Official Transcript of Proceedings NUCLEAR REGULATORY COMMISSION
Title:
Advisory Committee on Reactor Safeguards
Docket Number:
(n/a)
Location:
Rockville, Maryland
Date:
Thursday, December 6, 2007
Work Order No.:
NRC-1898
Pages 1-168 Closed Session: Pages 1-41
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 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION + + + + + ADVISORY COMMITTEE ON REACTOR SAFEGUARDS (ACRS) + + + + + 548TH MEETING + + + + + THURSDAY, DECEMBER 6, 2007 + + + + + The meeting was held in Room T-2B3, Two White Flint North, 11545 Rockville Pike, Rockville, Maryland, at 8:30 a.m., William J. Shack, Chairman, presiding. MEMBERS PRESENT: WILLIAM J. SHACK MARIO V. BONACA SAID ABDEL-KHALIK JOHN W. STETKAR OTTO L. MAYNARD DENNIS C. BLEY MICHAEL CORRADINI Chairman Vice Chairman Member Member Member Member Member
GEORGE E. APOSTOLAKIS Member DANA A. POWERS J. SAM ARMIJO JOHN D. SIEBER Member Member Member-At-Large NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
NRC STAFF PRESENT: SAM DURAISWAMY, Designated Federal Official ROBERT LEE TREGONING LEE ABRAMSON NILESH CHOKSHI KHALID SHAUKAT RICHARD DUDLEY TIM COLLINS GREG CRANSTON TAI HUANG ROBERT PRATO FAROUK ELTAWILA JIMI YEROKUN MIKE CHEOK JOCELYN MITCHELL DON DUBE RICH SHERRY ALSO PRESENT: DOUG PRUITT YOUSEF FARAWILA EDWIN LYMEN
3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
TABLE OF CONTENTS PAGE Opening Remarks . . . . . . . . . . . . . . . . . Draft Final NUREG-1829: Dr. George Apostolakis .. . . . . . . . . . Robert L. Tregoning . . . . . . . . . . . . 9 9 4
Nilesh Chokshi .. . . . . . . . . . . . . . 43 AREVA Enhanced Option III Long-Term Stability Solution: Said Abdel-Khalik . . . . . . . . . . . . . 77 Tai Huang . . . . . . . . . . . . . . . . . 78 Dr. Yousef Farawila . . . . . . . . . . . . 90
4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 of the order. This is the first day of the 548th meeting Advisory Committee on Reactor Safeguards. CHAIRMAN SHACK: P R O C E E D I N G S (8:28 a.m.) The meeting will come to
During today's meeting, the Committee will consider the following: Draft final NUREG-1829, estimating loss of coolant accident frequencies through the elicitation process; And a draft NUREG on seismic
considerations for the transition break size; The AREVA enhanced Option III long-term stability solution; The state-of-the-art reactor consequence analysis, SOARCA, which will be a part open and part closed meeting; A draft ACRS report on the NRC Safety Research Program; And preparation of ACRS reports. A portion of this meeting may be closed to discuss safeguards and national security information related to the SOARCA project. This meeting is being conducted in
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 interest. accordance with the provisions of the Federal Advisory Committee Act. Federal meeting. We have received no written comments from members of the public regarding today's session. We Mr. Sam Duraiswamy is the Designated for the initial portion of the
Official
have received a request from Dr. Edwin Lymen, Union of Concerned Scientists, for time to make oral statements regarding the SOARCA project. A transcript of portions of the meeting is being kept, and it is requested that the speakers use one of the microphones, identify themselves and speak with sufficient clarity and volume so that they can be readily heard. I will begin with some items of current The members are scheduled to interview a We'll be handing
candidate today during lunchtime. out a resume. group. Other information.
It's one candidate so we'll do it as
Ms. Barbara Jo White,
who has been with the ACRS office for almost 40 years is retiring on January 3rd, 2008. All of these years
she has provided outstanding administrative support to the members. She has always ensured that the members
have a good place to stay when they attend ACRS NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 endeavors. (Applause.) CHAIRMAN SHACK: In addition to a retiree, we have some new additions to the ACRS staff. Dr. meetings in town or out of town. She has been exceptional in assuring that the federal register notices for the subcommittee and full committee meeting have been issued consistent with FACA requirements. Her outstanding administrative support to members, hard work, dedication, professional attitude in dealing with no only the members and staff, but also the public are very much appreciated. Thank you, and good luck in your future
Harold Vander Mollen will be joining the ACRS staff as a senior staff engineer on December 24th. He will be
the responsible engineer for the Subcommittees on Reliability Practices. He came to the AEC regulatory staff from the National Bureau of Standards in 1974. He spent 13 years in several technical branches in NRR working on reactor physics, accident and transient analysis, technical specifications, generic issues program, and PRA issues. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
and
PRA
and
Regulatory
Policies
and
7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 In 1987, he and his section were
transferred from the PRA branch in NRR to the PRA branch in the Office of Nuclear Regulatory Research, just in time to work on the NUREG-1150 project. When it was finished, he was put in charge of PRA methods development. In 1999, after 12 years
in the PRA branch in RES, he took over the generic program issues program again. Welcome aboard. (Applause.) CHAIRMAN SHACK: Ms. Kendra Freeland
joined the ACRS/ACNW&M staff on October 22nd as an administrative assistant. She will be handling travel authorization, vouchers and compensation for the
members, one of our most important concerns. (Laughter.) CHAIRMAN SHACK: Prior to joining the ACRS ACNW&M staff, she served as secretary for the Division of Contracts in the Office of Administration. Kendra received a Bachelor of Arts degree in corporate and broadcast communications from Elon University, degree in Elon, North Carolina, from and a Master's Pacific
communications
Hawaii
University, Honolulu, Hawaii. She knows how to pick a graduate school. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Sounds good. Well, then we can move to our business today, and our first item of business is the draft NUREG on estimating loss of coolant accident, LOCA, frequencies through the elicitation process, and Dr. Apostolakis will lead us through that. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
(Laughter.) CHAIRMAN SHACK: Welcome aboard.
Ms. Guita Irani joined the ACRS/ACNW&M staff on November 13th, 2007, as an information
technology specialist. NRC.
She is a new member of the
Guita started her career in information technology working as a DOD contractor for the Joint Spectrum Center in 2000. In 2003, she moved to the
Pension Benefit Guaranty Corporation to support their federal contracts. Guita holds a Master's degree in
information technology from the University of Maryland and has been involved with software development and IT support throughout her career. Welcome aboard. MR. DURAISWAMY: you can do that tomorrow. CHAIRMAN SHACK: Okay. We'll hold. Janet is not here. So
9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MEMBER APOSTOLAKIS: Thank you, Bill.
We had a subcommittee meeting on the 27th of November when we heard from the staff on both studies. One is on the expert judgment elicitation
process and results, and the other one was more focused studies on seismic issues. There were no issues that were raised by the subcommittee. The members appear to be -- well,
actually they were -- pleased with what they heard. The staff also presented their responses to public comments on the elicitation process. So we asked them to come back today and give a shortened performance so that the members will form an opinion. And we are expected to write a letter at this meeting. Rob? MR. TREGONING: Yes. Okay. So with that, I should turn to you,
MEMBER APOSTOLAKIS:
MR. TREGONING: Thank you, Dr. Apostolakis and Mr. Chairman. My name is Rob Tregoning from the Office of Research, and to my right is Lee Abramson, and we will be leading you through the first abridged
presentation on the development of NUREG-1829, on passive system LOCA frequency development for riskNEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 had informed revision of 10 CFR 5046. Why did we get into this work? Well, our So
bosses essentially told us we needed to do this.
this work was done in response to Commission direction provided by SRM-02-0057, and a couple of quotes there. "The staff should provide the Commission a
comprehensive LOCA failure analysis and frequency estimation that is realistically conservative and amenable to decision-making with appropriate margins for uncertainty." So that was our edict. That was our
direction. Also, in the same SRM, the Commission said the staff should use expert elicitation to converge whenever possible service data and PFM results. So
those are our marching orders. That's what we set off to do. And we're here today, as Dr. Apostolakis indicated, requesting a letter or an ACRS Our
recommendation to publish the study, NUREG-1829. opinion is it sufficiently meets the
Commission
direction, satisfies that and should be published as a result. A brief executive summary. We used the
formal elicitation process to develop estimates of generic BWR and PWR passive system LOCA frequencies NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 results. agreement associated with material degradation. of 12 panelists. We had a group
They provided us with quantitative They
estimates supported by qualitative rationale.
did this individually in individual elicitations for underlying technical issues that were developed as a group. We on had the very good or generally good
qualitative
LOCA
contributing
factors. However, as you've seen in the report, there was large individual uncertainty and also large panel variability in actually quantifying the estimates. So coming up with frequency estimates associated with the phenomena that they were predicting. That wasn't surprising, of course. We
expected that, and that was the reason that we chose to do elicitation and a to begin with, to provide with a
framework
mechanism
for
dealing
the
expected large uncertainty in panel variability. The bottom line, we developed group
So we aggregated the individual estimates
for the LOCA frequency distribution parameters. So we didn't determine distributions per se, but we
determined certain parameters of the distribution, the 50th, 95th and the mean. We used a number of
different aggregation schemes. One scheme we used was NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 considering bottom line looked at the geometric mean. We thought those results were
consistent with the elicitation objective in structure and they're also generally comparable with the NUREG/ CR-5750 estimates. That study was the prior study that was used to develop LOCA frequency estimates back in the mid-'90s. It was not done using elicitation. It was
done by just simply evaluating service experience. As mentioned in this last bullet, and we
other
aggregation
schemes
other
aggregation schemes can give you quite a bit different results, and typically these other schemes that we looked at did result in higher LOCA frequency. We show the results here. results for BWR and These are the PWR. Generic
frequencies, you see three curves on each of those. The black curves are the medians, the reds are the mean, and then the green are the 95th. The the center best points estimate, are and what then we're their
confidence bounds, the error bars represent 90 percent confidence bounds. So a five percent and a 95 percent upper and lower confidence bound about that best estimate. These particular results, we did a modest NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 adjustment for overconfidence. We adjusted error
factors or uncertainty ranges in some expert opinions to coincide with a well-known elicitation. want to call it a fact, that but a finding tend I don't in to many be
elicitation
studies
experts
overconfident. We didn't see as strong a bias for
overconfidence in these results.
So that's why a
correction ended up only being relatively modest in this case. These 90 percent confidence bounds, it says 95, but it's really 90 percent -- they're used to represent or reflect the diversity or the differences among individual panelists' opinion, and then the difference between the medians and the 95th really reflect the individual panelists' uncertainties. So
there's two types of uncertainty or variability that we're trying to capture. MEMBER CORRADINI: you just said? repeat, please? MR. TREGONING: The confidence bounds Could you repeat what
You said that the very -- could you
about any individual value here, either about the mean, median or 95th percentile, they represent the spread or the difference among the individual
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
14 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 estimates that we got from the panel. For each
panelist, we asked for them for each answer that they provided, we asked for a best guess, essentially a median value, and then we asked high and low
estimates, which we interpreted as fifth and 95th percentile estimates about that mid-value. MEMBER CORRADINI: MR. TREGONING: Thank you.
I guess this is the only It was
new slide that the subcommittee hasn't seen.
put in at a request from Professor Apostolakis at the subcommittee meeting. He wanted to see what the So we did a very and these are
distribution shape looked like. simple exercise to create
these,
essentially -- all I did was take a simple, lognormal, not split or anything, just the full lognormal, and I fit them to the 95th and the mean because those are the two parameters that we're most interested in using. So I forced it to go through the mean and the 95th, and then the question was, well, how well does it estimate the median and the fifth. And the
fifth, really a lot of extrapolation to get down to the fifth. And both Lee and I were quite surprised at how well the fits tended to be. So in the medians in
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 there. MEMBER APOSTOLAKIS: So what does this all cases, there was less than 30 percent error, which again, for LOCA frequencies, considering the
variability we have, is pretty darn good. And even in the fifth percentile it was less than 50 percent error, except in one case where we had a percent error of 200. And I've picked four plots here. The blue plot and the red plot are actually the worst fit of all the distributions that we fit to this thing. So
these are the worst, and the green and the black are more representative of the types of fits you would see. And I just summarized the percent error
mean? Let me understand the curve. So the blue curve is for BWR-5; is that what you're saying? MR. TREGONING: Yes. So that's the BWR So -So you are only
LOCA frequency at LOCA Category 5. MEMBER APOSTOLAKIS:
showing mean curves. MR. distributions.
All of these are mean curves? No, these are
TREGONING:
So plotted on these are all of the
percentiles of the fit distribution. MEMBER APOSTOLAKIS: Oh, I'm sorry. Yes.
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MR. TREGONING: our actual values. MEMBER APOSTOLAKIS: MR. TREGONING: The means. I And then the points are
Those are our values.
identified where the means are because the other thing you can see here is that the means in all cases are a relatively high percentile, not surprising, but the means vary anywhere from about the 70th to even as high the 85th percentile on the distribution,
depending on which parameter you're looking at. MEMBER CORRADINI: So what you're plotting here is a fit shape to the three points that we saw in the previous curve on some break sizes. MR. TREGONING: the previous one -MEMBER CORRADINI: MR. TREGONING: That's fine. Right, right. If I go to
-- where we didn't show
the fifth, that was the four points that we developed. MEMBER CORRADINI: Got it. Now, the scale, I
MEMBER APOSTOLAKIS:
guess, distorts a little bit what is happening because they rise too steeply, don't they? I mean the curves. MEMBER BLEY: of magnitude. MEMBER APOSTOLAKIS: The green one is the Goes over about four orders
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 only one. MEMBER BLEY: And the other one goes over
at least two orders of magnitude. CHAIRMAN SHACK: There's a good spread in
those curves, and the spread is sort of what you think. For small breaks it's narrow, where for big
breaks, it's very wide. MR. TREGONING: small break. So the black one is the
So you can see they're in order of
increasing break size, obviously. MEMBER APOSTOLAKIS: So if I did a PRA
tomorrow and I needed the frequency of various LOCAs, I could use this one, although this is based -- these curves are based on what you call a baseline approach, right? MR. TREGONING: aggregated results. These were geometric mean
Yes, they were. So I guess since I
MEMBER CORRADINI:
didn't read in detail the report, I interpret the difference between the Ps and the Bs as primarily a pressure effect, not a materials effect and not a chemistry effect. What do the experts say relative to that in terms of their -- because if I remember this process, you have to elicit not just a number, but a NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 intermediate reason for the number. MR. TREGONING: Yes, and I didn't overlay
in this presentation Ps with Bs, but what you see is the Ps have higher small break frequencies. MEMBER CORRADINI: at that just from the numbers. MR. TREGONING: Right. Well, it's not I show two Right. I was looking
clear from this because I show two Ps. small Ps and then I show two large Bs. MEMBER CORRADINI: previous one I was looking at. MR. TREGONING: Right. Right.
It's just the
So the Ps are
higher at small break, and then they're actually lower in intermediate breaks, and then at the biggest breaks the Ps get higher again. The Ps are higher at small break primarily due to the fact steam generator tube rupture failures and concern for PWSCC issues related to CRDM, other small tube piping. The Bs tend to of get higher in the by
break
because
largely
driven
remaining IGSCC issues just due to the fact that BWRs, a lot of the large piping still retains flaws that were generated earlier under normal water chemistry the documents you see. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 will -MR. TREGONING: CHAIRMAN SHACK: MR. TREGONING: But even if I take the -It's still true. -- I don't show it here. And while the experts believe that they have been mitigated to the sense that the frequencies were relatively low, they still raised up or rose up to be one of the highest risk contributors. So even
though they have been mitigated, they still were the largest risk contributor. And then when you get down to the highest frequencies, PWRs dominate, again, and that's more of a population issue. PWRs have larger pipes, more,
bigger non-piping components that could fail and lead to a LOCA. So there wasn't anything unique that was
driving that other than the increased population. CHAIRMAN SHACK: But these frequencies
still include the steam generator tubes -MR. TREGONING: CHAIRMAN SHACK: Yes. -- for the PWRs which
If I take the steam generator out, Ps are still higher. CHAIRMAN SHACK: Now, you said the
lognormal plots were the baseline or are they the error factor corrected? NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 and low. MEMBER APOSTOLAKIS: MR. TREGONING: And low. higher. MR. TREGONING: They would be broader high yes. MEMBER APOSTOLAKIS: Now, if you included corrected. CHAIRMAN SHACK: Corrected. Okay. That's sort of your best estimate curve. MR. TREGONING: That's what we would call, MR. TREGONING: These are error factor
in this the multiple distribution, what do you call that? MR. TREGONING: The mixture distribution? MEMBER distribution. APOSTOLAKIS: Yes, mixture Would
How would these curves change?
they be broader? MR. TREGONING: Yes. Yes.
MEMBER APOSTOLAKIS: They would be broader on the high side especially or -MR. TREGONING: high, broader to the high. CHAIRMAN SHACK: And the means would be They would be broader
And low, and then the
means would be shift obviously. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 that. MEMBER APOSTOLAKIS: You could develop MEMBER BLEY: same probably? MR. TREGONING: No. Again, it's a totally different way to aggregate. So, no, the medians would be -- I'm not sure how the medians would work out. CHAIRMAN SHACK: They're higher. I mean, Medians would be about the
that's the way they work out when you look at the numbers. MR. TREGONING: CHAIRMAN SHACK: can read the table. MR. TREGONING: Yes, I haven't looked at Okay. I mean, just looking I
curves like this using that other method. MR. TREGONING: Yes. Yes, you could. But then ultimately
MEMBER APOSTOLAKIS:
you might want to combine the curves. MR. distribution, itself. TREGONING: come up with But the the mixture
you
distribution
So you wouldn't turn around and fit it as we
have here. MEMBER APOSTOLAKIS: I understand that,
but you can always change those distributions based on insights you got here. I mean, ultimately what
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 believe me. MR. TREGONING: with all of my heart. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
matters is what you learn from the whole exercise, right? I would make an analogy with licensing a reactor. We have been told many times in this room by the staff our decision is not based on a single analysis. It's the result of a process. So here, you know, you might say at the end I want a distribution which says, you know, I've been through this. I've done it ten different ways.
This is what I think it is. Now, that takes guts. (Laughter.) MEMBER APOSTOLAKIS: Anyway, it's okay.
I did want to inquire, but this is the biggest problem Bayesian methods have. MR. TREGONING: Okay. I think George I think
could give this presentation at this point. he already has. So we'll --
(Laughter.) MR. TREGONING: -- spend the rest of the
time -- we'll continue to move through. MEMBER APOSTOLAKIS: And you still don't
I believe you, you know,
23 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 there. MR. TREGONING: Yes, yes. For that one using So we did a number of sensitivity
analyses, and Professor Apostolakis alluded to one of those, and all I'm going to touch on is one because it's the one that's the most interesting. It's the
one that's the most controversial, and that's looking at different ways to aggregate individual results. So what I had shown before is essentially the blue curves here, and these are the means. So if
I go back to this plot, those blue curves correspond to the red curves on this plot. changing colors on you guys. And the red curves here, they represent either arithmetic mean to aggregate the I apologize for
individual expert estimates or analogously, at least for determining the mean, actually creating a mixture distribution from the result. MEMBER APOSTOLAKIS: Wait a minute now.
Isn't it true that they can be an arithmetic mean where the percentile is not the same as the mixture distribution. MR. TREGONING: looking at the mean it is. MEMBER APOSTOLAKIS: Oh, you may be right Right, but when you're
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 quantity. CHAIRMAN SHACK: MEMBER ARMIJO: It is. I don't understand why arithmetic parameter. MEMBER BLEY: about that one. MR. TREGONING: mean Well, the way we did we just took the I'd have to think real hard
aggregation,
arithmetic mean of all the percentile estimates of the estimates. So for the mean estimates that they gave us, it's just the arithmetic. the mean. It's just the mean of
So the mixture distribution, when you work
through it, that mean is also the mean in the middle. MEMBER APOSTOLAKIS: The mean is a funny
there's such a big difference between the mean and the mean of the mixture for LOCA Category 4 on the BWR. I mean, those two curves are very different compared to the PWR. What go that? CHAIRMAN SHACK: Essentially we had one of the panelists that if you look at the red curves, they're weighted by one panelist result. So
essentially one panelist was very much higher than the rest in their predictions for BWRs. So that's why the curve shape looks like that, and that's why it's so NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 right. CHAIRMAN SHACK: If you see a six it's not a pipe break anymore, you know. MR. TREGONING: This isn't a pipe break. You different than the geometric mean aggregate. Now, that one panelist, his model was -again, his biggest risk driver was IGSCC, and this was a PFM approach, and the PFM model was essentially giving him the result that, you know, a large break LOCA has about the same frequency as a much smaller LOCA in that large piping. So that's why his results look so flat, but they were very different. They were different
than everyone else's results, and that's one of the reasons for the big difference between or the primary reason driving the big difference between the
arithmetic mean and the geometric mean aggregated results. MEMBER ARMIJO: But that same person, when you got to the Category 5 and 6s was pretty much consistent with the rest of the -MR. TREGONING: Yes, they go back down,
So it's apples and oranges, and five for BWR.
really needed a complete rupture of the prime recirc. piping to get that. So when it came down to complete, NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 that they you know, these guys are balancing for all the pipes. They have to consider complete ruptures of smallest pipes leading up to that LOCA category, as well as partial ruptures of bigger pipes. MEMBER MAYNARD: Well, they all didn't
provide data for every category either, did they? MR. TREGONING: didn't all They were consistent in us BWR and PWR
provide
information.
Some of the experts only felt qualified But once they gave us
to give us BWR information.
information, they gave us information from all the categories from one plant type, and that was required because of the way we structured the elicitation. We
needed that to be so that they could develop selfconsistent estimates. Now, some of their estimates were very like if their qualitative response said I don't think the pump casings are a significant risk driver, right, they didn't necessarily need to give us quantitative estimates at that point. You know, we can take that
information and say, okay, I just need to make sure that these don't contribute to your final risk
profile. MEMBER ABDEL-KHALIK: What is the smallest size sample in all of these categories? NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 review. MR. TREGONING: We had eight BWR estimates and nine PWR estimates. So of the 12 we had one
expert that didn't provide any estimates for anything. So then we had 11 that gave us estimates, and eight of those gave us BWRs, and nine of them gave us Ps. So
we had two people that didn't give us Ps and three that didn't give us Bs. I wanted to talk a little bit about the We've had quite a bit of review. We started
with the panel itself.
We did a lot of Q&A and
feedback on the individual responses that they gave us. They gave us pieces. To develop one set of
frequencies for an expert it took about 100, 200 questions that they had to answer. So they didn't
necessarily see what their final outcome was when they were giving us a testimony. So when they were giving us the testimony, we were checking to make sure their rationale and the numbers they were giving us makes sense, and that was actually the most extensive part of the process So the
because quite often those things didn't match up. we had quite a lot of to feedback make of each that of
individual
experts
sure
their
quantitative numbers did support their qualitative rationale. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 They also reviewed the calculations and analysis that we did on their individual results to make sure it was accurate within the framework that we had provided them, and again, as I mentioned here, once the draft NUREG was put together, they also reviewed the general qualitative and quantitative findings and conclusions. Did you have a question? MEMBER CORRADINI: Well, I was going to -I was looking through the -- so they provided their analyses or their bases for their judgments and they spoke with each other and discussed it as part of it. Was there interplay between the experts? that's what I'm asking. And then did they reevaluate it and give you another set of numbers? MR. TREGONING: There was, and there were The way we I guess
chances for them to do the reevaluation.
structured it is we brought them together as a group to develop all of the issues and brainstorm and identify the things they were going to be evaluating. Then we did some background analyses which I'm not going to go into, but essentially the base case analyses, and we brought them together again and discussed that. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 sorry. helpful. document. MR. TREGONING: Okay. I'm sorry. Thank you. that Then they went off and did their
elicitations individually.
Then, once we got all of
the results, we brought them back together as a group again and said, "Here are your individual results and here are your results with respect to the group, and we focused on that meeting and that was about a threeday meeting. We were looking at differences because
you're always looking for, you know, if one expert is different than the other. You're looking to see if
there's qualitative reasons that the other ones hadn't thought about. And they were given the opportunity after meeting if they so chose to revise their
estimates, but to be honest, nobody did.
So even
though they were informed, no one felt strongly enough about the new information that they thought they needed to go back and redo their estimates. Again, we've had a lot of group exchange prior to that as well. MEMBER CORRADINI: On the flow chart, I'm
I didn't mean to take -- but that was very I was looking for the flow chart in the
MEMBER CORRADINI:
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 helpful. MR. TREGONING: So, again, we had a lot of feedback with the panel. We also had a small external peer review. We had two people with decision analysts and a statistician, and we asked them to look at the structure of the elicitation, the analysis procedure framework, how we did aggregation, and those review reports are publicly available. The external peer review was quite
It helped us refine our analysis technique.
We've had a large number of ACRS interactions that we thank you. this point. And then we've had internal staff review, NRR as well as people in the Office of Research, and finally we went through public review and comment. I'll briefly touch on here in the next few slides the public comments that we got. draft NUREG-1829 in June of 2005. We issued I think this is our 13th or maybe 14th at
We opened the
public comment period, and then we closed it on November 2005. We had 29 comments from the public. had nice diversity of comments. of the elicitation panelists We
We actually had one himself that felt
compelled to comment.
That was interesting.
We got
some comments from academia and -NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 okay. MEMBER APOSTOLAKIS: MR. TREGONING: them to do that. opportunity and That's good. dynamics. MEMBER APOSTOLAKIS: Did he ever explain
to you why he didn't raise his concerns during the elicitation process? MR. TREGONING: Oh, he did. Oh, he did? He just wanted to put
MEMBER APOSTOLAKIS: MEMBER CORRADINI: it on the record. MR. TREGONING: (Laughter.) MR. TREGONING: You
Yes, he did.
know
these
group
We discussed his comments and issues as a
group, and then the group -CHAIRMAN SHACK: MR. TREGONING: Didn't see the light. No, no one, but that's
And, in fact, I encourage
I said, you know, there is an just the fact that you were an
elicitation panelist, that shouldn't stop you from commenting as well as it shouldn't stop anybody from commenting. So he did that. MEMBER APOSTOLAKIS: you commented though. (Laughter.) NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
It would be funny if
32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MEMBER APOSTOLAKIS: MR. ABRAMSON: Right? Or Lee.
We're a united team.
MR. TREGONING: I don't know if that would be unprecedented, but probably close to it. And then we've got a number of comments from industry, owners groups, individual licensees. At the same time that we went out for public comment we were anxious to get this out and get some comment, but the document was being reviewed internally by NRR as well, and so we got a large number of comments from the NRR staff, and in fact, the document we provided to you has the NRR comments commingled with the public comments, and we grouped the comments topically just so ACRS -- we would be able to avail you of that information so you could consider all the comments that we got, and in total we got about 101 separate comments. So in general, to summarize the public comments, you know, public comments were generally useful. They identified some additions and
clarifications, that we went forward to hopefully improve the exposition, as well as facilitate the use of these results. the author's None of the comments certainly in and hopefully the responses
mind,
document that, presented a significant challenge to NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 the appropriateness of the objective approach,
analysis or results. With the public comments as well, you can see the most passionate controversy is still the proper method for aggregating individual estimates to produce group estimates. MEMBER ARMIJO: Was there any particular
-- and I would know this, I guess -- from the BWR owners on that discrepancy, was that a big, big issue? MR. TREGONING: You mean on the
discrepancy between the one expert and the others? MEMBER ARMIJO: MR. TREGONING: Right. No, we didn't get a
comment on that. The comment that we got from the BWR owners groups or at least one comment, and you've heard these, is they were concerned that we didn't appropriately credit mitigation of IGSCC. And we did
change some of the language in the report, but it didn't change any of the estimates, and you know, we documented in the report as well as in the response about how we considered mitigation, not just of IGSCC, but for all of these mechanisms that people were considering. IGSCC was probably the most unique case because a lot of the service experience that you have NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
34 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 estimates for BWRs is colored by that IGSCC experience and then anything -MEMBER ARMIJO: Anything prior to water
chemistry, pre-mitigation and all of that. MR. TREGONING: Anything before 1983, you
know, and we had a lot of discussion with the group about that. In fact, a lot of the service history
estimates showed pre-1983 precursor events, post-1983, and we actually then did sensitivity studies, both from a service history perspective and then a PFM perspective on the effect of different IGSCC
mitigations on the failure frequency. So it was something that we had discussed quite a bit in the elicitation. So I just wanted to give you an example of one public comment here, recognizing that we don't have time to go into a lot of them. be happy to take questions on any. But there was one comment that our SB LOCA were too high and that they weren't Of course, I'll
representative of operating experience. said, you know, that approximately
The comment order of
one
magnitude and then the NUREG/CR-5750 results. The implication is that we should be
having one SB LOCA every four years and that using NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 these estimates, at least the small break LOCA
estimates in existing PRAs would lead to unwarranted impacts that are not supported by operational
experience. MEMBER CORRADINI: that would be? MR. TREGONING: Yes, so we thought about So your thought about
that and responded, and I think the main thing that we thought was a good idea is we didn't have a comparison within 1829 on how the results compared with service experience. So we added this section. We had a section on how it compared with prior studies, and a lot of those prior studies had shown how they compared with service experience, but we thought a fresh look at service experience would be useful. And when we say "service experience," we're really limiting it to the small break LOCAs because that's where we have -- you can actually argue that we've had a couple of events. Certainly we've
had steam generator events, and we've had a few pipe breaks in Class 1 systems that border on the small break LOCA threshold. other than zero events. making those comparisons. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
So we actually had some data So we felt most comfortable
36 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 years. MR. TREGONING: About ten years now. 5750? This is the basis or sort of the basic response or the fundamental points in the response that we made to this one. It's at least the author's
opinion that the SB LOCA and the 5750 estimates are generally consistent. The steam generator tube
rupture estimates are virtually identical.
In fact,
they're actually a little bit lower, and that mainly is reflective of the fact that we've had additional service experience since 5750 came out, but there are about -MEMBER APOSTOLAKIS: I don't remember. MR. TREGONING: It was published, I think, in '97 or '98, but a lot of the events, most of the events were analyzed up to about '96. MEMBER APOSTOLAKIS: That's about ten What was the year of
The BWR SB LOCA estimates are actually quite similar to 5750, within about 20 percent. The
big discrepancies are the PWR SB LOCA estimates. They're higher than the 1829 study, about a factor of five, and again, the experts supported that with, again, there was a lot of concern at the time about the effects of PWSCC on small break LOCAs. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
So their
37 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 of things. right? MR. TREGONING: Well, there were a couple concern was reflected in this increase. We also -MEMBER APOSTOLAKIS: Excuse me, Rob. So
this is the estimate that would lead to one small break LOCA per four years? statement of that nature -MR. TREGONING: Right. -- in your slide, I mean they made a
MEMBER APOSTOLAKIS:
When we published the draft NUREG, we had Then we
the steam generator estimates separately.
just had the LOCA estimates that had combined the steam generator and the small break LOCA estimates. Okay? So when they did their estimates, they did a simple subtraction, and the way we aggregated, you can't really do a simple subtraction to get the results. So what we did is we went back and looked at
each individual set of results and for each of those individual results, we subtracted their steam
generator risk contribution from all the others, and then we re-aggregated. So we analyzed in a way that was
consistent with how we analyzed the rest of the NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 that? MR. TREGONING: It's not quite as high as results, and because it's not a linear analysis, you don't get the same answer as you would if you simply subtract them. So in the new 1829, we actually
published the small break LOCA estimates without steam generator contributions as well so that people can see what they are. So that's in addition. We added those, as well as we did -MEMBER answer to this? MR. TREGONING: MEMBER What do you mean? Is it a true APOSTOLAKIS: But what is the
APOSTOLAKIS:
statement that your estimate leads to an average of one small break LOCA every four years? MEMBER ARMIJO: Sanity check. MEMBER APOSTOLAKIS: Are you still doing A simple --
It doesn't make sense.
that, but you know, you're one in four, one in five, but, again, you have to look at -- these are not average. These frequencies are never intended to
represent averages over the entire operating fleet, right? They were meant to be snapshots of where we
are now, given concerns, and they were concerns about NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 saying. the effect of PWSCC and PWRs on what those frequencies were. We looked at all of the pipe breaks that we had and Lee helped us. We did a Poisson-type of
analysis, and you might want to, and you might want to comment on this, and showed that the estimates that we had as well as the uncertainty about those estimates, even though they were elevated, they were still
consistent with operating experience or they weren't inconsistent with operating experience. MEMBER APOSTOLAKIS: So is operating
experience telling us then that we have something we can call a small break LOCA once every four or five years? Is that what you're saying? MR. TREGONING: No, that's not what I'm
I'm saying the current frequencies that we
have are higher for PWR SB LOCAs. MEMBER CORRADINI: But if I could just
read on page 750 of the report, specifically the paragraph here that you guys have is that you point out that for small breaks, the current elicitation is lower than the pilot, but it is higher than -- as you state, "However, the current elicitation concerns for PWSCC cracking and BWR CRDM nozzles results in
additional increases." NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 it's Is that -- I'm looking for a physical reason why, if the expert judgment is larger than the service experience, does it come down to those sorts of -MR. TREGONING: Yes. Okay, all right. Yes, that was the
MEMBER CORRADINI: MR. TREGONING:
qualitative rationale driving it, and the expert said even though, again, when we did the study, when we did the results or when we did the elicitation, it was 2003. Oconee, So you know, we had had Davis-Besse. We had V.C. Summer. We had
We were still in the
process of attempting to develop mitigation strategies or we hadn't even started it yet, to be honest with you, for PWSCC. We're really starting that now.
And many of those same experts said while elevated now, the expectation is that once
mitigation has been fully implemented, that those frequencies will decrease again. MEMBER CORRADINI: Okay. That's fine. I
just wanted to understand because this is not my area. It's a material. So I'm always looking for the
physical reason underlying why an estimate might be different than the service experience. CHAIRMAN SHACK: Just Lee's Poisson
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 analysis gave him a resulting range of .7 to five breaks, and you've had one basically. MR. TREGONING: Yes. So the third red
MEMBER APOSTOLAKIS:
bullet, BWR small break LOCA estimates are higher by approximately a factor of five, but because NUREG-5750 is kind of old, they think this is reasonable. That's the implication there? MR. TREGONING: Yes. Okay.
MEMBER APOSTOLAKIS:
MR. TREGONING: And it's not the fact that NUREG/CR-5750 is old. It's the fact that, again, the
elicitation -- these estimates were supported by, you know, expectations for higher frequencies due to PWSCC cracking. So that's the third bullet. You know, the differences that do exist are supported by qualitative rationale, and we made a number of modifications. We
have provided the separate steam generator tube and small break LOCA estimates as I mentioned. much more extensive results, comparison and then we between also We have a 1829 and
historical
have
these
operating -MEMBER APOSTOLAKIS: So that frequency
then is roughly two or so, ten to the minus three, an NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
42 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 minus four. MEMBER APOSTOLAKIS: So if you have five average number? CHAIRMAN SHACK: Five times ten to the
minus four for small break LOCAs. MEMBER APOSTOLAKIS: minus four we mean by it. CHAIRMAN SHACK: Which in 5750 is one. Is what? Five, ten to the
MEMBER APOSTOLAKIS: CHAIRMAN SHACK:
A one times ten to the
ten to the minus four and you have how many PWRs? MR. TREGONING: Sixty-nine. Multiply that by 69. I
MEMBER APOSTOLAKIS:
Do I get this number of four or five per year? guess I -MR. TREGONING: No, no, no, no, no.
You
get one every four years is what the commenter -MEMBER APOSTOLAKIS: One every four years. So 69 multiplied by four. MEMBER MAYNARD: Well, another factor is
you want this to be a tool that's useful in the future, not necessarily reflecting exactly where we are today. It completes aging, and so the numbers
that you're giving and the tools that you're putting out there need to be a good five or ten years from now NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Let's go on. MR. TREGONING: That's it. as well as today. MR. TREGONING: Right. We did ask for
estimates. We asked for three time periods. We asked for current snapshot. Then we asked for 15 years from now and then we asked for another 20 years past that. So we did provide multiple estimates, but you know, there's a realization, too, that you know, your
epistemic knowledge state is changing as you go along, too. So certainly the further you asked people to
prognosticate, you know, we had enough uncertainty with the current day. So when you try to
prognosticate out further, you have more uncertainty, more variability. MEMBER APOSTOLAKIS: All right, fine.
MEMBER APOSTOLAKIS: Any questions before we move on to seismic? That's the next one, right?
(No response.) MEMBER APOSTOLAKIS: you taking over? MR. CHOKSHI: Yes. Thank you very much, Okay. Nilesh, are
MEMBER APOSTOLAKIS: by the way, as an afterthought. Tell us who you
are
and
why
you're
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
44 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 qualified to address this distinguished group. MR. CHOKSHI: Yes, I will. He drew the short straw.
CHAIRMAN SHACK: (Laughter.) MR. CHOKSHI: name is Nilesh Chokshi.
Okay.
Good morning.
My
I'm Deputy Director of the
Division of Environmental and Site Reviews, Office of New Reactors. And if you wonder why I'm here giving this presentation, I was in Research when this study was conducted two years back. So that's the reason I'm
here, and as you see from the list of names, this was an interoffice team, including seismic expertise, piping design, fracture mechanics, seismic risk, and also the people involved in the rulemaking. So this
was, you know, a substantial and also very large contractor support. In fact, Dr. Gery Wilkowski and I and his organization, EMC2, made the floor piping analysis, and he was here at the subcommittee presentation, and I think he's available on the phone also. good for me. So that's
He can answer some of the questions. So with that, let me -MEMBER APOSTOLAKIS: Who's the gentleman
on your right? NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 cover the we sorry. MR. CHOKSHI: This is Mr. No, Gery is not here. Shaukat. He's Oh, from
Khalid
Research, and you'll see him again when we move forward with this study, but he was the project manager of this study. Okay. basic did So now what I intend to do is to objective. discuss in At the subcommittee technical
meeting
detail
the
approach and rationale behind this.
I think during
the short time here I'm going to focus on some of the key research and findings and not as much on the methodology. And then what I want to do is towards the end I'll summarize the response from the industry on specific questions where we are asking the proposed rulemaking and where we are and what factors we need to consider as we move forward in this rulemaking process, but as pertains to this particular issue. So let me start with the objective.
Instead of directly estimating the seismic and use break frequencies as it was done for the expert elicitation, we decided to concentrate on a different question, conditions and and the question was: what would, are under the the
likelihood
which
seismic-induced loading, which would be incompatible NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 question, with the proposed TBS? In other words, would the seismic-induced breaks, would they be larger than the TBS and would have frequencies of ten to the minus five or more? And given, I think though, that's that a very object germane was to
the
provide this information so people can comment and respond to questions. In the context of the proposed
rule, this was a direct question and also within the time period it's something you can do, you know, estimating absolute frequencies given seismic events. It's a much larger undertaking. In order to answer this question we took six activities. piping, piping We looked at unflawed piping; flawed that has cracks or degradations;
indirect failures; review of past earthquake, past PRAS; and then there was a study conducted in the early '80s in connection with GDC4, which was to answer the question whether the LOCA and the
earthquake load seems to be combined, and this was a full-blown probabilistic fracture mechanics analysis. The first three are the different --
mechanisms, how the piping and piping system can fail and would have, you know, an impact on TBS. The review of past earthquake experience NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 for -MR. CHOKSHI: No. How long we'll talk and past PRAs were to get additional insight; also to calibrate system level analytical study CRD, what we see in the earthquake experience, and what we find analytically; is this consistent or not; and then this, the last, was an important study, and a lot of decisions were based on this study. It also provides
a direct way for us to calculate the mean direct failures. So we used the modified history to build a short current-day hazard and use that for the indirect figures. Now -MEMBER ARMIJO: Were these analyses only
done for the transition break size pipes? MR. CHOKSHI: No. Just for that size or
MEMBER ARMIJO:
about that, the next slide, the scope of the study. And we used basically a combination of deterministic and probabilistic approach. For the
unflawed piping and indirect failures, it's pretty much probabilistic approach, and we did not estimate the four distributions, but it was a probabilistic approach. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 On the flawed piping, the seismic loading was based on the probabilistic hazard, but the rest of the calculations were deterministic. To address some of the variabilities, we did a large number of samples, and I'll try to tell you about how many piping systems we looked at, and we also conducted some additional sensitivity analysis to look at the effects of key assumptions. Here is, I think, to answer your question, I'll move on to this viewgraph. One of the biggest challenges in
performing this type of analysis is the availability of the design information because that was our
starting point.
We needed normal operating stresses, I'm talking about design stresses,
seismic stresses.
material properties, and a few other things so we can do our calculations. One of the databases which had captured this information is the leak before break application database, and which basically applies to PWRs. So we
had these data available for PWRs, and that's why one of the reasons was material evaluations for BWRs. Having said that, there's nothing inherent in these matters or conclusions which does not apply to BWRs or other situation. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
49 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Now, the tool I mentioned about, that we wanted to look at a large sample to cover the range of materials properties, range of the seismic stresses, and the site conditions. We selected 27 PWRs, 24 on Large size We
the large side, three on the small sides.
from the seismic perspective is not critical. generally get higher stresses.
The second issue was what hazard curve to use, and this was a question because, two years back, this is when the Early Site Permit applications were coming in, and they were using new estimates. Without considering all factors, we thought that for the 27 sites the research we had available was that Livermore has those, and we will study the sensitivity of alternate hazard in a different way. So we decided to use the Livermore hazard curve. Now, I think to answer your question, because we see what's the effect on TBS, we selected piping systems larger than the TBS. We did examine
one or two cases with the TBS diameter, but more as a calibration, but which meant that we were looking at hot leg, cold leg, and crossover leg, and we selected 52 systems from 27 PWRs. We tried to capture the Okay?
highest trace locations and materials.
Now, one of the key, in this kind of NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 evaluation, you need to do a realistic estimate of seismic stresses. The design stresses is a starting
point, but as you go up in the earthquake, it's difficult; to seismic PRA, you have to do more
realistic estimate of seismic stresses at the higher level, and we used basically an approach commonly used in the seismic PRA and seismic margin to estimate those spaces. So that was a common approach, I would say, in all three, that we were trying to estimate realistic estimates at higher level of earthquakes, and for the flawed piping we selected two discrete levels, ten to the minus five probability of accidents and ten to the minus six probability of accidents. For direct and indirect, we can basically use, then, the entire probabilistic hazard code. So this is what I'm talking about matters and not too much more. In flawed piping and indirect, I'll do a little bit more, but at the subcommittee we showed some quantitative research on the unflawed piping, but I think other cases are a lot more
interesting.
So I'm going to -The key finding is that unflawed piping,
in order to get a seismic-induced failure, you have to have a lot of flaw. It just doesn't happen. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
In fact,
51 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 our numerical results would be an order of mean failure ten to the minus nine or less, okay, for the cases we looked at. I think it's important to a little bit talk about now earthquake experience. We have looked
at a number of industrial facilities and fossil power plants, most of the data come from, but all of the welded piping systems which are engineered actually behave very well in the earthquakes. We are looking
here at experience data up to .5g ground acceleration, and where we see failure, there's a severe
degradation, either support failure, again, associated with severe degradation; there is missing anchor bolts or corroded plates. We see relative motion. When you have an
inflexible pipe and there is a support, and in fact, at a recent earthquake in Japan, we saw, I think, all seven plants. There was a vent. Vent was connected
to a stack, which was in a different foundation, and other support was in a different foundation. All six
identically failed because of this anchor motion. And then things falling over the piping. So I think this result is consistent. So I think this case, that unflawed piping, unflawed piping is
basically a piping which meets the assumptions used in NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 through-wall questions. design and would fail by a particular shutting down than a collapse or a tear-through. Okay. Now, the flawed piping, I think Again, I have been
that's the more interesting.
conducting, you know, crude probabilistic analysis. We decided to look at design conditions and the conditions at the higher level. We wanted to look at
what are the critical flaws at ten to the minus five or ten to the minus six seismic range, and either they're large or small compared with the crude
allowable flow evaluation. So we performed all the normal operating conditions and earthquake, ASME inspection/evaluation criteria for circumferential surface flaw, and we also used the LBB procedure. wall flaws and how What would be the throughwould compare with the
they
critical flaws at ten to the minus five up rate and ten to the minus six up rate? And this was basically to answer two
Will ASME surface flaw criteria at normal
stresses find flaws that are smaller than the ten to the minus six or ten to the minus five? implies some inherent safety. And would the LBB are procedure smaller find than the the Because this
flaws
that
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
53 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 All right. particular flaws? And also I think, as I mentioned with the experience data, the question itself, how large these flaws have to be, is important. So that was the
parameter, you know, how large these flaws have to be. Because if you were really looking at the total probabilistically, then you will have to look at probabilities of existing flaws, probabilities of detecting flaws were they to link before. So I think
this information was very -- that's why I said that in the right context, the proposed TBS, this information was germane. So I'm going to now go to the resource. Let me first -- these are the two results This is the two
for the surface flaw evaluation.
systems from the 52 systems we examined, large, and the plot on the left, I believe it's in a hot leg looking at a ten to the minus five earthquake stresses from a Westinghouse PWR. what you are looking at. The X axis is the flaw length. Okay? And as you go from the extent of circumferential flaw length. On the ordinate is the flow depth ratio, is So But let me first explain
the ratio of through-wall to the pipe thickness.
as the flaws get smaller, circumferentially you have NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
54 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 to have a larger depth for them to become critical. As you move toward the right with the larger flaws, then you know, these critical depth ratios are becoming smaller. An interesting thing is
that you see that after a certain .5 or something or .6, you start approaching basically asymptotic value. You are also seeing the ASME code limit, that basically this requires evaluation of flaw. If
it's smaller than that, I think you can continue operation. .75 here. This is a typical case. this is called Category A. In the report In no case, you can go tolerate more than
The red line is the
critical ten to the minus five critical flaw length. Yellow is using the ASME code strength and procedure, and this is the same as the ASME procedure, but using actual strength. The code allows that. In this case
the critical flaws associated with the earthquake, large earthquake traces, is much larger and this is a typical ten to the minus five. Now, we also have what we call Category C, a few cases, and I'll show you the overall. In this
case, the critical flaw is smaller than what the ASME code would allow, but I think the one important thing is that in all cases there are very high -- issues. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
55 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 The cracks have to be pretty great. Now, what I'm going to show you now, the results for the ten to the minus six slightly different form. in a little
This is the calculation The ten to the
performed for a large earthquake.
minus five was sort of a starting point for the TBS, but we also wanted to look at what happens at larger earthquakes. One of the things, it answers the
question of an alternate hazard as well as whether there is a sharp transition somewhere, you know. And what you are seeing here is that upon a very large flaw, for a different seismic -- for ten to the minus six seismic stresses, what are the critical flaw depth values? And you see that .3 is
the smallest value. It's somewhat material-dependent, but in many cases, you wanted this stress level is much larger. So you have to have a very significant,
large cracks in the pipes before you get to the seismic in these breaks. Here is the summary of the 52 cases, and
as I mentioned, for eight times ten to the minus five, yes, 48 cases the critical crack sizes will be larger than the ASME code. In one case it was larger than
the core evaluation using the core values, but smaller using the actual strength values, and in three cases NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
56 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 size, the was the second case I showed that's called Category C. But, again, I think to me the -- well, I'll come back to the final message. So this was the
results of the surface flaw evaluation. The second thing we wanted to look at, the leak-before-break behavior, and we wanted to see that if you applied LBB analysis as currently with the factors of safety of ten on the leak rate and on the flow size factor of safety of two, and how would that compare when you do the same calculations, but using higher seismic stresses and also examine the
sensitivity to different factors of safety and also maybe different assumptions on the leak detection capabilities or the leak rate. So I'll show you the results from the sensitivity studies. This is also a factor of the
crack morphology, and I'm going to -- so we looked at three crack morphologies, one with a very smooth crack, a PWSCC type crack, and corrosion-free. Okay. So here are the results. On the X
axis side is the ratio of normal to normal plus higher side mixed traces at ten to the minus five. go left implies higher seismic stresses. This is the leakage over critical flow leakage flow size using the current So as you
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
57 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 says five. MR. CHOKSHI: Well, yes. The five gpm, procedures, and get the design basis stresses, SSC stresses, and this is the critical flaw sizes at ten to the minus five. These are the results of fire systems for all different plants. What it shows, that if you
consider the factor of safety of 1.5 instead of two, in the report there are other results, but if you do these calculations, you keep the liquid requirement the same as factor of safety of ten on the detection capability of one gpm. You will find some cases where the critical flaw size would be smaller than the LBB. But if you look at an alternate leak detection capability, and I understand that some LBBs use this, of .5 gpm, keep the same factor of safety, you can see that there. And you know, this was not to draw the conclusions, but to provide information so people can evaluate in all of the proper contexts. MEMBER ARMIJO: You said .5, but the chart
you calculate your break size and the flaw size based on the certain rate. Okay? Five gpm is used in this
calculation, but the current procedure requires that if you have a flow liquid of five gpm, your detection NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
58 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 capability should have a factor of ten safety, safety factor of ten. So detection to be .5 gpm. Okay?
Basically this is a sensitivity study on two different detection capabilities. results from the leak before break. So this is the So here is the
summary of the flawed piping. I think to me these two viewgraphs really are critical. The critical crack
slips are larger than 40 percent for the ten to the minus five and larger than the 30 percent thickness for ten to the minus six. substantial, large flaws. So you're talking about Again, I think to me it
seems to be consistent with what we are seeing. On the LBB flaw size, again, we see in many cases that the LBB flaw size will be smaller than the critical, and for the better appreciation of under what conditions the LBB could be, you know, at least a viable consideration, we need some sensitivity
studies. Now I am going to move to the indirect failure, and this is the failure mechanism which PRAs include. Seismic PRAs traditionally, and I think for
good reasons, have not included piping failure as an initiator of LOCAs, other than small LOCAs, but some of the PRAs have included this, and to give an
example, the 1150 study plant, there was a failure NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 mode where the steam generator support failed, and what happens is that those sequences occur in very large earthquakes. but core They they damage dominate not the release the an
contributions, contributing
are
dominating you make
because
assumption that if I'm going to feel a support of a steam generator that is going to allow movement, I'm going to fail containment also at the same time, and I'm going to have a LOCA which probably is not
possible to mitigate. And so this is the assumption also. The
assumption is that the failure frequency of support is the same as if we left concealed break. Now, let me now talk a little bit about the original Livermore study we just conducted in the early to mid-'80s. They grouped the plants in various renderings. They are Westinghouse, CE, BMW, and then
they also looked at one BWR plant. They selected the one pilot case, and then they looked at data across the fleet. They used the
generic hazard curve for east of the Rockies. What we had to do was to primarily update the hazard
information. Now, we have a plant specific or site specific hazard information. So that was the major
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 modification to the work they did. of studies. in the We selected two cases. case in their They had a number One was what was analysis of a
bounding
Westinghouse, and another was CE, and I'll show these in a minute, but in Livermore study they had estimate of the fragility of the support, and we modified that to reflect the site specific information. CHAIRMAN SHACK: Now, did you look at
IPEEE-2 to see if those fragilities from the Livermore study were still -- you know, when people looked at them presumably at their individual plant they did a little better job. MR. CHOKSHI: And I think what happened
when IPEEE, that seismic sequences were basically governed by other failures. inside, but when we did So there was little the seismic margin
development, we had looked at this, and I think only two components from the seismic margin you examined below .5g is the pressurizer support and the vessel support for the BWRs. Most of the other components are very high capacity, and the results reflect that. So basically
we completed the failure probability of the support with the site specific hazard and modification of the Livermore study. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
61 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 effect. MR. CHOKSHI: Cascading, right. You have the beams, MEMBER SIEBER: How detailed is the
calculations involving support fragility for large components like -MR. CHOKSHI: MEMBER SIEBER: MR. CHOKSHI: because Unit 2, you It says --- steam generators? It's fairly significant
know -Not just examination of forth, you know. My
MEMBER SIEBER: cradles and so
experience is that bolting is a critical issue in those large supports. MR. CHOKSHI: Right. Yes, we need to look at a variety of failure modes and see. You know, they have combined them so that you could arrive at those. MEMBER SIEBER: I guess you can draw the It's the bolts
same conclusion about pipe supports. that fail first.
MR. CHOKSHI: Yes, yes. And in the Diablo Canyon PRA, that was the mechanism they included, and it's more than one support in order to fail a pipe. You need to fail -MEMBER SIEBER: You get a cascading
to have at least I think, if I remember right, in the NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
62 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Diablo, we looked at five supports. MEMBER SIEBER: Thank you.
MR. CHOKSHI: Now, maybe if I show you the results from the Livermore regional study, and here you can see that -- let me concentrate on, let's say, median values. These are ten to the minus six, ten to the minus seven order, and what they had on the mean perspective, this was the lowest capacity plant.
Ninety percent was two times ten to the minus five. Making corrections to the fragility and using the site specific casuals, two times ten the minus six. when I That's so we wanted to compare that, and at the -and I believe for the
look
Westinghouse, the lowest capacity, three time ten to the minus six. So it was still an order of less than
ten to the minus five. Now, EPRI, as a part of the response to questions, did some additional calculations using the EPRI latest hazard coverage, and they examined three cases, and I'm going to report on the results, but we haven't reviewed. They're basically the same
approach, modified fragility, but they do add some additional factors, and their results range from six times ten to the minus six to five time ten to the minus eight, and one of them was a BWR. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
I think five
63 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 know, times ten to the minus eight was BWR. Anyway, so the bottom line, I think is that you still are away from the ten to the minus five type of pressure. So now let me go to the overall
summary of the story. Unflawed piping systems have very -- you seismic frequencies are small. Critical
suppressed floor and through-wall, you know, LBB, you have to have large flaws to have seismic induced failures. And then indirect piping failure, the things we looked at, it still seems like an order of ten to the minus six per year. overall summary. Now, I'm going to switch to the approval and questions associated, which are included in the draft code, and responses. The proposed rule So this was the
contained extensive discussion. You know, it observed that the expert elicitation had not included explicit consideration of seismic induced failure, and here is a large uncertainty, and there was still a question whether a plant specific assessment would be required or not, and there were three specific questions that were posed. One was to comment on the evaluations of
the study and, you know, if any comment they had on NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
64 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 that; effects on the five-day degradation on seismic in the LOCA frequencies and, you know, from their own information; and then also the other one was to potential policies and options to address this issue other than what, you know, we have put forth. The comments primarily were from the
industry, and I'll summarize quickly.
Basically they
had no really comments on the study itself. They said we agree with the study's findings and that TBS is not adversely affected from the seismic consideration. And I'll go to the next slide. This is important. This data risk or the
change in the risk due to seismic is considered low, and our basic argument was that components in the piping in the primary loop and supports generally have a much large capability or capacity. It's a lower
fragility compared to the rest of the plant, and so the risk is general dominated by the other previous scenarios, and so the seismic to be that. And then I mentioned within that failure, EPRI gave us additional resource which we'll have to look at, and then their bottom line conclusion was that plant specific assessment cannot be required. This is now my last slide. I think moving forward we have an ACRS recommendation. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
There is an
65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 SRM requirement, and to address this issue we need to consider a number of things before we you know. First, I think we need to look at the response to questions, and I think my three bullets actually. only. Changing the risk, I think probably that's important, that we fully understand that. One of the important, I think, The next two of those, one of the response
considerations will be from Commission SRM and ACRS recommendations, how this will get -- addresses the defense-in-depth and mitigation recommendations.
Because that will have an effect on any of the risks under any risks. We need to understand fully whether the seismic -- is that under what conditions the seismic risk could be affected, and I think it is my -- this is mine now -- that if the seismic risk comes from the structure type failures, unless plant modifications are made. Now, what the rule and mitigation plays and nonseismic failure plays, we need to look at and we need to better understand what we do here, but I think unless, you know, there is a significant change to the supports or something, I think I -- and then -NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
66 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 this has don't state. MR. CHOKSHI: Okay. MEMBER APOSTOLAKIS: MEMBER CORRADINI: I didn't get that. We don't understand
you. I 'm trying to understand you. MEMBER APOSTOLAKIS: Your conclusion you
MEMBER APOSTOLAKIS: Unless -- keep going. MR. CHOKSHI: Okay. What I was trying to
say, that unless somebody modifies the supports or something, unless there are physical changes to the seismic capacities are less. that occurring, you know. Now, there may be a system and operating condition changes which we have to evaluate after we understand what defense-in-depth and mitigation. Structural changes I think is, you know, somebody's postulating. on seismic risk. And then finally also I think, you know, to be reviewed to understand what the Then it will have an impact You know, I don't see
Commission has asked that we develop guidance on the issue of applicability of 1829, and I think to me this has some of the things which we may think
independently in the seismic, you know, some of the regulatory considerations and things like that. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
67 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 and see So at this point I think we have to wait how these things go before we make the
decisions or, you know, how do we deal with these issues. So that's the end of my presentation. VICE CHAIRMAN BONACA: Yeah, I repeat what I said during the subcommittee. You know, I would
like to see sensitivity of the results that you are presenting to that multiplier you used to eliminate the excess conservatism, as they call it, associated with the design stresses. I agree that we have to use a reduction, and I'm not proposing that you would use the design values, but that's a significant multiplier. I mean
you are using a .6 or something like that if I remember now that was in the report, and I would like to understand the sensitivity and its conclusions to that multiplier. MR. CHOKSHI: I think it's a good -- you
know,, and what drove us to the looking, also the ten to the minus six, you know, a number of things, what happens with automatic hazards, what happens in the seismic stresses, you know. It does not answer fully the question I think you asked, but -VICE CHAIRMAN BONACA: NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
Well, I mean, you
68 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Thank you. MEMBER APOSTOLAKIS: We were also told at made the point that, you know, you cannot tolerate the design values. They're excessive. Right. So there is some
MR. CHOKSHI:
VICE CHAIRMAN BONACA:
place between the design value and the reduced value, and the approach you're using to scale it down seems to be pretty empirical. MR. CHOKSHI: I mean, it just -I think maybe the one thing
I didn't mention, I think I agree with you, but there are a number of factors one can consider, and we basically selected adjustment of the seismic spaces using the concentrator on the site specific hazard information. We were, I would say, considerably biased in that selection. We did not use all of the factors, but you are right. It was qualitative, but we could
have examined it quantitatively. VICE CHAIRMAN BONACA: Yeah, to get a
sense again of the sensitivity, what is the margin it should have and -MR. CHOKSHI: Absolutely, we can do that.
the subcommittee, a member of NRR, that this rule, the priority of this rule has been reduced, right? NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
69 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MR. CHOKSHI: MR. DUDLEY: Well, then you have person. Yes. When the Commission
reviewed the staff's paper that addressed the ACRS recommendations, the Commission agreed with the
staff's recommendation that the priority of this rule should be reduced from a high priority rule to a medium priority rule. So we are proceeding forward with a rule, but on a little slower basis and our next due date is to provide a schedule to the Commission for completing this rule, and the schedule is due by March 31st, 2008. MEMBER APOSTOLAKIS: At the subcommittee
you told us that the reason or one of the reasons, I guess is that there are insignificant safety benefits of this rule, but the benefits really are negativities because they won't be able to raise the power. that a correct statement of what you said? MR. DUDLEY: Well, that was the ACRS, The staff pretty much was included in the Is
included in the ACRS' letter. agreed with that, and that
Commission paper. MEMBER APOSTOLAKIS: But isn't part of
risk informing the regulations to remove unnecessary regulatory burden? NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
70 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 staff. We agree with that, and that's why I think the rule was not killed as a whole. MEMBER APOSTOLAKIS: MR. COLLINS: Oh, okay. MR. DUDLEY: Yes, it is. And wouldn't this
MEMBER APOSTOLAKIS: rule remove such burden?
I mean we are not looking for safety benefits that haven't been risk informed. MR. COLLINS: This is Tim Collins of the
I mean, it was initially
considered high priority because of the potential for safety benefits. That's got its high priority. Now,
when we seem to come to the realization that there wasn't a whole lot necessarily there, then its
priority got reduced to medium because there was still the potential for reducing unnecessary burden. MEMBER APOSTOLAKIS: Yeah, that's helpful. MEMBER ARMIJO: If I recall, some of the
industry people were negative because the transition break size values were too high and implied that they wouldn't use a rule or they didn't expect too many people to want to use that rule. case? MEMBER SIEBER: It doesn't make any Is that still the
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
71 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 difference. MR. COLLINS: folks are here. themselves. Well, I think the industry
Maybe they could answer that for
I think that's still true. MR. DUDLEY: Right. We haven't heard any
indication from industry. MEMBER ARMIJO: So the priority, it's not
high priority for the Commission for safety benefit, and it's not much value to the industry. I think it's a good study, but I think the medium of priority for pursuing it is probably the right thing to do or even less. MEMBER MAYNARD: I think the value to the
industry or to a reduction in burden depends on what the form of the final rule comes out to be. What are
the transition break sizes and what are the mitigating requirements? I think it's something that can be of benefit and a reduction, or it can be something that provides really no benefit or no reduction, depending on really those two primary things, transition break size and what's required for mitigation. MEMBER SIEBER: Well, it allows more
realistic calculations in 5046 space, and to me that's a significant benefit. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
72 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 would be. MR. CHOKSHI: And a number of people, asking the MEMBER APOSTOLAKIS: So if it's of medium
priority, that means when are we going to see you again. MR. DUDLEY: I guess it will depend on the schedule that we provide to the Commission on March 31st. MEMBER APOSTOLAKIS: Okay. Your schedule. MR. DUDLEY: That's correct, and some of
that depends on the work that you're hearing today. So we really can't provide you a schedule at this point in time. CHAIRMAN SHACK: NDE people for Nilesh, did you think of what they think the
detectable crack size. You know, you've given me ASME code limits. You know, that's wonderful. I can't
find a ten percent crack very reliably. MR. CHOKSHI: I'm looking, but what I can
tell you is that there was extensive discussion about that issue. CHAIRMAN SHACK: I would expect there
including NRR resources. The best way to summarize is that we were starting basically can we put it, say, probability of detection, dealing directly. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
73 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 why I CHAIRMAN SHACK: And so what you're really relying on is the probability that you're not going to have a 30 or 40 percent deep, long crack, which, you know, is pretty small. MR. CHOKSHI: think it's a Implicitly I think that's way to present this
good
information, so people think about those factors. You know, what's the probability of having this size? How will it grow into the service, you know. to me it brings the focus. We were trying not to draw conclusions because it's hard to come to without any kind of probability. CHAIRMAN SHACK: You still have to make So I think
that judgment on how likely those cracks are. MR. CHOKSHI: And the report says that,
you know, that given this, you know, that's why we looked at ten to the minus five but actual probability of failures, you know, but if you consider all of these factors, you know, it's obviously small. MEMBER SIEBER: I think there's some
comfort if you take something like centrifugally cast, austenitic stainless steels, it's pretty hard to find; it's not as easy to find flaws in that as other -CHAIRMAN SHACK: That's a mild statement.
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
74 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 staff. I demonstration would almost view They those tell as you a the MEMBER SIEBER: It's not a curiosity. On
the other hand, for the large pipes made of that material, it usually before a break, which tells us you something, too, and so I don't think that we are left without assurance. CHAIRMAN SHACK: Oh, no, no, no. Those
are big cracks under any circumstance. MEMBER SIEBER: Absolutely. Well, this is what you want to avoid. MR. CHOKSHI: And I think then there are
many recent studies will even lock in a few more insights into what's more likely, but when we were doing it, I think, this was to present the information so people can make an informed judgment. MR. TREGONING: This is Rob Tregoning from
requirement.
performance that you have to have and then it would be up to maybe the reg. guide or even licensees that want to use 5046 to provide some sort of demonstration that their piping will meet that performance. flawed, you know. CHAIRMAN SHACK: But I think they're going to have to make that argument not on NDE, but on the NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
It's not
75 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 fact that they have water chemistry and materials that are not suspectable to this kind of -MR. TREGONING: Well, there are a variety
of ways that you can make that argument. CHAIRMAN SHACK: to make that. MR. TREGONING: NDE is a piece of it, but But I agree. They have
I would agree that you probably don't want to -that's not your sole argument. CHAIRMAN SHACK: I keep losing my argument that you actually have to demonstrate leak before break for these piping systems before you can take credit,b ut I still think that's a good idea. MR. CHOKSHI: But to me I think for a risk informed rule, this is really a key question, and we need to understand the potential changes and will they have an effect on seismic risk. Seismic risk is different than anything else and its common cause effects and are you really affecting this, you know? It may be dominated by some other things, and may not have a really -- you know, the redundance doesn't have the same effect from a mitigation point of view of difference in depth point of view on the seismic. If you put tow identical
systems it doesn't buy you much. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
76 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 much. MR. CHOKSHI: I hope he's there. MR. WILKOWSKI: (Laughter.) MEMBER APOSTOLAKIS: And, Mr. Chairman, 25 minutes early. CHAIRMAN early, George. MEMBER APOSTOLAKIS: I want to use that up in future meetings. CHAIRMAN SHACK: MEMBER SIEBER: CHAIRMAN SHACK: break now until 10:45. (Whereupon, the foregoing matter went off the record at 10:05 a.m. and went back on the record at 10:46 a.m.) CHAIRMAN SHACK: It's time to come back A credit. Eliminate one of them. I think we will take a SHACK: Twenty-five minutes Yeah, I'm here. Thank you. Thanks, Gery. So we, I think, need to ultimately be able to answer this question, you know. MEMBER APOSTOLAKIS: Any other comments or questions from the members? (No response.) MEMBER APOSTOLAKIS: Well, thank you very
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
77 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 the staff Chairman. The Thermal Hydraulics Subcommittee held a meeting on November 14 to review AREVA's detect and suppress stability solution and methodology. We heard presentation by AREVA and the staff regarding two licensing topical reports, ANP-262P, Rev. 0, entitled "Enhanced Option III, Long-term Stability Solution," and BAW-10255P, Rev. 2, entitled "Cycle Specific DIVOM Methodology Using the RAMONA5 Code." Subsequent to the subcommittee meeting, issued revised draft safety evaluation into session. Our next topic will be on the AREVA
Enhanced Option III long-term stability solution, a topical report, and Said will be leading us through that. MEMBER ABDEL-KHALIK: Thank you, Mr.
reports on November 27th. At this time we will hear presentations by AREVA and the staff. Parts of this presentation will
be closed because of the proprietary nature of the material to be presented, and at this time I'd like to call on Dr. Tai Huang of the NRC staff to begin the presentation. MR. CRANSTON: Let me interject just
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
78 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 point out quickly. My name is Greg Cranston, Reactor Systems
Branch Chief. Before we introduce Tai, I also want to that Jose March-Leuba from Oak Ridge
National Laboratory also participated in preparing this, is unable to attend today, and Dr. Tai Huang will be making the presentation. Thank you. DR. HUANG: Okay. I'm Tai Huang from
Reactor System Branch, and I'm the original reviewer, technical reviewer for the AREVA BWR Owners' Group long-term stability solution, including ATWS LOOP and instability, and like today the Chairman says that we have two topic reports, and these regarded to
stability.
One is Enhanced Option III and second
would be the cycle-specific DIVOM methodology using RAMONA5-FA code. These two topical reports are really interrelated, to support each other. And as you see today, because the industry demand on that extended operating domain, so you see these three because of this demand for this extended operating domain which pose new challenges to In
stability as shown in this power flow map there.
this, back in the old day, we starting with the original licensing thermal power, and now into the NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
79 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MELLLA EPU condition, and beyond that, they have a MELLLA+ region here because this stability boundary over here, and during here that, or the two up pump at trip these
situation, endpoints.
either
they
end
In this region it would be up there, and
this would be much thicker beyond this stability boundary region. So that instability, why they post these, the new kind of instability, as you see in this power flow map, and then what to do then. You know, the
staff and industry has developed and reviewed, and under this committee approved that they are generic solution for the BWR Owners" Group solution and to handle this region, and then after review, extended good up to the region here they're called BW owners group long-term stability solutions. However, in this region there are two measure authority. One of them has been approved.
The other one is today's, the under committee review. So you see this is a BW owners group approved
solution right there, that neither Document 319608, and give us all kind of solution. There are three options, E1A, 1B and
Option 2 and 3, and these are approved and documented in this document there or the U.S. BWR reactors have NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 implemented one of these solutions depend on their need today. And now because, like in the previous slides there to handle the MELLLA region, there are two methodologies. It's under review, and one of
them, GE DSS detect-suppress solution, density has been reviewed and approved for MELLLA+, and today one of their topical reports on AREVA, they're called EOIII, under review right now. So what is what they call EO-III and what is difference between EO-III and enhanced Option III, and as you see previously, the owners group provision, they have an Option III. So the difference would be
still keep the Option III features and plus some are different from AREVA so that they become enhanced Option III. So enhanced Option III really is an
evolutionary step, rely on existing methodology and hardware for Solution III and what the difference is that EO-III introduced measures for addressing the review of stability associated with extended flow in all conditions and the higher probability of single channel hydraulic instability excitation. So the enhanced Option III have this kind of features over there. So the new element to use
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
81 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 he's asking. MEMBER APOSTOLAKIS: three slides back. DR. HUANG: This one? Yeah. Two slides back, enhancement to existing Option III solutions are such as they introduced, introduction of a calculated
exclusion region on the power flow mat designed to preclude single channel instability. Also, they have a calculation procedure, how to do it. Option III. Yes. MEMBER CORRADINI: Can you show us on the So this is different from the regular
diagram where the exclusion region is? DR. HUANG: closed session. MEMBER CORRADINI: CHAIRMAN SHACK: cartoon though? DR. HUANG: You want to show on cartoon? Okay. Can you show us on the Okay. That would be in the
CHAIRMAN SHACK: Yeah. I think that's all
MEMBER APOSTOLAKIS: DR. HUANG:
Basically I would say most
likely similar with this concept, but the detail be in the process, you know, to show you one the slides. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
82 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 CHAIRMAN SHACK: I think we just need to
deal with this in the closed session. MEMBER CORRADINI: Okay. Thank you.
DR. HUANG: Okay. So now because you have EO-III and you have to have the way to apply it, EOIII, so they need something they call Option III. If
you're aware of the Option III, they have OPRM system using the OPRM input to get the set point. need a DIVOM curve. So the second topical cycle specific DIVOM methodology from AREVA, and this time on curve really is a relationship between the hot bundle relative oxidation magnitude and the limiting fractional change in critical power ratio, and this is really a document in BW owners group solution, Needle 32465 document, and details go in there. And our review will be a trace, you know, like capability of the RAMONA5-FA system core to model neutron oxidation of the regional mode pipe and also that range of input data defined that set points within the reload cycle for which diamond curve is generated, and to the end they have to summarize what is the result of these calculations and come out with a time on curve. start review So that would be, you know, the area for that time on So they
coverage
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
83 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 EO-III is methodology. And our conclusion under EO-III, we said an acceptable authority to detect and
suppress oscillation should they occur, which means a DVC-12 design criteria, design criteria 12. So the
EO-III solution features provide protection up to and including the end of MELLLA conditions. will be included in cross-section. Now, let's go into the conclusion for the time on curve. The time on category called there, The detail
this is AREVA mass authority, is consistent with previous approved BWR owners" group mass authority document in Needle 32465 document. RAMONA5 is an integral part of AREVA time on methodology, and they're using RAMONA5 and the staff review, and RAMONA5 is capable computing power flow and void oxidation with consistent phase lag and of a frequency that presented the unstable oxidations, and they can estimate the loss of critical power radio induced by this oxidation, and also AREVA has commit to support the staff review of RAMONA5-FA for time on calculation, and on top of this because the staff only make these limited reviews for this limiting
application for time on calculation, the detailed review will be filed in the future. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
84 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 And because this, so there is staff in the SER that has revised. there. They say RAMONA5-FA limitation
The first was in the MELLLA+ region, if you One condition is
want to apply this mass authority.
the application of RAMONA5-FA to calculate time on curve under extended flow window operating domain, such as MELLLA+, it restricted true stability
solution, having a scram protected exclusion region that substantially reduced the potential severity of power oxidation and why they have relieved that one there. In the cross-section we have a curve which
shows that region always protected, you know. And also, there's a penalty of ten percent must be added to time on slope calculated by RAMONA5FA for extended flow window operating domains, and this penalty is equivalent to penalty of ten percent added to calculated relative CPR response for even power oxidation magnitude, which means that they put more margin there. You cannot rely on this, you know. The endpoint would be the calculation of whether your final MCPR compared to the stability limit. So
there's penalty like equivalent to about close to ten percent because here is, say, from here and ten percent on top of that calculated there, and you put the same oxidation magnitude. You come out with CPR
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
85 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 over initial MCPR. So there's a penalty over there
and ten percent would be penalized for that. Then, you know, like a reason for this being important is because today's power operation, you need a higher radio power peaking. power flow ratio is higher. Also, your
So that means in the
MELLLA+ region you're exposed to this and start really quicker and then the probability is higher. So that's why, you know, start will be review these and fit these. If they want to get this ten percent penalty out, you have to review this line by line for the core. MEMBER ABDEL-KHALIK: I'd like to point
out that these two conditions were imposed by the staff after the subcommittee meeting on November 14th. So these were two new conditions that were included in the revised safety evaluation report that was issued on November 27th. MEMBER SIEBER: Let me ask a question. Yes. RAMONA5-FA been
MEMBER ABDEL-KHALIK: MEMBER SIEBER: Has
approved by the staff on its own merit as opposed to in conjunction with this application? DR. HUANG: Actually staff haven't
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
86 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 approved this. However, in that application portion
of the staff review, we review some of this, but not fully review for this RAMONA5-FA. MEMBER SIEBER: Okay. So RAMONA5-FA just
as a computer code has not been staff approved. DR. HUANG: Yes. And when you use RAMONA5-
MEMBER SIEBER:
FA for this application without that blanket approval, what alternate methods did the staff want to assure us and everyone else that RAMONA5 will give reasonably accurate results? DR. HUANG: Oh, okay. Ask staff.
MEMBER SIEBER: You don't have a code that will do that as far as I know. DR. HUANG: Yes. Staff really looked at
the RAMONA5A, what it can do for this limited use for the time on calculation, is try to learn that where the time on -RAMONA5A, they can confirm that
oxidation, you know, to that extent.
Also they can
confirm what the loads of CPR are, you know, how they're protected within the range of the uncertainty there. So staff looked at that and see this can
perform this limited application up to the MELLLA region, not MELLLA+, yeah. MEMBER SIEBER: Now, the restrictions that NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
87 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 the staff has recently proposed, ten percent and so forth, I take it somewhere along the line you're going to explain how those penalties somehow are related to the use of RAMONA5 and why that penalty is good enough to say that stability can be detected and suppressed. DR. HUANG: AREVA can support this one. They say five
The staff looked at proposed idea.
percent, for example, at beginning, and we say, well, this five percent penalty probably not good enough, and then we say, well, twice this five percent -- if we draw that line from that generic time on curve slope, it's about .05 slope. Now, we say ten percent penalize that one. We see about .5. You know, it's ten percent. If .5
sit up over there, equivalent to about ten percent of energy released and ten percent of CPR margin you lose. That's a lot of penalty. You know, you look at
and you compare that initial MCPR versus later CPR to come out with the set limit. It's kind of a big
penalty from this operation. So staff say, well, ten percent should cover these conditions. MEMBER SIEBER: I presume that some place
in the presentation you will elaborate on that. DR. HUANG: Yes.
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
88 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 about -MEMBER ABDEL-KHALIK: This remains as a MEMBER SIEBER: Because it looks to me
like there's some pulling of numbers out of the air and saying we don't think this is good enough, but there is no basis. But this ought to be okay. Yes.
DR. HUANG:
MEMBER SIEBER: And that may be because of my lack of full understanding. that's the way it appears. DR. HUANG: Yeah, okay. I will think On the other hand,
major concern inasmuch as it appears to be -- you know, the adequacy of this penalty has not been fully justified and/or documented, and hopefully we'll hear some information as to why this gives us adequate assurance that this is okay until the staff completes its review of RAMONA5-FA. MEMBER SIEBER: Well, the staff doesn't
have the analytical tools to do that right now I don't think. MEMBER ABDEL-KHALIK: You know, we would
like to wait and hear what they have to say as to justification for the adequacy for such a penalty. MEMBER ARMIJO: But in effect, if this
goes through, this would be a limited approval of that NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
89 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 code with some penalties that we get to determine whether it's justified, which seems to be a little bit backwards. It seems like you're going to approve the
entire code and then address its applicability to different problems. We're doing it backwards. The problem you've solved
MEMBER SIEBER:
now in the total review has to wait until additional analytical tools are available. DR. HUANG: Yeah, we're taking into
consideration it's ten percent penalty equivalent to MCP and they say .02, .01, some kind of number like that. know. MEMBER ABDEL-KHALIK: We'll probably get So perhaps with the So we justify why this ten percent is, you
more information in the closed session. what we ought to do is just move on
presentation. Thank you, Dr. Huang. At this time we'd like to move on to the AREVA open part of the presentation before we get to the closed session. (Pause in proceedings.) MEMBER ABDEL-KHALIK: Let's proceed with
the hard copies until visual aids are returned. MEMBER ARMIJO: We're working off of this
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
90 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 that. (Laughter.) MEMBER APOSTOLAKIS: single digits. DR. FARAWILA: Okay. Chairman, members of the ACRS Committee, my name is Yousef Farawila. I Well, you know, it's for this session? MEMBER ABDEL-KHALIK: MR. FARAWILA: the slide numbers. MEMBER APOSTOLAKIS: Sure. We can manage Yes.
And I will be giving you
will be presenting an overview of AREVA's Enhanced Option III long-term stability solution and associated DIVOM methodology using RAMONA5-FA. Slide 3. Just a quick road map of the presentation. First, I present a quick overview of the original Option III detect and suppress solution and talk about Part 21 report against it and the recovery from the Part 21 both in the short term and in the long term. And after that in closed session we will present enhanced Option III solution, which depends on excluding single channel hydraulic instability, and we will also mention the codes and methods that support that option, and then welcome your questions. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
91 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 that is LPRMs. monitors. MEMBER APOSTOLAKIS: Say it again. know. DR. FARAWILA: Oh, OPR, local power range In page 4, you will see a sketch
summarizing the original Option III, which is detect and suppress solutions. So if you look to your left
where the core sketch is, you will see a closely spaced LPRM strings and signals coming from them at different elevations. MEMBER APOSTOLAKIS: LPRM, OPRM? I don't
DR. FARAWILA: Local power range monitors, They are closely spaced so that they can
detect regional oscillations, not only core-wide, and for the sake of redundancy, you have several of these composite signals, and for each one of them, you collect signals from different LPRM elevations. When you sum them up, you get a signal called OPRM for oscillation power range
monitor.
That signal can be oscillatory, noisy, and
before you process it first, it is filtered to remove high frequency noise, and it's also normalized. And
the filtered and normalized signal goes to a period based detection algorithm, the PBDA. The function of the period based detection NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
92 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 understand. algorithm is to examine the periodicity of the signal. If the periodicity is confirmed by having several successive periods within a tolerance range, then you suspect that you have an oscillation and you want to examine the amplitude of that oscillation against a preset set point. If that is the case, if passed that test as well, that means you will get a trip signal. There is a trip logic that requires more than one trip signal order to actually scram and protect the
reactor. Next page. MEMBER CORRADINI: Let me make sure I
And then not only is it the amplitude, You look
but the number of times it crosses, right? for a number, not just -DR. FARAWILA: Okay.
I have two tests
here. One is the periodicity, and so you see a number of confirmations. It could be 12. It could be --
MEMBER CORRADINI: DR. FARAWILA:
That's where the N is.
Right. Okay. The second one is an
MEMBER CORRADINI: DR. FARAWILA:
amplitude, not periodicity. MEMBER CORRADINI: Don't worry. You're
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
93 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 part. (Laughter.) MEMBER CORRADINI: fancy new ones. DR. FARAWILA: Okay. Next page. Easier to see than the fine. (Laughter.) DR. FARAWILA: A big effort making this
All right. Because the system is designed to suppress the oscillation to protect the CPR safety limit, so inherently there is required a relationship between that oscillation and the CPR response. relationship is called the DIVOM curve. That
It is based
on time domain code calculation of a regional mode oscillation and the output is closest -- for each oscillation you get the relative oscillation
magnitude, and you see the corresponding loss of CPR margin, and you plug these against each other to generate a DIVOM curve. Originally, in the original Option III, that DIVOM curve is generic. It's calculated once,
and it covers all plans and all cycles and all field designs. Next slide. We come to mention, as the agenda
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
94 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 enough for requires, the Part 21 report against the Option III solution and how it was resolved. General Electric
filed for a Part 21 report in August of 2001, which states that the generic DIVOM curve is not always conservative, and not conservative meaning higher than stated CPR response, which is equivalent to saying that it has higher DIVOM slope. And that condition occurs at high radial peaking or high power-to-flow ratio, and the change was not always in the smallest steps. It can be
sometimes rather high, up to probably doubling the generic value. The way this issue was resolved in the short term was through the BWR owners group collective efforts and the procedure was revised in order to prescribe cycle specific DIVOM calculations. So with
DIVOM being cycle specific, if such higher slopes are present, they would be taken into account. However, this short solution was not good most severe conditions that could be
expected from MELLLA+ and also the accounting for much higher DIVOM slopes would result in low setpoints, which makes the system more susceptible to noise, and you have a probability of spurious scrams. very undesirable thing. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
That's a
95 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 another one. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
So in the long term, a solution that takes care of this DIVOM problems that's not susceptible to it would be also applicable to MELLLA+. So if we
focus on application to MELLLA+, it automatically covers the other operating regimes that are minor in comparison. AREVA's long-term stability solution is enhanced Option III. I will just give you a quick You
example of what a well-behaved DIVOM curve is.
see it's fairly linear and very well defined slope. This one we calculated with RAMONA5-FA. We wanted to
examine what is that elevated slope DIVOM curve. We'll go in closed session very shortly. So if you want to look at the other side, the ill behaved ones, you probably want to borrow progressing by focus from your neighbor. Then you
will see like on the next page, you will see how a not well defined DIVOM curve could look like, and in the next two or three viewgraphs we are looking at the same exact reactor state. Just we changed the initial perturbation to show you that the calculated DIVOM may not just simply have a sometimes higher slope, but it's not really that well defined. Can we show the next one? Another one,
96 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 up. Farouk. MR. ELTAWILA: slight cold. Good morning. I have a session. Okay. So you could see there that
DIVOM -- last time when we were here addressing the subcommittee we presented something similar with a reduced order model. This one we are actually showing the RAMONA calculations. The rest of the presentation should we go to the closed session. MEMBER ABDEL-KHALIK: Okay. For the
reporter, we are now switching to a closed session. (Whereupon, the foregoing matter went off the record at 11:23 a.m. to reconvene in closed session and went back on the
record at 1:15 p.m. in open session.) CHAIRMAN SHACK: We can come back into
Our next topic is the State-of-the-Art We
Reactor Consequence Analysis, the SOARCA project.
met with the staff in a subcommittee meeting and discussed, essentially, their approach to the problem, and some preliminary results that they'd received. And they're now going to update the Full Committee on the process, and their current status. MR. PRATO: My Division Director will open
I got it from Bill Shack, that he
invited us to Oregon, and just blasted us with the NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
97 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 winds over there that we could not survive. As you know, the State-of-the-Art
Consequence Analysis, I'm going to call it SOARCA for brevity from now on, is a voluntary effort, so we have to rely on the involvement of the industry. So I
would like to start by thanking Surry and Peach Bottom for their cooperation, and providing us with the information that enabled us to do this analysis. We met with the ACRS in the summer of 2006, and we met with the Subcommittee last week. And we have completed the baseline calculation for both Surry and Peach Bottom. We started the SOARCA, because as you are aware, that has been the -- the Sandia Siting Study has been called into a different arena, and people are using it out of context, so we decided to do this analysis to try to update and replace the Sandia Siting Study. And we were motivated by a lot of
things, among them, improvement in plant operation and maintenance; all the accomplishment that has been gained, or the insight that has been gained over the past 20 years from severe accident research, and additional regulatory requirement that was either imposed by NRC, or voluntarily implemented by the utility that improved plant operation and performance. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
98 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 If you look at all the study, you find that there were many driven, particularly the early fatality, mainly driven by scenario that lead into early containment failure. And if you look at our
history of dealing with this issue, we have resolved the main two or three issues that deal with early containment failure for pressurized water reactor containment heating. low probability. We're concluding that's a very
Alpha mode failure, we concluded
that is low probability, and physically impossible. So if you look at it from a phenomenological point of view, we have eliminated all the early containment failure. By that, that by itself, you can conclude
that there will be no early fatalities, because there are enough time to allow for the evacuation and implement emergency preparedness. So we are --
although, the analysis, we are still doing the sensitivity analysis right now, we're confident that the result at the end, there will be no early
fatalities. And I venture to say that even for latent cancer fatality, the result will be significantly improved over previous analysis. With that, I would like to ask Bob Prato to start the discussion. MR. PRATO: Thank you. Good afternoon. I'm Bob
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
99 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Prato. I'm the Program Manager for SOARCA, and I want to thank you for this opportunity. As Farouk
mentioned, the last time we were in front of the Full Committee was more than a year ago, and a lot of work has been accomplished over this past year and a half. And the team is looking forward to your feedback. As we discussed with the Subcommittee we are basically going to be covering process. requested by the a we Subcommittee, high-level began the we do But as have of a a
demonstration, sequence that
demonstration SOARCA process
with,
exercising the process itself. If you turn to the agenda, we're going to start with a project overview, and one of the slides for the project overview is a full diagram of the process, and we're going to get into great detail for each one of the boxes in that flow diagram. going to cover system accident states, sequence mitigative So we're
selection, measures, and peer
containment MELCOR,
MACCS2,
emergency
preparedness,
review in relatively good detail, hopefully to give the Full Committee an understanding of the SOARCA process, itself. And then we're going to cover a
sample sequence, and we're going to update you on the status of reporting latent cancer fatalities. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
100 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 accurate. MR. PRATO: that specifically? CHAIRMAN SHACK: Just you are still moving ahead now with the study on Sequoia as the next step NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
The SOARCA objective; it's really twofold. We are developing a State-of-the-Art, more evaluation and of progression, radiological frequency And we are of
realistic release,
off-site
consequences
for
dominated core damage accident sequences. going to provide a more accurate
assessment
potential off-site consequences to replace previous consequence analysis, such as NUREG-2239, which is entitled, "The Technical Guide for Citing Criteria Development", which was issued more than 25 years ago, in November of 1982. That cite is more commonly
referred to as the Sandia Siting Study. MEMBER APOSTOLAKIS: first objective. MR. PRATO: Yes, sir, as you requested. Suggested. You did change the
MEMBER CORRADINI: MR. PRATO:
Suggested. Suggested.
MEMBER APOSTOLAKIS: MR. PRATO:
And is it more accurate? I think it's more
MEMBER APOSTOLAKIS:
Anybody have any questions on
101 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 and again. MR. PRATO: As you well know, Charlie Charlie has the it. CHAIRMAN SHACK: We keep asking it again effort. in the process? MR. PRATO: We are going to be contacting They have agreed
-- we've contacted them previously. to volunteer.
They went into a refueling outage, so
we are going to contact them again next week and set up a schedule, and we plan to start somewhere in the February time frame, interacting with them. MR. ELTAWILA: Okay?
Again, it's a voluntary
We don't know if they are going to -- so we
are negotiating with different utilities, too. MR. PRATO: MEMBER Severe accident -- sir? I guess we have
APOSTOLAKIS:
discussed this ad nauseam, but why aren't you doing a Level 3 PRA? Is there a short answer for that? MR. PRATO: There isn't a short answer to
presented the staff's view on that.
most integrated knowledge of all the pieces. Charlie, unfortunately, is not here, but he did make several key points. MEMBER APOSTOLAKIS: What's his last name? MR. PRATO: Tinkler.
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
102 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MEMBER APOSTOLAKIS: MR. PRATO: Okay? Yes, we know him. The key points, and We believe that the
I'll cover the key points first.
Level 1 PRA has done an outstanding job at this point of identifying what is important with regards to sequences, both from a CDF perspective, and from a LERF perspective. Second, and one of the underlying
premises of the project is that the Level 2 and Level 3 deserve more attention, and more rigorous
quantification. It is also our view that the use of an integrated method, such as MELCOR and MACCS, together with an uncertainty analysis, was a better approach for this application, versus trying to quantify
thousands of sequences, and it would help to shed some insights on risk. The other thing is, is that with MACCS and MELCOR, if there is a problem with the analysis, we can attack the particular model in a more direct manner. no And in Charlie's words, "the information is in a sea of numbers for which it is
buried
difficult to extract this kind of information." So why are we using CDS as our screening criteria? Well, from the start, there is a historical emphasis at the NRC on CDS, as well as an abundance of NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
103 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 information on CDS. We have our updated benchmark
SPAR models as an internal source for CDF information. And, remember, we have a high confidence in the Level 1 PRAs, as well, so -CHAIRMAN SHACK: MR. PRATO: Internal events.
Excuse me? Internal events.
CHAIRMAN SHACK: MR. PRATO:
For internal events, correct. At full power.
CHAIRMAN SHACK: MR. PRATO:
Yes, sir. Too many.
MEMBER APOSTOLAKIS: MR. PRATO:
In addition, the NRC uses CDF
as its criteria for risk-significance in Reg Guide 1.174. This Reg Guide uses a CDF of 10 to the minus We use the same
6, and a LERF of 10 to the minus 7.
Reg Guide 1.174 criteria for CDF, and if you believe that the conditional containment failure probability is approximately 0.1, then we meet the criteria for LERF, as well. And, therefore, we captured the risk
significance based on that criteria. The only other question remaining is, are we capturing all the significant contributors to LERF by using CDF, as opposed to using LERF. PWRs, there really shouldn't be any Again, for significant
dispute that early conditional containment failure NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
104 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 confidence. MR. PRATO: Mike Yerokun. Let me try to -- a direct plan right probabilities are less than or equal to .01. As for
BWRs, in other studies initial results for station blackout events indicated that vessel failure does not occur for more than eight hours into the event. And
the customary definition for early is four hours, so we believe that we're on the right track for BWRs, as well. And although it's site-specific and sequence-
specific, we are paying very close attention to the timing of the release, and we are making sure that it is beyond the early criteria. MEMBER APOSTOLAKIS: Now NUREG-1150 and
some other studies that are done by the industry, they did go all the way to Level 3. Have you compared what
you have found with the findings of those studies? MR. PRATO: Not yet, sir.
MEMBER APOSTOLAKIS: But you will do that? MR. PRATO: now, but I'm not sure if that's our the results are relatively
preliminary. MEMBER APOSTOLAKIS: It would add
MR. YEROKUN:
response to the question of do you plan to compare your results with NUREG-1150, I mean, yes, whatever NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
105 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Level 2. MR. ELTAWILA: I'm going to jump here and comes out from SOARCA, we definitely will see what insights we can derive compared to all the previous studies. Obviously, the Siting Studies, but also
NUREG-1150 to see what knowledge we gain from the approach we've used for SOARCA, and what that really means for the risk approach that was used for NUREG1150. MEMBER APOSTOLAKIS: MR. YEROKUN: Level 3 results.
I'm sorry? Well, even full scope
MEMBER STETKAR:
say I don't know what benefit we will gain out of comparing the SOARCA study with NUREG-1150 study. I
think we believe that these previous studies are very conservatively done, and did not represent the plants as operated, and design, and improvement that have been to the plants, so we will not be comparing apples with apples. I think that -- I appreciate your
question, but I will prefer to do a Level 3 for a plant and compare it to a SOARCA study, but to try to compare the SOARCA with the NUREG-1150, it's not going to be a viable comparison. MEMBER APOSTOLAKIS: It would be nice to
know why there are differences. If you find different NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
106 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 -- if I go to the Peach Bottom evaluation in 1150, they give me -- well, all five plants, actually. They give me fatality curves, latent cancer curves, and so on, and they give me the dominant contributors. mean, even if it's not part of your I
objective,
wouldn't you be curious to know whether your results are different? are different? And if they are different, why they You may come back and say because we
did a better job, but to say I'm not even going to look at it, it's kind of -- doesn't make sense to me. MR. PRATO: Well, there was one other
point Charlie wanted to make, or Charlie made at our last meeting. And he said, "With MELCOR, we do
believe that additional large benefit is derived in looking at mitigating measures that has not yet been addressed in PRA, such as SAMGs, and other severe accident mitigation guidelines." MEMBER APOSTOLAKIS: Well, you can always I'm not saying
say those things when you compare.
don't say it, but at least, I mean, tell us how the results are different. MR. PRATO: And I think as you see -- as
we go through the sample analysis, you'll see how it becomes measures, obvious all how the considering mitigative the mitigative have a
measures
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
107 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 significant benefit to the outcome. MR. CHEOP: try to address that. review process, as This is Mike Cheop. Let me
I think as part of the peer we are looking at accident
sequences, we do ask ourselves why are we different from, let's say, 1150. the reasons are. And if you're different, what
And we will convince ourselves what And as we go forth into the
the differences are.
Level 2 and Level 3 space, again, we do introduce a lot more, as Bob said, mitigative equipment. And we
can't explain a lot of the differences through the different strategies that we're using, and the
differences. We may not make a formal comparison, but we do, as part of the peer review, and our internal review process, try to convince ourselves as to what the differences are, and what's causing the
differences. MEMBER CORRADINI: clarification. MS. MITCHELL: This is Jocelyn Mitchell I just wanted to remind Could I just ask for
from the Office of Research.
you that the Level 2 part of 1150 was done using the EXOR codes, like the PBSOAR, and the SRSOAR code, where they took a tiny handful of source term code package runs and spread them out into hundreds of NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
108 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 sequences, and then subsequently collapsed into 17 for Surry, and I don't know the number for Peach Bottom, of release categories. I think it would be an
exercise in futility to try to go back and say here is this integrated MELCOR analysis, and why did it change from expanding, collapsing, and basing on just a few runs. MEMBER CORRADINI: Just to expand, I
guess, what George is saying, make sure I understand the staff's position. So I think my way of saying it
in some sense coming up with the same result that George is, if you took, and I'm going to pick Peach Bottom and Surry because they have an interesting historical, you can essentially take that and explain the differences. And I think that's kind of what I
get from George is after, is explain the evolution of your insights, both in terms of modeling, in terms of additional measures that have been taken care of, and you can go all the way from WASH-1400 through 1150, through - and I was going to ask something about that, through a current, if they had, or if they do have a Level 3, and really then show what you've done, both in terms of methodology, models, and improvements. And that, I think, would help drive home the
improvements that you have with SOARCA. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701
I guess
(202) 234-4433
109 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 that's the way I view -MEMBER APOSTOLAKIS: MEMBER CORRADINI: That's part of it. That's another
motivation to do it, as George is suggesting. MEMBER SIEBER: But how does that benefit
the overall science of what it is they're doing here? For example, methods has and codes have changed, have the
reliability changed,
data
changed,
assumptions And to make
scope
has
changed.
comparison, you're going to list a lot of changes. And it's not going to -- you aren't going to be able to draw a conclusion from it, other than this one is liberal, realistic this one is conservative, and not liberal, have
versus
conservative,
methods
changed over the years. MEMBER APOSTOLAKIS: questions that come to mind. Well, there are two
First of all, I don't Is this such a
even know why we're discussing this.
big effort that the staff is resisting, too many resources? It wouldn't look like that to me. But
second, and what you're saying is that the curves that I see in terms of public consequences in NUREG-1150, and other Level 3 PRAs, have been completely
invalidated, that this SOARCA thing now says don't believe any of that any more? NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
110 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MEMBER SIEBER: MEMBER BLEY: That's how I took it.
If that's true, I guess I'd
really want to understand why. MEMBER APOSTOLAKIS: want to understand why. Me too. I really
I mean, they have very nice They tell you what
curves there, kind of smooth. dominates.
It would be nice to say yes, we are
consistent with those guides, but we're doing a better job. CHAIRMAN SHACK: I mean, we do have this
bifurcation where we select what we examine by looking at frequency, and then we examine the risk-
significance of what's left.
Why don't we just look
at risk-significance in the first place? MEMBER APOSTOLAKIS: Right.
MEMBER STETKAR: Well, that has to do with ----- having gone through a few recent, not 25 years old, not 15 to 20 years old, but within the last 10 years, full scope Level 2 risk assessments sponsored by the industry, not the NRC, that have included things like SAMGs, that have concluded that the most important contributors to off-site releases, and I'll stop it there, because these were not Level 3 risk assessments, are Level 1 core damage sequences that are in the noise level for core damage frequency. The NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
111 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 most important contributors to off-site releases are sequences that are a small percentage, very small percentage in total of the core damage frequency. In other words, it probably wouldn't even make the pie chart when you look at contributors to core damage. However, that has been the result of
detailed analyses using reasonably refined Level 2 codes, taking credit for existing SAMGs, existing whatever you want to call them, beyond core damage operating procedures. And the concern, I think --
part of George's concern is, is the 10 to the minus 6 screening criterion basically missing most of the things that current studies, current industry-
sponsored studies show, indeed, are most important to the issues that, indeed, you're examining, the Level 2, Level 3-type issues. In other words, are you
missing those sequences by your screening process? MEMBER SIEBER: Well, if you divide it up
enough, you're going to miss a lot in the aggregate that are going to mean something. MEMBER APOSTOLAKIS: That is additional
evidence from that EPRI report that concluded that you go to very low frequencies in order to -CHAIRMAN SHACK: Well, even if you look in 1150 and you cut it off at 10 to the minus 6, not a NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
112 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 accident. MEMBER APOSTOLAKIS: So what you are doing is fine. You do have deaths. It sheds a lot of whole lot happens. MEMBER APOSTOLAKIS: It's an unusual
situation in the sense that you have -- say you went to go to some end state, and you have in-between some important release. stuff, like core damage frequency and
In traditional PRA when you say dominant
contributors to something at the end, you calculate that something, and then you identify the dominant contributors. Here we are using an intermediate
state, core damage, to identify what we call dominant contributors, and then we see what their consequences are. It's a little different thing, which is useful I'm not saying it's not
by itself, by the way. useful.
It's very useful, but the question remains,
why not go all the way, and bring into an additional dimension. When you communicate to the public now,
when you say there are zero deaths, what does that do to your credibility? CHAIRMAN SHACK: It's no worse than a bus
light into what can happen for these sequences, and so on. But it's a mystery to me why there is such I mean, are we talking
resistance to go all the way.
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
113 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 those lower about doubling the effort or what? MR. ELTAWILA: effort, George. I don't think it's an
I think it is just if you keep
refining the sequence further and further, and go down to a lower frequency event, I think that all what you're generating is number, but they don't mean anything. And because in most of these situations,
you know what you can -- how we can deal with these scenarios, so it always will become to an accident management and improvement in evacuation, and
improvement in the plant operation.
So by just going
down in the frequency domain to a very low frequency, yes, you can get an answer, yes, you can get -- that answer might show you that it's risk dominant, but what is the meaning of that? MEMBER STETKAR: frequencies, The meaning is that at may be initiating
there
events and consequential failures that also completely disable all of those mitigating systems, and operator actions that you're talking about. The nature of the
consequences changes at those very low frequencies. MR. ELTAWILA: minus 9, but are -MEMBER STETKAR: Or 10 to the minus 7. Very low frequency, 10 to
MR. ELTAWILA: No, I don't think you -- we NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
114 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 correct. looked at 10 to minus 7, and we -MEMBER STETKAR: In the context of the
limited models that you had to deal with, in the context of that, those limited models, but did not look at very clearly external events, seismic events, things like that. MR. CHEOP: Well, I think that's not quite I mean, in the 10 to the minus 7 range, we
did look at the external events, the seismic and the fire events, so in that sense, those are included in our 10 to the minus 7 look. And I guess in addressing
Dr. Sieber's earlier comment, we actually have not tried to parse out the sequences to such a point where we can eliminate them from the screening process. We
did try to keep groups together, so that we do not parse them out so that they are below the screening criteria, but I guess, to answer the broader question as to why we don't do a risk analysis versus a frequency dominant cutoff, that would be, I would imagine, changing the objective of the study. I mean, the objective of the study is to look at CDF dominant sequences, and that's the way they're going at this point. MEMBER CORRADINI: I mean, if I could just - if I might. But I understand you've been directed NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
115 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 wanted to. MR. ELTAWILA: We are going -- but if understand. say this? MEMBER CORRADINI: MEMBER Yes. The Commission a certain way. And I understand when one is directed, one follows the directions, but I'm asking -MEMBER APOSTOLAKIS: Did the Commission
APOSTOLAKIS:
directed you to look at -MEMBER CORRADINI: Yes. So that I
That's why I guess in some sense why I'm
just suggesting for you to consider the staff to think of it from the historical perspective, and use the fact of the historical perspective to at least lay the explanation out as to what you're seeing, and the insights you're getting. MR. ELTAWILA: Yes. If my answer at the
beginning sounds like we're not going to do that, I apologize for that. MEMBER CORRADINI: I sensed that you
anybody asked me to try to quantify every single differences, this will be impossible, but we will try to -- what are the plant improvement that led into the lowering of the frequency? What is the
phenomenological understanding that help us addressing NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
116 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 accident on. We this issue? That's part of the SOARCA report, so
there is no doubt about that. MEMBER SIEBER: I agree with Dr.
Apostolakis, that this has to be -- this comparison needs to be done to a certain extent because this will be an important public document, and a lot of
questions are going to be asked.
And for sure, they
will point out the differences, and if you aren't prepared to answer that in a public forum, then you haven't done the job right. CHAIRMAN SHACK: have other things I think we'd better move to cover, which will
undoubtedly lead to discussion. (Off the record comments.) MR. PRATO: We got past this question. I
think we can move forward in a reasonable -MEMBER SIEBER: MR. PRATO: improvement, Do every other slide. We're on slide, severe 4. Severe accident
Okay. slide
improvements that is, in part, the motivation behind SOARCA project, included improvements such as the 25 years, and literally millions of dollars that have been spent on national and international research that provides a better understanding of severe accidents, and the basis to conclude that some presumed early NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
117 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 contributed accident containment failure modes have been or shown other to be
phenomenologically accidents that are
not
feasible, but
severe been
feasible,
that
have
demonstrated to be preventable by accident mitigation. Second item, regulatory improvement that reduced the likelihood of severe accidents, rules such as ATWS, Station Blackout Rule, and the Maintenance Rule, all of these have contributed to improved modeling
accident
management,
improved
computer
capabilities, such as MACCS and MELCOR.
Keep in mind
that in 1982, when the 1982 study came out, there was nothing like MELCOR that was used or available at the time. And for MACCS, there was a much more primitive
model, and there has been significant improvements, not only in the modeling, but in the computer
technology that allows us to use computer modeling. Enhancements in plant design, such as the TMI initial modifications, and the modifications that continued beyond the post-TMI modifications during the early 80s and late 90s, things that resulted in the installation of additional emergency diesels, for example. Other to plant improvements the likelihood improvement that of in have severe plant
reducing
include
general
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
118 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 process. performance. In the early and late 80s, capacity
factors were in the 60s and the 70s, today they're in the 80s and the 90s. Emergency preparedness
guidelines are available. They've been developed, and evolved, and tested very frequently. And mitigative
measures, as you will see, will play a big role. This next slide is an overview of the We're going to cover each one of the boxes
in a lot more detail, but this just shows how the process flows, and how we come to -- how the SOARCA process works, in general. The SOARCA approach. SOARCA is the only
kind of accidents we're considering of full power operation. We are not considering low power, We
shutdown, or spent fuel pool-type of accidents.
are using a plant-specific sequence truncation of CDF of greater than or equal to 10 to the 6th, and a CDF greater than or equal to 10 to the 7th for bypass events. MEMBER CORRADINI: MR. PRATO: Yes, Minus you mean, right? minus. I'm sorry.
They're in there. external events.
I apologize.
We did consider
We considered all of the mitigative
measures that were available to the licensees. We did and we're doing sensitivity analysis to assess the NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
119 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 I'll wait. MR. PRATO: SOARCA insights. Okay. effectiveness of the different safety measures. A
State-of-the Art Accident Progression Modeling based on 25 years of research to provide a best estimate of accident progression, containment performance, time of release, and fission product behavior. We are using
a more realistic off-site dispersion model, and we are doing site-specific evaluation of public evacuation based on site-specific updated emergency plans. MEMBER CORRADINI: Can I repeat something
that you guys said in the question and answer earlier, just so I put it in this context for the second bullet? So even though your cutoff, as directed, was
that, you went down another order of magnitude and surveyed what you saw at the 10 to the minus 7 cutoff. MR. PRATO: We're going to get into that. Okay. show you how the
MEMBER CORRADINI: MR. PRATO:
We'll
sequence selection was -MEMBER CORRADINI: MR. PRATO: The answer is yes.
But the -Okay. That's fine.
MEMBER CORRADINI:
Sequences are dominated by external events, primarily large seismic events that play out similar to a NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
120 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 station blackout. We also identified two additional
PWR bypass events that were within the scope of SOARCA. Previously used sequences have
significantly lower probability of occurrence, or are not considered feasible, and that includes the alpha mode, the high pressure melt injection, and ATWS. The first two are considered not feasible, and the ATWS is a much lower CDF than was considered in 1982. Mitigative measures are proven to be
effective at preventing core damage or containment failure. MEMBER SIEBER: Did you examine the effect of large seismic events on the effectiveness of the emergency planning? MR. PRATO: No, sir, we have not. Bridges knocked down,
MEMBER SIEBER:
roads closed, flooding, whatever. MR. PRATO: We are doing sensitivity
analysis to address that, sir. Sequence screening process. Okay. It's
important to mention at this time that SOARCA was never intended to be a risk study. However, the staff wanted its initial focus for SOARCA to include
sequences of greatest interest. NEAL R. GROSS
(202) 234-4433
Therefore, as the
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701
(202) 234-4433
121 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 initial input into SOARCA, the staff used the enhanced SPAR model to identify the sequences that are most likely to occur. Using SPAR, we applied a screening
criteria for the sequences included within the scope of SOARCA to identify those sequence or sequence groupings that have a CDF of greater than or equal to 1.0 E to the minus 6 to identify those sequences which are most likely to occur. In addition, we wanted to pay more
attention to those sequences that are potentially more severe, but that have a little lower likelihood of occurring. For example, interface system LOCAs that Therefore, we lowered the
bypass the containment.
screening criteria for inter-system LOCAs to a CDF of greater than or equal to 1.0E to the minus 7. These are the steps that are used to implement the screening criteria. initial screening. We started with an
We used enhanced SPAR model to
screen out low CDF sequences with an overall CDF of less than or equal to 1.0E to the minus 7, and sequences with a CDF of less than 1.0E to the minus 8 for bypass events. This step we estimated eliminated
less than 10 percent of the overall CDF, approximately 5 percent is what it typically ended up being. CHAIRMAN SHACK: Typically for two cases.
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
122 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 evaluated MR. PRATO: For the two plants, correct.
The sequence evaluation, we identified and dominant cut sets for the remaining
sequences, and we determined systems and equipment availability, unavailability, and accident sequence common to those sequences. We grouped the sequences
together that had similar times to core damage, and similar equipment availability. bounding sequences based on And then we selected the most limiting
mitigative measures available. For external events, we performed limited reviews of existing external event studies, and data to identify dominant externally initiated event
sequences for each plant of interest.
And where
available, we specifically identified the dominant accident sequences for those plants using the
following steps. First, we identified dominant externally initiated event sequences for external events, such as fire, seismic, flooding, wind. And based upon
available probabilistic assessment documentation, like NUREG-1150, the IPEEE submittals, as well as any additional available supporting documentation. We did not use seismic margins assessment because it lacked the risk information necessary. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
123 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 slide. MR. PRATO: Oh, I'm sorry. The resulting We identified potential mapping between dominant external events, and internally initiated events identified by the SPAR analysis. Where mapping between external and internal events are not possible or appropriate, a unique external initiating event or sensitivity study was recommended, and the resulting limit -CHAIRMAN SHACK: You need to click your
limited set of scenarios obtained for each SOARCA plant was used for subsequent accident progression and consequence analysis. Containment system states. The objective of this process is to identify the availability of engineering systems that can impact post-core damage containment accident progression, containment failure, and radionuclide release using the following steps. We determined and the anticipated availability systems of not We
containment
containment
support
considered in the Level 1 core damage analysis.
did this by determining the availability of front line systems using cut set information. If all support
systems were considered in the Level 1 analysis, availability was determined based on the cut set NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
124 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 that. MR. PRATO: weren't doing that. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
information alone.
If they were not, we constructed
a system dependency table showing the support systems for performance of the targeted front line system. We then determined the availability of the front line system using engineering judgment. example, if the necessary support systems For were
determined to be available or unavailable based on engineering judgment, of the then the availability line system or was
unavailability determined.
front
MEMBER APOSTOLAKIS: bit misunderstood. When you
This may be a little say determine the
availability, I believe what the study did was assume that the system was working or not. MR. PRATO: Right?
That's correct, sir.
MEMBER APOSTOLAKIS: Depending on what has been lost. MR. PRATO: That's correct, sir. In PRA space, the
MEMBER APOSTOLAKIS:
availability will be a probability. MR. PRATO: That's correct. And you didn't do
MEMBER APOSTOLAKIS:
That's correct, because we
125 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MEMBER APOSTOLAKIS: Okay. So that's
another thing that's different from traditional PRAs. The containment systems are either there or they are not, and you have some logical criteria to decide that. MR. PRATO: That's correct. But there is always
MEMBER APOSTOLAKIS:
a possibility of a random failure, or whatever the PRA does to come up with the unavailability number. MR. PRATO: We did not do an HRA. They didn't do that.
MEMBER APOSTOLAKIS: I mean, that's -MEMBER SIEBER: MEMBER
That may be more -Determine the
APOSTOLAKIS:
availability can be interpreted in different ways. Okay. That's just a clarification. CHAIRMAN SHACK: But, again, their
argument is their additional random failures would be lowering the frequency. MR. PRATO: That's correct. Yes. This cutoff is
MEMBER APOSTOLAKIS: always running our lives here. MR. PRATO: Okay.
In addition -Wait a minute now.
MEMBER APOSTOLAKIS:
The cutoff is for core damage frequencies. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
126 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 sequence. CHAIRMAN SHACK: But it's an overall
The cutoff is really on releases, and they
-- well, that's the directive of the SECY. MEMBER APOSTOLAKIS: MR. PRATO: Wait a minute now.
It was release frequency. Release frequency.
CHAIRMAN SHACK: MR. PRATO:
10 to the minus -Right. And the staff
CHAIRMAN SHACK:
took a conservative approach by going to the -MR. PRATO: Core damage. -- core damage, because,
CHAIRMAN SHACK:
again, you're not going to get a release without core damage. MEMBER APOSTOLAKIS: That's what was done. CHAIRMAN SHACK: MR. PRATO: That's what was done. In addition, the
Okay.
availability of containment systems determine the availability of systems such as the low pressure injection, and that can potentially impact containment accident progression. For example, cooling debris in
the reactor cavity, or cooling reactor vessel after the core damage, or prior to vessel failure. Those
are the systems we also considered for containment system states, as well. Mitigative measures analysis. The
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
127 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 mitigative measure analysis that we performed were qualitative sequence-specific system and operational analysis based on licensee identified mitigative
measures from EOPs, SAMGs, and other severe accident guidelines that were determined to be applicable to and available during a specific sequence, whose
availability, capability, and timing were utilized as inputs into the MELCOR analysis. CHAIRMAN SHACK: What does it mean by the
qualitative part, since you really used these to set the boundary conditions for your MELCOR analysis, as I understand. MR. PRATO: In other words, we didn't That
quantify it, and we didn't assign a risk to it. was the -CHAIRMAN SHACK: that it would be done. MR. PRATO: That's correct, sir.
You mean a probability
CHAIRMAN SHACK: the procedure -MR. PRATO:
You assumed if it was in
Well, we did more than that. We
We verified that the equipment was available.
verified that there was no reason to believe that it was not accessible. We insured that we took
consideration for communications, resources. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701
We did
(202) 234-4433
128 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 correct. The process that we used to do this, for those dominant sequence or sequence groups within the scope of SOARCA, we determined the potential a very extensive -CHAIRMAN SHACK: assign numbers to those. MR. PRATO: That's correct. That's Okay. You just didn't
availability of mitigative measures.
We performed a
system and operational analysis based on the initial condition, and the anticipated subsequent failures. We determined the anticipated availability, And we put
capability, and time to implementation.
all of that information into MELCOR to determine the effectiveness of those mitigative measures. We never, as part of the mitigative measures, assessed its effectiveness. of the modeling. MEMBER BLEY: Let me back you up to that So you, We let MELCOR determine that as part
last question, just make sure I'm following.
essentially, or as you said earlier, you did no HRA. You, essentially, said if the equipment this
scenario, if the equipment is there that could work, and the procedures would make it work. We'll say it You've
works and put it into the MELCOR analysis. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701
(202) 234-4433
129 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 essentially assumed people perform perfectly given the equipment's available. MR. PRATO: In general. However, we
assign very conservative times to the implementation. We did try to consider accessibility, availability in not only support systems, but support equipment that was needed. Was it on hand, was it pre-staged? We
looked at a lot of the parameters that you would consider in an HRA. MEMBER BLEY: If there's time to do it,
and if the equipment works, it will be used and it will work. MR. PRATO: That's correct. We looked at
availability of capacity.
Are the storage tanks
available, the bottom of the storage tanks. MR. DUBE: Don Dube, NRO. I want to just
add that for most of these sequences, the MELCOR analysis was done with and without the mitigative measure. Right? MR. PRATO: That's correct. We are doing
sensitivity analysis both with and without -MR. DUBE: I know probability was
assigned, and it failed to -MEMBER BLEY: MR. PRATO: But we have both results. Yes, sir.
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
130 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 the MEMBER BLEY: Before you got here, I want When
to understand the statement you made earlier.
you did the basic scenario, did you look at -- you didn't do an HRA there, either. Is that right?
Effectively, the same thing, if the equipment is there, you assume the equipment will -MR. PRATO: MR. CHEOP: analysis, we did CDF you include -To get to the CDF portion of HRA as part of the CDF
calculation. MEMBER BLEY: Okay. What there isn't, isn't
MEMBER STETKAR:
a conditional dependent HRA for the Level 2, Level 3 mitigative functions. They're -- HRA ends at Level 1, basically. MR. PRATO: The structural analysis. The
objective of the structural analysis was to evaluate the behavior of containment structure under
unmitigated severe accident conditions and to predict the following criteria; and that is, functional
failure due to pressure, the structural failure due to pressure, and to develop leak rates, and leak areas as a function of internal pressure. As a result of the structural analysis, the dominant cause for containment failure at Peach NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
131 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Bottom is the -- we determined to be the head flange bolts strained under gradual increasing internal
pressure.
And for Surry, it was cracking around the
equipment and personnel hatch. MELCOR analysis. As stated earlier,
MELCOR or a similar model wasn't available in 1982. MELCOR, since that time, has been developed, and it has evolved, and we have implemented significant
improvements to get it to where it is today, which we believe is the state-of-the-art. More recently, and
specific to SOARCA, we improved the MACCS output interface. We implemented fuel collapse model logic.
We updated MELCOR defaults, and we added approved model. For each of the analysis that have been completed, we developed a site-specific model, and we performed the accident progression for each plant using MELCOR computer code to determine source term, potential containment failure states, and time of release as an input into the MACCS analysis. Similar with MACCS-2, MACCS was around in 1982. It has evolved significantly. For the purposes of SOARCA, we implemented a significant number of improvements to bring it up-to-date, and to make it state-of-the-art. I'll go over just a couple of NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
132 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 previous these. We increased the number of evacuation cohorts
from three cohorts, which was previously, up to 20. We have the capability of dividing up into 20
different cohorts.
We increased angular resolution
from the typical 16 compass points up to 64 segments. We added more plume segments, and we included KI ingestion model, as well. couple. MEMBER BLEY: say "more plume segments"? MEMBER SIEBER: MR. PRATO: What do you mean when you Can you describe that? It is finer. And these are just a
Jocelyn, do you want to --
MEMBER BLEY: Are you letting them move as the weather changes? MR. PRATO: No.
MS. MITCHELL: We usually break it up. In analyses, there were usually one puff
release, and then a long tail, so you had two plume segments. And now we take this very long, drawn out
release, and break it up into typically one-hour releases, and MACCS will pick up for the release of the second plume. If the weather has changed, the
wind speed has changed, then it will pick up a new weather sequence, so it's -MEMBER BLEY: Opens up then.
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
133 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MS. MITCHELL: Yes, right. Right. But it is a one-hour, typically one-hour releases. MR. PRATO: We performed the consequence
analysis for each plant and each sequence using the MACCS-2 computer code to determine early fatalities and latent cancer effects. Some of the MACCS-2 assumptions that we used, we assumed that no contaminated food or water would be consumed. dose 12 We used the latest in federal Federal to
guideline Guidelines
conversion and 13 for
factors specific
isotopes,
specific organs, given different specific pathways. We assumed KI ingestion by half of the 10mile population, and we used sub-optimal timing. That sub-optimal timing results in a fraction efficacy and if you do the KI ingestion at just the right time, it's approximately in the mid-090s range. for the efficacization. We used medium values from the U.S. and European study for uncertainty for non-site-specific parameters. This study used expert elicitation for We used 70
approximately 140 different points important to offsite calculations, and resulted in the distribution of the response for each of those inputs. And we used
the mean as the input into MACCS for these things. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
134 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 wasn't it? MR. PRATO: MEMBER BLEY: MS. MITCHELL: Yes, sir. Oh. This is Jocelyn Mitchell. that? MR. PRATO: That was a separate study done for not only our modeling applications, but for the Europeans, and the Asians, as well. It was a And it includes things, such as the dry deposition velocity, the wet deposition velocity, some of the parameters in the food chain model, a set for those type of parameters are included in those 140 different inputs. MEMBER BLEY: Who were your experts for
combination of U.S. and -MEMBER APOSTOLAKIS: It was years ago,
The study was done about 10 years ago, and had six different panels that worked on different disciplines that were necessary, and they had usually eight
experts on the panel, four from the U.S., and four from the EC countries. MEMBER BLEY: George just reminded me.
This is the one using Roger Cook's approach. MEMBER APOSTOLAKIS: MEMBER BLEY: Yes.
Where it calibrates the
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
135 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 experts. MS. MITCHELL: The expert elicitation
itself was conducted by the U.S., and the European part of it was to take the discrepant distributions that come out of it, and evaluate them. We have
redone that, because we found some problems with how it was done, and so we have re-sampled it, and have a distribution that encompasses the eight expert views, and we take the median, the 50th percentile from that resulting distribution. MR. PRATO: MEMBER Okay? Just a quick question,
BLEY:
because two or three slides earlier you talked about -- you ran -- I thought I saw uncertainties on the parameters, but I'm not so sure I'm hearing that any more. You looked at the uncertainties, and then you
picked mean values or medians, and ran them through the -MS. MITCHELL: The difference is that the
MACCS code itself now has a user-friendly front end that enables a relatively easy parameter uncertainty consideration, so you could put in a range of values, and a degree of belief, and it would sample out of that range and degree decks, of run belief, them and construct and
multiple
MACCS
sequentially,
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
136 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 evaluate the grand averages when you get finished. That model is available for what -- and we intend to exercise it, but we have not, as yet. We intend to do
it, but for a point estimate for the stuff that we have done to-date, we have taken that range of values and degrees of belief, and have taken the 50th
percentile, and put it in as our point estimate. CHAIRMAN SHACK: But you do the
meteorological stuff statistically, and take a mean value. MS. MITCHELL: The meteorological stuff is sampled in a stratified random sampling method. This
would be for other things, like the dry deposition velocity, and a whole bunch of other things. MR. PRATO: For each site, we use siteWe use an
specific population meteorological data.
assumed projected relocation dose, and time for the area beyond the evacuation zone during the seven-day emergency period of 5 rem and one day for relocation, and for 2 rem, two days for relocation. Return
criteria at Peach Bottom we use .5 rem, which is EPAspecific, and for Surry we used 4 rem and five years, which is EPA-specific. In general, releases are
divided into one-hour plumes, as Jocelyn explained previously. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
137 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 the answer? MR. PRATO: Just a high level discussion. population, Emergency preparedness. We modeled the
protective response afforded by current site-specific emergency preparedness programs to improve realism. We used site-specific evacuation time, time estimates for evacuation of the EPZ. We used a new program
called OREMs, which is Oak Ridge Evacuation Model, to model evacuation of the 10 to 20 mile area. We modeled cohort time, data, such as and
evacuation
travel
speeds,
roadway networks. And the data was used in MACCS-2 to develop consequence estimates. MEMBER uncertainties. MR. PRATO: MEMBER Excuse me, sir? Were there any APOSTOLAKIS: Peer review. So there was no
APOSTOLAKIS:
uncertainties in these evaluations? MR. PRATO: That's coming up, sir. Coming up.
MEMBER APOSTOLAKIS: MR. PRATO: Yes, sir.
MEMBER APOSTOLAKIS: MR. PRATO: Okay.
Good.
MEMBER APOSTOLAKIS:
The discussion, or
We plan to do an uncertainty analysis, and a peer NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
138 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 this? MR. PRATO: No, we're working on it. We couple of review in the next calendar year. Peer reviews. staff peer Okay?
Internal, we've done a including the PRA
reviews,
aspects that's used within SOARCA.
We have recently
went to ACNW and given them an overview of SOARCA, as well as discussed the dose threshold issue. And we've been here a number of times. MEMBER APOSTOLAKIS: and international experts? names? MR. PRATO: now as we speak. MEMBER APOSTOLAKIS: Oh, you haven't done We're working on that right Who are the national
Can you give us a few
plan to do it in the next calendar year. MEMBER SIEBER: We are here.
MR. PRATO: For Peach Bottom, Peach Bottom accident sequences. The PRA model -But, you see, when
MEMBER APOSTOLAKIS:
you select the experts, I mean, an expert can come in there. He's an expert say on MACCS. He will look at
what you've done. Would any of
He'd probably say it's very good. experts dare question your I
these
objectives, the same way this committee is doing? NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701
(202) 234-4433
139 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 of it. CHAIRMAN SHACK: Onward. don't know. to answer. MR. PRATO: CHAIRMAN I have no answer. SHACK: Many of George's I doubt it. It's okay. You don't have
questions you don't have to answer. MEMBER SIEBER: MEMBER POWERS: We are available. But all of George's
questions should be appropriately considered. MEMBER APOSTOLAKIS: It's one of the rare
occasions where the fact that I can't hear you is good. (Laughter.) CHAIRMAN SHACK: you said about him, George. MEMBER repeat it, either. MEMBER SIEBER: But George didn't hear any CORRADINI: He's not going to That's the nicest thing
MR. PRATO: The PRA models indicate a core damage frequency is dominated by seismic events, which functionally work this way out as a long-term station blackout. We did consider fire and flooding, as well, but when we looked at the general damage, the general availability of other systems to be able to mitigate NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
140 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 it, the seismic event had much more widespread damage, and we determined that the seismic -- we evaluated the seismic event, we would bound the other two. MEMBER BLEY: From your statement there, It's
it's functionally a long-term station blackout.
an earthquake big enough to cause that, but not to damage equipment in the plant? MR. PRATO: The only thing we considered,
we assumed that was undamaged was containment and the RCS, initially. Okay? If we had questions about
mitigative measures, we did an additional evaluation to determine if the seismic event would result in that, and damage of that equipment. MEMBER STETKAR: Let me follow-up on it,
because I was in the Subcommittee meeting, and this was one -- this follows up on something Jack brought up earlier, and something you brought up earlier. We
don't have the details of this sequence, but it was described in the Subcommittee meeting. First of all,
this frequency seismic event corresponds to something, I think I remember something in the 1g acceleration rate, so this is a 7 to 8 magnitude on the Richter scale earthquake. The analysis of this scenario from the HRA perspective takes full credit for operators manually, NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
141 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 locally, mechanically controlling RCIC flow at the RCIC turbine. Think of that. And it does not include possible seismic effects on population evacuation, emergency response. Is that correct, both of those? Correct. Okay. Now on the HRA
MR. PRATO:
MEMBER STETKAR:
perspective, it seems somewhat optimistic to say that the operator, under these conditions -MEMBER BLEY: It does not take advantage
of what people have -- unusual behavior we've seen in people under very, very large earthquakes. And for
these people, that are very, very large earthquakes. MR. PRATO: Can I caution us not to get
into the details of any of the results from this meeting? It's too preliminary. MEMBER BLEY: you model people? MR. PRATO: MEMBER BLEY: Do we model people, sir? The thing John was pointing Okay. Not results, but do
out is, if one did an HRA of human performance after this earthquake, one had better understand the
psychological impact of such an earthquake on people. It's not something you can look up in THERP, or SPARH, or any of the existing HRA methods. he's saying. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
That's what
142 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 on. MR. PRATO: Internal events were all less any more. MEMBER APOSTOLAKIS: The impact of the MR. PRATO: Understand. I think that point was
MEMBER CORRADINI: made at the Subcommittee. MEMBER SIEBER:
Not only the operators, They
but also the people who live around the plant. go --
MEMBER BLEY: They're not living in houses
earthquake on the evacuation itself. MEMBER SIEBER: Yes, bridges down, holes.
MEMBER APOSTOLAKIS: Not only just people, but also damage to bridges. MEMBER BLEY: And I think I saw something
about half the people taking -CHAIRMAN SHACK: I think we better move
than 10 to the minus 6, and bypass events were very, very low frequency, much less than 10 to the minus 7. For Surry, the events included a long-term and a short-term station blackout initiated by a large seismic event. And in addition to that, we had an
inter-system LOCA, a LOCA on a low pressure injection system, and we had a steam generator tube rupture. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
143 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 SBO events are due to seismic, flood, and fire
initiators, and are modeled as seismic events, again for the same reason, because there was much more widespread damage, and a lot more equipment that was unavailable. MEMBER CORRADINI: So in a similar fashion as you said in Peach Bottom, the seismic encompasses what might have been a fire or flood event? MR. PRATO: That's correct. It bounds it. The IS LOCA and steam generator tube rupture are due to random equipment failure, and then by a number of operator errors. This is a sample of sequences high level, but hopefully it will give you a feel for how this process was implemented. loss of a vital AC bus. This sample sequence is a This sequence was selected
and assessed for demonstration purposes only. MEMBER BLEY: last two view graphs? Can I reflect back on your
If I got it right, in general,
we're saying we're using, although you're not doing it here, a cutoff of 10 to the minus 6 per year on sequences, and yet our dominant sequence is about 10 to the minus 6 per year. MR. PRATO: MEMBER BLEY: Is that right? That's correct. Have I got the story right?
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
144 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 that claim go ahead. MR. SHERRY: your question. I guess I don't understand it? MR. SHERRY: MEMBER BLEY: Yes. That seems troublesome, but MR. PRATO: Rich, that's correct, isn't
In what sense were you referring -I understand his point. If the biggest thing there
MR. PRATO: MEMBER BLEY:
is is about one times 10 to the minus 6, and you don't look at anything else, it makes me uncomfortable that there might be other things that would add up to substantially more than that. MR. SHERRY: In the external events, or
internal events, or across the board? MEMBER BLEY: MR. SHERRY: Any events. I think I can say fairly
confidently for the internal events, that we probably captured the risk dominant sequences for the PWR for Surry. Okay? And I suspect -- well, I really can't
make that statement for Peach Bottom. For external events, it's harder to make because we, essentially, obtained our
sequences by looking at older studies.
We really
didn't use up-to-date -- we didn't have up-to-date NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
145 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 seismic PRA or other external event PRA results for these plants, so we essentially looked at past
studies, like NUREG-1150, whatever information we could get from the IPEEEs, and made judgments about what the dominant sequence characteristics would be. And, typically, station blackout was a typical
frequency dominant sequence for seismic events. CHAIRMAN SHACK: SPAR external events models. MR. SHERRY: There's references to Just what are they?
For a limited number of
plants, there have been a number of SPAR models which have been upgraded to include, to a limited extent, external events. Seismic -Does that include Surry
CHAIRMAN SHACK: and Peach Bottom? MR. SHERRY: Yes.
CHAIRMAN SHACK: PRA of some sort. MR. SHERRY:
So you do have a seismic
Except that these haven't Okay?
really been validated in any sense. MR. PRATO: MEMBER BLEY: MR. PRATO: Okay? Go ahead. As I
said,
sequence
was
selected and assessed for demonstration purposes as not within the scope of SOARCA, because the CDF is NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
146 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 purposes. process, this is? MR. PRATO: It's just for demonstration this. really less than 10 to the minus 6. The MELCOR
analysis showed that this event can be mitigated. Okay? So even though the sequence indicated that core damage can be achieved, we were able to mitigate this event. MEMBER APOSTOLAKIS: I don't understand
You're showing us an analysis of a sequence
that should have been screened out? MR. PRATO: Correct, sir. And the purpose of
MEMBER APOSTOLAKIS:
When we first started up SOARCA in the and working with MELCOR, we selected a
sequence, and we performed some analysis on it.
And
we kept it because we felt that it had demonstration value. MEMBER APOSTOLAKIS: But this is not going to convince us that -MR. PRATO: It's not intended to try to Okay? Well said. Nor could it ever.
convince you of anything, sir. MEMBER CORRADINI: MEMBER STETKAR:
(Laughter.) MEMBER APOSTOLAKIS: NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
Very well put, sir.
147 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 off. MR. YEROKUN: If I may try to fend that
For the purpose of this discussion, we said
earlier we're not prepared to discuss the preliminary results of this, sorry. So we've gone through the By the way,
process, and all these technical steps. my name is Jimi Yerokun.
But for the benefit of the
Full Committee, we thought it would be appropriate to at least give you some demonstration of how this process works through, some hypothetical sequence, and go through the whole analysis. example where it's not within And this is one the scope of the
analysis, but it portrays how we step through using some other sequences. MEMBER APOSTOLAKIS: But you have produced results of this sequence? MR. PRATO: Yes, sir. And you're not going
MEMBER APOSTOLAKIS: to show them? MR. PRATO: what the outcome is.
Well, we show you generally I'm going to go through that,
but we don't have the graphs, and the slides, and everything that went with it. No, sir. Okay.
MEMBER APOSTOLAKIS: MR. PRATO: Okay?
MEMBER APOSTOLAKIS: NEAL R. GROSS
(202) 234-4433
All right.
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701
(202) 234-4433
148 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MR. PRATO: The initiator was a loss of
Division IV DC power resulting in a scram, MSIV closure, started, and and containment one CRD isolation, was a RCIC and autostill
pump
active
available.
The initial operator actions, the load
shed to maximize duration of the DC power, they maximize the flow of the single CRD pump. The CRD
pump at this facility has a range of 110 to 180 gallons per minute depending on RCS pressure, but that 110, there is some throttling involved, so they go down and they open it up, and that 110 basically turns into 140 at normal RCS pressure. They depressurized the RCS in about an hour and a half, which with the CRD and RCIC flow they had to secure the CRD from four to seven hours to prevent reactor pressure vessel over-fill. The
capacity of make-up was sufficient to prevent core damage, even though core damage was predicted by the Level 1 PRA. MEMBER BLEY: I take it this plant's Level 1 PRA did not take advantage of the CRD pumps? MR. PRATO: the next slide. MEMBER BLEY: MR. PRATO: Oh, sorry. Sufficient injection Sir, I'm -- that's right on
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
149 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 actions. capability, there was sufficient injection capability to prevent core damage. CRD for coolant makeup. The SPAR didn't credit the Reactor pressure vessel
depressurization, and maximization of CRD flow are important to operator actions to optimize recovery. One other thing that we didn't consider, that wasn't considered was standby liquid control, was also
available for high pressure injection at about 50 gallons per minute. And battery duration was
determined to be important for RCIC operation and instrumentation. MEMBER BLEY: I'm sorry for so many. Let me interrupt you again.
You thanked the utilities for
cooperating. Did they cooperate in the performance of the analysis, or in allowing their plant and their PRA to be used? MR. PRATO: They were involved in the
performance analysis. We did make a site visit. They did review our sequence truncation and verified it, and we had some exchanges on that. two, and they've taken one away. MEMBER BLEY: This one had a lot of human We added one or
Were they involved in the quantification of
the human actions? MR. PRATO: They were involved in the
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
150 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 trees, and mitigative measures analysis. input and insights. MEMBER BLEY: I mean, on things like They provided a lot of
whatever you came up with on -- or is that as shedding load and that sort of thing on the DC part of what you're talking about? MR. PRATO: We had a lot of procedures for this facility, and we, basically, assessed -- yes. MEMBER BLEY: And I know you said this For the Level 1 I mean,
earlier, I just want to confirm. PRA, you used the SPAR PRA.
Is that right?
you used that as the basis for coming up with your scenario. MR. PRATO: MEMBER BLEY: That's correct. Okay. Do those now really
have a thorough coupling of all the dependencies among systems in those models? I know not too long ago they did not. MR. CHEOP: event We have the full set of fault to couple all different
trees
dependent failures, and human errors. MEMBER BLEY: So that is pretty thorough
and the utilities will have -MR. CHEOP: MEMBER BLEY: That's correct. Okay. Thanks.
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
151 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 MR. PRATO: Okay. And that brings us to
the conclusion on the sequence sample. The only other thing left is the report on the status of the latent cancer fatality reporting. The history behind this is one of the key and in objectives for the of SOARCA is risk risk to
communication, communication
non-NRC does
staff, not
this
context
refer
communicating PRA information. It's a methodology and of
It's more general. very a technical unified
reporting up
information,
coming
with
understanding of the information, coming up with a single outcome of the information. And, initially, we proposed to do a range of doses, and we felt that a range of doses conflicted with the risk communication aspect of SOARCA, which is one of our more important criteria, so what we decided to do is we decided to take a look at other options for reporting latent cancer fatalities. And we're putting together a
Commission paper to inform the Commission, and right now what you see here is we have three options. Very
recently, we had other considerations thrown at us. The three options we are considering on this slide is range of threshold, linear no-threshold, or an
estimated point value based on a Health Physics paper. We are also looking at other things, like just
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
152 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 anything. ongoing. reporting dose. We are looking at potentially
truncating distance.
There's a number, a variety of
variations, and our overall objective is to come up with one single answer for each sequence at each site, instead of coming up with multiple. MEMBER conservative. Right? Yes. This process is still SIEBER: These are very
MR. PRATO:
We believe that it's going to come to
conclusion relatively in the near future, but right now, we're still in the final development stage of the Commission paper. MEMBER SIEBER: I was thinking if these
are realistic, I'm in trouble. MR. PRATO: Excuse me, sir? If these are realistic,
MEMBER SIEBER:
I'm in trouble after of SCAT scans. MR. PRATO: Me too. Me too. That
completes the presentation. MEMBER BLEY:
Any other questions?
As you left that last one,
what kind of a point value are you leaning toward? MR. PRATO: We are not leaning towards
The staff is working out the options, and And we may
we will probably evaluate the options.
make a recommendation, we may let the Commission NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
153 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 you. CHAIRMAN SHACK: Thank you very much. Mr. Lymen, I think you wanted to make some remarks. MR. LYMEN: Yes, if I may. decide. The final form of this Commission paper has
not been decided yet. MEMBER BLEY: Have you got a
recommendation from the other Committee? MR. PRATO: MEMBER BLEY: MR. PRATO: formal recommendation. The ACNW? Yes. The ACNW did not give us a They asked us to consider We are
dose, and they asked us to consider risk. looking at them as potential options. MEMBER BLEY: Thank you.
CHAIRMAN SHACK:
Any more questions?
MR. PRATO: Any questions, anybody? Thank
(Off the record comments.) MR. LYMEN: Can you hear me? I
appreciate, as usual, the opportunity to make a few remarks here. And, actually, when I walked in and
heard the opening discussion, I was wondering if it was even necessary, because I think many of the comments I heard from the Committee already reflect a lot of our concerns. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
154 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 I'd say the bottom line is that my
organization, the Union of Concerned Scientists, is very supportive of an authoritative and independent study that would approve the technical credibility and accuracy of analyses of consequences of severe
accidents. And two issues, in particular, is one, can protective actions be improved based on better
information, and can better siting decisions be made in the event that new reactors are actually located around the country, which is going to become
increasingly important.
But our view, at this point,
is that SOARCA is not on track to fulfill this role, and that's because the political goals from the outset threaten to overwhelm the technical part. What is the real point of SOARCA, it seems to be knocking down a strawman, which was the 1982 CRAC2 study. And as you already discussed, a lot of
the public rationale for SOARCA seems to ignore the fact that NRC staff, and contractors, and licensees have been severe accident analyses for more than 20 years in the interim, including NUREG-1150, source term is based and on MELCOR, at least used its in initial
formulation, applications.
are
commonly
regulatory
So there is a whole body of work since
then, and to say we haven't done anything since 1982 NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
155 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 really makes -- exaggerates the importance of this project. It's really an incremental improvement over
what's already been done. And looking at some of the statements that have been made about it, the SOARCA project may show that a large early release may not credibly exist, for example, raises the concern that the real point is to rehabilitate severe accident analysis to eliminate the most risk-significant sequences on the basis that they're low probability. Now one concern we have is the
inappropriate focus on risk communication.
That has
been part of this project, in the forefront from the beginning. Results will be presented documented risk techniques to achieve public
communication
understanding, which is a little Orwellian in my view. The fact is, a risk communication plan has already been developed for this project years before the study's results are even going to be available, and that raises suspicions if the PR aspects of this project are predominant. And we say really, "Just the facts, ma'am." If you really want to achieve public
understanding, the best way to be clear about all your assumptions and arguments in a step-wise fashion, so that the impacts of the various changes to previous NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
156 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 studies, like CRAC2 or NUREG-1150, can be readily observed and understood. And we would strongly advise against bundling everything you're doing to a black box and spitting out a best estimate, and forcing the public to unravel, to the extent we can, exactly what's been going on here. And I think the discussion earlier makes clear that there is a lot going to this recipe, and it would be better if we did understand those changes incrementally so we can see what are the significant differences. Now with regard to risk communication, going into a little history, there's a reason why CRAC2 made the front page of the "Washington Post" in 1982, was not because it was necessarily so
frightening, it was because NRC was originally only planning to release the mean values across the
meteorological distribution that is generated by these codes, and someone leaked the files that showed the maximum, or peak consequences for in worst case, where the scenarios were evaluated, and it was that fact, I think, that led to the reason why it got as much press as it did, which should be a lesson, that if you're open about what you're doing, you don't try to parse the results for public perception that you're going to be better off in the long run. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
157 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 things. public? MR. LYMEN: No. And I asked for it. I This excessive secrecy marking this
project, which is a concern.
The original SECY and
SRM are still being withheld from the public in their entirety, despite the objection of one of the
Commissioners. SOARCA have
And many of the meetings discussing been closed, in some cases with,
apparently, inappropriate rationales. And one example was the ACNW meeting a few weeks ago where I was closed on the basis that it would be pre-decisional, meaning it would have a severe impact on a regulatory decision, except that SOARCA, by definition, has no regulatory application, and is a project plan, so it looked pretty ridiculous to say that that was predecisional with regard to any regulatory decision, and the meeting was opened after that was challenged. MEMBER APOSTOLAKIS: The SRM is not
sent a letter and was told it's sensitive. Now with regard to SOARCA, there are good And to the extent MELCOR, accident
progression and source term development can be updated using reactor-specific data, input decks and
experimental insights that have been achieved through Phebus and other severe accident studies, that's a NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
158 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 good thing. And understanding, if there is more data
to better understand containment performance, and I'm not sure there actually is, but to the extent there is, and that can be fed in, that seems to be a big uncertainty in my review of NUREG-1150, and the way expert elicitation was used to gloss over things that weren't known about containment performance. would be very beneficial. And, again, if you can model protective actions better and with more accuracy, and you can come up with better results in what you do, those are good. But one of the bad things is what appears to be That
a totally inconsistent way of truncating these low CDF sequences, and I think if you're screening out five, or even 10 percent of a CDF, that is not an
insignificant chunk. In fact, if you look at what the LERF would be for those plants, you can see that it would be less than 10 percent, usually, so you would be clearly taking a big bite out of a LERF. seems like external events are being And it treated
inconsistently, low powering shutdown risks are being treated at all, which is absurd, because it's the shutdown risk that may dominate early releases if the containment is open. the truncation is And it just seems like the way being done, you're introducing
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
159 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 round-off error that will round-off and eliminate exactly the things that you should be looking at, and it's absurd, circular reasoning. If you take out the
events that would lead to most severe consequences and you find the consequences are less than you thought, that's circular reasoning. Also, credit for unregulated measures, like SAMGs, seems to be inappropriate. what I heard earlier is that I think that have been
these
demonstrated, accidents have been demonstrated to be preventable where containment failure or whatever through SAMGs. I don't think SAMGs have ever
demonstrated anything. and they're not
Those are voluntary measures, not tested in any
regulated,
verifiable way, and they should not be credited, at least, to the extent they are. If, again, as I heard
earlier, both the results with and without SAMGs have been evaluated, and present both of those to the public, and explain why you believe the SAMG one is more credible. Finally, the "ugly" is the use of dose thresholds in direct contradiction to recommendations of the BIER VII Committee, among others, and I won't belabor this point, but NRC is going to have to do a lot of work to justify why this recommendation should NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
160 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 be ignored in the study, if it goes in that direction. And I think that would severely undermine its
credibility. Now what was so bad about CRAC2? Well, I
went back and looked at CRAC2, and it seems like it, obviously, isn't applicable today, but the reasons are different from what we've heard. For instance, CRAC2
uses 1970 census data. It actually assumed the entire EPZ would be completely evacuated within six hours after the warning was issued, when current evacuation time estimates are generally much longer than that, so it was actually more conservative than appropriate in some cases. victims of It assumed medical treatment for all acute radiation exposure would be
aggressive. It used a BIER III correlation for cancer fatalities, which is out of date, and underestimates by a factor of four, compared to ICRP-60. And it only sampled a handful of weather sequences, because it at that time, it took a long time to run a weather sequence. Today, you can run an entire year's worth
of weather times 16 wind directions in a matter of minutes, so there's no reason to use sampling any more. You can use the entire year's worth of weather
data, and you get a much larger number of results for your consequence distribution. I found comparing that NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
161 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 sampling only 100 weather sequences compared to the full year underestimates peak consequences by about 30 percent. So we've been using MACCS-2 for a long time, and find, in particular for Indian Point, the CRAC2 was -actually gives fairly good results And we
compared to what you can do with NUREG-1465.
used NUREG-1465 as a starting point for source terms, because it is recommended, has been vetted, expert panel reviewed it for applicability to high burn-up fuel only a few years ago, and basically confirmed it was appropriate. It's being used for design-basis
applications by a number of licensees, but it's not being used for any severe accident applications, even though it does actually cover severe accidents. the question is why? And
That's generally because if you
use it for design-basis applications, it gives the licensees a benefit, while actually make things worse if you look at the full severe accident term, so just showing the source term that I used based on NUREG1465 truncated after about two hours, so there's a tail that we need to consider, comparing that to, let's say, what Entergy used derived from MAAP for Indian Point in their license renewal application, a source term which has a lot of odd things about it, NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
162 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 22.9 hour duration, and release fractions which are generally lower in most categories than for NUREG1465. And the consequences are dramatic in
comparison, and so the results within 50 miles, the mean consequences for Indian Point using the 2034 population density are 860 early fatalities, mean latent cancer is 38,500, the peak early fatalities 70,800, peak latent cancer is almost 700,000. So
since NUREG-1465 corresponds to a low pressure event like large break LOCA, you can see why staff from these results may not want to consider large break LOCAs, if they can avoid it. So my conclusions are if the main impact of SOARCA is to reduce severe accident consequences by eliminating consideration of large early releases, that's circular reasoning. The more reasonable
approach would be to group sequences, not to truncate at the CDF level, to truncate at the Level 2 so that you have the frequencies of the various release
classes, and then do your screen at that point, which was an approach that was rejected from the beginning. The inclusion of thresholds without authoritative technical justification is going to undermine the credibility of the results. And finally, to get an apples-to-apples NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
163 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 slides? MR. LYMEN: Absolutely. Lymen? MEMBER STETKAR: Can we get a copy of the comparison, the way to do that is when changes are made, that people understand the impact of each
particular change, and not simply the entire package handed to you without a clear understanding of what went into it. So without that, we can't understand
what's due to better data, improvement of the codes, better technical understanding, what's due to simply changing sequence the are ground and rules aren't by what accidents So or my
considered.
recommendation, I'm glad to hear there's going to be an external peer review, that is one of my
recommendations, and I believe the best way to -- for the credibility of the study to be insured is to publish the results in a peer review journal. So with that, I thank you. CHAIRMAN SHACK: Any questions for Mr.
MEMBER POWERS: Ed, let me ask a question. The Health Physics Society seems - and I'm being cautious in what I say - try to reproduce what they've said, but, in effect, they've said look, you get doses below about 100 millirem per year, don't try to do NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
164 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 you're anything with it, because it's difficult to calculate down there, it's hard to measure things and whatnot. So for the purposes of assessing risk, cut it off at 100 millirem for something like that. your take on all that? MR. LYMEN: going to make Well, my take is that if a radiation protection I mean, what's
recommendation, that it has to be based really on the most authoritative sources. And the Health Physics
statement is not really supported by peer review references, as was BIER VII. BIER VII had the
opportunity to consider, and they did at length, the literature supporting the notion for thresholds, and like you said, it wasn't. consider is that and well, low if dose, But the other thing to you're that talking actually about does
uncertainties
coincide with the uncertainties in doing atmospheric modeling beyond a certain radius, so to avoid having to say you're using a dose threshold, a better
approach would simply be to say that we don't think the MACCS-2, the validity of the code is really -- is that useful beyond 50 miles radius, correspondence between lower doses and longer distances, so that would be a way to dance around what the actual -MEMBER POWERS: Yes, but that's another
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
165 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 take. By the time you get out there, but I was just
interested in the view on the Health Physics. The other thing they emphasize in BIER VII in their view, is there is a wide perception that somehow linear no-threshold comes from people not knowing how you draw a straight line, they know it's based on a biochemical model. And you really can't
contest it unless you have a contesting biochemical model. And you struggle heroically to come up with a
biochemical model that for low LET radiation could yield a threshold, because it involves double breaks to the DNA strands, which are not easily repaired. And the analogy drawn to chemical effects is
inappropriate, because chemical effects are all single breaks to DNA strands. And so, yes, BIER VII is a
fairly authoritative, but now you get down to the practicality of doing calculations. And like you say, I mean, it seems to me the argument to make, one argument certainly you could appeal to Health Physics, but the other one is, there's a point where MACCS just breaks down as a useful computation -MR. LYMEN: Right. And that may not
coincide with the dose threshold you're talking about. That's the only thing. But right, I mean the
conclusion of BIER VII was that a single DNA lesion NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
166 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 argument. MEMBER ARMIJO: Well, I think the -- I I've from the -MEMBER POWERS: It's a probabilistic can lead to cancer, and on that basis, there is no model that would justify -MEMBER ARMIJO: Well, that kind of flies
in the face of a lot of research at the cell level that directly refutes that assumption, that one single lesion instantly leads, or without doubt will lead to cancer. by -MEMBER POWERS: MEMBER ARMIJO: BIER VII didn't say that. Well, that's what I heard And I've been trying to read up on this work
guess I'm more impressed by experimental work.
been reading some of the work of Dr. Mitchell of AECL and others, and I think there's just a preponderance of information that tells us that a threshold does exist. And I think it's -- to say we mustn't think
about it, we mustn't talk about it, because that's perhaps politically incorrect, just bothers me. MR. LYMEN: No, that's not what I'm
saying. Think about it, talk about it, but support it with argument. protection, it And in the context of radiation really has to be weight of the
NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
167 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 very much. NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
evidence, and if you don't -- to those who believe the data reads that radiation protection standards and models should be changed, there simply isn't enough of a coherent body of evidence to support that change. And BIER VII, had the opportunity to say that, and they rejected it, so at this point, it does not make
sense to include a threshold in these models, and it is going to undermine the credibility of the result, a three-year study and all the effort that's going into all the various aspects of it should not be undermined on that point. CHAIRMAN SHACK: comments? MEMBER POWERS: It's also fair to say that BIER VII did a comprehensive examination of the Further questions or
biochemical evidence.
And, interestingly, they came
back saying that there's pretty good evidence that there's no threshold in the case of neutron damage. And it's only the LET, Low Energy Transfer, radiation where this uncertainty exists. MR. radiation, high LYMEN: LET Right, because cause high LET
particle
can
multiply
damaged sites. CHAIRMAN SHACK: Okay. Well, thank you
168 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 NEAL R. GROSS
(202) 234-4433 COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. WASHINGTON, D.C. 20005-3701 (202) 234-4433
MR. LYMEN:
Thank you.
I appreciate it.
CHAIRMAN SHACK: break now. Thank the
We're going to take a staff again for their
presentation. 3:15.
Good discussion.
Take a break until
We're off the record. (Whereupon, the proceedings went off the
record at 3:01:19 p.m.)