put in the tracing 8 model back on to the reference, but we can go back and 9 look at it more.
10 I'm just saying that we cannot resolve it 11 here because we don't have the right people, we don't 12 have them here.
13 MEMBER BALLINGER: So, you're on record 14 now as saying you're going to run this to ground?
15 CHAIR STETKAR: They said they'd get back 16 to us. From an academic sense, not that I understand 17 any of this stuff, I'm really interested to understand 18 why the code is blowing up.
19 From a bigger picture regarding this study 20 and its report, the study and its report does not 21 highlight the conversation we just had.
22 It basically says look, look, look, we did 23 this Focus Study, which is a really good idea.
24 The results from the -- we focused the 25 sample space on a regime that we really wanted 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
58 1 test, that was a) blowing up more frequently in the 2 baseline study, and b) generally more important for 3 offsite health consequences, early containment failure 4 anyway.
5 It has a good chance, but then we observe 6 the same behavior, that 40 percent of the runs didn't 7 run to completion. That's mentioned, it's documented, 8 but it's just sort of mentioned in passing.
9 And the thing that concerns me most from 10 the perspective of the study is the inference that 17 11 percent of that 40 percent of the runs that did not 12 complete would have behaved the same as the 60 percent 13 that did complete.
14 Ergo, yes, we might pick up two or three 15 more early containment failures from the runs that 16 didn't complete.
17 And therefore, our overall study results 18 and conclusions and insights aren't affected by those 19 run failures.
20 That latter point is the thing that 21 bothers me, and drawing that inference, I believe, is 22 misleading.
23 I think that you should just own up to the 24 fact that the runs didn't run to completion and you 25 don't really know what fraction of those incomplete NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
59 1 runs may have gone into early containment failure.
2 You know from the baseline study that it 3 can't be -- because I've got confidence in the 4 baseline study sampling -- you know that it can't be 5 more than I think 17 additional runs in the baseline 6 study.
7 So, it can't be 200 more samples in the 8 baseline study that would have gone into an early 9 containment failure.
10 But you don't really know what fraction of 11 those additional runs can go to early containment 12 failure, and I think the overall study should 13 highlight that.
14 It's just something we don't know right 15 now.
16 MS. GHOSH: Okay, so I think we understand 17 the comment. The one other thing I'll point out again 18 is this Figure 1-4, which is in the Appendix.
19 This was our attempt to, at least from a 20 statistical standpoint, to look at is there a region 21 of the inputs that we're putting in?
22 So, that's only just the initial 23 conditions. It doesn't get to what's happening 24 midstream at all.
25 But at least, was there a region of 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
60 1 this focus input space where we were getting excessive 2 failures.
3 So, the I-4, the one I think that Trey 4 summarizes, is, once again, the one area where you see 5 a little bit of a correlation in terms of the 6 failures.
7 And we already knew this, because we knew 8 this from the leger studies, that if you have very few 9 cycles dealing with big, open areas, you continue to 10 be more likely to have an incompletion rate.
11 But if you look at the successful 12 realizations from those Figures which are shown in 13 red, the blues were the samples that were attempted, 14 and then the red ones are -- so the blue is the CDF of 15 what we attempted to do.
16 The red is the cumulative distribution 17 function of what actually succeeded in completing.
18 You can see that what Casey was saying, 19 there's very good coverage across the range of the 20 inputs, and for most of the input and variables, we 21 get a very good match.
22 But one exception, again, is there's a 23 little bit -- you can see a little bit of a difference 24 in the CDFs for the number of cycles, and that's kind 25 of pushed to success for a higher number of cycles.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
61 1 So, you do see that same correlation.
2 But we, at least from a statistical 3 standpoint, not having had been able to do the full 4 forensics on these realizations, from a statistic 5 standpoint, we felt comfortable to say that we thought 6 it was likely that in the runs that didn't complete, 7 you would see the same percentage of runs going into 8 early containment failure.
9 So, again, from a purely statistical 10 standpoint with the input space --
11 MEMBER MARCH-LEUBA: And that's why I 12 wanted to get a little time, because I do know a 13 little bit about this topic.
14 Based on the comments we receive here, I 15 see likely unconcerned that the Staff is going to go 16 into a five-year effort into de-bagging all those 40 17 percent run.
18 I'm thinking, using engineering judgment 19 of the guys that know what's going on, can you look at 20 the run and it fail, but I see the passage going, this 21 is going to be okay.
22 And then I have to look at that one.
23 CHAIR STETKAR: That would be really good, 24 wouldn't it?
25 MEMBER MARCH-LEUBA: Our guys who 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
62 1 looked at the other 362 can look at one and say, yes, 2 it crashed, but this isn't going anywhere.
3 Good. Don't go into a five-year effort.
4 CHAIR STETKAR: And, Tina, I'm sorry, but 5 don't muddy it up with statistics on sampled 6 parameters. Make it an engineering problem.
7 MEMBER MARCH-LEUBA: In fact, it would be 8 perfectly good for this exercise. And you know it, 9 you look at the run and say I'm wasting CPU.
10 CHAIR STETKAR: At 239 runs, 57 of them 11 had this flavor, 52 of them had this flavor.
12 MEMBER BLEY: That would be much more 13 convincing than just saying it's got to be the same 14 percentage.
15 CHAIR STETKAR: Okay, I think we beat that 16 horse. That's not a dead horse, it's beaten.
17 DR. HATHAWAY: So, I'm going to be doing 18 the next part. The next Figures present the CDF, so 19 the hydrogen-generated, because that was sort of that 20 interest.
21 Really, when I first started out, I was 22 just kind of interested in what was the early 23 containment failure potential in this region.
24 So, just looking at the data a little bit 25 more, we went in and pulled out the mass of hydrogen NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
63 1 generated up to the time of first deflagration, and 2 the mass of hydrogen reaching the dome up to the time 3 of the deflagration reaching the dome.
4 Because those tend to be the important 5 aspects of the -- that tends to be one of the 6 important insights from the study.
7 So, what this Figure shows, the dots in 8 blue represent the BOC realizations, the dots in 9 purple represent the MOC, and the dots in green 10 represent the EOC.
11 In these Figures, the dots that are 12 highlighted in red represent the early containment 13 rupture cases, and what you can see looking at 14 hydrogen generation, there was slightly more hydrogen 15 generated in the BOC realizations.
16 But the EOC and MOC realizations overlap 17 pretty well.
18 But then when you look at the amount 19 reaching the dome, that was consistent across all the 20 realizations, and also, what you see is when you have 21 a larger fraction of hydrogen reaching the dome, those 22 tended to have a higher fraction of the cases leading 23 to early containment failure.
24 MEMBER BROWN: Excuse me, for the 25 uninitiated, what do you mean by early? I looked NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
64 1 here, and you see the hydrogen buildup, is that over a 2 6-hour, 10-hour, 20, 1 day, period? After?
3 DR. HATHAWAY: No, it's within the first 4 15 hours1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br /> of what the simulation.
5 What I meant by early is if you have a 6 failure due to the hydrogen deflagration, not 7 overpressurization, which is lead failure.
8 MEMBER BROWN: Okay, so it's about 15, in 9 the ballpark of 15 hours1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br /> for the hydrogen, for the 10 deflagration --
11 DR. HATHAWAY: So, if you looked at the 12 previous study, the times are presented on the right.
13 So, six up to ten hours is what was the result.
14 MEMBER BROWN: I couldn't connect the dots 15 after I listened for 30 minutes or 45 minutes on that 16 last interchange of why the runs didn't work.
17 So, I lost the ball on this one. So, 18 thank you, I appreciate it.
19 DR. HATHAWAY: -- the first 15 hours1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br /> for 20 this part of the analysis.
21 So, the next thing we kind of looked at is 22 what is the pressure response in the dome?
23 So, what this Figure shows, the blue dots 24 represent realizations and the redundant Focus Study.
25 So, you can see that those have the early valve NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
65 1 failures on that range of 1 to 65.
2 The purple represent the results from the 3 full UA, and what this demonstrates is in the full UA, 4 the axis on the right represents the difference 5 between the peak dome pressure and the sample 6 fragility.
7 So, what that really shows you is how 8 close that early burn gets to actually rupturing the 9 containment due to the hydrogen burn.
10 So, you can see that in the full UA, a lot 11 of the results don't get within 25 PSI of the sample 12 fragility, so they had smaller burns initially.
13 But in the Focus UA, you kind of fill out 14 that region between the -- well, going back, a lot of 15 the full UA studies have a hotleg failures, what the 16 initiating event is first. Or what the 17 depressurization then is.
18 So, when we were de-pressurizing before, 19 you would kind of fill out that Region between that 20 hotleg failure and the admission event.
21 And what you also see in the Uncertain 22 Figure is a lot of the burns were still -- when they 23 failed containment, it was when the sample fragility 24 was less than the mode of the distribution.
25 But also, just looking back, if you 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
66 1 at the distribution, 42 percent of the fragility 2 distribution is above the mode, but only 13 percent of 3 the cases gave you a burn large enough to get into 4 that region of the distribution.
5 So, after we did this, and really, this 6 was after the last ACRS Meeting, we looked at it and I 7 tried to look at the consequences analysis.
8 So, what we did is we went back and 9 grabbed all the cases.
10 Since we truncated these cases to 15 11 hours1.273148e-4 days <br />0.00306 hours <br />1.818783e-5 weeks <br />4.1855e-6 months <br />, we went back and ran them all from the 12 beginning and extended the problem 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />, to really 13 be able to look at what the consequences of these 14 cases are.
15 So, we did this for two sets. We looked 16 at all the early-containment rupture cases, but we 17 also looked at a subset of the late-containment 18 rupture cases.
19 And we randomly selected late-containment 20 rupture cases but tried to keep the same breakdown as 21 the time cycle.
22 So, if there were 11 percent in the 23 sample, try to pick 11 percent of the BOC cases, for 24 example.
25 So, what this Figure is, you can kind 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
67 1 see where the BOC realizations clustered around six 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />, but the MOC and the EOC kind of spread apart.
3 Also, these late-containment failures in 4 the UA, a lot of them were smaller releases.
5 But we kind of looked at this, this came 6 about very late, that some of these, when you had 7 valve failures, produced larger late releases.
8 So, that's why we selected a subset of 9 these late releases.
10 And this is just really showing the 11 results, the average cesium release for the early-12 containment ruptures was 2.2, and the iodine was 6.3.
13 The late-containment ruptures were 0.004 and 0.021.
14 MEMBER REMPE: So, can you go back to that 15 slide?
16 I was going to bring this up later, but if 17 you had done a sensitivity study where you assumed at 18 the beginning of the cycle you had a higher eutectic 19 melt relocation temperature, how do you think it would 20 have affected all these results?
21 I mean, everything always says, okay, the 22 cesium release fraction's kind of similar if it's all 23 at the beginning of cycle.
24 And, again, if you do kind of a thought 25 exercise on how all these things are interrelated 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
68 1 MELCOR, what would have happened?
2 Because you would have had a longer time 3 for the fuel to heat up, it would have gone to higher 4 temperatures, you would have had a lot of more 5 release.
6 And it would have really kind of thrown 7 things off kilter, and I was really disappointed that 8 you didn't do a sensitivity study.
9 I mean, you've got this 2500 plus or minus 10 83, instead of 200, which is what the Europeans 11 recommend.
12 And I kind of think that's an important 13 sensitivity technology, some of the limitations and 14 uncertainties associated with this plan.
15 And all plans, frankly. But any comments 16 on that?
17 MR. WAGNER: So, it's a timely comment 18 because we're getting ready to do Surry and we 19 struggled with your comment on how do we address it?
20 It makes sense that BOC is going to have a 21 higher melting temperature, and maybe that's something 22 we could look at in Surry, certainly at the 23 sensitivity at a minimum.
24 But I agree with your conclusions but the 25 releases would probably be a little bit higher at BOC NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
69 1 potentially because --
2 MEMBER REMPE: Hydrogen generation could 3 be higher. It could really kind of be an outlier 4 there, that it's not being explored.
5 And so, again, it's kind of like you get 6 up in the Executive Summary when you come up with 7 these general insights from these results.
8 I'd really like to see some limitations to 9 those insights, because of things like that, so that 10 you don't get yourself over a bind, in this case, 11 because there's some uncertainties.
12 In the case of the consequential tube 13 rupture, your scope of your study is bringing you to a 14 certain conclusion.
15 Whereas, if you consider other ongoing 16 studies even, there's some knowledge that you guys 17 have that ought to be acknowledged as a limitation.
18 And so that's why I kind of wanted to 19 bring it up here because --
20 MR. WAGNER: To your comments there, maybe 21 our standard deviation about that is too narrow. The 22 inverse is maybe 200 and we really ought to go back 23 and rethink that BOC again.
24 MEMBER REMPE: Yes, at this point, I'm not 25 trying to say do more calculations, although, 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
70 1 like to see what that result looks like because of the 2 way the models have been constructed.
3 But just to acknowledge the limitations 4 and uncertainties.
5 MEMBER POWERS: I've tried diligently not 6 to make comments during this, but I can't anymore.
7 Do you really, really honestly think that 8 melt relocation depends at all on how long the fuel's 9 been burned up?
10 MR. WAGNER: I'd be looking for Randy at 11 this moment. I thought so.
12 MEMBER POWERS: Do you have any solid 13 evidence that the accumulation-efficient products as 14 the fuel burns up changes the melting point of the 15 fuel significantly?
16 MR. WAGNER: That's why we weren't able to 17 address Joy's comment. We said we agreed with the 18 hypothesis but we didn't know, we didn't have data to 19 --
20 MEMBER POWERS: Well, there are data in 21 the literature.
22 There are several people who have tried to 23 measure the melting point of irradiated fuel, and they 24 usually come back with results that are not wildly 25 different from the melting point of unburnt fuel.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
71 1 The fact of the matter is that in that 2 particular range for LWR fuel, there's sufficient 3 uncertainty and variability just by small changes in 4 straight geometry and whatnot, that it's very hard to 5 say.
6 But the overall phenomenon of melt 7 relocation really doesn't depend very much on the 8 melting point of the fuel. It really has to do with 9 liquefaction with the clad interacting with the fuel.
10 And it seems to me that if you were to 11 hypothesize burn-up has an effect, it would be because 12 of something happening on the inside of the cladding 13 to give you a larger region of oxide formation there, 14 that inhibited liquefaction.
15 It would have had nothing really to do 16 with burn-up, but it had to do with internal oxidation 17 of the cladding.
18 MEMBER BALLINGER: But that is a function 19 of the lambda, right?
20 MEMBER POWERS: It can be depending on 21 what's happening on the inside --
22 MEMBER BALLINGER: So, the film on the 23 inside of the cladding is going to be a function of 24 the potential?
25 MEMBER POWERS: It depends on how the fuel NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
72 1 chooses to behave.
2 And in particular, what it seems to depend 3 most upon is how much of metallic modules you'd get 4 in.
5 Because the molybdenum acts as a buffer 6 for the accumulation of oxygen, and the availability 7 of the oxygen to go oxidize that cladding.
8 So, it depends on a lot of things, but not 9 which are sensitive enough, since the code works in 10 nodes that are like this, as opposed to individual 11 fuel rods, you're averaging over a lot of things here 12 to get liquefaction.
13 And liquefaction is treated in the code in 14 a relatively non-mechanistic fashion.
15 So, you do what they've done.
16 It's they take some breadth of that 17 relocation number based on matching a variety of 18 experiments, and they say, okay, you match these 19 experiments and we take some temperature range as a 20 criterion for relocation.
21 Honing that number down based on some 22 hypothesis about what's going on strikes me as kind of 23 futile.
24 MEMBER BALLINGER: They say about the 25 great tragedy of science, the slaying of a beautiful NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
73 1 hypothesis by an ugly fact.
2 MEMBER REMPE: So, there's also thermal 3 cracking, there is cracking during the radiation. But 4 the thing is that --
5 MEMBER POWERS: Joy, what happens --
6 MEMBER REMPE: Let me finish just for a 7 minute, sir.
8 There's uncertainty with the phenomena but 9 there's also what's embedded in the code, and because, 10 again, there's uncertainty in the phenomena, I think 11 it's a useful exercise to bump that thing up a bit.
12 Because, again, they only had plus or 13 minus 83 degrees, and I've been reading up a lot of 14 stuff from France, where they're saying there needs to 15 be a larger uncertainty.
16 But just to see how it's all connected, 17 because I would challenge or query that I think that 18 if you did bump it up, you might see that because of 19 the way the models are interspersed in that code, 20 you're going to see some significant differences.
21 MEMBER POWERS: And presumably, if I take 22 the ranges large enough, I will get some pretty 23 radical differences. I'm sure of that.
24 MEMBER REMPE: Yes.
25 MEMBER POWERS: That one I'm confident 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
74 1 MEMBER REMPE: And I would like to see the 2 sensitivity to explore that because they're getting 3 some very similar things.
4 And I think it would be worthwhile to 5 understand if it's realistic if you go a bit higher, 6 even if your range is up --
7 MEMBER POWERS: I guess I don't understand 8 how increasing your understanding improves your --
9 increasing your uncertainty improves your 10 understanding.
11 A little bit of a mystery to me but I'm --
12 MEMBER REMPE: But, again, they've limited 13 it to VERCORS for their plus and minus 83, and if you 14 consider other things like PHEBUS, the French are 15 saying go to plus or minus 200.
16 MS. GHOSH: Can I just clarify --
17 MEMBER POWERS: I think they're centered 18 around PHEBUS.
19 MEMBER REMPE: The report indicates that 20 they've centered a lot on VERCORS.
21 MS. GHOSH: That's true, but I do want to 22 clarify --
23 MEMBER POWERS: VERCORS didn't involve any 24 relocation at all.
25 MEMBER REMPE: It's their eutectic melt NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
75 1 temperature, and that's what they based it on.
2 MS. GHOSH: I just want to clarify that 83 3 is the standard deviation.
4 MEMBER REMPE: Right.
5 MS. GHOSH: So, our bound's actually go up 6 to 200-plus, up and above. You just have less samples 7 in that area.
8 MEMBER REMPE: Right. But I think an 9 uncertainty would be worthwhile to explore.
10 MEMBER POWERS: I don't understand at all 11 now. If the sigma is 83, then 3 sigma is huge.
12 MEMBER REMPE: I don't think we ever got 13 up to that Figure. What was the little red dot?
14 Was it 2800 or something, and you never 15 had any sampling up there at all?
16 MEMBER POWERS: The 3 sigma would...
17 CHAIR STETKAR: But again --
18 MEMBER POWERS: ...violate the --
19 CHAIR STETKAR: Please.
20 MEMBER POWERS: -- annihilate the 21 molybdenum issue, annihilate everything.
22 CHAIR STETKAR: This is an Uncertainty 23 Analysis, and if believe the uncertainty distribution, 24 if you get a small number of samples out in that tail, 25 you get a small number of samples out in that tail.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
76 1 As you should, if you believe the 2 distribution.
3 MEMBER REMPE: But then the conclusions on 4 the Regression Analysis are that the temperature is 5 very important, that it affects hydrogen, and 6 eventually, it becomes pretty important in the follow-7 on documentation.
8 And that's why I am emphasizing it.
9 MEMBER POWERS: I guess I'm still at a 10 loss that you concluded it's important.
11 MEMBER REMPE: But the models are --
12 CHAIR STETKAR: Well, it is important.
13 Tina, I don't remember your Regression Analysis. I 14 mean, it showed up in the mix.
15 Was it very important?
16 MS. GHOSH: The eutectic melt temperature 17 is clearly very important for hydrogen generation.
18 We've see that in all the studies.
19 It shows up in combination with other 20 variables for things like cesium release.
21 So, it's got high interaction effects, 22 again, for obvious reasons, because it has a big role 23 to play in the hydrogen generation, which then affects 24 early containment failure, et cetera, et cetera.
25 MEMBER POWERS: Also, I would think 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
77 1 the analyses of transport wouldn't mediate the 2 importance of fuel relocation on cesium release to the 3 containment.
4 Just because cesium escapes from the fuel 5 does not ipso facto mean that it goes into the 6 containment.
7 There are a lot of slips towards the cup 8 in the lip on that, and the uncertainties inherent in 9 that transport process might work to mitigate the 10 importance of fuel relocation.
11 Fuel relocation is very important for 12 hydrogen because you cook things for a long time, and 13 that does give you high cesium releases from the fuel, 14 but doesn't necessarily affect the transport process.
15 So, it doesn't lead -- there's not a one-16 to-one correspondence between every location, and 17 cesium release, just because of the transport process.
18 And quite frankly, you're putting a lot of 19 heat in the fuel and you haven't got heat to put into 20 the piping system. So, you end up with a cool piping 21 system which leads to high deposition rates.
22 MR. WAGNER: Our volatiles are pretty 23 close to 100 percent release.
24 MEMBER POWERS: You can't get --
25 (Simultaneous Speaking.)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
78 1 MEMBER POWERS: That's released from the 2 fuel, but it's not released into the containment.
3 MR. WAGNER: Right.
4 So, over most of these cases, we do get 5 the volatiles out, but maybe I misspoke that our 6 releases might change a lot.
7 Our hydrogen could change and that has 8 dramatic effects, but maybe not so much what gets off 9 the fuel.
10 The release coefficients that we have from 11 the fuel, but most of the volatiles come off.
12 MEMBER REMPE: I'll have to look where I 13 saw it, but I think there's a discussion somewhere on 14 the report that does kind of have a different 15 perspective.
16 But the higher relocation temperatures 17 would increase the releases from the fuels somewhere.
18 And I'd have to find where it is, but that may be 19 something to look for.
20 DR. HATHAWAY: Okay, so I'm just sort of 21 moving to the Consequence Analysis.
22 We took all of the source terms that were 23 run to 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />, and we put them in the uncertain 24 MACCS model of the site to see what the latent cancer 25 fatality risk for the zero to ten-mile interval.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
79 1 And again, as in the full study, these are 2 conditional on the occurrence of the assumed --
3 CHAIR STETKAR: Trey, I want to make sure 4 I understand both of these tables, but 11 is the one I 5 want to understand.
6 These results, to me, mirror the results 7 in the baseline study, which is encouraging in the 8 sense that you're showing a bimodal effect on the 9 conditional risk, the way the dots on the lower left 10 are clustered.
11 And the numerical ranges are similar.
12 The difference is that the first column in 13 11 that you called really Containment Rupture is a 14 subset of the scenarios in Figure -- I'll give you the 15 Figure number -- 6.1 and 6.3 -- so that the right-hand 16 side of those Figures, the 87 percent that goes to 17 early containment failure.
18 The first column in this table is a subset 19 of them, right?
20 The second column, the thing that's called 21 late containment rupture, is a combination of so-22 called late containment failures that are also in the 23 right-hand side of those figures in Chapter 6.
24 And the small set that never go to 25 containment failure which is the far left-hand side, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
80 1 where the risks are really, really small, the other 2 mode of that bimodal behavior.
3 And therefore, comparing the numbers in 4 this table to numbers that appear in other sections of 5 the report is a little bit misleading. I mean, you 6 really have to kind of look through that stuff.
7 I was surprised that you didn't cast these 8 results in the same way as the results in Chapter 6, 9 so you could see that, yes, when we turned up the 10 microscope in this area, we basically got the same 11 shape, we got the same fractions in the bimodal, and 12 roughly the same ranges in the conditional latent 13 cancer fatality risk, coming through the zero to ten-14 mile snapshot here.
15 You can kind of see it from this, but you 16 really have to, I think you really have to, study it.
17 Have I mischaracterized anything, Tina? I 18 mean, you're really familiar with all of those curves.
19 MS. GHOSH: I think it's a good comment 20 and we can supplement the --
21 CHAIR STETKAR: I mean, basically, on the 22 Figure on the lower left, if I take a horizontal slice 23 in that Figure at about 1e to the -6, everything above 24 that horizontal slice, is what's on I'll call it the 25 right-hand side of the Figures in Chapter 6.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
81 1 Everything below the horizontal slice is on the 2 left-hand side of those Figures approximately, right?
3 I just want to make sure that I got it.
4 MS. GHOSH: Yes, I think that's right.
5 These results are spread out a little bit more.
6 CHAIR STETKAR: Oh, yes, they're going to 7 be spread out a little bit just because what you did 8 here. But in the big-picture concept --
9 MS. GHOSH: Yes, I think that's a good 10 idea. We can add a paragraph to compare back to the -
11 -
12 CHAIR STETKAR: Because I like what you 13 did in Chapter 6, by the way.
14 We talked about that bimodal behavior, and 15 then you split it out to more explicitly show -- and I 16 really liked that, that kind of -- emphasize some of 17 the points you were trying to make in the text.
18 And you can draw the same conclusions from 19 this but it's a lot more difficult. I mean, you have 20 to really study this. Okay.
21 DR. HATHAWAY: Okay, I probably won't 22 dwell much on this.
23 The top table just is the sort of 24 statistics on the total risk where the bottom is sort 25 of statistics but looking at the emergency-phase NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
82 1 contribution, and the long-term phase and 2 intermediate-phase contribution individually.
3 So, finally, because we had more of these 4 early containment ruptures, we were interested in the 5 early fatality risk.
6 Previously, we only had four realizations 7 that had early fatality. So, we just looked at these 8 results. Past four miles, there was zero calculated 9 early fatality risk.
10 And here are just the statistics, and 11 again, this is a conditional long-term event 12 occurring.
13 And what we tried to do, and this sort of 14 all is dependent on the discussions from earlier, but 15 we tried to scale these results to consider the full 16 distribution because everything outside of this 17 parameter space would be to zero.
18 So, we tried to just scale it out to 19 consider the entire distribution, and that's what that 20 second column is.
21 It would be what the risk would 22 potentially be if you were to essentially add a lot of 23 cases that had late containment.
24 CHAIR STETKAR: A comment on Chapter I 25 think it's 6, where these similar tabulations are NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
83 1 shown.
2 It would help a lot of people to provide a 3 little bit more of a tutorial in that section about 4 how one can have a mean value that's something on the 5 order of ten to the minus, I don't know, small number, 6 eight, or whatever those numbers are, nine.
7 And yet, have precisely zero values for 8 the median 5th and 95th percentile.
9 The only reason I dwell on that is that 10 people who use very broad and very skewed 11 distributions routinely understand that.
12 Not necessarily everyone who reads this 13 study will have that level of understanding, and a 14 little bit of a tutorial will help to those folks to 15 not get the impression that this is simply some sort 16 of statistical mumbo-jumbo that has no basis in 17 reality, kind of anchoring them back to say that, yes, 18 I can get 95 percent of my samples to have precisely 19 zero, and yet, the mean value can still be non-zero.
20 And show them why that is.
21 Otherwise, the naysayers will say, well, 22 obviously, this is just statistics and numbers and 23 numbers games.
24 So, I think that would help in Chapter 6.
25 Here also, but here, at least, most of them, the 95th NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
84 1 percentile is above the mean, and that's reassuring to 2 a lot of people.
3 DR. HATHAWAY: So, that was all I was 4 going to cover in this.
5 CHAIR STETKAR: We're at a break time 6 here. No, any more questions, first of all, on 7 Appendix I? Because we're going to shift gears here 8 entirely.
9 If not, let's take a break, and I'm going 10 to try to hold us to ten minutes. I'll probably fail 11 but try to get back here by 10:20; we'll recess until 12 then.
13 (Whereupon, the above-entitled matter went 14 off the record at 10:08 a.m. and resumed at 10:21 15 a.m.)
16 CHAIR STETKAR: We're back in session. We 17 are back in session. Trey, Tina, I don't know, it 18 says Casey. Anybody?
19 MEMBER BLEY: I don't know when to ask the 20 -- thank you. I don't know when to ask this, but I'm 21 going to ask it now, as we move into more general 22 things. Sure.
23 (Laughter.)
24 MEMBER BLEY: But it's kind of the same 25 cast of characters here and who we meet in the Level 3 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
85 1 PRA work. And over in the Level 3 PRA work, I think 2 we've learned that some of the things you're doing in 3 SOARCA aren't making its way over there, because it's 4 not state of the practice.
5 But something comes up over there, I'm 6 curious about it, and over there, in our recent review 7 of the Level 2 PRA, I find a whole section of their 8 report on model uncertainties that's really pretty 9 thoroughly laid out, at least the thinking of things 10 that might not work the way we expect them to work and 11 ways that they might consider them in the report.
12 I don't find anything quite as thorough, I 13 mean, you do a tremendous job on the parametric 14 uncertainties, but a few sensitivity studies that are 15 kind of picked here and there, but nothing as 16 systematically thorough, of saying, where could there 17 be things about the model that aren't right, that I 18 see in this new part of the Level 3 PRA.
19 Have you thought about that? And probably 20 the work you've done here led to how that evolved, but 21 I would sure like to see it here. It seems that part 22 of the uncertainty analysis that we've always kind of 23 talked about and never gone very far on over here.
24 MR. FULLER: Yes. So, Dennis, 25 unfortunately, Don and Hossein, who were, along with NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
86 1 myself --
2 MEMBER BLEY: They were here all morning.
3 MR. FULLER: They just walked in and they 4 missed your whole question.
5 (Laughter.)
6 MEMBER BLEY: You want them to handle it, 7 you don't want to handle it.
8 MR. FULLER: So, what I will say is this, 9 the team that worked on that document and that work, 10 obviously Don Helton is the lead for all of that.
11 MEMBER BLEY: Right.
12 MR. FULLER: It's the same team, so you 13 could say that --
14 MEMBER BLEY: It kind of seemed that way to 15 me.
16 MR. FULLER: Yes. So, the portion of the 17 SOARCA team that's been thinking about, we kind of 18 focused our efforts in supporting Don in the Level 3 19 PRA study for that more systematic thinking. And, of 20 course, the scope is much broader for the Level 3 PRA.
21 MEMBER BLEY: It is, but --
22 MR. FULLER: Yes.
23 MEMBER BLEY: -- SOARCA makes a claim that, 24 one, it finally acknowledges that it isn't a risk 25 assessment and you pick out particular scenarios. But NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
87 1 then you do a thorough consequence analysis and now, 2 you've added to that a thorough uncertainty analysis.
3 And, in fact, the parametric uncertainties 4 affect the results substantially, and we see that.
5 But the potential of the modeling uncertainty is to 6 overwhelm what we see in the parametric uncertainties 7 has always been an issue and we've talked about it 8 occasionally.
9 But this new part of the Level 3 PRA on 10 model uncertainties and the Level 2 analysis is the 11 first attempt at being thorough in a qualitative 12 examination of the things that might affect it.
13 It seems to me that at least some 14 acknowledgment of that approach and that it hasn't 15 been done in the SOARCA analysis would add to the 16 comfort level of people who have been uncomfortable 17 with SOARCA to this point.
18 MR. FULLER: I think -- thank you for that 19 comment. I think it's a good comment and we can 20 certainly add pointers to that work in our executive 21 summary and the introduction and when we talk about 22 the scope of our work being limited to the parameter 23 uncertainty.
24 We can certainly refer to that work as an 25 example of more thorough, at least qualitative NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
88 1 thinking through of all the modeling uncertainties 2 that we're not explicitly addressing.
3 MEMBER BLEY: And that's the first step.
4 Eventually, one would like to be quantitative about 5 it, but in SOARCA, we haven't even been qualitative 6 about it, in a thorough way.
7 MR. FULLER: Okay. Yes, thanks for the 8 comment. And I see Don is at the mic.
9 MR. HELTON: I just want to -- Don Helton, 10 Office of Nuclear Regulatory Research. I do just want 11 to point out, though, that that work is non-public and 12 will remain non-public for, at least the current plan, 13 well after this document would be finalized. So, 14 there is --
15 MEMBER BLEY: The engineering thinking --
16 MR. HELTON: -- a challenge there.
17 MEMBER BLEY: -- beyond it has been around 18 a long time. The application hasn't. And --
19 MR. HELTON: Yes, I mean, it can certainly 20 --
21 MEMBER BLEY: -- maybe I shouldn't have 22 mentioned it, because that is not --
23 MR. HELTON: No, I'm not worried about you 24 mentioning it and it can be described from a 25 conceptual standpoint as what was 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
89 1 MEMBER BLEY: That's what I'm after.
2 MR. HELTON: I'm just reacting to the idea 3 that it's --
4 MEMBER BLEY: To have a pointer to it.
5 MR. HELTON: -- it can be cross-referenced 6 --
7 MEMBER BLEY: Yes.
8 MR. HELTON: -- and that just presents some 9 challenges, that I'm sure we can work through. But --
10 I don't know.
11 MR. WAGNER: Casey Wagner, Dycoda. So, I'm 12 going to go through, not all, but we thoroughly went 13 through the comments from the last meeting and I'm 14 going to address some of the highlights of the more 15 important ones and how we addressed them.
16 So, we're on the first slide here. First 17 one had to do with the fabric seal failure criteria.
18 And we had a range of different types of comments that 19 were relatively extensive discussion on that.
20 And they concerned the error itself, a 21 confusing description of the failure parameter, and 22 then, there was also some questions about how we were 23 handling the seal failure error. We tried to address 24 these by adding new documentation to the next version 25 of the report.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
90 1 And in particular, we did a little bit 2 more work on the impact of the fabric seal on early 3 containment failure. And added some new text in 4 there. In particular, the flow through the ice chest 5 and the other leakage pathways allowed hydrogen to 6 move around, whether that seal had failed or not.
7 So, there were other mechanisms for, over 8 the time frames that we were looking at, that hydrogen 9 could be moving throughout the containment. So, 10 certainly, if the seal had failed, which with the 11 error introduced, it failing more early, it flowed a 12 little bit more thoroughly.
13 But nevertheless, there was other leakage 14 paths that allowed that to occur. And so, there's 15 some discussion about that. In particular, the 16 focused value study had many cases and in all those 17 calculations, the fabric seal had been corrected.
18 And so, we looked at the response of those 19 and tried to compare that to the UA calculations and 20 drew conclusions that the responses were similar on 21 key metrics.
22 And then, finally, Hossein had presented 23 last meeting a comparison to the 2016 UA results, 24 which did have the correct -- well, it had a different 25 distribution on the fabric seal failure, but it was NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
91 1 not prone to failing early. And that discussion is 2 also in there.
3 CHAIR STETKAR: Casey, when I read that, 4 and I -- the discussion of this issue, by the way, is 5 much, much improved. But I know nothing about the 6 physics, but I thought that in the current study, when 7 you looked at the limited cases that you ran for long-8 term, that you saw some -- I think a couple of those 9 kind of transitioned to an early containment failure.
10 So, the question that I had is, how 11 significant is that in the context of the full study?
12 I mean, what fraction -- if you had the corrected 13 model in all 600 realizations in the real study, do 14 you expect a similar number to transition to early 15 containment failure?
16 Because we're not finding many early 17 containment failures, so, like two or three can double 18 the number that we've seen. And I think I understand 19 why that's happening, but I'm trying to get a sense of 20 looking at the full spectrum of the baseline study.
21 You only ran a small number of these long-22 term or late containment failure scenarios, right? In 23 the --
24 MR. WAGNER: Yes.
25 CHAIR STETKAR: -- focus study, there'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
92 1 MR. WAGNER: Yes.
2 CHAIR STETKAR: -- about 18 or 20 or 3 something like that, that you ran, to examine it.
4 MR. WAGNER: Right. If you combine that 5 with the 2016 calculations that had a more robust seal 6 --
7 CHAIR STETKAR: But a different --
8 MR. WAGNER: Yes, it's --
9 CHAIR STETKAR: The reason why I dislike 10 this notion of the comparisons is that the uncertainty 11 difference -- the differences on the uncertainty 12 distributions make a difference.
13 MR. WAGNER: They do, yes.
14 CHAIR STETKAR: So, you have to kind of 15 focus in the regime of the 2016 study, where --
16 MR. WAGNER: This was about a -- in the 17 couple cases I looked at, it was about a two to three 18 PSI difference if the seal was there during a burn 19 that was -- the ones that flipped over --
20 CHAIR STETKAR: Before the --
21 MR. WAGNER: -- that were just below the 22 failure pressure and if the seal was intact, it was 23 just above.
24 CHAIR STETKAR: Okay.
25 MR. WAGNER: And so, you had to be 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
93 1 close. So, we judged --
2 CHAIR STETKAR: Yes, but I was going to 3 say, but there are two. If I ran an infinite number 4 of samples in the baseline study, such that you had 5 the full distribution for everything, I'm trying to 6 pulse what fraction, a guesstimate. Is it one 7 percent, ten percent --
8 MR. WAGNER: Okay.
9 CHAIR STETKAR: -- half of a percent might 10 transition?
11 MR. WAGNER: I certainly can't answer --
12 CHAIR STETKAR: Yes, okay.
13 MR. WAGNER: -- the percentage, but this 14 plot might be kind of useful. So, we're looking at --
15 and unfortunately, the blue ones have the corrected 16 seal behavior --
17 CHAIR STETKAR: Yes.
18 MR. WAGNER: -- versus the purple. But the 19 purple ones that are within a couple PSI could have 20 flipped over. So, I see maybe two there.
21 CHAIR STETKAR: And --
22 MR. WAGNER: I think that --
23 CHAIR STETKAR: But, again, they're -- when 24 you say within a couple PSI, you're still limited by 25 the number of samples from both the failure pressure NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
94 1 distribution.
2 And, if you're not hitting many samples, 3 that could either be fortuitous or -- I mean, you 4 would expect those samples to be somewhere in the 5 middle of the distribution, if there's only a small 6 number of samples, right?
7 MR. FULLER: Yes, so, I guess just --
8 CHAIR STETKAR: The containment failure 9 probability --
10 MR. FULLER: Right.
11 CHAIR STETKAR: -- distribution is what I'm 12 talking about.
13 MR. FULLER: Right. So, if we get back to 14 -- so, the focused study was looking at the range 15 where you could have an early containment failure.
16 So, you saw in 80-some percentage, it still went to 17 the late containment failure versus the early. And 18 what gives us some confidence -- and all of those have 19 the fabric seal failure criteria fixed.
20 So, what gives us some confidence is that, 21 in those realizations, the 360-something, with the 22 fixed seal pressure, we still saw the same, roughly 23 the same percentage of those cases where it's 24 possible, actually going to early containment failure.
25 So, that's what -- that's kind of where we're --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
95 1 CHAIR STETKAR: That might, yes, that might 2 be reassuring. I have to think about it that way, 3 thanks. Thanks.
4 MR. WAGNER: I verbally gave you some 5 information on late containment failure, because in 6 the last set of documentation in June, you didn't have 7 that and so, I documented that.
8 The Committee expressed some confusion on 9 this parameter in general, because there was a thermal 10 part and then, there was sampling. And so, that was 11 rewritten and hopefully, that's much clearer now. And 12 --
13 CHAIR STETKAR: On the record, it is.
14 MR. WAGNER: Great, great. It was a good 15 comment and I'm glad we could improve the 16 documentation. And then, there was also some 17 questions about the seal failure area related to that 18 comment, because we have some where it's sort of a 19 cold mechanical failure and then, some where there's a 20 thermal presence that's maybe more widespread and 21 maybe that should have a different failure area.
22 And we hadn't thought about that and that 23 was -- it was good that it was pointed out. It wasn't 24 a controlling factor on flow, if you look at the K 25 over A-squared and you look at the thermal hydraulic NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
96 1 resistance, even the small holes that we were sampling 2 allowed a pretty robust natural circulation to occur.
3 So, we concurred that that probably was 4 likely, that there's maybe a wider, larger area that 5 we should use, but it wouldn't impact the results, was 6 my engineering judgment.
7 And then, if you look at the seal, it 8 winds around the nooks and some of it's well-shielded 9 from thermal radiation effects. And so, depending on 10 where the fireball was, it -- there certainly is some 11 uncertainty on what parts would be affected.
12 Next slide. The ice condenser doors, and 13 we're going to call this AJAR parameter. The AJAR 14 parameter in context of the next couple of bullets, is 15 whether we fully opened up the doors to form the 16 hinges somehow and they stuck open. We satisfied that 17 criteria.
18 CHAIR STETKAR: Casey, just for the benefit 19 of everybody else, make sure, these are the lower 20 condense inlet doors?
21 MR. WAGNER: That's right.
22 CHAIR STETKAR: Okay.
23 MR. WAGNER: Exactly. And that's the only 24 place we had done the sampling, because we thought 25 those were by far the most important. And just 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
97 1 little background, those are modeled with five MELCOR 2 flow paths, representing, I think it's about 20 doors.
3 They take about 46 pascals to fully open, 4 so it's not a lot of pressure, but surprisingly, that 5 took a bit of a pressure pulse in order to open them, 6 all of them, because they open up such a large flow 7 area, about ten square meters each. And so, you get 8 these opening up and they start relieving a lot of 9 pressure.
10 And the general comment was, we didn't 11 have much discussion about how they behaved. We had 12 the sample parameter, so we added discussion about how 13 the valves or how the doors behaved and FSEM plots and 14 tried to look at a range of different responses.
15 Where we might have had a hydrogen burn 16 early, prior, due to the PRT, one that might have been 17 delayed until hot leg failure. And then, one where 18 there wasn't even enough hydrogen there, we had hot 19 leg failure and looked at the response.
20 And the interesting thing that we also 21 got, Trey was able to dig out a few more diagnostics 22 out of his study, was that if we had a hot leg 23 failure, we almost got 100 percent, all five doors 24 sticking open.
25 So, that pressurization from the hot leg NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
98 1 failure was enough. The pressurization from the PRT 2 wasn't. And that is just a smaller tank, that's 3 blowing down.
4 And even though I can imagine that's 5 pretty violent down there, it was triggering some of 6 the doors sticking all the way open, but not all of 7 them. And in general, I call that a partial AJAR, 8 because it's not all five MELCOR flow paths fully 9 jammed open.
10 And so, we were able to get some really 11 interesting insights, digging into the door behavior a 12 little bit more. It's all documented in the report.
13 We talked a bit about ACRS member comment on or it 14 came up that -- we talk about burn direction, inside 15 the report.
16 And we have known ignition sources and you 17 said, we know the geometry, we know where they're 18 located, does it make sense that we are randomly 19 sampling the direction? That made maybe more sense 20 when we had the random sampling, a couple studies ago.
21 And I'm going to try and make a 22 distinction to you that, although we know the location 23 of those ignition sources, they're coming from the hot 24 leg or the PRT or from the ex-vessel debris, the 25 distinction I want to make is where the ignition NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
99 1 actually takes place may be a little ways away from 2 there and elsewhere inside the compartment, because of 3 the hot plume has to come out and it's got to mix with 4 some oxygen.
5 It's coming out as just pure hydrogen and 6 steam and that's not combustible until it mixes and 7 maybe ends up in different locations. And so, there 8 are -- or it's aerosols and hot things coming out of 9 the ex-vessel debris. And so, we were -- we 10 recognized the point that you were trying to make, but 11 we felt like there is still uncertainty where ignition 12 takes place.
13 And then, we kind of go one step further, 14 recognizing the comment that you had is that, let me 15 just read that last line, identifies the flame 16 directions as it propagates within the compartments as 17 surrogates for uncertainties in the ignition location.
18 So --
19 MEMBER CORRADINI: So, I interpret that to 20 mean you didn't change anything?
21 MR. WAGNER: We didn't change the words 22 that we used. The upward, horizontal, and downward is 23 pervasive through the document. And we tried to give 24 some motivation that there's some uncertainty on where 25 it was located.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
100 1 CHAIR STETKAR: They -- as I read it, you 2 changed nothing in the model --
3 MEMBER CORRADINI: Right.
4 CHAIR STETKAR: -- you added some words in 5 the text to --
6 MEMBER CORRADINI: Justify what you're 7 doing.
8 CHAIR STETKAR: -- impart this notion of 9 perhaps a more randomization of the ignition location, 10 despite the fact that you had three nominal point 11 sources.
12 MR. WAGNER: We --
13 CHAIR STETKAR: Is that -- that's what I 14 kind of hear you saying this morning, that --
15 MR. WAGNER: Yes, the --
16 MEMBER CORRADINI: I --
17 MR. WAGNER: -- that --
18 MEMBER CORRADINI: I'm sorry.
19 MR. WAGNER: Go ahead.
20 MEMBER CORRADINI: No, you finish, I'm 21 sorry.
22 MR. WAGNER: We understood, we could --
23 that was, I think, Mike's comment at the time. We 24 know where these things are, shouldn't we -- but 25 that's where the hot gas leaves from. And then, 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
101 1 isn't ignited yet, it's -- you have to have adequate 2 oxygen, adequate hydrogen, and you have to have that 3 mixing and the source, the high temperature, in order 4 for it to ignite. And so --
5 MEMBER CORRADINI: But -- okay. So, I'm 6 not going to -- I mean, I don't think this is -- well, 7 I don't know if this is a big effect, it may not be a 8 big effect.
9 But in my simple mind, if I know where 10 it's coming out and I know the geometry, and we're 11 talking about a specific reactor, then I'm going to 12 essentially entrain the constituents so that I could 13 estimate when I'm going to get the right concentration 14 and, therefore, decide whether it's up, down, or 15 sideways, versus letting it be just a casting.
16 So, that's kind of the source of my --
17 right? If I'm dumping it out from hot gas to the hot 18 leg, then I know how it's going to, not know, but I 19 can, from an engineering judgment standpoint, decide 20 how it's going to mix.
21 I'm -- I seem to remember, my memory is 22 not so good, I seem to remember, I was just very 23 confused about the directionality issue. And it just 24 seems to add confusion, particularly when you know 25 where it's 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
102 1 So, I'll ask it a different way. Did you 2 do a side calculation to confirm that it didn't make a 3 big effect or was nothing done calculational, you just 4 re-explained it? Re-rationalized it, excuse my 5 English.
6 MR. WAGNER: It was the later.
7 MEMBER CORRADINI: Okay.
8 MR. WAGNER: It's coming out at sonic 9 speeds, it's not combustible, what's leading --
10 MEMBER CORRADINI: Well, I mean, there's an 11 old study -- so, I only remember things that are old.
12 There was a study about 35 years ago from Spalding, 13 which has a very set of simple experiments of a gas 14 jet, whether it's a light gas or a heavy gas, he does 15 everything.
16 And he basically does a very beautiful way 17 of showing how far down the pathway, how the 18 concentrations mixed with something, and you can use 19 Spalding's study as a way to estimate, if I have the 20 hot debris coming up PRT, it's so many L over Ds and, 21 by God, that's where it's going to go and it's going 22 to propagate in that direction and then, we're done.
23 MR. WAGNER: Yes.
24 MEMBER CORRADINI: Versus doing it 25 stochastically. That's kind of what was going through NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
103 1 my mind.
2 MR. WAGNER: Yes. Our dominate source is 3 the hot leg failure, which we have a lot of 4 uncertainty on --
5 MEMBER CORRADINI: Where it's going to 6 happen.
7 MR. WAGNER: -- the working -- and the PRT, 8 I think that probably makes more sense on. But most 9 of our ignitions came from the hot leg. And then, the 10 cavity is a little bit different animal, because it's 11 hot debris --
12 MEMBER CORRADINI: I think --
13 MR. WAGNER: -- it's kind of closed up --
14 MEMBER CORRADINI: -- the reason I focused 15 on this is, I have to admit, I still don't understand 16 the logic of the directionality. Maybe other people 17 do.
18 CHAIR STETKAR: I thought it got it when we 19 had a uniformly distributed set of random ignition 20 sources.
21 MEMBER CORRADINI: Because then they --
22 CHAIR STETKAR: Because they could --
23 MEMBER CORRADINI: -- can go anywhere in 24 the volume.
25 CHAIR STETKAR: -- go anywhere 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
104 1 volume.
2 MEMBER CORRADINI: But when I know where it 3 is in the volume, it seems to me, I've narrowed my 4 possibilities. That's, in my --
5 CHAIR STETKAR: I think the truth is, 6 there's still uncertainty, but it, in my opinion, it's 7 much more constrained. It ought to be much more 8 constrained in the current model. But what you said 9 earlier, does the difference -- would it make a 10 difference? I can't, I certainly can't answer that 11 question, I have no idea. I have absolutely no idea.
12 MEMBER CORRADINI: I mean, I know you have 13 -- sorry, my green light's on. I know you have 14 limited resources and the uncertainty study is fairly 15 complex, but I guess a side calculation, just to 16 verify that directionality doesn't make a big 17 difference once I have the known location of the 18 source, would be an interesting way to kill this guy 19 off. And I see we have a helper.
20 MR. ESMAILI: Let me -- well, I don't know.
21 CHAIR STETKAR: Hossein? Identify yourself 22 first.
23 MR. ESMAILI: Let me --
24 CHAIR STETKAR: Hossein? We don't know 25 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
105 1 MR. ESMAILI: Oh, okay. Hossein Esmaili --
2 CHAIR STETKAR: Thank you.
3 MR. ESMAILI: -- Office of Nuclear 4 Regulatory Research. So, I hope I'm not going to 5 confuse you further. That's why we wrote down here, 6 the location of ignition within the compartment. So, 7 in other words, this is tied to the flammability 8 limit.
9 So, we have this control volume, and you 10 can shake your head if I'm going in the right 11 direction or not, so, we have this control volume, 12 right? Hot gases are coming from one source, right?
13 I don't know whether my flame -- the way we are 14 modeling it, we are modeling it as a flame, right?
15 Propagating either, within this control 16 volume, not going to other control volume, whether 17 it's going to go down, going sideways, or going 18 upwards, right? So, this is one way of capturing 19 where that flame is going to propagate from and 20 whether I have satisfied my flammability criteria.
21 Typically, what we do is that, by default, 22 in medical, for example, we say, if in a control 23 volume, if you have more than ten percent hydrogen and 24 other flammabilities are satisfied, then you have 25 ignition. Or if the ignitor is available, it's seven NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
106 1 percent.
2 Here, we don't know, because we don't know 3 how this gas, as they are coming out of this hot leg 4 or as they are going upwards, they are mixing within 5 that control volume, so I don't know exactly where 6 that frame front is propagating from.
7 So, this is not only the location of the 8 ignition sources, but it's tied to how I am doing the 9 flammability limit within that control volume. This 10 is not propagating from this control volume to that 11 control volume. That one, we know already.
12 MEMBER CORRADINI: Okay. So, within the 13 volume, we're speaking?
14 MR. ESMAILI: Within the volume, we're 15 speaking. So, as Casey said, this gases are coming up 16 right at the location where the hot leg is or where 17 the PRT is, that does not -- that may not satisfy the 18 flammability limit, because he says that it is mostly 19 hydrogen and steam. It has to mix with this volume.
20 And we are working within a control volume 21 approach, right? So, what we really do is that, we 22 are, in the control volume, we are trying to model a 23 flame that propagates in that control volume and for 24 that, we are just changing the ten percent to four 25 percent, six percent, nine percent, because 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
107 1 know where it occurs. Did I --
2 MR. WAGNER: Yes. It originally went back 3 to the Kumar data and if you ignite at ten percent, it 4 can go downward. If you ignite at seven percent, it 5 can go sideways and upward. And if you ignite at four 6 percent, it only -- you're the expert, you can --
7 (Laughter.)
8 MEMBER POWERS: I simply relate an example 9 I have actually used in a class I taught, where having 10 explained how well we know these ignition directions 11 and whatnot. That at 12 percent, you get kind of a 12 uniform expansion of a flame. At nine percent, it has 13 a tendency to go up. At seven percent, it has a 14 profound tendency to go up.
15 And then, I show an example of a seven 16 percent ignition and it does exactly as I said, the 17 flame propagates upwards. Everybody watches and 18 amazed that I'm so accurate in my prognostications.
19 And promptly, the flame comes down. And it's because 20 the gas gets heated up and pressurized.
21 Directionality of flame once it's ignited 22 is an extremely complicated thing to model and doing 23 it in a lump node code, where you've got all kinds of 24 structures actually there, that are vacuous in the 25 code, it seems to me to be braver than I am at these NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
108 1 things, especially when you're talking about 2 relatively lean flames.
3 It -- directionality is easy and 4 reasonably concentrated, where reasonably concentrated 5 is like ten to 11 percent. Once you get down to lean, 6 around seven, it's going to depend on things other 7 than just the concentration of the hydrogen in there, 8 everything else is going to make a difference, too.
9 MR. ESMAILI: This is Hossein Esmaili, 10 again, from Office of Research. One thing we want to 11 make clear is that, when we say within the 12 compartment, we don't mean every compartment that we 13 model.
14 This one, we only imposed this in the 15 lower compartment region, correct, Casey? Because 16 that is where the location of the hot leg is. So, 17 this only means that we do this uncertainty in the 18 lower compartment, this control. So, the other 19 compartments are still allowed to propagate. So, yes, 20 this was another --
21 MR. WAGNER: Yes. That's a good point to 22 be made. After we ignite it, in some ways, you could 23 look at this as surveying a range of ignition 24 strengths and that first place that it ignites. And 25 then, it will propagate according to the default NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
109 1 criteria, which uses Kumar also. I understand your 2 point, Mike.
3 MR. FULLER: Okay. So, we'll move on to 4 the next one. I don't know about you all, but after 5 the last few years of struggling with the safety valve 6 modeling, every time I see those words together, 7 safety valve, I have this sense of dissatisfaction 8 deep inside my core.
9 But we took your comments from the last 10 Subcommittee meeting and we tried to address them as 11 best we could. Oh, you missed my -- dissatisfaction 12 deep in my core, every time I see those words 13 together, safety valves.
14 (Laughter.)
15 MR. FULLER: But we're trying, we keep 16 trying.
17 CHAIR STETKAR: You're very trying.
18 MEMBER POWERS: You're an obsessive-19 compulsive, that's all there is to it.
20 MR. FULLER: You're right, it takes a 21 certain amount of OCD to keep going and soldiering on.
22 We keep trying.
23 (Laughter.)
24 MEMBER POWERS: This is a psychosis that 25 can be treated.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
110 1 CHAIR STETKAR: Tina?
2 MR. FULLER: Yes?
3 CHAIR STETKAR: In all seriousness --
4 MR. FULLER: Yes.
5 CHAIR STETKAR: -- I'm really disappointed.
6 I mean, I'm really, really disappointed. I asked, in 7 June, a simple question. In June, we said, well, we 8 had this data -- I'm sorry, I'm not even going to call 9 it data anymore. We had these numbers from NUREG/CR-10 7037.
11 MR. FULLER: Yes.
12 CHAIR STETKAR: Those numbers are 15 13 failures in a total of 769 demands. That's all it's 14 in. Then, you said, well, okay, NUREG/CR-7037 only 15 goes up through September of 2007, so, gee, we went 16 back to see what more recent experience is.
17 So, we looked at experience from October 18 of 2007 through March 2016 and, lo and behold, we come 19 up now with 16 failures, one more failure, in a total 20 of -- I have to look up my numbers here, so that 21 they're on the record. I can't find them and I'm not 22 going to waste my time. A total of 75 more demands, 23 which you've subdivided into initial demands and 24 subsequent demands.
25 So, my question in June was, tell me how NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
111 1 you estimated the additional 75 demands and where the 2 additional failure came from, because that, at least, 3 was done by people who are still living today and 4 perhaps have a memory, and probably documented what 5 they did, so that we could understand, at least from 6 that snapshot, the thought process.
7 That does not at all come through in the 8 current report. In the current report, all it says 9 is, we needed to use engineering judgment and, yes, 10 there's still a lot of uncertainty on what the failure 11 rate should be. That's all it says. That's all it 12 says.
13 So, the report does not answer any of my 14 questions. So, I'm hoping that some physical body 15 sitting in this room can answer my questions.
16 MR. FULLER: Yes.
17 CHAIR STETKAR: Because if you can't, I'm 18 willing to say that the entire study is flawed. This 19 is a big issue, you make assertions in the report 20 that, oh, yes, well, things could slide to the left, 21 things could slide to the right, but you see the same 22 thermohydraulic behavior whether a valve fails to 23 open.
24 That's great if you're doing a conceptual 25 study. It's not good if you're drawing insights 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
112 1 conclusions about the behavior of the Sequoyah plant.
2 MR. FULLER: Okay. So, what we did add to 3 the report was a summary paragraph. I realize, again, 4 it's not satisfying, because a lot of engineering 5 judgment does have to go in from the people who are 6 reading these reports and capturing both the demands 7 and the failures in the database.
8 So, we tried our best. We did talk to the 9 people who actually do this, the coding of the 10 operating experience, to figure out what they -- how 11 they decide. So, everything is reliant on licensee 12 reporting.
13 CHAIR STETKAR: I --
14 MR. FULLER: So, they have a requirement to 15 report failures of the codes they see. They certainly 16 don't have to report all the demands.
17 But typically, the operating events, these 18 are post-scram events, you can try to guess how many 19 demands were placed on the safety valves, because you 20 do have at least information that the scram event 21 happened and then, you can read the description of 22 what happened.
23 Sometimes, there's information about 24 system pressures, sometimes there isn't. The level of 25 information just varies a lot from LER report to LER NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
113 1 report.
2 So, in terms of the numerator, at least we 3 have more confidence that, because of the reporting 4 requirement, that we know how many failure events 5 happened for the safety valves. For the denominator, 6 the number of demands, we don't have a definitive 7 answer.
8 And I think what we said in the report is 9 that we may be undercounting the demands, because we 10 have to use some judgment on --
11 CHAIR STETKAR: But --
12 MR. FULLER: -- when those valves were 13 demanded and how many.
14 CHAIR STETKAR: Let me read verbatim what's 15 in the report, so it's on the record. Many LERs 16 describing the plant response to scram report the 17 operation of SVs, but a significant portion of the 18 LERs just use phrases like, all systems operated as 19 expected. Hence, for the number of demands, expert 20 judgment must be used to interpret information 21 provided in the LERs and may be undercounted.
22 My assertion is that you may have 23 substantially over counted the number of demands, 24 because it's very, very surprising to me that between 25 October of 2007 and March of 2016, we had a 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
114 1 scrams that resulted in 75 demands to open main steam 2 safety valves and/or pressurizer safety valves, but I 3 would be really surprised if we opened any of them.
4 So, my assertion is that you may have 5 substantially undercounted the denominator, making the 6 whole failure rate substantially optimistic. And if 7 the same process was used in NUREG/CR-7037, the 8 numbers from that report may be substantially 9 optimistic, the failure rate.
10 So, what I was trying to -- what I was 11 hoping to achieve, and apparently will fail, is to get 12 someone who actually did the work from 2007 to 2016 to 13 tell us what criteria were used when you examine the 14 operating experience to say, this event I judge today 15 to result in a demand to open the main steam safety 16 valves at Plant X and I assumed that n number of 17 safety valves were demanded to open.
18 I'm not hearing that answer. And without 19 that answer, I have absolutely no confidence in the 20 failure rates that are used. And I firmly believe 21 that they're optimistic. I can be convinced 22 otherwise, but I need to be convinced otherwise by 23 somebody who tells me what they did.
24 MR. FULLER: Yes, okay. So, I apologize.
25 As I said, I summarized the input we got from the team NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
115 1 in that one paragraph we inserted, but maybe we can 2 arrange a side conversation with the folks --
3 CHAIR STETKAR: No, no, we don't do things 4 --
5 MR. FULLER: -- who actually --
6 CHAIR STETKAR: -- in side conversations, 7 this is a -- the Subcommittee has raised this --
8 MEMBER BLEY: At least, not technical 9 information.
10 MR. FULLER: Right, okay.
11 CHAIR STETKAR: Not technical information.
12 MR. FULLER: So, maybe what we can do is 13 try to set aside some time at the full Committee 14 meeting to --
15 CHAIR STETKAR: Well, we have to decide on 16 --
17 MR. FULLER: Okay.
18 CHAIR STETKAR: -- the problem with the 19 full Committee meeting is that we have pretty good 20 representation of Committee Members here and have had 21 --
22 MR. FULLER: Yes.
23 CHAIR STETKAR: -- in previous meetings.
24 But I don't think that we've had the full complement 25 of the full Committee. And the Members who haven'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
116 1 attended the Subcommittee meetings need to get briefed 2 on the whole scope of this study.
3 MR. FULLER: Yes.
4 CHAIR STETKAR: And even those of us who 5 have attended the Subcommittee meetings have to have a 6 better feel for how you're presenting the full study 7 to the full Committee.
8 MEMBER BLEY: Other than an informal 9 communication, which we can't do, if you wrote 10 something up and sent it to us, that we could look at 11 ahead of time.
12 MR. FULLER: Oh, okay, I see.
13 CHAIR STETKAR: Yes, if you want to write a 14 little report --
15 MR. FULLER: I see --
16 CHAIR STETKAR: -- we can do that.
17 MR. FULLER: -- and then put it --
18 CHAIR STETKAR: But you --
19 MEMBER BLEY: A memo.
20 CHAIR STETKAR: A memo.
21 MR. FULLER: -- put it in the public 22 record, yes. Okay. Actually, that's a good --
23 CHAIR STETKAR: And that might be a way, if 24 we can get it -- the full Committee meeting is in less 25 than two weeks.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
117 1 MR. FULLER: Yes. Okay, yes. So, yes.
2 MEMBER SUNSERI: While you --
3 CHAIR STETKAR: But we can't do it 4 informally.
5 MEMBER SUNSERI: While you --
6 MR. FULLER: Okay.
7 MEMBER SUNSERI: While you're thinking 8 about that question, the question I had in association 9 with this is, does the failure analysis include 10 testing? I mean, you test these safety valves every 11 refueling outage --
12 MR. FULLER: Yes. So, we had -- yes. We 13 had previously talked about that quite a bit. We 14 discovered through talking to the folks who actually 15 conduct the testing that the testing setup doesn't 16 quite test the reclosing probability, which is what we 17 care about.
18 The testing requirements are focused on 19 relieving pressure under design-basis overpressure 20 accidents. So, they test very well the ability of 21 that valve to open when you demand it to open, but 22 they don't really test it under the conditions that 23 we're looking for, repeatedly passing hot steam and so 24 on. So, there is a --
25 MEMBER SUNSERI: Okay, so you don't NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
118 1 consider a valve that may be in an as-found test 2 condition low on its setting to be a valve that would 3 be open during an accident?
4 MR. FULLER: So, yes. So, the setpoint 5 drift is a separate issue and you're right that we 6 don't worry so much about the setpoint drifts that are 7 typically found during testing.
8 It's true that a valve may open at a 9 slightly lower pressure than it's supposed to, but in 10 the grand scheme of things, we had looked a bit about 11 the effect of that in the -- through sensitivity 12 studies in the Peach Bottom study, which of course is 13 a different plant and we were looking for different 14 things.
15 But the setpoint drift issue is not as 16 important as all the other things that are going on in 17 the integrated analysis.
18 MEMBER SUNSERI: Okay, thank you.
19 MR. FULLER: And so, yes. So, we are not 20 using the testing data, because if you look at the 21 distributions are very different from the operating 22 experience data. And we did kind of talk to the 23 testers, we figured out that it's not applicable.
24 That's why there's so much focus on just using the 25 operating experience data, which is from actual scram NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
119 1 events.
2 CHAIR STETKAR: Tina, if you -- the other 3 thing that we brought up back in June is that, and I 4 don't want to get into details of the uncertainty 5 distribution, but you -- we've revised the uncertainty 6 distribution for the open fraction, I forgot the 7 parameter value, but given the valve sticks open, how 8 big is the area that it sticks open, that was revised 9 from 2016 to current time.
10 That distribution says that there is a 30 11 percent probability, for example, that a valve remains 12 open more than 90 percent. And that -- the shape of 13 the distribution seems intuitive, the distribution and 14 the discussion of the distribution in the report is 15 quite good.
16 It says that most likely it's just going 17 to weep, it's going to stick open just a little bit.
18 There's a non-zero, in this case 30 percent, chance 19 that it could stick open big. And then, there's kind 20 of a small probability that it could stick open 21 somewhere in-between, and that equal probability was 22 distributed over those open area fractions.
23 That all sounds really good, if indeed we 24 have information about the numerator, because people 25 report these things in LERs --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
120 1 MR. FULLER: Right.
2 CHAIR STETKAR: -- and we have something on 3 the order of 15 or 16 failure events, 30 percent of 4 those would be something on the order of four, five, 5 six events where we would expect to see a valve stuck 6 open pretty big.
7 So, we would also ask the question, is the 8 operating experience, even in the numerator, the 9 things that you have more confidence in, because you 10 can go read those things, does that operating 11 experience support the uncertainty distribution for 12 the open fraction?
13 MR. FULLER: Yes.
14 CHAIR STETKAR: And you said, yes, we're 15 going to go back and look at that.
16 MR. FULLER: Yes. So, that's what my 17 second bullet is about. We did go back and reread all 18 of the LERs for those events and we actually had a 19 valve subject matter expert also review both the LERs, 20 as well as we had additional information from a 21 proprietary database, which we can't get into the 22 details of that in the study, but we explained in the 23 report that we have access to this and we looked at 24 it, but we can't explicitly refer to the information 25 from 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
121 1 Unfortunately, in relooking at all of 2 that, there was nothing new that we could tease out to 3 kind of add additional details in the report that 4 would be worthwhile. We continued to rationalize, I 5 guess, the distributions that we had come up with.
6 And for the open area fraction, it's even more 7 frustrating because some of the reports had no 8 information on it.
9 But, however, I will say for that large 10 open area fraction, there were a handful of events, I 11 can't remember if it was four or five, I want to say 12 on the order of four, that were due to latent 13 conditions, assembling errors, maintenance errors 14 where the valve was left in the position that, when it 15 was open, the internals got completely misaligned, 16 something fell out of place, and it was basically 17 stuck open with a pretty large area.
18 So, the mechanisms for the failing with 19 the large open area versus the weeping area are kind 20 of different. And what we saw in some of these events 21 was, when there was a maintenance or assembly error, 22 sometimes it got stuck open with a very large area and 23 couldn't be recovered, even by lowering the pressure.
24 So, we had data, again, hard to quantify 25 very specifically what percentage should fall in these NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
122 1 different bins, but we did have data to support these 2 two ends of the distribution.
3 MEMBER BLEY: You give, in a couple places, 4 I think, a description of qualitatively the kind of 5 people you talked to and how you tried to go at this 6 and the data you looked at. I didn't see anything 7 real clear on how you picked the 90 percent and the 30 8 percent.
9 I have a question about the design of the 10 valve. I think the design of the valve stem is such 11 that if you're 30 percent open on this kind of valve, 12 but I'm not sure of this, that you get something like 13 30 percent mass flow.
14 I don't know if that's true. How did you 15 translate valve open fraction into mass flow rate for 16 your calculations?
17 CHAIR STETKAR: That's a MELCOR question, 18 right?
19 MEMBER BLEY: Well, I --
20 MR. FULLER: It's a --
21 MEMBER BLEY: -- don't know if it's an 22 input to MELCOR or if it's something done inside the 23 code. I don't know if that aligns with how these 24 particular valves are designed for flow rate versus 25 percent open.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
123 1 MR. FULLER: Yes.
2 CHAIR STETKAR: And in particular, take 3 that very, very unlikely thing in the center, where 4 it's stuck open 37 percent.
5 MEMBER BLEY: That could be a whole lot --
6 CHAIR STETKAR: That's --
7 MEMBER BLEY: That could be almost full 8 flow or it could be fairly low flow, depending on 9 exactly how that valve --
10 MEMBER MARCH-LEUBA: If it's high pressure 11 --
12 MEMBER BLEY: -- seeping surface is 13 designed.
14 MEMBER MARCH-LEUBA: If it's high pressure, 15 it's likely to choke flow and be linear with area.
16 And at this pressures, I would say it's choked flow, 17 right?
18 MR. WAGNER: That's a true statement.
19 Sometimes, you get what you're getting to, you get 20 stem position versus flow area. And so, it will 21 change and it can be nonlinear.
22 MEMBER BLEY: Well, no, actually, this 23 isn't one, but if you have a gate valve and it opens, 24 it only has to get a little bit open before you get 25 nearly full flow through 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
124 1 If you have a carefully designed throttle 2 valve, then the flow rate, ignoring the point you just 3 brought up, can be proportional or some other 4 relationship to the position of the valve stem. So, 5 there's a lot of detail here and I don't know if the 6 detail has been worked into the model, that's kind of 7 what I'm asking.
8 MR. WAGNER: No, we have a simple linear 9 representation of that.
10 MEMBER BLEY: Okay. Which would be sort of 11 appropriate, if it's a good throttle valve, if that's 12 the design.
13 MR. WAGNER: Yes, it --
14 MEMBER BLEY: I don't know that relief 15 valves are designed to do that. They're designed to 16 pop open and then, pop shut.
17 MR. WAGNER: Yes. Well, we're --
18 MEMBER BLEY: So, 30 percent open might be 19 damned near full flow, I don't know.
20 MR. WAGNER: We don't have many samples in 21 the middle there, we're kind of bimodal distribution 22 of weeping or full --
23 MEMBER BLEY: Well, you got a --
24 CHAIR STETKAR: But, again, to answer 25 Dennis's question, that's why --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
125 1 MEMBER BLEY: Something near ten percent 2 chance of 30 percent open.
3 CHAIR STETKAR: Yes. Well, but --
4 MEMBER BALLINGER: So, you're satisfied 5 that 30 percent is the right number? I mean, that's 6 an awful lot of really bad screw-ups in calibrating 7 those valves, because those valves have to be --
8 they're tested quite frequently. It just seems like 9 30 percent of them --
10 CHAIR STETKAR: No, this is 30 -- given the 11 fact that it's stuck open --
12 MEMBER BALLINGER: Oh, okay.
13 CHAIR STETKAR: -- what is the fraction of 14 the area that it's stuck open at?
15 MEMBER BALLINGER: Okay.
16 MR. FULLER: So, I think --
17 CHAIR STETKAR: It -- what I'm -- my 18 initial question was, how likely is it that it sticks 19 open? And that number, right now, is on the first 20 demand, 2.65 times ten to the minus two, I think, 21 based on 16 failures in -- I've forgotten the -- I 22 don't have the table in front of me here, but --
23 MEMBER BALLINGER: Okay.
24 CHAIR STETKAR: -- 16 failures in some 25 couple hundred demands, 600 demands or something like NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
126 1 that.
2 MR. FULLER: So, I think the one -- the 3 reason we gave that ten to 30 percent a little bit 4 more weight than the 30 to 90 percent is, again, we're 5 trying to squeeze as much as we can out of a very 6 small data set.
7 And in one of the events, we were able to 8 calculate, they had enough information that we could 9 calculate that there was 20 percent flow area that had 10 opened up. So, that kind of said, well, so it does 11 happen, maybe -- so, that was our -- we had that one 12 data point that --
13 MEMBER BLEY: Just an overview.
14 MR. FULLER: Yes.
15 MEMBER BLEY: What you've done is a lot 16 more physically pleasing than 2016 was, which wasn't 17 at all physically pleasing. So, in general, the idea 18 makes sense. I don't know how far to push this. I 19 didn't find much to give me, in what's written, to 20 give me confidence that where you ended up made a lot 21 of sense either. But it's at least in the right 22 direction, I would think.
23 CHAIR STETKAR: I think that, I mean, we've 24 talked about this a few times, you're clear on our 25 sort of concerns. My personal thought is that 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
127 1 issue, for the purposes of this particular study, 2 ought to be met head-on.
3 Acknowledge the fact that perhaps you do 4 not have good justification for that failure rate or 5 perhaps even the specific probabilities that's shown 6 for the uncertainty distribution, although the shape 7 of the uncertainty distribution is -- makes a lot of 8 engineering sense.
9 And that indeed will have implications on 10 the pedigree of, again, I'll call them numbers, I will 11 not call them data, I refuse to do that if nobody can 12 tell me where it came from, on the numbers in 13 NUREG/CR-7037.
14 And I personally would have absolutely no 15 problem calling into question the fact that they are 16 suspect and that they have not been corroborated as 17 part of this effort. That doesn't say that the 18 overall study is fatally flawed, it just says, use my 19 data, get my results, and you used this data and you 20 got your results.
21 Going forward -- and we raised this 22 question back, I looked back through my notes, in the 23 Surry study about the valve failure rate. And at that 24 point, we said, well, it wasn't all that very much 25 important, so get over it, it's from 7037.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
128 1 Anything you do from here on out in the 2 future, ought to, in my personal opinion, recreate a 3 defensible failure rate for those valves. Even if you 4 have to go back and start with a piece of paper that 5 looks like a blank piece of paper and create your new 6 failure rate, because it has to be justified somehow.
7 MR. FULLER: So, I think we will beef up 8 the discussion in the report. What I did try to say 9 in the executive summary and the conclusions was that, 10 we recognize that what you assume for the safety valve 11 modeling is very critical to the outcomes and that 12 there's still considerable uncertainty in our state of 13 knowledge.
14 And I think that's an accurate statement.
15 I mean, even if you --
16 CHAIR STETKAR: It is, Tina, but it -- if 17 I'm a reader of the executive summary, I get these 18 words uncertainty thrown in and I get confused, 19 because what you're talking about is, we have no 20 confidence, I'll be blunt, we have no confidence in 21 the numbers that come out of that NUREG.
22 We have uncertainty distributions about 23 those numbers that we use and that we talk a lot 24 about, because we do this parametric uncertainty 25 analysis and we went back and revisited the open area NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
129 1 fraction and we have an uncertainty distribution.
2 It gets muddied to a reader of the 3 executive summary, because all you say is, yes, 4 there's uncertainty about what the failure rate should 5 be, in the executive summary. And I say, yes, you 6 have an uncertainty distribution about the failure 7 rate as a parameter in your model.
8 So, all you're doing, to me, as kind of a 9 dispassionate reader of the executive summary, all 10 you're doing is reinforcing the fact that you looked 11 at uncertainty and that parameter. That's the way I 12 came across, reading the executive summary --
13 MR. FULLER: Okay.
14 CHAIR STETKAR: -- rather than saying, 15 look, we used the failure rate from this source, which 16 is the best failure rate that we had at the time. We 17 updated it with additional operating experience, but 18 we still have questions about what the number is.
19 And that a takeaway from the study is 20 somebody really needs to go back and reexamine the 21 technical bases and operational history for not only 22 the failure rate, but given a failure, what the 23 relative fraction of the stuck-open area is.
24 MR. FULLER: Yes, okay. Yes. I --
25 CHAIR STETKAR: That's kind of what -- 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
130 1 the sense of a lessons learned. It isn't -- I'm 2 coming across as kind of condemning the study, and I 3 would condemn the study if it says, these are the data 4 that we are going to hang our hat on, as far as 5 drawing insights and conclusions about risks.
6 MR. FULLER: Okay. So, I think we are 7 taking your comments again, we'll try to do a better 8 job of framing the discussion. The only other thing I 9 can offer is that, we've had pretty wide review of 10 this at this point, because of all the questions.
11 We've presented the work at an ASME Pump 12 and Valve Symposium, we've talked to more valve 13 subject matter experts. Nobody has looked at the 14 approach and said, you're doing this totally wrong, 15 there's something a lot better out there. So, we 16 haven't -- I understand your point.
17 Maybe we have to reframe how we present 18 the work a little bit better, given that there is less 19 confidence in maybe one of the key parameters here.
20 So, again, as I said, every time I see the words 21 safety valve on the paper, I have a deep sense of 22 dissatisfaction.
23 CHAIR STETKAR: No, it's -- but honestly --
24 MR. FULLER: It's just kind of where we 25 are.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
131 1 CHAIR STETKAR: Having collected a ton of 2 data in a previous life, it is absolutely true that 3 there, in many cases, must -- in all cases, 4 engineering judgment is involved, because in some 5 cases, you can't even divine whether something was a 6 failure or not.
7 So, there's even engineering judgment in 8 how you throw the numerator into a box for the 9 numerator. And very often, very often, there's more 10 engineering judgment in calculating the denominator.
11 That being said, a responsible data 12 analyst will document the judgment that was used to 13 calculate the denominator. Today, I used the 14 following criteria, based on my knowledge, for these 15 types of events, will result in a safety valve demand.
16 So, that's the type of event.
17 And then, if the event occurs at Plant X, 18 this is the number of safety valves at Plant X that I 19 included. Now, why is the second part important?
20 Well, if you know something about the plants, for 21 example, Plant X might have five safety valves, Plant 22 Y might have three, typically not all of the safety 23 valves will be demanded.
24 So, out of a population of five, you might 25 get one or two. So, if I, today, assumed that all NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
132 1 five were demanded at Plant X every time I achieved 2 these criteria, I'm probably being very, very 3 optimistic in the denominator.
4 But it's incumbent on me as a data analyst 5 to document that engineering basis for that 6 denominator. And if I don't do that, nobody should 7 have any confidence in the numbers that I put forth, I 8 have failed as a data analyst. I'm not a valve 9 expert, I'm not somebody who attends a conference, I'm 10 somebody who analyzes experience.
11 And I may have uncertainty about that 12 denominator, there might be plus or minus a factor of 13 two or three on that, which is fine, I can handle 14 that. Okay?
15 MR. FULLER: Okay.
16 CHAIR STETKAR: We've beat that one up 17 enough --
18 MR. FULLER: Yes.
19 CHAIR STETKAR: -- we do have to be 20 cognizant of the time here.
21 MR. FULLER: The final bullet on that slide 22 was just to note that we did add the discussion for 23 why we didn't include the failure-to-open failure mode 24 and hopefully that was convincing. We showed that 25 there is a couple orders of magnitude less NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
133 1 probability.
2 CHAIR STETKAR: I'll only make a statement 3 that I have about a page and a half written up on 4 that. I agree with you that the failure-to-open curve 5 that you show in the report will lie below the 6 failure-to-close curve.
7 I believe that the margin is considerably 8 less than is shown there, because even the day --
9 sorry, I almost used that word again, let the record 10 show I said day, not data -- even the numbers in 11 NUREG/CR-7037 for the failure-to-open failure mode 12 give you a higher curve than what you show.
13 And it's not clear at all to me how you 14 accounted for common-cause failures. Because I -- the 15 curve, I backed out a, if you want to call it an alpha 16 factor, or a beta gamma factor combined, for all three 17 valves of something on the order of couple times ten 18 to the minus five.
19 And if I look at NUREG, the latest version 20 of whatever the NUREG for common-cause failure 21 parameters for safety valves, this is a -- for that 22 combination of three spring-loaded safety valves 23 failing to open, I get a couple of orders of magnitude 24 higher than that.
25 I get kind of two or three or four, I've NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
134 1 got the numbers here, but it's not worth whining about 2 them right now, higher than that, so that if I apply 3 the individual failure rate from NUREG/CR-7037 and the 4 composite common-cause failure fraction from this 5 other NUREG, I get a couple of orders of magnitude on 6 the failure-to-open on the first demand, if you will.
7 MR. FULLER: Okay. No, that's a fair 8 point.
9 CHAIR STETKAR: And that's -- again, the 10 mark, it's still below --
11 MR. FULLER: Yes.
12 CHAIR STETKAR: -- and I think, 13 qualitatively, you can still draw that conclusion.
14 MR. FULLER: Yes.
15 CHAIR STETKAR: But the curves that are 16 plotted, they're --
17 MR. FULLER: No, that's --
18 CHAIR STETKAR: -- a little snarky.
19 MR. FULLER: That's a fair point, I think 20 we have to discuss the common-cause aspect. So, we'll 21 do that.
22 CHAIR STETKAR: The only reason I hang up 23 on this is, I have absolute -- what happens, Casey 24 would know this, what happens if none of them open at 25 time -- on the first demand? Things get a lot more NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
135 1 interesting, don't they?
2 MR. WAGNER: Yes.
3 CHAIR STETKAR: Okay.
4 MR. FULLER: Yes, so we haven't modeled 5 that lower probability scenario. We did --
6 CHAIR STETKAR: Right.
7 MR. FULLER: -- do a sensitivity in Surry, 8 but --
9 CHAIR STETKAR: Yes.
10 MR. FULLER: -- we haven't done that 11 sensitivity here.
12 CHAIR STETKAR: Yes.
13 MR. FULLER: And you're right, it's a 14 completely different --
15 CHAIR STETKAR: It's things -- so, for the 16 purpose of this study, I think it's clear that you 17 haven't modeled it. I think the rationale for why 18 you've not modeled it is -- can be punched up a little 19 bit.
20 MR. FULLER: Okay.
21 CHAIR STETKAR: Especially because somebody 22 like me is going to go pull those numbers out and do 23 that calculation. Okay.
24 MR. WAGNER: Do you have any guidance on 25 time? We have about two slides left.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
136 1 CHAIR STETKAR: We're going to try to get 2 through as quickly as we can here. I'm going to make 3 an executive decision just because --
4 SPEAKER: Control yourself.
5 CHAIR STETKAR: I cannot do that. I've 6 tried in the past.
7 If we have to run a little bit long, we'll 8 run a little bit long, but not much. So, let's try to 9 get through these and I'll try to be good.
10 MR. WAGNER: The next area that we 11 addressed is time in cycle sampling. And we had a 12 lively discussion on that one, too.
13 We have a two-part approach to addressing 14 the concerns that were brought up. The first part is 15 we improved the discussion of the distribution between 16 BOC, MOC and EOC in the current documentation. No 17 changes were made, but we improved the documentation, 18 the justification why we went into the three areas.
19 The second approach is we took that to 20 heart and for the Surry UA we're going to do a much 21 more continuous approach of using many more ORIGEN 22 results across the full cycle.
23 CHAIR STETKAR: I saw that. I still, you 24 know, it is what it is. The discussion is much 25 improved. I understand the rationale here.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
137 1 I still don't know why you picked 200 days 2 and then said there must be 50 percent of the 3 probabilities, you know, equally distributed on either 4 side of that 200.
5 Why you didn't pick 200 and whatever the 6 midpoint is, why you didn't pick 275 days, say, you 7 know, 250 percent, but it is now explained.
8 Somebody can read through there and say, 9 okay, I might take issue about your selection, but I 10 understand the rationale that you used.
11 MR. WAGNER: To briefly go over the 12 rationale, more ORIGEN work was done to develop the 13 rationale. We looked at -- we had three points when 14 we presented last time.
15 We did about 10 or 15 more ORIGEN 16 calculations to kind of put those points in context 17 and give us some insights on the buildup of the short-18 lived radionuclides.
19 And so, the -- try to get justification 20 for the BOC space is for that buildup to the secular 21 equilibrium of iodine-131, that was our surrogate for 22 short-lived, and how the decay heat changes over that 23 60 days.
24 And then the EOC, the justification on 25 that you can -- we had more data points on the Cesium-NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
138 1 137. And so, that was a definition of why we put an 2 EOC boundary there.
3 And then the sampling looked -- the 4 particular value that was used, looked at load decay 5 heat. So --
6 CHAIR STETKAR: No, the story hangs 7 together much, much better.
8 MR. WAGNER: Next slide.
9 And a question was on the stability --
10 statistical stability analysis conducted for the 11 figures of merit. That was performed and that was 12 added to Appendix A.
13 We recognized the possible impact of 14 eutectic melt temperature on burnup. We didn't change 15 anything, but we have some --
16 MEMBER POWERS: Since there is no eutectic 17 in the system, I really wonder what that is. There's 18 a monotectic.
19 MR. WAGNER: We use eutectic as our 20 conglomerate melt and sort of dissociation temperature 21 for the field.
22 MEMBER BALLINGER: In that case, eutectic 23 should be put in quotes?
24 MR. WAGNER: Oh, sure. Yeah. Yeah.
25 MEMBER BALLINGER: Okay.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
139 1 MEMBER REMPE: Just to follow up on the 2 earlier discussion very briefly, I'm looking at a 3 paper from IRSN and Karlsruhe, Comparison of Core 4 Degradation Phenomena for Various Experiments.
5 There is a quote in there that clearly 6 it's referring to some PHEBUS data and it has 7 "irradiated fuel is dissolved more quickly than 8 unirradiated fuel." It cites the PHEBUS-FP tests.
9 And in that whole paragraph they're 10 comparing differences between irradiated fuel and 11 temperatures versus unirradiated fuel. And so, 12 again, it might be a good place to start looking for 13 data.
14 MR. WAGNER: Yeah, took some notes on 15 that. Thank you.
16 There was some added discussion on our 17 approach for failed computer runs. It is already 18 noted that we --
19 CHAIR STETKAR: Yes, there was. We've 20 discussed that. Next bullet.
21 MR. WAGNER: I wanted to go on. There was 22 some discussion of ice bed response during seismic 23 event. Improved vessel failure discussion to address 24 the cited issues.
25 MEMBER REMPE: So, again, a nitpick, but NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
140 1 it -- the text there needs -- first of all, what is 2 "gross failure"? It's better than variable failure, 3 but you have the word "gross failure."
4 To me, that implies it unzipped and the 5 whole head fell off. I don't think that's what you 6 assumed. You assumed a limited area, specific area, 7 right? And so, find some nicer words that are more 8 representative of what was done.
9 Also, there's another place where it used 10 to talk about the drain line and John harangued you to 11 get rid of that. And now it has something about that 12 -- about penetration -- instrument failures or other 13 failures. And you have in quotes, "instrument 14 failures."
15 And so, look at the words there and just 16 fix it, because it's -- you're having "instrument" 17 both times and it's just --
18 MR. WAGNER: Oh, I see. Okay.
19 MEMBER REMPE: It's an editing thing. But 20 I look for those things, because we have had some 21 comments about it.
22 MR. WAGNER: Yeah. Yeah. The Committee 23 pointed out that there was some confusing discussion 24 on the long-term station blackout on specific 25 generators versus symmetric and asymmetric responses.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
141 1 And I tried to rewrite that to address that.
2 CHAIR STETKAR: That reads -- there's 3 still obviously a numbering difference, but it's okay 4 now.
5 MR. WAGNER: Great.
6 CHAIR STETKAR: As far as I'm concerned, 7 it states what's done.
8 MR. WAGNER: Okay. Great.
9 MEMBER REMPE: Could I go back to the very 10 beginning since it has the phrase "steam generator,"
11 or the acronym there, about our discussion earlier 12 about the limited scope and not considering 13 consequential steam generator tube rupture.
14 And you did consider countercurrent flows, 15 so I bet your MELCOR runs actually have temperatures 16 of the hot leg and temperatures of the tubes in it.
17 And so, again, my earlier point where I 18 was slapped down because I was saying, can you use 19 some insights based on other evaluations you've done 20 so that up front you -- you're basically looking at 21 different pump seal leakage rates and you're making 22 comments about releases into the containment and 23 things like that.
24 I bet you could even get those insights 25 that are plant-specific based on the runs you have on NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
142 1 -- and infer that the temperatures of this tube versus 2 the temperature of the hot legs, you might be able to 3 come up with a conclusion about what would have been 4 more likely or not, even.
5 And it would be plant-specific; right?
6 MR. WAGNER: Uh-huh. Yeah. I've noticed 7 that when I've implemented the natural circulation 8 modeling, three loop versus four loop. There are some 9 differences.
10 MEMBER REMPE: Okay. Thank you.
11 MR. WAGNER: Oh, we looked at RCP leakage 12 as an ignition source, the gases that might be coming 13 out of the RCP and added a discussion.
14 CHAIR STETKAR: That's an interesting 15 discussion. And that's all I'm going to say. That's 16 an interesting discussion.
17 I have some questions about that, but we 18 don't have enough time for it right now. There were a 19 couple of curiosities that I --
20 MR. WAGNER: Okay.
21 CHAIR STETKAR: But it's an interesting 22 discussion.
23 MR. WAGNER: The very last slide is just 24 to point out that we added two things that we talked 25 to you about. Hossein's discussion from last meeting NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
143 1 got into the report, and then Trey's focused 2 pressurizer safety valve study is now fully documented 3 in the report.
4 On to MACCS.
5 MR. BIXLER: Okay. On the -- there were 6 several comments also in the MACCS, the consequence 7 analysis portion of the work from last time.
8 And I think I picked the more major things 9 here and tried to respond to those and explain what we 10 have done to address them.
11 The first thing here is a question on 12 emergency response and triggering by -- in our case, 13 we had some emergency response triggered by SAE siren.
14 And there was a question, what is that? Does that 15 really happen? Is that what TEMA would do?
16 And so, we went back and looked at that.
17 And I think we offered at the time of the last 18 meeting, that we would take a more careful look at 19 that.
20 And what we found, we found a couple of 21 things. One is that there's a documentation by FEMA 22 in 2004, of an exercise for the Sequoyah Nuclear Power 23 Plant that shows pretty clearly a time line of events.
24 And one of the earlier things on the time 25 line is the SAE declaration. And then following 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
144 1 is a first siren.
2 And it notes there that that would trigger 3 school relocation, that buses would be -- bus drivers 4 would be notified to deploy, et cetera. A little bit 5 more is included on the first SAE -- or the SAE 6 message -- the EAS messaging.
7 And then following that, at a later time 8 is the GE declaration. And then following that is --
9 it clearly specifies a second siren and a second EAS 10 message.
11 So, that definitely corroborates what we 12 were modeling the way we had modeled schools and some 13 of the other cohorts that were also related to that.
14 MEMBER SKILLMAN: Nate.
15 MR. BIXLER: Yeah.
16 MEMBER SKILLMAN: This is Dick Skillman.
17 MR. BIXLER: Yeah.
18 MEMBER SKILLMAN: I asked that question.
19 MR. BIXLER: Yeah.
20 MEMBER SKILLMAN: And I asked that 21 question, because the gentleman from -- when 22 challenged, the gentleman from TVA said, "Yes, we 23 really do move the children at a site versus a 24 general," which is where, at least in my experience, 25 the bulk of the evacuations occur.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
145 1 MR. BIXLER: Uh-huh.
2 MEMBER SKILLMAN: And this has 3 specifically to do with Cohort 2.
4 MR. BIXLER: Yes.
5 MEMBER SKILLMAN: And this is actually 20 6 percent of the population --
7 MR. BIXLER: Yes.
8 MEMBER SKILLMAN: -- of the area.
9 MR. BIXLER: Yeah.
10 MEMBER SKILLMAN: But what's interesting, 11 and I'm going to go back to what either John or Dennis 12 said, the study is a specific study for this site.
13 And the study for this site is tied very closely to 14 TEMA, to how Tennessee Emergency Management and TVA 15 have agreed in their emergency plan.
16 Because what they do is at an alert, they 17 dispatch the buses. So, the buses are staged before 18 the site occurs, if the timing is --
19 MR. BIXLER: Right.
20 MEMBER SKILLMAN: So, from that 21 perspective, one must be aware that this is not, for 22 instance, applicable to McGuire, Catawba, TMI, Oyster 23 Creek --
24 MR. BIXLER: Okay.
25 MEMBER SKILLMAN: -- pick your plant.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
146 1 MR. BIXLER: Yeah. Okay.
2 MEMBER SKILLMAN: Because this is, at 3 least in my experience, unusual that an alert --
4 actually, the control room calls the school.
5 MR. BIXLER: Uh-huh.
6 MEMBER SKILLMAN: That's what the 7 gentleman from TVA said.
8 MR. BIXLER: Okay.
9 MEMBER SKILLMAN: So, I mean, the analysis 10 is what it is if that agreement is as stated. I have 11 no reason to believe it's not --
12 MR. BIXLER: Right.
13 MEMBER SKILLMAN: -- but one must be clear 14 in one's own mind to recognize this is different.
15 MR. BIXLER: Okay. Yeah, and we -- I 16 don't think we tried to draw any conclusions for other 17 plants, just for this one.
18 MEMBER SKILLMAN: But this is 20 percent 19 of the population.
20 MR. BIXLER: That's right.
21 MEMBER SKILLMAN: So, this is not an 22 inconsequential --
23 MR. BIXLER: Yeah.
24 MEMBER SKILLMAN: -- uniqueness here.
25 MR. BIXLER: Yeah.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
147 1 MEMBER SKILLMAN: This is important.
2 MR. BIXLER: Yeah. Okay.
3 MEMBER SKILLMAN: Thank you.
4 MR. BIXLER; And I think I can probably 5 conclude with that. One additional thing I'd like to 6 say, though, is that even for this case where we --
7 and we believe faithfully triggered the beginning of 8 the emergency response by SAE, it's still fairly late.
9 It doesn't happen early on the overall 10 time line, because we assume that there'd be quite a 11 bit of delay for the buses to be able to get to the 12 schools.
13 And so, it didn't end up being a 14 particularly early evacuation of the school kids, it 15 was just triggered earlier. It started the chain of 16 events earlier, but it still took a while to 17 accomplish by our time line.
18 MEMBER SKILLMAN: And one probably should 19 recognize that there could be hours or minutes between 20 the site and the general.
21 MR. BIXLER: Yeah. Yeah.
22 MEMBER SKILLMAN: And so, that's really 23 the important piece here. When you tell the children 24 "Go" at a site, they may be in the next county by the 25 time their parents are told to leave.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
148 1 So, there are some practical implications 2 associated with this assumption.
3 MR. BIXLER: Yeah.
4 MEMBER SKILLMAN: Thank you.
5 MEMBER BLEY: Nate.
6 MR. BIXLER: Yeah.
7 MEMBER BLEY: There was one thing I think 8 we talked about the last time, and I wonder if you 9 folks thought more about it.
10 You do have an assumption that bridges 11 within the ten miles are unusable --
12 MR. BIXLER: Right.
13 MEMBER BLEY: -- and you shoulder people 14 until you find a route out for them, but I think we 15 asked if -- what about the people who don't wait for 16 that and who jump out on the roads and then find 17 themselves cut off?
18 I know that area kind of well and there's 19 a lot of bridges in a lot of place nearby there.
20 MR. BIXLER: Yeah.
21 MEMBER BLEY: Do you think about people 22 who get stranded for a fair amount of time because 23 they can't find their way back out again when they 24 keep running into dead bridges?
25 MR. BIXLER: We did --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
149 1 MEMBER BLEY: Did you leave a fraction of 2 those who get caught outside rather than sitting 3 inside?
4 MR. BIXLER: I think we ended up doubling 5 the amount of evacuation time to account for the fact 6 that people might reach dead ends and then have to 7 retrace their steps and try something else.
8 MEMBER BLEY: Okay.
9 MR. BIXLER: So, we did try to account for 10 that in our evacuation time line, hopefully, in a 11 reasonable way.
12 MEMBER BLEY: Okay. Yeah, I guess that 13 makes sense.
14 MR. BIXLER: Yeah.
15 MEMBER BLEY: Thank you.
16 MR. BIXLER: Okay. I think -- next slide.
17 Okay. There's also a question about how 18 MACCS treats deposition under humid conditions, which 19 I imagine you're likely to get in the Tennessee Valley 20 area there. And that was a good question.
21 And at the time, we responded that that 22 would have an impact -- in principal, it could have an 23 impact, at least, if you have hygroscopic aerosols 24 being released that are not fully saturated with 25 water. They could grow to a greater size, deposit NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
150 1 faster, et cetera.
2 That's something we don't specifically 3 treat in MACCS at this point. There's no accounting 4 for relative humidity in the area when you're doing a 5 release.
6 But what we ended up doing in response to 7 that, was to do a bounding estimate of how much 8 difference that could make on the deposition velocity 9 of the aerosols.
10 And I think that's reasonably well-11 documented now in the new version of the document, but 12 what we ended up finding is that it wouldn't even be a 13 factor of 2 in deposition velocity.
14 And our uncertainty range covered a range 15 of ten plus or minus the square root of three -- or 16 square root of ten, roughly a factor of 3 up and down.
17 So, I think we covered that uncertainty 18 already in our uncertainty range, but it is -- it's 19 something that we thought was worth adding some 20 additional description/discussion in the document.
21 So, we went ahead and did that.
22 Okay. Next slide. And last comment -- or 23 set of comments was on the presentation of some of the 24 risks that were in the -- with the earlier version of 25 the document, both EF, early fatality, and LCF risks.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
151 1 So, I basically agree that at the time 2 that we met last time in June, that we would go back 3 and reconsider that and add to the discussion. And 4 so, we've done that now.
5 There's a specific discussion on the 6 bimodal nature of the LCF risk and what that stems 7 from and we reduced the amount of presentation on EF 8 risk.
9 Basically, there used to be a table and a 10 figure there that we eliminated and now there's just a 11 minimal amount of discussion saying that it only 12 occurs in a few of the realizations that we had.
13 And even there it's trivial overall that 14 there's almost no chance of an early fatality. So, 15 those are the things we did to address the comments 16 from last time.
17 CHAIR STETKAR: I commented earlier on the 18 early fatality. There's still a tabulation in there 19 that shows the mean value and zeroes for all the other 20 percentiles.
21 I think -- again, I think a tutorial will 22 help.
23 MR. BIXLER: That might be in the 24 appendix, are you thinking, or --
25 CHAIR STETKAR: I don't -- no, there's 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
152 1 in the appendix, but there's -- the one in the 2 appendix has mean values and 95th percentiles, as Trey 3 showed --
4 MR. BIXLER: Right.
5 CHAIR STETKAR: -- for four of the five 6 distances that indeed are higher than the mean value, 7 which is reassuring to a lot of people.
8 There is a table in the main body of the 9 report and I, you know, if you want me to --
10 MR. BIXLER: In Chapter 6?
11 CHAIR STETKAR: I don't remember whether 12 it's Chapter 6.
13 MR. BIXLER: Okay.
14 CHAIR STETKAR: It probably is. I think 15 that does show non-zero mean values and zero values 16 for the median, fifth and 95th percentiles.
17 MR. BIXLER: I think the intention was to 18 take that table out.
19 CHAIR STETKAR: Well, it's still in there 20 and it -- you can argue one way or the other about 21 whether it's useful. It might be useful.
22 If it's in there, as I said earlier, I 23 think that it would be very useful to have a bit of a 24 tutorial for folks --
25 MR. BIXLER: Okay.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
153 1 CHAIR STETKAR: -- about how that type of 2 behavior makes sense --
3 MR. BIXLER: Right.
4 CHAIR STETKAR: -- because most people 5 don't get it. And most people -- what I've found is 6 many people do not support the notion of quantitative 7 uncertainty analysis and are too easy to dismiss it as 8 simply an abstract mathematical, statistical, 9 nonphysical exercise that you push a button at the end 10 of a study and something comes out.
11 So, I think that table is actually useful.
12 It shows that the numbers are small --
13 MR. BIXLER: Okay.
14 CHAIR STETKAR: -- but it needs a bit of 15 discussion.
16 Now, regarding the bimodal nature of the 17 distribution, in the executive summary, the notion of 18 a bimodal distribution is introduced. You show the 19 cumulative curves. You still have this notion that 20 risk decreases as a function of distance from the 21 site, which is not entirely true.
22 The discussion in Chapter 6 is much, much 23 better. You essentially - now, remember an executive 24 summary has to be written not for people who did the 25 work. It has to be written for people who want to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
154 1 understand what you did and what you're trying to 2 convey with the results so that when you say a bimodal 3 distribution, most people won't understand what the 4 heck that is even from the shape of the cumulative.
5 They just won't.
6 If you say, well, if we turn up the 7 microscope on this funny-looking thing and show what's 8 shown in Chapter 6, and then say, well, look, for the 9 80 percent -- 80, whatever the heck it is, seven 10 percent of the time when we get an early containment 11 failure; A, it's 87 percent of the time and here's 12 what the risk -- conditional risk profile looks at for 13 that 87 percent. And, look, this one's from the site, 14 doesn't make any difference.
15 But the other 13 percent when you 16 basically don't get a containment failure and have a 17 very small, protracted release, the risk from that is 18 really, really, really small. And in that case, 19 distance from the site makes a big difference.
20 MR. BIXLER: Right.
21 CHAIR STETKAR: And I think telling that 22 story that way in the executive summary, to most 23 people who read it, is a lot more beneficial than just 24 the words that are in there and the figures, because 25 people won't get those figures.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
155 1 MR. BIXLER: Okay.
2 CHAIR STETKAR: And the same way as 3 publishing the results as a range from something on 4 the order of ten to the minus ninth to something on 5 the order of ten to the minus three is misleading 6 without that corroborating story.
7 MR. BIXLER: Okay.
8 CHAIR STETKAR: Especially in the 9 executive summary.
10 MR. BIXLER: Okay. All right. We'll take 11 that comment to heart and go back and reword that a 12 little better.
13 CHAIR STETKAR: Because it's really -- I 14 think to a lot of people it's unexpected. That shape 15 is unexpected and the fact that that for in this 16 particular site, in this particular study, distance 17 from the site for the vast majority of the things that 18 you're looking at doesn't make any difference is also 19 counterintuitive to a lot of people.
20 MR. BIXLER: Okay.
21 MS. SANTIAGO: So, that brings us back to 22 talking perhaps about the topics that we should focus 23 on for the November 2nd full committee presentation.
24 MS. GHOSH: Yeah. We didn't have any 25 slides on this, because basically we're soliciting NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
156 1 your input on what we should prepare and focus on.
2 CHAIR STETKAR: Well, in my opinion, I 3 think recognize that there are some members of the 4 Committee who I believe have not heard any of this 5 story.
6 I've tried to go back and see who attended 7 what meetings, but let's go on the presumption that 8 there are at least one or more members of the 9 Committee who have heard nothing of this.
10 That being said, I think that you need to 11 provide an overview of the objectives of the study.
12 The fact that it was tailored, in a sense, to examine 13 hydrogen issues focusing primarily on early 14 containment failure, I would not dwell too much on the 15 long-term station blackout despite the fact that it's 16 part of the study.
17 I think you don't have enough time to draw 18 that distinction. I would keep focused on the short-19 term station blackout exclusively and said that's the 20 -- that's where we did all of our analysis, that's 21 where the integrated uncertainty analysis was done and 22 we just don't have enough time to kind of talk about 23 the others -- the other part of it.
24 I would show the -- certainly show the 25 results and explain the results, the bimodal effect 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
157 1 the results, you know, where they're coming from so 2 that other members can understand, you know, 3 essentially the sources of risk and the kind of 4 phenomena that are contributing to them.
5 And then certainly, you know, in my 6 opinion, and I'll ask for other help here, I would 7 grapple -- I would grasp the demon by the horns and 8 say, you know, what are some of the lessons that we've 9 learned from doing this study?
10 I think one of the lessons that we learned 11 is that the -- paying careful attention to the 12 uncertainties as you do the study rather than as an 13 afterthought is; A, essential, and; B, can have a 14 measurable effect on the results.
15 You may want to show a couple of examples 16 of that in things like the open area fraction and the 17 fabric seal type of things, but only as an 18 illustration.
19 You don't want to get bogged down in 90 20 minutes in a full committee discussion with people who 21 don't understand how the machine works and don't 22 understand the study in terms of detail.
23 So, I' m not sure, you know, think about 24 that, but I think part of the message is that the 25 uncertainty analysis is not an afterthought and it's NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
158 1 important that it's not an afterthought.
2 I would -- this is -- address the issues 3 that we have discussed, the non-completion of MELCOR, 4 the concerns about the pedigree of the numbers for the 5 valve failure rate and, you know, you've already 6 established the fact that that's an important 7 parameter in the study. And perhaps one or two other 8 issues, you know, that we've discussed.
9 I think that -- here's one area, and think 10 about this, again, it's -- you're going to have to put 11 together a lot of material in a 90-minute 12 presentation.
13 A lot of the really elegant stuff in this 14 study, both in terms of the modeling and how the 15 sampling algorithms are done, is really, really neat 16 stuff.
17 And all of the stuff that you did on 18 regression analyses to kind of ferret out important 19 contributors and how they may vary, are really, really 20 interesting.
21 I don't know how you jam that into 90 22 minutes or even whether you try. And that's where I'm 23 looking for -- oh, it's a lot of really neat stuff, 24 but I'm not sure whether it's worth trying to get at.
25 MEMBER BLEY: Well, what do you want from NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
159 1 the ACRS? I think that's what you ought to focus on.
2 And anything that you really like, but isn't related 3 to what you want from the Committee, you ought to put 4 aside.
5 For example, you've done a lot of work 6 recently on this focus safety valve study. I would 7 put most of that aside.
8 On something that John talked about, I 9 might go a little further, because when -- I like most 10 of the things you talked about, John, but I think you 11 ought to really narrow it down to the things that will 12 -- what did you learn, what's good?
13 Now, the one thing I hoped you learned was 14 when we started going with you through the uncertainty 15 analysis and said, gee, everything is pointing to this 16 open fraction of a safety valve, all the important 17 stuff was sitting there that led you to go back and 18 say, "Did we really do that right," and reevaluate it, 19 I think that's an important story to tell.
20 Great details about what you actually did, 21 I don't think, is that important, but that you learned 22 from the uncertainty analysis that some part of your 23 modeling was a lot more important than you probably 24 figured it would be when you started.
25 I might be wrong about that, but 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
160 1 would be my guess. So, that would be the one thing I 2 would focus on a little.
3 MEMBER SKILLMAN: I'd like to offer a 4 comment, if I could. The major take-away for me is 5 the uniqueness of the applicability of this study to 6 this plant.
7 John said it very well a couple hours ago.
8 One should not paint all ice containments with this 9 result.
10 McGuire and Catawba or Watts Bar, because 11 of their site and because of their emergency plan, 12 might have a very different outcome.
13 And so, in the abstract and in your --
14 what will become your executive summary, I think it's 15 very important to communicate the unique applicability 16 of this specific study for this specific plant.
17 That doesn't take away from the potential 18 use of some of the results, but the final conclusions 19 are unique and the users should be advised don't be so 20 hasty just to apply this anywhere.
21 And if that caveat isn't in the Peach and 22 the Surry analyses, it probably ought to be because 23 these are very highly-specific, complicated analyses.
24 It's a unique piece of -- it's a huge piece of work, 25 excellent piece of work, but it's unique.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
161 1 CHAIR STETKAR: Any other suggestions as 2 far as the full committee presentation from our 3 members?
4 MEMBER MARCH-LEUBA: Yes. For my benefit, 5 I would like to hear something about why we're doing 6 this and -- let me rephrase the question in a more 7 final way.
8 Have you guys worked ourselves out of a 9 job? the moment you write the report, we can fire you 10 because we're never going to use this again. Or what 11 are we going to use -- what benefit does the Agency 12 gain from this exercise? How are we going to apply 13 it?
14 Is it going to apply to licensing? Tell 15 me why we did this.
16 CHAIR STETKAR: Okay. Be careful about 17 that, because you haven't been here for the last 18 decade. The --
19 MEMBER MARCH-LEUBA: More reason to tell 20 me why we did this.
21 CHAIR STETKAR: Okay.
22 MEMBER CORRADINI: But I think Jose's 23 point is fair. Ten years ago, or thereabouts, this 24 was done for a particular reason. But now that you've 25 done it, how do you use it going forward? I think NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
162 1 that's what Jose is asking.
2 MEMBER BROWN: I would -- can I -- since 3 I'm uninitiated and I've been here for nine years and 4 I've sat in on several of these meetings, but not all 5 --
6 MEMBER BLEY: Started the year before you 7 joined us.
8 MEMBER BROWN: I know.
9 MEMBER BLEY: You're no longer 10 uninitiated.
11 MEMBER BROWN: Well, after the first 12 couple of meetings, I decided I would be more 13 selective in the meetings that I attended.
14 (Laughter.)
15 MEMBER BROWN: I phrase that very 16 succinctly. And I agree with Jose in that I think 17 it's important since this is kind of the end product, 18 to say why we did this and what are the conclusions we 19 should gather from this in terms of how we use the 20 results.
21 The ones I've sat in on hydrogen 22 deflagration and stuff like that, how does it apply to 23 the safety posture of the plants? How can it be 24 generally applied, if at all, to plants other than the 25 ice, you know, this specific plant?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
163 1 Because, otherwise, if you go through a 2 bunch of stuff where you show graphs and lots of ten 3 to the minus seventh, eighth and ninth, it's just --
4 it's a sleeper. You're going to fall asleep.
5 You'd really like -- at least I would, you 6 really want to know what conclusions do we draw out of 7 this in terms of is this a big, big problem, safety 8 posture, because it's just going to spread all over 9 the place, or did we get -- were we able to come look 10 at the results and say, look, yeah, this happens, but 11 here's the general configuration or the impact at the 12 site or the general areas, because that's -- it's the 13 safety posture that counts, not the fun and games of 14 playing with titivating models.
15 That's my personal opinion in terms of 16 what ought to come out of the final presentation at 17 the end. That's all I have, John.
18 CHAIR STETKAR: And something to keep in 19 mind kind of along these lines of why did you do it, 20 what are you going to do with it, etc.
21 The executive summary does touch on the 22 fact that some of this, a good fraction of it, is 23 developing methods, models and tools that can be used 24 throughout the Agency in many other applications. So, 25 it isn't just it, the study, and developing expertise NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
164 1 within house on those.
2 And I think that the full committee --
3 MEMBER MARCH-LEUBA: We need to answer the 4 question now.
5 CHAIR STETKAR: No, but, I mean, I wanted 6 to kind of alert them that it isn't just it, this 7 thing.
8 MS. SANTIAGO: Right. We've talked about 9 that.
10 CHAIR STETKAR: Yeah.
11 MS. SANTIAGO: And it's a good point.
12 CHAIR STETKAR: Anything else on stuff for 13 the full committee presentation? You have a 14 challenge, clearly.
15 If not, I will ask if there are any 16 members of the public in the room who would like to 17 make comments. Please come up to the microphone and 18 do so.
19 And not seeing a stampede, I'll ask if 20 there are any members of the public on the bridge line 21 who would like to make a comment. Please do so.
22 The bridge line should be open. Please 23 identify yourself and make a comment.
24 (Pause.)
25 CHAIR STETKAR: That's a good indication NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
165 1 the bridge line is open. So, not hearing anyone, if 2 there's anyone out there, again, please make a 3 comment. Okay. We'll get that reclosed again.
4 As we always do in a subcommittee meeting, 5 I'll go around the table and ask for final comments by 6 each of the members.
7 Ron.
8 MEMBER BALLINGER: Well, I think we've 9 pretty much beat the dead horse in understanding what 10 went wrong with the computer runs and things like 11 that. So, that would be my only comment.
12 CHAIR STETKAR: Matt.
13 MEMBER SUNSERI: I have no other comments.
14 Thank you.
15 CHAIR STETKAR: Dick.
16 MEMBER SKILLMAN: No further comments.
17 Thank you.
18 CHAIR STETKAR: Dana.
19 MEMBER POWERS: I have an organizational 20 conflict of interest here, so I don't make comments.
21 CHAIR STETKAR: I got to remember that.
22 MEMBER POWERS: I don't believe in 23 eutectics and the uranium-zirconium oxygen system 24 based on a huge number of experiments.
25 MEMBER CORRADINI: I just thank the staff.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
166 1 I do think, though, I guess my only emphasis would be 2 I would treat the other four or five members that 3 aren't here as your first shot at the intelligent 4 public.
5 So, cast it in that regard for them, 6 because you'll get "why" questions and what-do-you-7 use-it-for questions from some of the other members.
8 And I think that's kind of how I would at least try to 9 -- at least customize the beginning of your couple of 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> in front of us in full committee.
11 Other than that, I appreciate all that 12 you're trying to do to educate us for those that 13 forget what you've done in the past since it has been 14 a few years.
15 CHAIR STETKAR: Dennis.
16 MEMBER BLEY: Nothing more to add.
17 Thanks.
18 CHAIR STETKAR: Jose.
19 MEMBER MARCH-LEUBA: Nothing.
20 CHAIR STETKAR: Charlie.
21 MEMBER BROWN: Nothing more.
22 CHAIR STETKAR: Joy.
23 MEMBER REMPE: I also want to thank you 24 for your presentations and your efforts. I guess the 25 only point I'd like to emphasize is as you finalize NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
167 1 the executive summary, please consider caveats where 2 needed when you have conclusions and statements.
3 And I think I've tried to emphasize a 4 couple of examples today. Thank you.
5 CHAIR STETKAR: Now, as part of a -- let 6 me also say thanks. Thanks for all of the work that 7 you did between May or June or whenever and now. A 8 tremendous amount of stuff has been done.
9 I think the report benefits from it. I 10 think the understanding benefits from it. Although we 11 probably won't discuss it, I like Appendix I. I think 12 it's a confidence builder. And thanks for all of the 13 work.
14 I was going to say -- oh, the ACRS will 15 write a letter on this. The ACRS letter will be 16 written against the material we have and today. So, 17 there will be no update to the executive summary to 18 effect an ACRS letter or anything.
19 If you want to send us a little side 20 report to bolster confidence in the numbers for the 21 valves, that's fine, but we're not going to entertain 22 -- we can't because of our scheduling. So, we're 23 going to write a report on what we have and today.
24 I would ask, and I don't know how we're 25 going to do this given our other constraints, I really NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
168 1 want some input from the individual members on what 2 might go into a draft of our letter. And let's 3 discuss that offline.
4 With that, the meeting is adjourned.
5 (Whereupon, the above-entitled matter went 6 off the record at 12:07 p.m.)
7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
State-of-the-Art Reactor Consequence Analyses (SOARCA) Project: Sequoyah Integrated Deterministic and Uncertainty Analyses ACRS Subcommittee Briefing October 18, 2017 Patricia A Santiago, Chief Accident Analysis Branch Division of Systems Analysis NRC Office of Nuclear Regulatory Research
Schedule
- ACRS Full Committee on SOARCA Sequoyah Analyses -
November 2, 2017
- Transmit SOARCA Sequoyah NUREG/CR report to the Commission and NRC publications - November 30, 2017
- Updating SOARCA Surry Uncertainty Analysis - ongoing -
June 29, 2018 2
Core Team Members
- MELCOR and severe accident progression: Kyle Ross, Jeff Cardoni, Chris Faucett, Troy Haskin, Randy Gauntt (SNL);
Casey Wagner (dycoda); Hossein Esmaili, Trey Hathaway, Allen Notafrancesco, Salman Haq, Ed Fuller (NRC)
- MelMACCS: Nathan Bixler, Doug Osborn** (SNL); Trey Hathaway (NRC)
- MACCS, consequence analysis and emergency response:
Nathan Bixler, Matthew Dennis, Joe Jones, Doug Osborn**,
Fotini Walton (SNL); Trey Hathaway, Jonathan Barr, Keith Compton, Todd Smith, Edward Roach (NRC);
- Accident scenario development: Selim Sancaktar, Jose Pires (NRC)
- Co-leads 3
Outline
- Focused pressurizer safety valve study
- New Appendix I in draft report
- Updates in response to ACRS members comments from June 6, 2017 subcommittee meeting
- Discussion of topics for presentation to full committee 4
Focused Pressurizer Safety Valve Study Trey Hathaway Accident Analysis Branch NRC Office of Nuclear Regulatory Research
Focused Safety Valve Study
- Updated Sequoyah model run with sampled safety valve cycles less than 65 (13% of distribution) and open area fraction greater than 0.3 (40% of distribution)
- Most parameter distributions identical to UA
- Safety valve distribution constructed from sampled values of of UA and sampled on range of 1 to 65
- Open area fraction distribution sampled with lower bound of 0.3
- Calculation initially limited to 15 hours1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br />
- All early failures were less than 15 hr
- Identical model to UA
- Sampled fabric seal failure Figure I-1 Cumulative pressurizer safety valve open area pressure input corrected fraction versus total number of safety valve cycles, where early containment failures are shown in red**
- NOTE: This data was taken from the draft 2016 Sequoyah SOARCA analysis (ADAMS Accession No ML16096A374) 6
Early Containment Failure
- Approximately 17% of the Table I-1 Statistics on time (hours) of early containment rupture for BOC, realizations in focused MOC, and EOC realizations.
study resulted in early BOC MOC EOC containment failure Mean 6.0 6.6 6.6 Median 6.0 6.9 6.8
- BOC: ~15% 5th-Percentile 5.5 4.3 4.5
- MOC: ~16% 95th-Percentile 6.3 10.8 8.8
- EOC: ~19%
7
In-vessel Hydrogen Generation and Transport to Dome Up to First Deflagration
Hydrogen Deflagration Peak Pressure Figure I-9 Time to containment failure plotted against the containment rupture pressure.
Figure I-13 Difference between peak containment pressure and sampled containment fragility plotted against maximum hydrogen reaching the dome around the time of burn initiation in the dome. 9
Cesium and Iodine Release Fraction Figure I-14 Cesium release fraction versus time. Figure I-15 Iodine release fraction versus time.
Table I-10 Statistics on Cesium and Iodine release fraction at 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> for the early and late containment rupture realizations.
Early Containment Rupture Late Containment Rupture Cs release I release Cs release I release Mean 0.022 0.063 0.004 0.021 Median 0.022 0.06 0.002 0.014 5th Percentile 0.01 0.028 0.0003 0.004 95th Percentile 0.042 0.122 0.015 0.084 10
Latent Cancer Fatality Risk Statistics on Total Risk Table I-11 Statistics for the 0-10 mile individual LCF risk assuming LNT and conditional on the occurrence of a STSBO.
Early Containment Rupture Late Containment Rupture Mean 7.0E-04 8.3E-05 Median 5.8E-04 6.7E-05 5th Percentile 2.1E-04 3.2E-09 95th Percentile 1.5E-03 2.2E-04 Statistics on Emergency and Intermediate/Long-Term Risk Table I-12 Statistics for the emergency phase, and intermediate and long-term phase contributions to the 0-10 mile individual LCF risk assuming LNT and conditional on the occurrence of a STSBO.
Early Containment Rupture Late Containment Rupture Intermediate Intermediate Emergency Emergency and Long-Term and Long-Term Mean 2.4E-04 4.6E-04 7.6E-08 8.3E-05 Median 1.1E-04 4.2E-04 5.3E-08 6.7E-05 5th Percentile 8.5E-06 1.7E-04 2.5E-10 3.0E-09 Figure I-18 0-10 mile Individual LNT LCF 95th Percentile 1.0E-03 1.0E-03 2.4E-07 2.2E-04 risk conditional on the occurrence of a STSBO as a function of containment rupture time.
11
- NOTE: Only 20% of the late containment failure realizations were extended to 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />.
Early Fatality Risk Table I-14. Statistics for mean, individual, early fatality risk conditional on the occurrence of a STSBO and on an early containment failure (within 15 hr of accident initiation). Tabulated statistics are means over weather variability and express epistemic uncertainties in the MELCOR and MACCS input parameters.
0-1.0 mi 0-1.3 mi 0-2.0 mi 0-3.0 mi 0-4.0 mi Mean 2.6E-06 1.7E-06 1.1E-06 3.8E-07 2.2E-07 Median 1.9E-12 - - - -
5th-Percentile - - - - -
95th-Percentile 2.0E-05 1.3E-05 6.8E-06 2.0E-06 -
Table I-15. Statistics for mean, individual, early fatality risk conditional on the occurrence of a STSBO. Tabulated statistics are means over weather variability and express epistemic uncertainties in the MELCOR and MACCS input parameters.
0-1.0 mi 0-1.3 mi 0-2.0 mi 0-3.0 mi 0-4.0 mi Mean 2.3E-08 1.5E-08 9.7E-09 3.4E-09 2.0E-09 Median - - - - -
5th-Percentile - - - - -
95th-Percentile - - - - -
12
Updates in Response to ACRS Members MELCOR Comments Casey Wagner, dycoda, LLC Tina Ghosh, Accident Analysis Branch, NRC Office of Nuclear Regulatory Research
Fabric Seal Failure Criteria
- ACRS member comments concerned identified error, confusing description of the failure parameter, and seal failure area
- New documentation added on the impact on early containment
- Flow through the ice chest and other leakage pathways allow hydrogen transport to the dome prior to the first burn
- Focused valve study includes fabric seal correction
- Comparisons to draft 2016 UA results show continuous behavior
- New documentation added on impact on late containment failure
- Intact seal failed following the first hydrogen burn in the dome 14
Fabric Seal Failure Criteria
- Fabric seal parameter description was re-written to clarify UA parameter sampling versus the deterministic thermal-mechanical failure criteria
- Sampled UA parameter is for the cold failure differential pressure
- Deterministic non-sampled failure criteria for elevated temperature conditions
- ACRS member questioned whether seal failure area is different for mechanical over-pressure versus due to elevated temperature
- Range of failure areas allowed robust natural circulation
- Failure area was not important parameter in regression results 15
Ice Condenser Doors
- Sampled AJAR parameter controls lower ice chest door open area after fully opening
- Five separate flow paths (Only 46 Pa required to fully open the doors)
- AJAR flow paths are large and not the controlling flow resistance when
<5 flow paths are open
- New UA discussion illustrating range of door behaviors
- Analysis from the focused SV study and UA shows all 5 MELCOR flow paths usually satisfy AJAR criteria when there is a hot leg failure
- Partial (<5 flow paths) or no AJAR is the usual response when there is no hot leg failure 16
Hydrogen Ignition and Burn Direction
- Known ignition sources
- Hot gases from the hot leg and the PRT
- Ex-vessel debris
- Location of the ignition within the compartment has uncertainty
- May not be near the hot leg, PRT, or ex-vessel debris
- Dependent on local gas concentrations, mixing, and strength of the ignition source, etc.
- New discussion identifies flame directions as it propagates within the compartment as surrogates for uncertainties in the ignition location 17
Safety Valves
- Added discussion of how the number of demands (denominator) is calculated in NUREG/CR-7037 operating experience data
- Revisited publicly available information from Licensee Event Reports on the main steam safety valve failure-to-close events, and SMEs summary of these events, but did not add anything further to the report
- Added discussion on rationale for not including failure-to-open SV failure mode 18
Time in Cycle Sampling
- Two-part approach
- 1. Improved the discussion of the distribution between BOC, MOC, and EOC in the Sequoyah documentation
- 2. New Surry UA uses more continuous sampling of inventory &
decay heat throughout the fuel cycle
- New Sequoyah discussion includes
- New ORIGEN work to characterize selected Sequoyah values within the continuum of the fuel cycle
- BOC sample space defined with consideration of decay heat build-up and development of secular equilibrium inventory of I-131
- BOC investigates low decay heat response
- EOC sample space is developed with consideration of long-lived Cs-137 inventory
Other Additions
- Statistical stability analysis conducted for figures of merit and documentation added in Appendix A
- Recognized possible impact of eutectic melt temperature on burnup but also inadequate data to quantify
- Added discussion on approach for failed computer calculations
- Added discussion of ice bed response during seismic event
- Improved vessel failure discussion to address cited issues
- LTSBO discussion improved to address limitations on asymmetric SG response
- Added discussion/plots on RCP leakage as an ignition source 20
Other Additions
- Comparison of new UA results with draft 2016 results with different valve failure distributions
- Presented last ACRS meeting
- Focused pressurizer SV study results 21
Updates in Response to ACRS Members MACCS Comments Nathan Bixler, Sandia National Laboratories
Emergency Response Timing
- Question - Is emergency response normally triggered by GE?
- Response - Verified that emergency response modeling is faithful to TEMA planning
- Resolution
- Final Exercise Report Sequoyah Nuclear Power Plant (FEMA, 2004) timeline lists sequentially
- SAE declaration
- GE declaration
- Notes from conversation with TEMA state that upon SAE, schools are evacuated to paired schools outside EPZ
- Discussion added to final document 23
Humidity
- Question - How does MACCS treat deposition under humid conditions?
- Response - authors indicated that humid conditions could cause hygroscopic aerosols to grow and deposit faster, but MACCS does not treat this mechanism
- Resolution
- Performed bounding estimate to show that deposition velocity of hygroscopic aerosols could increase by less than a factor of 2 and this is within the uncertainty range
- Expanded discussion in section on uncertainty of deposition velocity 24
Presentation of Risks
- Comment - Improve presentation of EF and LCF risks
- Response - Added discussion of bimodal nature of LCF risk, minimal dependence of risk on distance, and reconsider discussion of EF risk
- Resolution
- Discussion of LCF risk was expanded to address comments
- Figure and table showing EF risk was eliminated 25
DISCUSSION OF TOPICS FOR FULL COMMITTEE PRESENTATION 26
References
- SECY-12-0092, State-of-the-Art Reactor Consequence Analyses -
Recommendation for Limited Additional Analysis (July 2012)
- NUREG-1935, State-of-the-Art Reactor Consequence Analyses (SOARCA)
Report (November 2012)
- NUREG/CR-7110, Vol. 1, SOARCA Project Peach Bottom Integrated Analysis, Rev. 1, (May 2013)
- NUREG/CR-7110, Vol. 2, SOARCA Project Surry Integrated Analysis, Rev. 1 (August 2013)
- NUREG/CR-7008, MELCOR Best Practices as Applied in the SOARCA Project (August 2014)
- NUREG/CR-7009, MACCS Best Practices as Applied in the SOARCA Project (August 2014)
- NUREG/CR-7155, SOARCA Project Uncertainty Analysis of the Unmitigated Long-Term Station Blackout of the Peach Bottom Atomic Power Station (May 2016)
- NUREG/BR-0359, Modeling Potential Reactor Accident Consequences, Rev. 1 (December 2012, update in progress) 27
Acronyms & Abbreviations AC Alternating Current MSIV Main Steam Isolation Valve BOC Beginning of Cycle NTTF Fukushima Near-Term Task Force CCDF Complementary Cumulative PDF Probability Density Function Distribution Function PGA Peak Ground Acceleration CCI Core Concrete Interactions PRA Probabilistic Risk Assessment CDF Core Damage Frequency PRT Pressurizer Relief Tank CST Condensate Storage Tank PZR Pressurizer DC Direct Current RCP Reactor Coolant Pump EOC End of Cycle RCS Reactor Coolant System EPZ Emergency Planning Zone RLZ Realization EF Early Fatality RPV Reactor Pressure Vessel GE General Emergency RtePM Real Time Evacuation Planning Model HL Hot Leg SBO Station Blackout FLEX Diverse and Flexible Coping Strategies SG Steam Generator FTC Failure to Close SAE Site Area Emergency FTO Failure to Open SIP Shelter in Place LCF Latent Cancer Fatality SME Subject Matter Expert LNT Linear No Threshold SNL Sandia National Laboratories LTSBO Long-Term Station Blackout SOARCA State-of-the-Art Reactor Consequence Analysis MACCS MELCOR Accident Consequence Code System STSBO Short-Term Station Blackout MCR Main Control Room SV Safety Valve MELCOR Not an acronym - accident TDAFW Turbine Driven Auxiliary Feedwater progression code System MelMACCS MELCOR to MACCS Source Term TVA Tennessee Valley Authority Converter UA Uncertainty Analysis MOC Middle of Cycle 28