477th Advisory Committee on Reactor Safeguards (ACRS) - November 2, 2000
1
1 UNITED STATES OF AMERICA
2 NUCLEAR REGULATORY COMMISSION
3
4 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
5 ***
6 MEETING: 477th ADVISORY ON REACTOR SAFEGUARDS
7
8 U.S. NRC
9 11545 Rockville Pike
10 Two White Flint North
11 Rockville, MD
12
13 Thursday, November 2, 2000
14
15 The committee met, pursuant to notice, at 8:30
16 a.m.
17 MEMBERS PRESENT:
18 DANA A. POWERS, Chairman
19 GEORGE APOSTOLAKIS, Vice-Chairman
20 MARIO V. BONACA
21 THOMAS S. KRESS
22 GRAHAM M. LEITCH
23 ROBERT L. SEALE
24 WILLIAM J. SHACK
25 JOHN D. SIEBER
. 2
1 GRAHAM B. WALLIS
2
3 ALSO PRESENT:
4 MEDHAT EL-ZEFTAWY, Federal Official
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
. 3
1 P R O C E E D I N G S
2 [8:30 a.m.]
3 DR. POWERS: The meeting will now come to order.
4 This is the first day of the 477th meeting of the
5 Advisory Committee on Reactor Safeguards. During today's
6 meeting the Committee will consider a revised report of the
7 final technical study of spent-fuel pool accident risk in
8 decommissioning nuclear powerplants, risk-informed
9 regulation implementation plan, and proposed framework for
10 risk-informed changes to technical requirements of 10 CFR
11 Part 50.
12 The Committee will also examine the activities of
13 its ad hoc Subcommittee on the Differing Professional
14 Opinion on Steam Generator Tube Integrity and will examine
15 some proposed ACRS reports.
16 The meeting is being conducted in accordance with
17 the provisions of the Federal Advisory Committee Act. Dr.
18 John T. Larkins is the designated federal official for the
19 initial portion of the meeting.
20 We have received no written comments from members
21 of the public regarding today's session. A transcript of
22 portions of the meeting is being kept, and it is requested
23 that speakers use one of the microphones, identify
24 themselves, and speak with sufficient clarity and volume so
25 they can be readily heard.
. 4
1 I'll begin this session today by calling members'
2 attention to a couple items of interest. In your package
3 you have a presentation Dr. Meserve gave concerning the
4 revised reactor oversight process and some of his views on
5 the role of NRC in regulating nuclear powerplants.
6 You also have a presentation Commissioner
7 Merrifield gave at the water reactor safety information
8 meeting. I found that particularly illuminating, and am
9 very much enamored with his four challenges that he offered
10 the staff on developing their research program. I have had
11 a chance to discuss that some with Commissioner Merrifield,
12 and asked if we could cop some of his language, and he was
13 generous in allowing us to plagiarize from him.
14 Are there any comments that members would like to
15 make before we begin today's session?
16 Seeing none, I will turn to the first item of
17 business, which is to discuss the spent-fuel pool accident
18 risk, and Dr. Kress, I believe you'll take the lead on this.
19 DR. KRESS: Thank you, Chairman Powers.
20 I'll remind the Committee that we heard
21 presentations on an earlier draft version of this technical
22 study, and the idea was that spent-fuel pools for
23 decommissioning plants, as time goes by one would expect the
24 risk to decrease due to the decay of the fission products so
25 at some point one might consider a rule that would relax
. 5
1 some of the requirements, particularly the requirements for
2 emergency planning and maybe safeguard and maybe even
3 insurance-related activities.
4 As you recall, we found a number of technical
5 problems with that earlier study, and so the staff went back
6 and basically addressed our problems, and have come back
7 with a revised version, and I'll remind you our problems
8 were with the use of Reg Guide 14174, LERF acceptance
9 criteria, in view of the different nature of the source
10 term. We had problems with the ignition temperature for
11 when you would set off a zirc-air fire, and we had problems
12 with the plume energy they used, which was for reactor
13 accidents and not fires. And we had problems with the fact
14 that they didn't deal with atmospheric dispersion
15 uncertainties in relating the safety goal to the LERF, and
16 we had a problem with the bounding nature of the seismic
17 analysis. You recall seismic was the dominant risk in this
18 problem.
19 So the staff addressed those and developed a
20 revised technical study, and now they're going to tell us
21 about the results and how they went about dealing with our
22 problem.
23 With that, I'll turn it over to Timothy Collins.
24 MR. COLLINS: Good morning. I'm Tim Collins. I'm
25 the Deputy Director of the Division of Systems and Safety
. 6
1 Analysis. We have three pieces to our presentation this
2 morning. I'm going to give an overview of the study. Jason
3 Schaperou will give you details on the revised consequence
4 analysis that we did. And Bob Palla will give you a
5 discussion of the revised risk assessment.
6 This presentation outline is my presentation.
7 Those guys will give you their own presentation outlines.
8 We have several other members of the staff with us
9 today to respond to other questions in areas outside the
10 work that Jason and Bob did. Joe Stademeyer is here for any
11 questions on thermohydraulic analysis, and Glenn Kelly is
12 here for any questions on the frequency assessments.
13 Frequency assessments haven't really changed relative to the
14 February report. We also have Dr. Robert Kennedy is here if
15 there's questions on some of the seismic issues.
16 What I intend to do is start at the findings of
17 the February report, which Dr. Kress has already started my
18 presentation for me. I was going to talk about the
19 significant comments that we received on that report, our
20 approach to comment resolution, results of the reanalysis,
21 and the technical conclusions that we reached.
22 The February report which was put out for formal
23 public comment concluded that the frequency of a zirconium
24 fire at a spent-fuel pool was low. It also concluded,
25 however, that the consequences of such a fire could be
. 7
1 serious and similar to those of a reactor accident large
2 early release, and that was why we were using the LERF
3 criteria of 1 times 10 to the minus 5 as a screening
4 criterion.
5 We found in the February report that seismic
6 events were dominant. To a large extent there were
7 commitments made by the industry, what we call in the report
8 IDC's, industry design commitments, and design assumptions
9 with regard to the pools, which when these are implemented
10 the likelihood of any events related to like loss of cooling
11 or draindowns absent a rapid fracture basically went into
12 the mud. So seismic events were left.
13 The report also concluded that relaxation of EP
14 was supportable about a year after decommissioning, and the
15 basis for that was that there was enough time available
16 after a year's decay for implementation of ad hoc measures
17 to effect a reasonable evacuation.
18 The February report also concluded that because of
19 the large fission product in the pool, that you would have
20 to consider security for as long as you maintained fuel in
21 the pool.
22 The relaxation of insurance was a little squishier
23 in the February report. There was a conclusion that after
24 about five years it looked like air cooling would be
25 sufficient to preclude reaching the zirconium ignition
. 8
1 temperature. However, there are so many plant-specific
2 assumptions in that analysis that the report really
3 concluded that you had to have plant-specific information.
4 It was hard to draw a generic conclusion.
5 DR. LEITCH: Could you say a word about what is
6 meant by EP relaxation, that is, how relaxed is relaxed?
7 MR. COLLINS: This would be only with offsite
8 consideration. Sirens would be taken down, there would be
9 no exercises, no official radiological preplanning. The
10 offsite organizations would be disbanded.
11 DR. LEITCH: And as far as the onsite
12 organization, would the emergency response organization
13 still be fully staffed?
14 MR. COLLINS: No, I don't believe. We didn't
15 assume a fully staffed -- basically all we really needed was
16 someone to identify that an event had happened and to notify
17 offsite authorities. You need much more in your onsite
18 emergency plan.
19 If Dave Bars is here, if there's something more he
20 would like to add, he can jump in --
21 MR. BARS: Dan Bars.
22 MR. COLLINS: Dan Bars. Sorry.
23 DR. KRESS: Please introduce yourself and your
24 affiliation.
25 MR. BARS: Dan Bars, emergency preparedness
. 9
1 specialist in NRR.
2 As far as the relaxation of the offsite emergency
3 planning requirements, in the previous rulemaking effort,
4 which we attempted and sent up to the Commission, the
5 assumptions we made were that they could take the ad hoc
6 actions based on the fact that there had been an emergency
7 plan at the site for 20, 30, 40 years, and there's a
8 residual knowledge that would exist for, you know, a certain
9 time frame, and that the offsite officials, you know, that
10 capability and response capability wouldn't disappear Day 1
11 or overnight. There would be some residual knowledge. They
12 also have the ability to respond to other emergencies,
13 chemical spills, fires, tornadoes, whatever it may be,
14 hurricanes. So there is some inherent emergency planning in
15 the community that would exist, and the idea was that they
16 could tap into that if necessary for the radiological
17 concerns after the fact.
18 Since then, as we've discussed with FEMA, there is
19 a concern from FEMA that there probably still needs to be
20 some level of emergency planning. So when we say relaxation
21 of offsite emergency planning, it's not going to probably be
22 a complete relaxation as early as maybe anticipated. As we
23 go forward and negotiate with them and other stakeholders
24 we'll determine what that planning level will be. But
25 certainly as the technical study shows us, as time goes on
. 10
1 we gain the factor of having additional time to take action,
2 and there is in communities inherent capabilities which we
3 will depend on and rely on.
4 DR. LEITCH: And as far as the onsite facilities,
5 that is, staffing can be reduced, the emergency response
6 centers are no longer required?
7 MR. BARS: Yes. We're not sure --
8 MR. COLLINS: We need to be a little careful. We
9 weren't making recommendations for the rule itself at this
10 point, we were making assumptions for the purposes of the
11 analysis.
12 DR. LEITCH: I see.
13 MR. COLLINS: And all we simply did was presume
14 that there would be a capability to rapidly notify an
15 offsite authority in the event that the pool had been
16 drained. We did not do specific removal of this piece,
17 removal of this piece type of analysis.
18 DR. KRESS: Yes. That's a clarification I meant
19 to make. This is a technical study to be used as input to
20 the decision makers that will maybe develop a rule, and all
21 it does is give them the technical input.
22 MR. COLLINS: Yes, this is to provide some risk
23 insights to a process which is going to have to involve some
24 very significant other policy considerations,
25 public-confidence issues. There's a lot of other players in
. 11
1 the rulemaking process. This is just the risk-insights
2 aspect of it.
3 DR. SEALE: There's the rule of unexpected
4 consequences, like institutionalizing the support of an
5 emergency response organization for 40 years at a plant site
6 by the utility and the public relations impact you would
7 have. If you suddenly withdrew that support, you might find
8 people sitting on their hands if an emergency came up
9 because of resentment over having lost that support.
10 DR. KRESS: How do you deal with that in a
11 technical study?
12 DR. SEALE: You don't.
13 MR. COLLINS: We say we'll deal with it in the
14 rulemaking.
15 DR. SEALE: But still it's not a zero concern.
16 MR. COLLINS: We received lots of reports on the
17 February draft which caused us to go back and do a whole lot
18 of work. The first comment -- this is from the Committee
19 itself -- was that the source term that we used may be
20 nonconservative because of the oxidation of fuel in air as
21 opposed to typical in a reactor environment you have an
22 air-starved environment. And that oxidation could result in
23 release of a large amount of ruthenium, which has very
24 significant health effects.
25 We also received comments from the industry and
. 12
1 from the Committee as well that the seismic hazard estimates
2 that we had used in the February report were too
3 conservative. In the February report we used the Livermore
4 hazard estimates. The Committee also suggested that the
5 zirconium ignition temperature might be too high, that we
6 could have a significant fission-product release at lower
7 temperatures than we had assumed.
8 We received a comment from the public that the
9 partial-draindown scenario should receive more attention.
10 This is a case where the fuel doesn't get completely
11 uncovered. In the February report we had emphasized cases
12 where there was a complete uncovery of the fuel and
13 subsequent heatup. And the commentor indicated that you
14 might have a more rapid heatup because of the loss of the
15 cooling flowpath if you only had a partial draindown.
16 Another commentor indicated that our study said
17 that EP could be relaxed at one year, but that we didn't say
18 how much earlier than that it could be relaxed as well. We
19 simply did like a one-year point and suggested that as soon
20 as the iodine was decayed, that our results would be just as
21 applicable, and that would take it back to about 60 days.
22 We also received a comment which wasn't actually
23 on the report but was related to the integrated rulemaking
24 plan that subsequently went to the Commission. That plan
25 went up in SECY-00145, and it had potential rule changes in
. 13
1 it, and the industry indicated that that rulemaking plan was
2 not very risk-informed. We subsequently modified our study
3 to give more support in a risk-informed nature to the next
4 rulemaking attempt.
5 That 00145, for example, had said that you could
6 relax EP at a year because ad hoc measures may be sufficient
7 to substitute for formal EP measures, and that was viewed
8 as, well, there's nothing risky about that, that's just a
9 tradeoff of one thing for another. And the insurance
10 recommendation indicated that because there was no chance of
11 a fire after a certain amount of time that you could relax
12 insurance requirements. And again, that was a zero
13 threshold, that's hardly a risk-informed approach.
14 So our approach to addressing these comments, we
15 expanded our consequence analysis quite a bit. We included
16 a ruthenium and fuel fines portion to the source term, used
17 a very large ruthenium release fraction in our sensitivity
18 analyses.
19 We did 75 percent with each fraction, and I think
20 we even did a case with a hundred percent release of
21 ruthenium. We did sensitivity studies, taking into account
22 the Committee's comments on the plume parameters, and we
23 also expanded the consequence analyses to take into account,
24 times from about 60 days or 30 days after shutdown to ten
25 years after shutdown.
. 14
1 DR. KRESS: I might note that they carried the
2 losses out to full Level Three, rather than stop at a LERF.
3 It finesses our problem with whether or not to use 1.174
4 LERF. They go directly to the safety goal.
5 MR. COLLINS: Right.
6 DR. KRESS: Which is a nice way to do it.
7 MR. COLLINS: Didn't need the surrogate anymore.
8 DR. KRESS: Right, didn't have to worry about
9 whether it was right or not.
10 MR. COLLINS: To address the concern with regard
11 to the conservatism in the seismic hazard curves, we had a
12 meeting with the -- a public meeting with NEI back in August
13 to discuss the concerns.
14 And they indicated, the NEI indicated that they
15 thought the use of an EPRI curve was sufficiently
16 conservative to use in the analysis.
17 And after that meeting, we concluded that there
18 was really no basis to exclude either the Livermore curves
19 or the EPRI curves, so we redid the entire risk analysis,
20 incorporating the EPRI curves as well.
21 So, our results in the report are shown using both
22 the EPRI curves and the Livermore curves.
23 DR. KRESS: And how do you expect the
24 decisionmakers to deal with those two sets of results?
25 MR. COLLINS: Well, as it turns out, the risk is
. 15
1 low.
2 DR. KRESS: With either one?
3 MR. COLLINS: With either one, right.
4 DR. KRESS: You come out good with --
5 MR. COLLINS: Hopefully we can dodge that specific
6 problem.
7 DR. KRESS: For this issue?
8 MR. COLLINS: Right, for this issue, yes.
9 DR. KRESS: And that maybe just gives an idea of
10 the margins they might have?
11 MR. COLLINS: Sure. Yes, the risk is either low
12 or lower.
13 DR. KRESS: What if the two curves had straddled
14 the acceptance criteria? What would you have done?
15 MR. COLLINS: We would have asked for an
16 extension.
17 [Laughter.]
18 DR. KRESS: Okay.
19 MR. COLLINS: And to address the concern that the
20 rulemaking was not sufficiently risk-informed, we expanded
21 our analysis to do a small change in risk analysis with
22 regard to emergency planning.
23 We couldn't use the same approach for insurance,
24 because insurance isn't affected by either probability or
25 consequences. So, we did a small change in risk analysis
. 16
1 for emergency planning.
2 We looked into how each of the sequences might be
3 impacted by flow blockage. This is the partial draindown
4 concern.
5 There are other ways to -- the problem with the
6 partial draindown is that the airflow path is interfered
7 with. And you don't need just a partial draindown for that
8 to happen; you could have a piece of the roof fall onto the
9 pool.
10 The configuration of the assemblies themselves
11 could be changed if you have a major earthquake, so we
12 concluded that we needed to look at each one of the
13 sequences and see, is it likely that you could get a partial
14 draindown.
15 We also looked at the impact of the lower
16 temperature criterion. Charlie Tinkler did some real good
17 work going through the literature and studying that to
18 death, and as it turns out, it doesn't have a big impact,
19 because when we did the analysis in shorter times, the
20 heatup to the temperature was so short that even a lower
21 temperature criterion isn't going to make any difference and
22 the risk was still low. So, thank you, Charlie.
23 DR. KRESS: And I thought that that was a good
24 insight that he had, that temperatures well below, you
25 expect this thing to take off, the hydrides go into
. 17
1 solution, are not possibly readily available to do the
2 ignition. I thought that was a good insight.
3 MR. COLLINS: When we finished the revised
4 analysis, we found that the consequences with the ruthenium
5 included the large ruthenium release for action, and fuel
6 fines, we used about three and a half percent, I believe, of
7 fuel fines in our analysis.
8 We found that there was a notable increase in
9 consequences, particularly fatalities, but they were still
10 within the range of consequence calculations that were done
11 for like NUREG 1150.
12 So, if we needed to use the PPG for something in
13 the future, we think that's still a reasonable guideline.
14 DR. KRESS: The thing that took away from that,
15 though, was that the risk and the consequences of the spent
16 fuel pool fire are on par with that of an operating reactor.
17 MR. COLLINS: It could be, yes. I think that when
18 you see -- I think one of Bob's plots is going to show the
19 risk numbers, and it shows the sidebar, the results from
20 NUREG 1150.
21 DR. KRESS: I'm not sure PRAs currently include
22 those, and it may be one of the things we need to think
23 about later on.
24 CHAIRMAN POWERS: When you do these analyses, you
25 have a distribution of core ages, fuel ages, in the pool.
. 18
1 Did you propagate the fire even into the very old fuel that
2 might be there?
3 MR. COLLINS: We let the fire propagate into the
4 equivalent of three and a half cores, which is fuel, I
5 think, that is ten years old. The last assembly that burned
6 would have been about ten years old, yes.
7 So, our analysis indicated that the risk was low,
8 but it could still be in the ball park of operating
9 reactors.
10 The use of the -- all the debates about the
11 seismic curves, as it turns out, changed the risk by a
12 factor of four. There's a table --
13 DR. WALLIS: I'm sorry, that ball park is not a
14 technical term. What does it mean?
15 DR. KRESS: Equivalent to.
16 DR. WALLIS: Do you mean within a factor of ten,
17 100, 1,000? A ball park is a pretty large place.
18 MR. COLLINS: A ball park is a pretty large place.
19 It depends on whether you use the ruthenium source
20 term and the Livermore curve, or you use the EPRI curves and
21 the low ruthenium source term.
22 DR. WALLIS: A factor of ten?
23 MR. COLLINS: Have you got your plot there? Bob's
24 got these plots and he'll put it up there and we can see
25 what a ball park looks like on the plot.
. 19
1 DR. WALLIS: If this is a conclusion, it ought to
2 be more specific, perhaps.
3 DR. KRESS: It is in the report.
4 MR. PALLA: This is an early fatality risk measure
5 on a per-year basis. And this is the ball park, is up in
6 the right-hand corner there.
7 Generally, you find that with the highest seismic
8 hazard assumption from the Lawrence Livermore study, in
9 conjunction with the high ruthenium source term, and early
10 on, just following shutdown, you're in the range of the
11 results from NUREG 1150, based on just the -- at the Peach
12 Bottom results.
13 DR. WALLIS: So it's a pretty big ball park?
14 MR. PALLA: Yes.
15 MR. COLLINS: Yes.
16 MR. PALLA: And this is just two plants that we've
17 got. If you looked at others, I'm sure the range would get
18 broader. I'll have this information later.
19 MR. COLLINS: Now, the report also concluded that
20 the -- a relaxation of EP as early as 60 days was a small
21 change in risk, consistent with the guidelines of Reg Guide
22 1.174. Bob will give you all the details on that.
23 DR. WALLIS: I thought you were being colloquial
24 again when you talked about small change.
25 MR. COLLINS: Well, Reg Guide 1.174 --
. 20
1 DR. WALLIS: You mean to be technical in terms of
2 small change to --
3 DR. KRESS: LERF.
4 DR. WALLIS: Again, it's not --
5 MR. COLLINS: In accordance with Reg Guide 1.174
6 definitions.
7 DR. KRESS: Now, the reason for that bullet is,
8 number one, seismic dominated, so the things that were not
9 seismic, you would get some help from EP?
10 MR. COLLINS: Yes.
11 DR. KRESS: But they didn't dominate the risks.
12 MR. COLLINS: That's correct.
13 DR. KRESS: They were low enough even with no EP?
14 MR. COLLINS: Yes. Bob is going to walk through
15 this.
16 DR. KRESS: He's going to walk through this.
17 MR. COLLINS: Yes.
18 DR. KRESS: And the fact that seismic dominated,
19 told you that EP wouldn't be very effective anyway.
20 MR. COLLINS: We walked through each of the
21 sequences.
22 DR. KRESS: Yes.
23 MR. COLLINS: And we tried to decide, would
24 offsite planning have a significant impact for those
25 different sequences? For the seismic events, we needed a
. 21
1 seismic event of very large magnitude, and we would expect
2 --
3 DR. KRESS: In order to have this fire in the
4 first place?
5 MR. COLLINS: Yes.
6 DR. KRESS: If you had one --
7 MR. COLLINS: The fuel pools are very strong.
8 DR. KRESS: -- down in rubble, probably, if you
9 had one of that magnitude.
10 MR. COLLINS: It would certainly be damaged
11 significantly, we might expect. And we didn't believe that
12 the infrastructure would be in place for the formal EP to
13 have a significant impact.
14 The next sequence --
15 DR. KRESS: That was basically driven by your
16 assumption on how effective would be under seismic
17 conditions.
18 MR. COLLINS: I consider it a reasoning instead of
19 an assumption. We reasoned each of the sequences through to
20 decide --
21 DR. KRESS: The assumption.
22 MR. COLLINS: Sure.
23 DR. SEALE: Well, there are earthquakes that have
24 happened contemporaneously, which are small, compared to the
25 kind of earthquake you're talking about here, which have
. 22
1 demolished freeways systems and so forth.
2 So, basic elements of the emergency planning are
3 just not there.
4 MR. COLLINS: That was our reasoning. And the
5 next event for which the infrastructure would be in place
6 would be -- and the next highest frequency was the cask drop
7 event. That's around down to two times ten to the minus
8 seven.
9 DR. KRESS: It's low enough that it didn't matter.
10 DR. WALLIS: Well, the VP's not in place, and it
11 makes the consequences worse, presumably.
12 MR. COLLINS: That's the way we did the analysis,
13 right. We modeled it as either early evacuation or late
14 evacuation.
15 If you has successful early evacuation, the
16 consequences are significantly lower, and notably lower.
17 These are the things that Bob is going to walk through.
18 DR. WALLIS: But how would you do that? You'd
19 have to know the seismic event was coming, and then evacuate
20 people ahead of time.
21 DR. SEALE: Tricky.
22 MR. COLLINS: After the pool is drained, there's
23 time before the fuel heats up to the ignition point.
24 DR. WALLIS: But you were just telling us the
25 seismic event was so large that people couldn't get out
. 23
1 anyway.
2 MR. PALLA: That's why you only have --
3 DR. WALLIS: So any evacuation would have to be
4 before the pool drained?
5 DR. KRESS: Or very late after. The issue is
6 small change. If I didn't know seismic was part of this,
7 and just focused on the other accidents, the loss of cooling
8 and the cask drop, then EP, the effect of EP on that could
9 not be described as small change; could it?
10 MR. COLLINS: Sure. Well, I mean, the frequency
11 of those other events is very, very low.
12 DR. KRESS: You're saying they are very low, but
13 there's a big difference whether you use EP or not there for
14 those particular sequences.
15 MR. COLLINS: I think the difference for that
16 sequence might be considered to be large, but it would be
17 small on an absolute scale.
18 And if you took the -- if you took as a figure of
19 merit like a LERF and said that the LERF would have to
20 increase by ten to the minus six, if you use that as a
21 limit, when you perturb that low frequency sequence, you're
22 not going to be changing the LERF in excess of that.
23 DR. KRESS: Ten to the minus seven or something
24 like that?
25 MR. COLLINS: Yes, it will be -- you'll
. 24
1 significantly change that ten to the minus seven sequence,
2 but you won't exceed any of the guidance in Reg Guide 1.174.
3 DR. KRESS: It wasn't meant for being on a
4 sequence basis, anyway.
5 MR. COLLINS: Right.
6 DR. KRESS: But, you know, that's one way to look
7 at it.
8 MR. COLLINS: With regard to the obstructed
9 airflow, the partial draindown event, this was important,
10 particularly with regard to insurance considerations.
11 Historically, we've considered that when air cooling was
12 sufficient to remove the decay heat, such that you wouldn't
13 reach a predefined criterion, that insurance could be
14 relaxed.
15 Well, after this study, we concluded that you just
16 can't on a generic basis, be assured that you're going to
17 have an airflow path. So there is going to have to be a
18 different criterion for insurance considerations in the
19 future.
20 DR. KRESS: Does NRC regulate insurance
21 requirements? Is that in the rule?
22 MR. COLLINS: There is a regulation which requires
23 that a licensee have a certain amount of insurance.
24 DR. KRESS: And that's an NRC regulation?
25 MR. COLLINS: It's an NRC regulation, yes.
. 25
1 The temperature criterion, the issue raised by the
2 committee, we concluded that it was interesting but not
3 really important to the decision-making, because the times,
4 the heatup times, are so short in the early times after
5 shutdown anyway that a little bit shorter isn't going to
6 make any difference.
7 DR. WALLIS: So you don't really need to establish
8 what this temperature criterion is?
9 MR. COLLINS: Well, we used the criterion of 800
10 degrees C.
11 DR. WALLIS: I guess we questioned the number.
12 MR. COLLINS: Yes.
13 DR. KRESS: They had a higher one than 800
14 previously and that's one of the questions.
15 MR. COLLINS: A higher one would give us longer
16 times, which helps in all cases.
17 In a shorter one, you can't get much shorter than
18 we already assumed in the analysis. It was a couple of
19 hours.
20 DR. WALLIS: You mean it's bad enough already, it
21 doesn't matter?
22 MR. COLLINS: Pardon me?
23 DR. WALLIS: It's bad enough already --
24 MR. COLLINS: Yes, yes -- short enough --
25 DR. WALLIS: -- so if it were 1000 it doesn't make
. 26
1 any difference?
2 MR. COLLINS: Well, a thousand times shorter
3 than --
4 DR. WALLIS: Your conclusions are not sensitive to
5 what you take as a temperature criterion?
6 MR. COLLINS: That's correct.
7 DR. WALLIS: This business about short times
8 confuses the conclusion for me because I don't quite know
9 what you mean by short time.
10 MR. COLLINS: The time I am talking about is once
11 the fuel is uncovered the time it heats up from its initial
12 temperature to the criterion for a fission product release.
13 DR. WALLIS: Well, if you can't do anything it
14 doesn't matter how long a time it is. You don't have
15 anything you can do about it. It's gone, so I don't see why
16 time is really important.
17 MR. COLLINS: Well, that is part of my conclusion.
18 Time is important with regard to whether or not you can take
19 credit for ad hoc EP.
20 DR. WALLIS: Time isn't important so temperature
21 then becomes important.
22 MR. COLLINS: Longer times -- time can go two
23 ways. It can be either shorter or longer than what we had
24 originally assumed, right? Okay? if it is longer, things
25 always get better. Our conclusions are the change is small
. 27
1 already -- if the time get longer it is still going to be
2 small.
3 DR. WALLIS: The important thing is does it burn?
4 Does it reach the temperature? Isn't that the important
5 criterion, not the time.
6 MR. COLLINS: Time is an important --
7 DR. WALLIS: Half a day or three-quarters of a
8 day, it still burns.
9 MR. COLLINS: Yes.
10 DR. WALLIS: That is the problem, isn't it? Does
11 it burn?
12 MR. COLLINS: That's the problem. If it doesn't
13 burn, we don't have a problem.
14 DR. WALLIS: And I would think temperature would
15 have to be important in determining whether or not it burns,
16 but perhaps you are going to explain all that.
17 MR. COLLINS: Well, no, we don't have a further
18 presentation on that issue.
19 DR. WALLIS: I just wondered if you were somehow
20 finessing the temperature problem, which really is the
21 problem --
22 MR. COLLINS: No.
23 DR. WALLIS: -- by bringing in something which
24 doesn't matter so much, which is all it takes.
25 MR. COLLINS: If the temperature is lower, then
. 28
1 the time it is going to take to heat up to that temperature
2 is going to be shorter.
3 That means that it would be less time to implement
4 ad hoc evacuation measures.
5 DR. WALLIS: Ah.
6 MR. COLLINS: And that would change the way we did
7 our comparison of the benefit of formal offsite planning.
8 However, the time is already so short in the
9 analysis that we did that making it short won't change any
10 of our conclusions.
11 DR. WALLIS: Well, if the temperature criterion
12 were high enough, the problem would go away, wouldn't it?
13 MR. COLLINS: If you would never get the fire.
14 DR. WALLIS: So there must be some influence of
15 temperature.
16 MR. KELLY: This is Glenn Kelly from the Staff.
17 When we performed the thermal hydraulic analysis,
18 we basically did it two ways. One was where we considered
19 that we had air flow to provide oxygen to the potential
20 oxidation of the fuel and also to provide cooling to the
21 fuel and the other one was we assumed that there might have
22 been flow blockage such that we had a near adiabatic heatup.
23 In the adiabatic heatup case effectively as long
24 as you have decay heat, you are going to eventually be able
25 to get the fuel temperature up to whatever is your criteria,
. 29
1 so when Mr. Collins was talking about that we can't on a
2 generic basis preclude the possibility of obstructive flow
3 cases, we're saying that in these cases as long as you have
4 decay heat and you have obstructed flow it might take a very
5 long time but the potential is there that fuel might
6 eventually reach ignition temperature.
7 DR. LEITCH: The 60 days -- what starts the timer?
8 Is that from reactor shutdown?
9 MR. COLLINS: From reactor shutdown.
10 DR. LEITCH: Of the most recent fuel?
11 MR. COLLINS: Yes, the most recent fuel.
12 DR. LEITCH: Thank you.
13 MR. COLLINS: All right.
14 Dr. Wallis, this is a calculation that we did of
15 how the temperature takes off, so a higher temperature
16 criterion is going to be reached very shortly afterward. If
17 it was 900 or 1000 --
18 DR. WALLIS: The time shifts to three hours
19 doesn't really make much difference?
20 MR. COLLINS: Right.
21 DR. WALLIS: The real question is does it take
22 off -- to me. Maybe you can explain all this later. In
23 adiabatic heating it is always going to take off eventually.
24 DR. KRESS: At one time they thought that --
25 MR. COLLINS: -- that it would reach --
. 30
1 DR. KRESS: -- yes, where our cooling would turn
2 this around and you'll never get to the takeoff point.
3 MR. COLLINS: The problem is that we can't assure
4 any specific geometry so we can't on a generic basis
5 conclude that we are ever going to limit the temperature.
6 That is exactly the problem.
7 DR. KRESS: That's the heat transfer problem.
8 MR. COLLINS: If you can't define the geometry you
9 can't do an analysis that shows the heat goes away.
10 DR. WALLIS: So are you saying you could never
11 assure us that it will not burn?
12 MR. COLLINS: That is correct.
13 DR. WALLIS: You always have to assume it will
14 burn, even after 10 years?
15 MR. COLLINS: We can't on a generic basis say it
16 is not going to burn.
17 DR. KRESS: The decay heat doesn't go away after
18 10 years.
19 MR. COLLINS: The decay heat is persistent.
20 DR. WALLIS: So it could still burn after 10
21 years?
22 DR. KRESS: It could if you can't cool it.
23 DR. SEALE: If you don't have a sink.
24 DR. KRESS: At some point the issue may come down
25 to you want to determine the probability of a burn and use
. 31
1 that in your analysis but that involves developing
2 probabilities of particular geometries and the probability
3 of cooling, and it is not easy to do that, but it could be
4 done.
5 Eventually they may come up with a way to decide
6 that you can do away with the requirement after a certain
7 time based on probabilities.
8 MR. COLLINS: In the final report the conclusions
9 that we have reached are that the risks at decommissioning
10 plants is low. It is well within the Commission's safety
11 goals, even in the consideration of a large ruthenium
12 component to the source term.
13 We found that relaxation of EP is consistent with
14 a small change in risk.
15 DR. WALLIS: Now 60 days after what?
16 MR. COLLINS: After the last -- after shutdown,
17 the last fuel offload.
18 DR. WALLIS: After the last fuel offload into the
19 pool.
20 MR. COLLINS: Right.
21 DR. WALLIS: And this is because?
22 MR. COLLINS: You mean the 60 days --
23 DR. KRESS: Basically the thing that causes the
24 consequences are the cesium and the ruthenium and --
25 DR. WALLIS: -- the risk has gone down. The
. 32
1 probability of a fire, you have just taken as one, no matter
2 what.
3 MR. COLLINS: That is correct. We have taken it
4 as -- given the fuel is uncovered.
5 DR. KRESS: Yes.
6 DR. WALLIS: So this is another example where the
7 technical problem is made to go away by bringing in risk
8 considerations? The problem of is there or is there not a
9 fire has sort of gone away because you have looked at the
10 risk consequences.
11 DR. KRESS: Absolutely.
12 DR. WALLIS: Right, okay.
13 MR. COLLINS: The idea was to provide risk
14 insights to this rulemaking, yes.
15 DR. KRESS: And the 60 days just allows the iodine
16 to go away, maybe some of the ruthenium but not much of it.
17 MR. COLLINS: Yes, the ruthenium has got a
18 half-life of a year so most of the ruthenium is still there.
19 DR. KRESS: Most of it is still there.
20 MR. COLLINS: Right.
21 We also concluded that insurance is going to have
22 to be viewed from a different perspective. There needs to
23 be some sort of a policy decision on how insurance is
24 considered because we can't assure a geometry which would
25 assure cooling and because of the large fission product
. 33
1 inventory which stays for such a long time that you have got
2 to consider security for as long as you have fuel in the
3 pool.
4 These conclusions with regard to the risk change
5 and stuff are independent of whether or not we use the
6 Livermore curves or the ruthenium source term. It just gets
7 better and better as the lower probability --
8 DR. KRESS: Fortunately.
9 MR. COLLINS: Yes, fortunately.
10 I will let Jason Schaperow walk you through the
11 consequence analysis now.
12 DR. WALLIS: So I guess your conclusion is no more
13 work needs to be done on analyzing these fires with ignition
14 criteria?
15 DR. KRESS: For the issue of decommissioning
16 relaxation.
17 MR. SCHAPEROU: Good morning. As Tim said, my
18 presentation describes in some detail, our consequence
19 assessment for spent fuel pool accidents at decommissioning
20 reactors.
21 To restate some of the highlights of what Tim
22 mentioned, is that the overall risk assessment for spent
23 fuel pool accidents is comprised of three elements:
24 Consideration of initiating event frequencies; the second
25 element being the hydraulic anlaysis to further refine the
. 34
1 events leading to fuel and -- heating.
2 And the third element, which my presentation
3 describes is a consequence assessment for the events which
4 in the earlier analyses led to loss of pool cooling
5 inventory, fuel heatup and degradation, and fission product
6 release.
7 Our consequence assessment focused on issues
8 important for spent fuel accidents. We looked at source
9 term, plume issues, and, of course, evacuation.
10 We examined these issues by performing
11 consequences calculations wiht our MACCS reactor accident
12 consequence code. We reassessed the source term and the
13 release fraction of fission products.
14 And as I said, we performed sensitivity
15 calculations varying with release fractions of the various
16 fission products, including ruthenium, cesium and fuel
17 fines.
18 We performed sensitivity calculations, evaluating
19 the effects of the reduced inventory for different decay
20 times as far out as ten years.
21 We updated plume spreading modeling, and plume
22 heat content associated with spent fuel pool accidents, as
23 opposed to reactor accidents.
24 We performed consequence calculations for both
25 early evacuation and late evacuation cases, to allow Bob to
. 35
1 do his assessment of risk. The results of this large number
2 of MACCS calculations, we used, as I said, in Bob's risk
3 assessment.
4 The first area I would like to talk about is the
5 effect of ruthenium, which you've heard a lot about already
6 today.
7 There have been a number of small scale fission
8 product release tests done by the Canadians, and also by Oak
9 Ridge, with an air environment. These tests have shown
10 significant ruthenium release, particularly the AECL tests
11 that showed that following cladding oxidation, you get an
12 early complete release of ruthenium.
13 We performed consequence calculations, assuming a
14 release of all of the ruthenium inventory.
15 Our calculations showed that in this case, we get
16 a large increase in early fatalities between a factor of 20
17 and 100. This is because this particular element in its
18 assume form of ruthenium oxide, has a very high dose
19 conversion factor.
20 It goes into the lungs and it stays there, and the
21 clearnace class is years, which is the longest clearance
22 class for the lung.
23 We also thought about what things might mitigate
24 this consequence increase from ruthenium, and we note that
25 it does have a one-year half-life, so after a few years, it
. 36
1 will decay away.
2 Also, we thought a little bit about fuel geometry.
3 If there is degradation of fuel geometry as a result of the
4 heatup and oxidation, this degradation could limit the air
5 ingress and limit the ruthenium release.
6 Finally, I'd like to note that there is a Phebus
7 test planned to examine this important effect on a larger
8 scale.
9 This table shows the results of some of the
10 calcuations we performed for ruthenium. As you can see, in
11 going from the first row with a very small ruthenium
12 release, to the second row, which has 100 percent ruthenium
13 release, we get a very large increase in the early
14 fatalities.
15 We also see a reduction in the early fatalities as
16 expected when we implement an early evacuation, that is, an
17 evacuation before the release of fissio products begins.
18 DR. WALLIS: Could you put this in perspective?
19 People probably don't die from ruthenium in their lung
20 within the first year anway, unless they have a huge amount.
21 What's the long-term fatalities from this
22 ruthenium in the lungs? How does this one compare with ten
23 year fatalities?
24 MR. SCHAPEROU: I'm not sure I understand your
25 question.
. 37
1 DR. KRESS: I think he's asking for latent
2 cancers.
3 DR. WALLIS: Late fatalities.
4 MR. SCHAPEROU: Latent cancers?
5 DR. KRESS: Yes.
6 MR. SCHAPEROU: Latent cancer fatalities are a
7 proportion of the societal dose. I bleieve the factor is
8 about ten to the minus four.
9 DR. WALLIS: I'm really asking whether early
10 fatalities in this case is a good measure of risk to
11 society.
12 CHAIRMAN POWERS: Graham, the early fatality
13 calculations involve looking at the Schein emersion dose, as
14 well as the comitted dose from inhalation. I suspect for
15 these analyses, that the Schein does is what gives you the
16 prompt fatality. It's death within 30 days, due to an acute
17 radiation exposure.
18 So it's because he's putting out so much
19 radioactivity material, not because it's going into the
20 lungs so much.
21 The long-term dose out there is the fatal cancer
22 incidence, and that probably does have to do a lot with the
23 radiological behavior of ruthenium.
24 DR. WALLIS: But do we have an estimate of how
25 much it is?
. 38
1 CHAIRMAN POWERS: How much? What --
2 DR. WALLIS: How many people are likely to die as
3 a result of this accident?
4 CHAIRMAN POWERS: If you take those numbers and
5 divide them by about 2,000, that will give you an estimate
6 of the fatalities.
7 DR. WALLIS: So we take the numbers on the right?
8 CHAIRMAN POWERS: And divide them by 2,000.
9 DR. WALLIS: And so we're saying maybe 2,000
10 people die from a societal dose, so that one is a somewhat
11 misleading number.
12 CHAIRMAN POWERS: Those -- one is the prompt
13 fatalities within 30 days.
14 DR. WALLIS: Yes, but the message to the public
15 would be how many people are going to die as a result of
16 this accident is in the thousands; it's not in the ones.
17 MR. SCHAPEROU: That's out of a population of
18 about three million in this, for this distance.
19 DR. WALLIS: I'm just asking for a measure which
20 is understandable and meaningful.
21 CHAIRMAN POWERS: Well, if you're going to press
22 for meaningful, then you also have to add on to the fact
23 that of those people that get exposed, roughly a third of
24 them will ultimately die of a cancer of some sort, whether
25 or not the accident occurs.
. 39
1 DR. KRESS: And a lot of that is related to the
2 linear no-threshold issue that kills a lot of those people.
3 DR. APOSTOLAKIS: But is that consistent with a
4 goal of 1/10th of one percent?
5 DR. KRESS: Well, that -- we have two goals, and
6 the early fatality limit up there is related to the one
7 goal; and societal dose is related to the other, and it's
8 consistent with the -- in fact, it meets the goal.
9 DR. APOSTOLAKIS: But I think that's Graham's
10 question.
11 DR. KRESS: Yes. It beats the goal.
12 DR. APOSTOLAKIS: In terms of cancers, not in
13 terms of --
14 MR. COLLINS: Excuse me, the goal is in terms of
15 risk to an indivudal. The safety goals are in terms of risk
16 to an individual, early fatality risk to an individual and
17 the latent cancer risk to an individual.
18 DR. APOSTOLAKIS: Okay.
19 MR. COLLINS: Those parameters, Bob is going to
20 put up as part of his presentation.
21 DR. KRESS: Unfortunately, we don't have a goal on
22 total deaths.
23 DR. APOSTOLAKIS: No, but you're going to divide,
24 in other words, the 2,000 cancer deaths by the three million
25 people; is that what you mean?
. 40
1 DR. KRESS: Yes, and that gives you an indvidual
2 risk for cancers, and then you're going to multiply that by
3 the frequency, and say that that meets the safety goal.
4 MR. SCHAPEROU: We did the calculations, and we
5 had a lot of output measures. We have societal dose, cancer
6 fatalities.
7 We had different distances; we had early
8 fatalities; we had cancer fatalities.
9 DR. KRESS: MACCS will give you all that.
10 MR. SCHAPEROU: Yes, what's shown here is the zero
11 to 100 miles, which I -- was used early on for our
12 comparisons with reactor accident consqeuences and we stuck
13 with that.
14 But when you see Bob's comparisons, we mainly
15 focused on the cancer fatalities within ten miles, which is
16 the safety goal.
17 DR. APOSTOLAKIS: So you will show the indivudal
18 risk?
19 MR. SCHAPEROU: Yes, and the early fatalities
20 within the one mile. Those were the two things that we
21 really focused on in the end.
22 The reason my presentation is as it is, is that we
23 started out with zero to 100 miles, and we kept looking at
24 those distances for a lot of our sensitivity studies.
25 DR. WALLIS: Do you agree with Dr. Powers that you
. 41
1 can divide this last number by about 200?
2 MR. SCHAPEROU: Two thousand, yes.
3 DR. WALLIS: Two thousand, about two thousand,
4 okay.
5 DR. KRESS: But the safety goals, which are what
6 we would normally compare with, one of them is up to one
7 mile, and that's the fatalities, and the other one is out to
8 ten miles for hte cancers. We have to keep that in mind
9 when you think about the safety goals.
10 DR. APOSTOLAKIS: Actually, the safety goal
11 statement is in terms of societal risk.
12 DR. KRESS: No.
13 DR. APOSTOLAKIS: We just happen to be calculating
14 it as individual risk.
15 DR. KRESS: Well, there are documents that
16 interpret that. It's been interpreted always in terms of
17 individual risk.
18 We had a whole --
19 MR. COLLINS: We're using the qualitative health
20 objectives -- quantitative health objectives, teh QHO.
21 MR. SCHAPEROW: I mean it's 95 percent evacuations
22 and --
23 DR. WALLIS: Is that a surprising conclusion?
24 MR. SCHAPEROW: No, it was not, but it stuck out.
25 It stuck out when we went over the results. We're
. 42
1 like oh, yes, of course, factor of 10.
2 As a result of ACRS comments and all the
3 sensitivity calculations I just showed you, we decided to
4 reassess a source term.
5 The source term that has historically been used
6 for the spent fuel pool consequence calculations was that
7 given in NUREG CR-4982. This NUREG study was performed
8 about 12 years ago and it was performed for a generic safety
9 issue 82. It was regarding spent fuel pool risk at
10 operating reactors.
11 The source term from this study, which is shown in
12 the first row of this table, has large release fractions
13 that involve isotopes, that is noble gases, iodine and
14 cesium, and small release fractions of the other fission
15 products and for some of the ones of importance like
16 ruthenium, very small release fractions.
17 We decided that the NUREG-1465 reactor accident
18 source term, which is based on more recent research, had
19 undergone significant peer review, was a better basis for
20 our offsite consequence calculations, so we proceeded to
21 perform the entire array of calculations from 30 days out to
22 10 years using the NUREG-1465 or, as it's more commonly
23 known, the revised reactor accident source term.
24 We also performed the same set of calculations
25 with a modified version of this.
. 43
1 As we discussed, there's a lot of uncertainty in
2 the ruthenium release fractions and the fuel fines release
3 fractions.
4 For ruthenium we decided to go up to the same
5 release fraction as we had for the volatiles, that is, 75
6 percent. For the lanthanum and cerium we chose the fuel
7 fines release fraction reported in a recent report on the
8 Chernobyl accident, 3.5 percent of all of the UO2 and
9 whatever fission products are embedded in it.
10 My next two slides give tabular results from these
11 calculations with these two source terms. I would like to
12 substitute a graph for this. This is a graph of societal
13 dose and it shows the important trends here that I have
14 discussed.
15 The first trend is that if you have a ruthenium
16 release, which is the two top curves, modified source term
17 and a large fuel fines release, you will have higher
18 consequences.
19 The second trend is the effective early
20 evacuation. You do get some benefit from an early
21 evacuation.
22 The third trend is the case where you don't have
23 the large ruthenium release. The consequences really don't
24 fall off much --
25 DR. WALLIS: So these used to be in color or
. 44
1 something?
2 MR. SCHAPEROW: Yes, I'm sorry --
3 DR. WALLIS: The documents look the same --
4 MR. SCHAPEROW: The top two are the modified --
5 DR. WALLIS: So the codes are in the order in the
6 table, are they?
7 DR. KRESS: Yes, they are in the order of the
8 table.
9 MR. SCHAPEROW: The top two are the high ruthenium
10 and the bottom two are the low ruthenium, and the very top
11 one is with late evacuation, which is the worst
12 consequences, and then as you go to early evacuation your
13 consequences get lower and the same for the low ruthenium
14 cases. As you go from late evacuation to early evacuation,
15 the consequences get lower, but in the case of the low
16 ruthenium release you are dominated by cesium and the
17 consequences don't really fall off as you go out in time
18 because of the 30-year halflife.
19 Another issue was involved with the
20 thermohydraulics, the question of how much fuel would heat
21 up and release its fission products.
22 Most of our work was based and the stuff you just
23 saw was based on the heatup of the entire spent fuel pool
24 inventory of the Millstone 1 reactor.
25 DR. KRESS: Now one of the questions we had was
. 45
1 how representative is that 3.5 cores of the class of
2 decommissioned plants that are out there.
3 MR. SCHAPEROW: I think Tim may have understated
4 it a little bit, the number of years. Three and a half
5 cores, if you have a refueling every year and a half that
6 would be one and a half times ten. It is about 15-20 years.
7 MR. COLLINS: No, no, the first three batches are
8 the same timeframe. The last core is three batches, gets
9 off at once and then it is a batch at a time. You have to
10 take off the refueling cycle time by batch, so it takes
11 every three batches is one core.
12 MR. SCHAPEROW: This is representative of the
13 Millstone 1 reactor as it stood in 1988, which is I guess
14 about 20 years, 15-20 years into its operations, so this is
15 fairly late in the life of a reactor.
16 DR. KRESS: The question is could it be more?
17 MR. SCHAPEROW: Yes, it could if Millstone
18 operated longer, which it did operate a little longer.
19 DR. KRESS: For decommissioning plants that are
20 out there now, the rule is going to apply to them.
21 MR. SCHAPEROW: That's correct. It could be more.
22 MR. STAUDENMEIER: This is Joe Staudenmeier from
23 the Staff.
24 One of the recent decommissioned plants, Zion,
25 shared one spent fuel pool for both reactors and I believe
. 46
1 that had about 10 reactor cores in that spent fuel pool.
2 DR. KRESS: For that particular plant you'll do
3 plant-specific thinking on whether to relax any
4 requirements?
5 MR. COLLINS: Throughout the report we have made a
6 lot of assumptions with regard to -- design assumptions and
7 industry commitments. Whenever the rulemaking comes forward
8 it is going to have to take into account how those things
9 will be reviewed as part of the licensing process. Somebody
10 which differed from any of those things would have to be
11 looked at on a plant-specific basis.
12 DR. KRESS: If I am a decision-maker can I take
13 your risk numbers and multiply it by the ratio of that to
14 the number of cores I have, or is it not that easy?
15 MR. PALLA: Well, I think you would not -- if you
16 just added additional cores, you would go from having the
17 oldest fuel assembly 20 years to 40 years. You are not
18 going to add any more ruthenium. You are going to just have
19 more cesium.
20 DR. KRESS: So all you are going to change is the
21 societal end of the thing, rather than the prompt
22 fatalities.
23 MR. COLLINS: The prompt should be changed. The
24 ruthenium, almost all the ruthenium is in the last core that
25 you offloaded.
. 47
1 MR. PALLA: So you might have a factor of 2 or 3
2 in the cesium effects, in the longterm effects. Ruthenium
3 would have been diminished anyway.
4 Now if you look at the margins that we show, you
5 might rationalize that you still would be below the safety
6 goals or below whatever figure of merit you want to use.
7 MR. SCHAPEROW: We actually rationalized that it
8 may be as little as one core involved in the heatup one year
9 as a result of the work done for Generic Safety Issue 82, so
10 when we did a sensitivity on the amount of fuel released in
11 fission products we went from 3.5 cores down to the final
12 core offload.
13 DR. WALLIS: So the plume from one core is the
14 same as the plume from ten cores burning?
15 MR. SCHAPEROW: In what fashion?
16 DR. WALLIS: I would think the plume would be
17 different if you burned more stuff.
18 MR. SCHAPEROW: The heat of the plume?
19 DR. WALLIS: Burning quicker -- and the fire you
20 get would be different so the whole, this would change the
21 physics of things as well as the total amount distributed.
22 You are still in the same plume presumably?
23 MR. SCHAPEROW: Yes, we assume -- I will get into
24 that in a few minutes. I have got some discussion of the
25 plume modeling that we looked at.
. 48
1 DR. WALLIS: In reality the plume might differ if
2 you burned more cores?
3 DR. SEALE: The energy that is in the fission
4 product decay takes you to ignition, if you get there, but
5 the energy that is in the clad and other oxidation once you
6 get to ignition is the thing that drives the plume.
7 DR. KRESS: Right.
8 MR. SCHAPEROW: That's correct, with one small
9 change. At later times when the decay heat is really low
10 the energy of decay heat just gets you up a little bit in
11 temperature until the oxidation reaction actually provides
12 the dominant heat source.
13 DR. WALLIS: That's right.
14 MR. SCHAPEROW: Even at the lower temperatures.
15 DR. WALLIS: So burning old cores is just as
16 effective as burning new cores as far as the zirconium goes.
17 MR. SCHAPEROW: Although once the zirconium is
18 burned, that is the end of it. It's not like a Chernobyl
19 type accident where you have got graphites sitting there for
20 long period of time burning.
21 DR. SEALE: There is not nearly as much charcoal
22 in there. Yes, I agree with you.
23 DR. KRESS: The way the dispersion codes treat the
24 plume is it has a temperature which is a density, related to
25 the density of it, and as it rises up it entrains cooling
. 49
1 air and it cools off as it rises until it reaches the point
2 where it is neutrally buoyant and then normally the wind
3 turns it over and it disperses that way, so it does matter
4 whether you are burning a little bit or a whole lot because
5 that influences how long it takes it to cool off this plume
6 as it rises.
7 The temperature starts out the same, so the
8 buoyancy driving force is the same, but how much fire --
9 DR. SEALE: How much lofting you get --
10 DR. KRESS: -- is involved does affect the
11 lofting.
12 DR. APOSTOLAKIS: Actually, Bob, I thought about
13 this latent risk business. I think the number of years over
14 which the fatalities occur should be a factor in this,
15 because the risk, the Commission's goal is one-tenth of one
16 percent of the rate of deaths per year, so you have to
17 really consider the number of years before the present year,
18 have the probability of accident, calculate the contribution
19 to the deaths of this year, so --
20 DR. KRESS: I think you're right.
21 DR. APOSTOLAKIS: But have you done that? Have
22 you calculated latent risk?
23 MR. PALLA: Well, to be honest with you, we used
24 the number that the code calculates.
25 The code has a built-in -- it automatically
. 50
1 accounts for the population in that bin and it spits out the
2 number that is used to compare to the safety goal so we
3 would have to go look closer at how that is done.
4 DR. APOSTOLAKIS: I mean to calculate the latent
5 risk for this year, you have to consider the contribution to
6 the deaths this year from a number of years back and add
7 them up, because that is what the Commission says.
8 MR. PALLA: Well, we believe we have done it
9 consistently with how it's been done in the past. Now
10 whether that is correct or not, that's a different question.
11 DR. KRESS: I think, George, that what we have is
12 a fixed number for the number of cancer deaths per year that
13 are background, and this is a fixed number of deaths that
14 occur over a number of years --
15 DR. APOSTOLAKIS: Right.
16 DR. KRESS: -- and you divide by that number of
17 years to get a cancer per year contribution out of that even
18 though part of it is earlier and part of it is later.
19 DR. APOSTOLAKIS: Yes, but then you would have to
20 add the contributions from each past year.
21 DR. KRESS: No, you don't do that.
22 DR. APOSTOLAKIS: Why not?
23 DR. SEALE: You only die once.
24 DR. KRESS: That's not the way the safety goals is
25 written.
. 51
1 DR. APOSTOLAKIS: Well, can we find out how this
2 is calculated?
3 DR. WALLIS: I guess when you show us all these
4 numbers it would be helpful if you said if this number on
5 the right were 10 to the 9th we would be in trouble or
6 something and give us something to scale it by.
7 MR. SCHAPEROW: I have to put that out for Bob.
8 DR. WALLIS: How big does that number have to be
9 before you worry about it?
10 DR. KRESS: You had a graph that showed the risk
11 versus the safety goal.
12 MR. PALLA: Yes.
13 MR. SCHAPEROW: You are going to see a bunch of
14 them --
15 DR. WALLIS: So you are going to get to that.
16 Just it would help at this stage when you are showing us all
17 these numbers if you would put them in perspective some way.
18 CHAIRMAN POWERS: But this is a conditional thing.
19 How can you --
20 DR. KRESS: You can't do that here with -- it's a
21 condition.
22 MR. SCHAPEROW: These are just consequences.
23 DR. WALLIS: Being naive, I just see a big number
24 there -- gee whiz.
25 DR. KRESS: What they are trying to do is show us
. 52
1 the effects of these issues we had if you dealt with them.
2 DR. WALLIS: But if the number you need to get to
3 is 10 to the 9th then all this is irrelevant. Maybe it's
4 not.
5 MR. COLLINS: It's not irrelevant. What it is
6 showing is it is not as bad as would be a problem.
7 DR. WALLIS: So it is important to get that number
8 right on the right within --
9 MR. COLLINS: We have to have that number in order
10 to get to our comparison with the safety goals.
11 DR. WALLIS: Okay.
12 MR. SCHAPEROW: This is just a step along the way.
13 DR. WALLIS: Yes, but you are giving us an awful
14 lot of numbers and I want to know if they are important or
15 not.
16 MR. PALLA: One way to think of that number in the
17 right column is that if you looked at a severe reactor
18 accident and it will depend on the site, but you could be
19 looking at numbers on the order of two times 10 to the 6th
20 to two times 10 to the seventh.
21 DR. WALLIS: So this is comparable.
22 MR. PALLA: So that is the way that I would
23 perceive it.
24 DR. WALLIS: And someone has to know that.
25 MR. PALLA: It's relative, what we're looking at
. 53
1 here is the sensitivity of that number to the different
2 changes --
3 DR. WALLIS: So it is within the range?
4 MR. SCHAPEROW: That's correct. We are
5 releasing -- these are large release fractions.
6 These are reactor accident release fractions.
7 The benefit is the short-lived isotopes will have
8 decayed away.
9 These are one year of decay.
10 The two sensitivities shown here -- one is for a
11 small release fraction of ruthenium and the other is for a
12 large. We do see a consequence reduction in each case and
13 we go down to one core of fission products.
14 The consequence reduction is not quite as big a
15 reduction for the large ruthenium case again because of its
16 halflife and that a lot of it if not most of it is in the
17 final core.
18 DR. APOSTOLAKIS: So this is the mean value over
19 the what?
20 MR. SCHAPEROU: Over the weather.
21 DR. APOSTOLAKIS: Over the weather.
22 MR. SCHAPEROU: The model we have does a sampling
23 of the weather.
24 DR. APOSTOLAKIS: If I go back to 11 and I
25 consider the various decay times -- back to slide 11.
. 54
1 MR. SCHAPEROU: All right.
2 DR. APOSTOLAKIS: Are you finding also the mean of
3 these values, in other words, multiply 192 by the
4 probability that the accident will occur within 30 days,
5 then 162 by the probability is 90 days, and find that mean
6 value?
7 MR. SCHAPEROU: We use mean value for the
8 frequency, and these are mean values based on the weather.
9 DR. APOSTOLAKIS: Right. And I'm asking
10 whether --
11 MR. SCHAPEROU: So we just took a mean times the
12 mean in calculating the risk. We didn't propagate any
13 uncertain events.
14 DR. APOSTOLAKIS: No, but the total mean will be
15 the mean with respect to the weather, and the weight that
16 some of these numbers you have there depending -- the weight
17 is the probability that you will have 30-day decay time,
18 90-day decay time.
19 MR. SCHAPEROU: We calculated it discretely at 30
20 days and at 90 days and at -- you know, we maintain the
21 constant probability of the event. We did not go in and say
22 the likelihood of having a spent-fuel pool accident at 30
23 days is -- we didn't try to account for how it might change
24 with extended --
25 DR. WALLIS: The earthquake doesn't know the state
. 55
1 of the fuel pool.
2 MR. PALLA: It doesn't know how long it's been
3 there. But other -- I mean, you could -- in theory
4 accidents that occur later have longer times to uncovery --
5 boildown sequences, for example -- so if you were doing a
6 very rigorous analysis with fancy human-reliability models,
7 you might take some additional credits for that. And also
8 in the earlier times right after shutdown you would likely
9 have the same systems available that you had while the plant
10 was operating. You wouldn't have started to remove things,
11 so it's kind of a moving target, what you really have at the
12 site.
13 DR. APOSTOLAKIS: So the ultimate comparison is
14 you take say the 30 days, 192, multiplied it by its
15 frequency and compared it to the goal?
16 MR. PALLA: Yes.
17 DR. APOSTOLAKIS: So you compare each
18 individual --
19 MR. PALLA: But not that measure. I mean, we -- a
20 measure that's comparable to that. It's the risk to an
21 average individual of an early fatality.
22 DR. APOSTOLAKIS: Yes.
23 MR. PALLA: I'll explain that.
24 MR. SCHAPEROU: The other area I wanted to speak
25 for a few minutes about was on the plume. I have a short
. 56
1 presentation here on plume spreading and on plume heat
2 content.
3 On the plume spreading, max does use a gaussian
4 plume model with the amount of spreading determined by sigma
5 Y and sigma Z, model primaries which you are probably all
6 familiar with that type of model. As part of an
7 international cooperative effort on consequence assessment
8 codes, the experts in this area did provide updated values
9 for these parameters. This work was done over the last
10 several years. The experts provided distributions for these
11 two parameters instead of point estimates. We went forth
12 and we did perform max calculations based on sampling from
13 these distributions.
14 CHAIRMAN POWERS: Mr. Schaperou, my recollection
15 of this study was that the experts were asked what was the
16 amount of material deposited at particular points away from
17 the plume, and that the authors of the study subsequently
18 turned those into sigma Y and sigma Z, and they did that
19 because the experts utilized tools they had available to
20 them. Some of those tools were not gaussian plume models.
21 Am I incorrect in my recollection on this?
22 MR. SCHAPEROU: It's been too long since I've
23 looked over that work. It sounds reasonable.
24 CHAIRMAN POWERS: My recollection of what the
25 experts -- and actually the innovation in the way they did
. 57
1 their solicitation was to ask the experts to provide the
2 equivalent of a chi over q at particular sets of distances,
3 and there were three of them I think they asked them.
4 MR. SCHAPEROU: That sounds right.
5 CHAIRMAN POWERS: And then they said if that were
6 the case, how would I have to change the max model and its
7 sigma Y sigma Z things to get that result.
8 MR. SCHAPEROU: Yes.
9 CHAIRMAN POWERS: And that resulted in having
10 something that was directly useful.
11 MR. SCHAPEROU: That is my recollection, and the
12 point I was trying to make here was that they provided
13 distributions, and that made an additional level of
14 complication for our analysis. We were able to as I said
15 carry out a sampling and form the calculations. We saw a
16 decrease in early fatalities as a result of these updated
17 parameters. The updated parameters -- the experts basically
18 said you are going to have more spreading than the earlier
19 model.
20 DR. WALLIS: Could you tell me something about how
21 big the fire is? I mean, is this something like the fire
22 I'd get if I burned a pot of brush, or is it like a college
23 football celebration fire, or how big a fire is it? What
24 are we talking about?
25 MR. SCHAPEROU: I have some estimates on the next
. 58
1 slide --
2 DR. WALLIS: It's a pretty small fire, isn't it?
3 MR. SCHAPEROU: Of the heat content of this plume.
4 DR. WALLIS: Let's put it in perspective. Is it
5 like a car catching fire or something? How big is it?
6 MR. SCHAPEROU: I can't answer that question right
7 now.
8 MR. COLLINS: He's trying to be precise at this
9 time by giving you it in heat.
10 MR. SCHAPEROU: I can give you a number, but I
11 don't know -- I don't have a comparison available --
12 DR. WALLIS: So I have difficulty putting it in
13 perspective. But there isn't all that much --
14 CHAIRMAN POWERS: But could I give you an idea
15 of -- if you'll help me and tell me what you would like as a
16 metric on fire, maybe I could give you --
17 DR. WALLIS: Well, let's say it --
18 DR. KRESS: How many maple trees?
19 DR. WALLIS: Is it a ton of dry redwood, or is
20 it --
21 CHAIRMAN POWERS: No, it's a little bit bigger
22 than that.
23 MR. SIEBER: How about a gasoline truck?
24 DR. KRESS: It's about like a gasoline truck.
25 CHAIRMAN POWERS: It's different because of the
. 59
1 high specific energy of gasoline, but it would -- when you
2 burn these zirconia clad, they are enormous aerosol
3 producers because they're going into breakaway oxidation,
4 and so you would see an enormous cloud of smoke coming out,
5 but maybe a faint glow, okay? You would not see roaring
6 flames. You would not see --
7 DR. WALLIS: Not very energetic.
8 CHAIRMAN POWERS: At the actual point of reaction
9 it is very energetic, but it's unlike a fire with a volatile
10 substance. It's not spread out over a combustion zone.
11 DR. WALLIS: So it would look like a rather smoky
12 brushfire burning.
13 CHAIRMAN POWERS: Very, very smoky. It would
14 completely fill the zone of the fuel pool, which is now
15 presumed to be drained, and it would look very smoky.
16 @@ CHAIRMAN POWERS: And then after you'd worked on
17 the early removed fuel, then you would see a slow
18 propagation as you moved -- as the fuel became older and
19 older clad.
20 DR. WALLIS: Now that's another -- I don't want to
21 get into this too much, but I'm trying to envisage what
22 happens. It's a slow propagation; it's not a rapid ignition
23 of the whole.
24 DR. KRESS: The heat transfer is by radiation
25 probably, and it may spread to the other parts of the pool
. 60
1 pretty fast. That's one of the issues, is how do you make
2 this heat-transfer calculation.
3 DR. WALLIS: That's what I wonder, how the experts
4 can make all these estimates unless they know just what kind
5 of a fire it is.
6 DR. SEALE: Like a stack of old tires.
7 DR. KRESS: Well, they just assumed the full 3.5
8 cores are burning all at the same time.
9 DR. WALLIS: Yes, but how fast they'd burn always
10 makes a big difference to the plume.
11 DR. KRESS: Sure.
12 DR. WALLIS: Okay. Well, go on.
13 MR. SCHAPEROU: If it burns very slowly, then
14 you're going to be able to move people out and you won't get
15 the consequences. So we typically assumed calculations it
16 burns in 30 minutes. It's fairly quick.
17 DR. WALLIS: You assumed?
18 MR. SCHAPEROU: That's correct.
19 DR. WALLIS: Is there any basis for 30 minutes?
20 MR. SCHAPEROU: That is -- I believe it takes
21 about 20 minutes to consume a core in a large-break LOCA
22 type situation from steam oxidation, so it's --
23 DR. KRESS: Anywhere up to an hour.
24 MR. SCHAPEROU: This is basically motion of a
25 flame --
. 61
1 DR. WALLIS: So there is an analysis behind it.
2 MR. SCHAPEROU: Not a detailed analysis. This may
3 be closer to an assumption.
4 Because of the potential for the plume content to
5 be higher than in a reactor accident involving a large early
6 release, because of this direct burning of the fuel and
7 release of the heat by that mechanism, we did some
8 sensitivity calculations using different plume heat
9 contents, and we used the model that Tom described in a lot
10 of detail. The base case we use in most of our calculations
11 was the plume heat content from the NUREG --
12 DR. WALLIS: Are those megawatt-hours or days or
13 what? The heat content is megawatts times some time, isn't
14 it?
15 MR. SCHAPEROU: No, that is the product times time
16 it comes out -- it's divided by 30 minutes.
17 DR. WALLIS: Oh, megawatts -- so it's -- 21-1/2
18 megawatt-hours.
19 MR. SCHAPEROU: It's the heat release divided
20 by -- the joules divided by 30 minutes.
21 DR. WALLIS: So it's a rate of feed release.
22 MR. SCHAPEROU: We made a bounding estimate of
23 plume heat content for this 30-minute period based on
24 oxidation of the most recent core in 30 minutes, released
25 all of that heat. And that was 256 megawatts. We had a
. 62
1 more detailed estimate of the plume heat content based on
2 the amount of heat that would be absorbed into the fuel, and
3 at some point basically saying the oxidation's over because
4 the thing has collapsed on itself or melted or collapsed
5 somehow, and that was about 43 megawatts.
6 We did sensitivity calculations for this range,
7 from 3.7 megawatts up to 256 megawatts, and we got what we
8 felt was an expected trend, which was on the early
9 fatalities, we saw those come down, because you're lifting
10 the plume higher and away from the people close in, which
11 are the ones that would get killed in the early-fatalities
12 column, and societal dose we saw basically no change, just
13 slight changes. This calculation is out to 100 miles, and
14 this additional heat content would carry the plume out
15 further and perhaps put these consequences further out.
16 DR. KRESS: You might have had to go further than
17 100 miles.
18 MR. SCHAPEROU: That basically concludes my
19 presentation. I guess I have the -- this is the original,
20 the first slide I had, and I could just reiterate at this
21 point that we did perform a large number of calculations to
22 try to cover the important issues for spent-fuel pool
23 accidents.
24 DR. SEALE: You're up.
25 DR. WALLIS: I think you burned about three tons
. 63
1 of dry lumber.
2 DR. KRESS: It's like about --
3 DR. WALLIS: It's not a very big fire.
4 DR. KRESS: About four feet of one maple tree. I
5 think you must have missed the calculations.
6 DR. WALLIS: You make about 100 gallons of sap
7 with that much heat -- maple sugar.
8 MR. PALLA: Bob Palla in the Probabilistic Safety
9 Assessment Branch of NRR. What I want to talk about here is
10 briefly the integration of the level 1 frequencies,
11 frequencies of spent-fuel uncovery, integration of that with
12 the consequence calculations that Jason just described, and
13 we did this to both establish a baseline for comparison with
14 the safety goals and other, you know, operating reactor
15 risks and also we wanted to look at the implications of
16 changes to emergency preparedness requirements and, you
17 know, how they would impact the various sequences of
18 significance to spent-fuel pools.
19 So, what we did was, we essentially used the
20 frequencies of the various spent fuel pool accidents from
21 the Level I. And just to summarize, for seismic events,
22 with the Lawrence Livermore curves, we're talking about an
23 uncovery frequency of two times ten to the minus six per
24 year. This is a mean value.
25 And it's two times ten to the minus, if you use
. 64
1 the EPRI seismic hazard study. Now, this is the mean of the
2 values reported for the population of sites which were mean
3 values, so I believe it's a mean of the mean values that are
4 reported.
5 So, in essence, a factor of ten difference in the
6 frequency of seismic, and that mean value bounds
7 approximately 70 percent of the sites.
8 Now, for the cask-drop accidents, we looked at
9 heavy-load drop and basically the cask-drop accident is the
10 next highest frequency at two times ten to the minus seven
11 per year.
12 And then boildown accidents such as would occur
13 with extended loss of station power, these are slightly less
14 at 1.8 times ten to the minus seven.
15 So we used those frequencies as our point of
16 reference for the risk analysis, and we then coupled that
17 with consequences taken from the research study.
18 DR. APOSTOLAKIS: That's where I have a question.
19 I mean, you have the frequency of the sequence.
20 Now, the consequences, as Jason showed us, assume
21 various decay times.
22 MR. PALLA: Right.
23 DR. APOSTOLAKIS: How did you factor the
24 probability that the decay time will be 30 days versus ten
25 years?
. 65
1 MR. PALLA: Well, what we do -- and I'll show you
2 in a subsequent curve -- is, we calculated these risk
3 measures at each time interval.
4 DR. APOSTOLAKIS: But actually, the risk should be
5 lower, because the conditional probability of -- given the
6 event, having it occur within 30 days, you know, is not
7 accounted for. You just did sensitivity studies.
8 MR. PALLA: Glen, do you want to mention why that
9 frequency -- what we did is, we assumed the same annual
10 frequency constant throughout the ten-year period.
11 DR. APOSTOLAKIS: I understand that. The
12 frequency of the occurrence of the initiator in the sequence
13 is constant.
14 MR. PALLA: Right, okay.
15 DR. APOSTOLAKIS: But it depends very much on
16 whether it occurs within 30 minutes or ten years, according
17 to these consequences.
18 MR. KELLY: That's correct.
19 DR. APOSTOLAKIS: And that conditional
20 probability, I don't think is accounted for, which would
21 reduce the risk even further.
22 MR. KELLY: This is Glenn Kelly from the Staff.
23 Dr. Apostolakis is exactly correct. What he's saying is
24 that when you look at the consequence results and it shows
25 what the consequences are at 30 days, you can't then say,
. 66
1 well, okay, six years later, the consequences are still
2 going to be the same as they are at 30 days.
3 Those consequence at 30 days are only good at 30
4 days, and if you show the risk numbers, which is the
5 convolution of the frequency with the consequences, if you
6 -- you have to understand that if you're showing it at 30
7 days, in essence, when you consider that value, you have to
8 understand that that's only good for a little window around
9 30 days.
10 MR. PALLA: Right.
11 DR. APOSTOLAKIS: In other words, the window, the
12 time window within which the constant frequency initiating
13 has occurred.
14 MR. PALLA: It's just that these are a series of
15 point estimates that we made. That's just a linear
16 interpolation.
17 But we did these consequences, assuming that the
18 plant was shut down, either 30 days, 90 days, one year, two
19 years, five years, ten years.
20 DR. APOSTOLAKIS: I understand that.
21 MR. PALLA: These are discrete calculations,
22 discrete consequence results for those, and we combine them
23 with constant frequencies of occurrence of the accident.
24 DR. APOSTOLAKIS: So what saves you is that even
25 in the worst case, you're still below the goal.
. 67
1 MR. PALLA: That saves us.
2 DR. APOSTOLAKIS: Otherwise, we would have to do
3 --
4 DR. WALLIS: Well, he hasn't shown us yet that
5 he's saved.
6 MR. PALLA: Well --
7 DR. WALLIS: I'm still waiting for him to show us
8 that.
9 DR. APOSTOLAKIS: Is there an equation someplace
10 that shows how this calculation is done?
11 MR. PALLA: We don't have an equation in the
12 report. I think it's fairly straightforward. We did not
13 put one in there. We didn't think it was necessary, because
14 we have basically taken the product of the frequency and the
15 time-dependent consequences.
16 DR. APOSTOLAKIS: Yes, but --
17 MR. KELLY: This is Glen Kelly from the Staff. We
18 did not have an integrated risk calculation giving you the
19 total risk over what we would consider to be the total risk
20 over the time that you might have some exposure.
21 What we have shown is the risk at a point in time,
22 at 30 days, 60 days, and really probably from my standpoint,
23 the number that's most -- because we did the risk assessment
24 for the -- the frequency numbers were for one year.
25 And that's where we have the most confidence in
. 68
1 our frequency estimates. We held it then constant for
2 whether it was longer periods that the fuel has been out of
3 the reactor or shorter periods. We held it constant in
4 there, and if you want to, I can go into the reasons for
5 that.
6 DR. APOSTOLAKIS: Well, I guess if I look at Slide
7 11 again, from Jason, it seems that after two years, you
8 have a dramatic decrease, you really have significant early
9 fatalities up to two years, 192, 62, 77, 19, and after that,
10 it goes down to 1 --
11 MR. COLLINS: You lose the ruthenium.
12 MR. PALLA: It's the effect of the ruthenium.
13 DR. APOSTOLAKIS: I'm looking at the ruthenium,
14 yes.
15 MR. PALLA: Okay? Implicit in our calculation is
16 the fact that uncovery actually leads to a fire. We realize
17 that there could be, depending on the scenario, the chance
18 that that does not occur.
19 But we assumed in these scenarios that it did
20 occur.
21 DR. APOSTOLAKIS: That's pretty significant,
22 though. I think the numbers will go down significantly if
23 you did what I suggested.
24 MR. PALLA: Well, one of the reasons we did that
25 is -- and let me just -- this slide is -- this is in the
. 69
1 report. It's not in the package, but --
2 DR. APOSTOLAKIS: Do we have the report, by the
3 way?
4 MR. PALLA: Looking at the time after shutdown,
5 the number of hours that you have until reaching the
6 oxidation temperature at which fission products would be
7 released, is not substantially different for PWRs or BWRs,
8 or for that matter, whether it's air cooled or adiabatic.
9 So, there's a fair degree of latitude in this
10 area. And in most of the calculations that we were doing,
11 we're looking at the first five years. And in these
12 calculations, you would generally conclude that you're going
13 to end up with a fire, regardless of whether it was air
14 cooled or adiabatic.
15 It's only after you reach about four years in our
16 air cooled that you see the potential for remaining below
17 this runaway oxidation temperature.
18 Okay, we looked at the -- we had available to us,
19 consequence results for early evacuation and late
20 evacuation. And when I say early evacuation, what I mean is
21 that the evacuation is initiated and completed, prior to the
22 release of the plume.
23 Late evacuation, the plume is released and passes,
24 and then the doses are absorbed and then the evacuation
25 would occur later.
. 70
1 What we did is, we had those results also for what
2 we call the high ruthenium source term. It was NUREG 1465
3 source term, modified to include additional fraction, 75
4 percent, of ruthenium, and fuel fines of like 3.5 percent.
5 We also used -- and we called it the low ruthenium
6 source term, but it was the NUREG 1465 source term as a
7 point of reference.
8 And it's noteworthy that the source term that was
9 used in the original generic issues study, this NUREG
10 CR4982, of about a decade ago, if you used that it would be
11 quite a bit lower than the NUREG 1465 rule or low ruthenium
12 source term.
13 So we only used the 1465 source term and the
14 modified 1465 with the high ruthenium source term.
15 We looked at the sequences that contributed to
16 uncovery of spent fuel, considered whether the evacuation
17 that would result from emergency planning or ad hoc
18 measures, would be effective in these various sequences.
19 And that's really a question of whether timely
20 notification would occur and whether the things like in the
21 seismic sequence, whether the infrastructure is intact that
22 would enable notification of the population and the ability
23 to actually evacuate.
24 And timing is another area that is an important
25 factor. If you have sufficient time, tens of hours, for
. 71
1 example, you can achieve the same, effectively the same
2 level of dose savings as a formal emergency plan. You can
3 achieve that via ad hoc measures, just simply because you
4 have enough time to do these, the notification and the
5 evacuation on an ad hoc basis.
6 DR. KRESS: Now, when you talk about evacuation,
7 you're talking about out to ten miles?
8 MR. PALLA: Out to ten miles is what we had
9 modeled.
10 DR. KRESS: Now, clarify a little more for me,
11 late versus early. Early is before the plume even starts?
12 MR. PALLA: Yes.
13 DR. KRESS: And you have time to get the ten mile
14 people out.
15 MR. PALLA: Everyone would start moving, I
16 believe, at the same time.
17 DR. KRESS: I'm a little confused on late, though,
18 because you said it was after the plume is gone. But the
19 longer you leave people in there, even after the plume is
20 gone, the more dose they'll get.
21 MR. PALLA: That's true.
22 DR. KRESS: So it does depend on a specific time.
23 And is that 24 hours?
24 MR. PALLA: No, it wasn't; it was --
25 MR. SCHAPEROU: It was the same assumption that
. 72
1 was assumption that was used in the large early release
2 calculations in NUREG 1150; that is, about an hour after the
3 plume has ended, the early plume has ended, people are
4 evacuated.
5 DR. KRESS: That's call late evacuation?
6 MR. SCHAPEROU: That's correct. That's what we're
7 calling our late evacuation case. That's the one where
8 everything happens very quickly in the reactor case, and in
9 about an hour after that, then you start moving people out.
10 MR. PALLA: I think it was the timing that you'd
11 have for like an interfacing system LOCA type of reactor
12 accident. You're starting everything right at time zero,
13 but it just takes time to make the notification and get
14 people moving.
15 In our model, I believe it was a radial evacuation
16 in all directions that started at all locations at the same
17 time and proceeded radially outward. I think in actual
18 reactor accidents, there might be more of a focused, keyway
19 type of an evacuation where, depending on wind directions --
20 DR. KRESS: You did this all for the Surrey site?
21 MR. PALLA: The Surrey site, yes.
22 Okay, and what we've done, I've summarized on the
23 last bullet, and described in more detail on the next slide,
24 for seismic events, we looked, for starters, at how -- what
25 was done in NUREG 1150.
. 73
1 There, they classified seismic events as either
2 high-G value or low-G value. For the low-G value
3 earthquakes, they assumed that the evacuation start time was
4 delayed and the speed was, I think, half the speed, so it
5 took longer to complete it.
6 And then for the high-G earthquakes, they said
7 there would be no effective evacuation, and those people
8 were basically left there for 24 hours and then relocated.
9 We think that the seismic events that would
10 threaten the pool would clearly be of the high-G value type.
11 The .6 G was -- peak ground motion, was used as the point of
12 demarcation in 1150,a nd we feel that looking at the
13 fragilities for the spent fuel pools, we'd have to have G
14 values of that value or higher in order to get spent fuel
15 pool failures.
16 DR. KRESS: Was the spent fuel pool is designed to
17 a safe shutdown earthquake for the site?
18 MR. PALLA: I believe it's seismically designed to
19 the -- yes, I would say yes, and maybe Glen or Goutam could
20 answer that, if I'm not right. But it's pretty robust.
21 There are also shielding considerations that come
22 into play, so it may be even more robust than it would need
23 to be, just to meet the seismic criteria, design criteria.
24 But it would take a substantial earthquake.
25 DR. KRESS: But you did some sort of fragility
. 74
1 estimate and came up with the .6?
2 MR. PALLA: Yes.
3 DR. KRESS: Are the spent fuel pools enough alike
4 that that's sort of generic number?
5 MR. KELLY: This is Glen Kelly from the Staff.
6 When we -- we don't have information on all of the
7 individual pools to make a determination generically. So
8 working with the industry, what we did was, in order to come
9 up with kind of a minimum capacity of the pools, we -- the
10 industry proposed that we enhance the checklist, the seismic
11 checklist where if a plant passes the checklist, we feel
12 that it would have a capacity, at least as high as 1.2 G
13 spectral acceleration.
14 Now, that's equal to about .5 G peak ground
15 acceleration. And we believe that that's -- many pools may
16 be considerably more robust than that, but by meeting this
17 checklist, we can at least assure that. And that's about
18 the same level that we would consider to be high G level,
19 and therefore you're not going to have much infrastructure
20 left.
21 DR. KRESS: Okay.
22 MR. PALLA: We also looked at some of the previous
23 Commission decisions regarding the need for emergency
24 preparedness to deal specifically with seismic events, and
25 the previous Commission decisions reached were that the
. 75
1 value of emergency preparedness in seismic events would have
2 a marginal benefit, because of the extensive damage offsite
3 to structures, bridges, roads, all of those elements that
4 are needed for effective evacuation.
5 And we then also consulted with our expert, Dr.
6 Kennedy, and he essentially confirmed these judgments that
7 there would be minimal impact of emergency planning in a
8 large seismic event of the type that we're talking about to
9 damage the pool.
10 DR. KRESS: Yes.
11 MR. PALLA: So, on that basis, we, in our
12 analysis, assumed that or rationalized that there wouldn't
13 be any effective taking away any of these offsite planning
14 requirements, so that the delta, with and without EP,
15 offsite would be negligible in our analysis.
16 DR. KRESS: Now --
17 CHAIRMAN POWERS: Dr. Kress, I wonder how long
18 we're going to -- we're doing damage to the schedule here.
19 DR. KRESS: I think we're only about two-thirds of
20 the way through. Would that be a good estimate.
21 DR. WALLIS: I'd like to see the bottom line.
22 MR. PALLA: Okay. There are a few plots in your
23 package that give you the risk values. I don't know that I
24 need to show them.
25 I could go right to the risk conclusion slides,
. 76
1 and you can refer to the figures, if you wish to confirm
2 them. I'm going to flash this figure pretty quickly, and
3 just indicate this is the contribution to the risk profile
4 from the castroph accident. The castroph accident is the
5 only accident that we considered to be impacted by emergency
6 planning. And what the top figure -- the top line there is
7 a consequence result if you had late evacuation. And the
8 bottom line is the risk if you would have early evacuation.
9 And the difference that -- we have a line that starts at the
10 top, the dotted line that drops quickly down to the lower
11 line as if you had full EP. We assumed that you had to have
12 in excess of at least an hour to -- I think maybe it was
13 four hours to--
14 MR. UHRIG: Four to five hours.
15 MR. PALLA: Four to five hours to have -- even
16 with full EP, to be effective to complete the evacuation.
17 So, but once you had that, and we think you have that at
18 about 90 -- at 90 days you don't quite have it, but at one
19 year, you do have it. This is just a linear interpolation.
20 And so we transfer down to the early evacuation curve after
21 the one year, at one year and beyond.
22 With--
23 MR. WALLIS: These are all details. I'm trying to
24 -- does it matter on what's on below, some number doesn't
25 matter. So, what is the scale on the left tell me? Is it
. 77
1 -- one ten thousandths of a person dying.
2 MR. PALLA: Yeah, that doesn't tell you a whole
3 lot.
4 MR. WALLIS: Is that insignificant or not?
5 MR. PALLA: No, it's not significant when you look
6 at it in terms of the overall picture. Now, this--
7 MR. WALLIS: It's about the same.
8 MR. PALLA: This is the roll up of the risk--
9 MR. WALLIS: Well, that's bigger.
10 MR. PALLA: From all of the contributors--the
11 seismic combined with the cask drop. In that previous curve
12 that I showed you it is reduced to the dotted -- the
13 difference between the solid line and the dotted line on
14 each of these curves.
15 And what we're showing here is we have the
16 Livermore seismic hazard curve with a high source term and
17 with the low source term. And we've got the EPRI seismic
18 half curves.
19 MR. WALLIS: And how big do these numbers have to
20 be before you worry. I mean, it's 1 e to the minus 3
21 important?
22 MR. PALLA: The purpose of this curve is just to
23 show how it compared with the risk levels from operating
24 plants.
25 MR. WALLIS: But does that make me secure or not?
. 78
1 How does it compare with some Commission criterion or
2 something?
3 MR. PALLA: Okay, let me -- before I leave the
4 curve, let me just say that early on with the highest source
5 term and the highest seismic, you're in the range of
6 operating reactors. If you had the lower--
7 MR. WALLIS: Well, that may tell me the peach
8 bottom's in trouble. I don't know. What's the criterion
9 for success?
10 MR. PALLA: You want me to go faster.
11 DR. KRESS: Have patience. We'll get there.
12 DR. APOSTOLAKIS: Still, though, it seems to me
13 that these are overestimates by a factor of four to ten in
14 the early years.
15 MR. PALLA: Which ones?
16 DR. APOSTOLAKIS: All of them. Because you have
17 not included the time window.
18 DR. KRESS: But there's other conservatisms, also.
19 MR. PALLA: But the time window is the--
20 DR. APOSTOLAKIS: Yeah, but the time window we're
21 doing in the shut down and low power.
22 DR. KRESS: There's conservatisms in seismic--
23 DR. APOSTOLAKIS: Situations. That's what brings
24 the risk down.
25 MR. WALLIS: See he has to average these in some
. 79
1 way or he has to weight these in some way.
2 DR. APOSTOLAKIS: That's right, because you see --
3 let's say I'm concerned only about two years. There is a
4 probability that the initiator will occur in those two
5 years. The frequency of the initiator times two. Then it's
6 equally likely you will look at anywhere, and what you're
7 really interested in is the 30 days or the 90 days. So if I
8 take that interval and divide by 365 times 2, I really get a
9 low number.
10 MR. WALLIS: So you just average these curves?
11 DR. KRESS: You don't want to do that because
12 these things are looking at options on the status of the
13 plant. So you're -- you can't compare one with the other
14 because you've got a different plant up to the 30 days, and
15 then it changes to a different status after that. So you
16 can't do just what you're saying.
17 DR. APOSTOLAKIS: No, but the occurrence of the
18 initiator is important.
19 CHAIRMAN POWERS: I think this is a thing that
20 will have to rest on our discussion. We have -- we are now
21 over a half an hour behind. Please conclude in two minutes.
22 MR. PALLA: Okay.
23 DR. KRESS: Please put the dotted lines up there,
24 Professor Wallace.
25 MR. PALLA: Yeah, I'm going to show you the
. 80
1 comparison to the safety goal. Here's what everybody's
2 waiting for. Okay, the point -- these -- the top solid
3 curve and the dotted line just below it is for the highest
4 -- high seismic hazard, high source term. The low curve is
5 the low seismic hazard and the low remedial source term.
6 And you can postulate other cases in between that are
7 different combinations. The point here -- this is early
8 fatalities. We're about a decade below the safety goals
9 even early on. And this increases as time goes on due to
10 the decay of fission products. Ruthenium is one of the low
11 -- this is a five-year time period, and you can see the drop
12 off. You're probably about a decade lower if you're at --
13 with the low source term and the low Ruthenium -- low
14 ruthenium and low seismic hazards.
15 MR. WALLIS: Now, why didn't you show that two
16 hours ago?
17 CHAIRMAN POWERS: I don't think it's germane. Why
18 he didn't. We need to conclude.
19 MR. WALLIS: But do we have to go through all
20 those perturbations for the--
21 CHAIRMAN POWERS: We can discuss that later.
22 MR. WALLIS: Okay.
23 MR. PALLA: I'm just trying to show you the range
24 that these results are coming out under with this -- with
25 the assumptions that we made. One of the key -- this is
. 81
1 latent cancer fatalities. The curves are more flat. These
2 are driven by others -- longer term -- long-lived
3 radionuclides. You can see in both the curves the
4 difference between the dotted line and solid lines are
5 really the effects of the EP relaxations that we were
6 talking about. It's quite a bit smaller. It doesn't really
7 affect the margins of the safety goal at all.
8 CHAIRMAN POWERS: I think you've satisfied.
9 MR. PALLA: I think that's the bottom line of the
10 whole thing is that we're -- in the range of operating
11 reactors early on and decreasing to various degrees
12 depending on assumptions. Assumptions on seismic could know
13 you down an order of magnitude. Likewise for source terms.
14 You -- combined effects of lower seismic hazards and low
15 source terms could give you a couple decades reduction, and
16 you've got substantial margins to the safety goals. I won't
17 even get into the slides that I've got. But what -- on the
18 comparison to the safety -- to Reg Guide 1174, but we did
19 look at each of the key safety principles--defense and
20 depth, and margins, et cetera. We found that -- we feel
21 that we have adequate defense and depth with the ad hoc
22 measures, given the extensive amounts of time that would
23 exist. And margins are substantial, but -- and margins
24 would be retained even when one would relax the off-site
25 requirements.
. 82
1 DR. KRESS: Dr. Powers, on the agenda, we had
2 about 20 minutes for a representative of the Nuclear Energy
3 Institute to give us their perspective, and another 15 or 20
4 for a representative of the Institute for Resource and
5 Security Studies.
6 CHAIRMAN POWERS: Yeah, I'm anxious to get those
7 perspectives.
8 DR. KRESS: Okay.
9 CHAIRMAN POWERS: So let's please continue on.
10 DR. KRESS: So we could move to the -- I think we
11 can move to that part of it now, and the Nuclear Energy
12 Institute is first.
13 CHAIRMAN POWERS: And I -- the speakers for
14 Resource Security Studies and the Nuclear Energy Institute
15 ought not feel pressured by time.
16 DR. KRESS: Thank you.
17 MS. LYNETTE: You're a gentleman and a scholar,
18 Dana.
19 CHAIRMAN POWERS: Flattery will get you anywhere
20 you want to.
21 MS. LYNETTE: Alright.
22 DR. APOSTOLAKIS: She didn't say that on the
23 record, though.
24 DR. KRESS: We're very familiar with Lynette
25 because she's been here before, but who's this stranger you
. 83
1 have with you?
2 MS. LYNETTE: Dr. Robert Henry with Salsky and
3 Associates. He's--
4 MR. SEALE: Bad company.
5 MS. LYNETTE: He's -- but he's here to give us
6 some good insights on the progression of the event, the
7 thermodynamics and some of the considerations that were
8 spoken about earlier. I understand we have between us about
9 half an hour, and I'm going to start with Dr. Henry, because
10 I think they'll probably be a lot of discussion about what
11 he has to stay.
12 MR. HENRY: Tom, maybe if I got a little hair
13 dyed, you might remember me.
14 DR. KRESS: Well, welcome back, Bob. How's Notre
15 Dame doing this year?
16 CHAIRMAN POWERS: Could have talked all day
17 without bringing that up.
18 MR. HENRY: Let's get to something that I can
19 discuss right now.
20 CHAIRMAN POWERS: This is our first test. He
21 qualifies as an engineer.
22 MR. HENRY: I will try to keep on schedule, Dana,
23 and even though I--
24 CHAIRMAN POWERS: Don't feel pressed. We just
25 forced Dr. Kress to stay after school.
. 84
1 MR. HENRY: I'd just like to discuss a few of the
2 issues that have been touched on in various aspects this
3 morning that would relate to the thermohydraulic response of
4 the pool under these fire conditions or whatever conditions
5 are postulated under these very low probability events.
6 The major points that -- the first four are pretty
7 straightforward. You've gone through these before. But
8 certainly given an accident with the loss of heat removal,
9 when you analyze these, you'd like to be able to analyze
10 them in the most realistic manner you can. So you like to
11 understand where you think the failures might be or the
12 spectrum of failures to be addressed.
13 Certainly with a -- if I stand up, you can't --
14 can you hear me?
15 CHAIRMAN POWERS: Well, no. We can give you a
16 mobile microphone, if you'd like.
17 MR. HENRY: Yeah, I'm more comfortable standing
18 up. I can see that.
19 CHAIRMAN POWERS: Yeah. Of the old schools.
20 MR. HENRY: You guys get to do this five days a
21 week. I only get to do it here. Right, George?
22 DR. APOSTOLAKIS: I don't know about the five
23 days.
24 MR. WALLIS: You're actually a bigger target when
25 you're standing up.
. 85
1 MR. HENRY: Graham, you're usually so accurate,
2 that's not an issue. Certainly with the spent fuel pool
3 inventory, when the pool is adequately cooled and you have
4 days, of course, even without a pool cooling function. The
5 water level decreases sufficiently to uncover the top of the
6 fuel bundles as a result of the accident condition. The
7 heat removed by boiling and steam flow is important, but the
8 power distribution really is not very important throughout
9 the pool and we'll touch briefly on that.
10 If, in the sequence of these low probability
11 events, it's assumed that the pool is going to eventually
12 dry out, then the fuel bundle configuration is somewhat
13 influential, and we'll talk briefly about that. Key to this
14 I think is that if the fuel pins become sufficiently hot
15 that oxidation becomes an event. The rate of oxidation is
16 comparable to the decay power, then obviously you're going
17 to be driven by the chemical energy, which was the point
18 made earlier.
19 But that we'll get to that. Just doing some
20 simple hand calculations tells you that that's going to look
21 pretty much like an in power or an at power case, except in
22 somewhat different time scale, and particularly that the
23 Zirkoy reaction is the thing that's going to get her the
24 oxygen, and that you would expect the same kind of geometry
25 changes just like you would have seen in analyzing cores
. 86
1 under normal -- in reactor accident conditions, in something
2 maybe like TMI in terms of relocation, in terms of what
3 limits the oxidation, and in terms of minimizing the
4 interfacial areas.
5 I'd like to touch briefly on those in a second.
6 When we do these evaluations, evaluations always are
7 beneficial if we can identify some kind of mechanistic
8 failure, so you understand exactly what it is that you're
9 addressing. It certainly helps the analysis, and it helps
10 these kinds of discussions. The evaluations should also
11 look at the results of potential recovery actions because
12 this is the other place that time really comes into play,
13 and we talked about time being influential with respect to
14 both EP and fission product decay. But this is the place
15 that has the greatest bang for the buck, if you will.
16 The evaluations should also obviously consider all
17 the cooling mechanisms and the energy generation mechanisms,
18 including natural circulation, which we'll talk about in
19 just a second.
20 The focus for analysis obviously different from
21 at power conditions. We're dealing with something which is
22 atmospheric pressure. The flow through the assemblies is
23 laminar, which means that we're not particularly limited by
24 any kind of changes in resistances. They're very well
25 characterized by standard representations. It's like 64,
. 87
1 the Reynolds number.
2 It is important to at least consider the fact that
3 there are openings in individual fuel assemblies that you
4 don't necessarily see in the core, and I'll show you a
5 couple of those. But they're really principally second
6 order in nature. But it should always be remembered that
7 they're there, because they are favored flow paths through
8 the assembly.
9 The fuel assembly distribution within the pool
10 isn't going to matter, as I said earlier, when there's a
11 reasonable amount of water in the pool, roughly if the core
12 is covered to about 70 percent of the fuel assembly height,
13 and I'll show you where that comes from.
14 And the fuel assembly distribution does matter if
15 we start dropping down below that.
16 I think this particular one is slightly out of
17 order, but you have this. This just shows you what -- for
18 PWR fuel assemblies--
19 MR. WALLIS: Do we have any of these, excuse me?
20 I don't think we have any of your slides.
21 MR. HENRY: I gave copies. I don't. Maybe they
22 got all added together. There's two sets.
23 LYNETTE: Yeah, yeah. I think there were two
24 separate handouts.
25 MR. HENRY: There's two separate handouts. That's
. 88
1 probably.
2 MR. WALLIS: Well, we got one very skinny one.
3 MR. HENRY: And there's--
4 MS. LYNETTE: We'll be making some more.
5 CHAIRMAN POWERS: It's clear so far.
6 MR. HENRY: I didn't think it would challenge you
7 guys too much.
8 MS. LYNETTE: Thanks, Graham.
9 MR. HENRY: Now we'll get to the guts of the
10 issue. Anyway, when we -- when these assemblies go into a
11 pool, of course, these holes that are there for the control
12 rods are open, so they are favored flow paths, because they
13 have roughly the hydraulic, twice the hydraulic diameter as
14 those between the fuel pins. So when we deal with laminar
15 natural convection, these are favored flow paths, as is the
16 central path, which is there for the in-core instrument,
17 which also doesn't have anything in it in the storage pool.
18 MR. SIEBER: We used to put burnable poisons and
19 flow limiters and source assemblies in those holes as a way
20 to store them so they aren't in every case open.
21 MR. SHACK: You're too big to have you stand in
22 front of the screen.
23 MR. HENRY: I'm sorry. I used to be slightly
24 thinner. No, and those are the specific things that need to
25 get looked at if those are practice for a particular plant.
. 89
1 And the other part that I didn't have any handy
2 information on is when you get into a fuel storage pool what
3 is the gap between the assemblies because obviously that has
4 some meaning, and we'll come to it in just a minute.
5 They're not quite as dense as they are in the
6 core. So just for some numbers. Am I in your way, Bob?
7 Are you able to see okay?
8 MR. SEALE: I got you.
9 MR. HENRY: Okay. When we look at the response to
10 boil down, if we assume we have an average power of about
11 five kilowatts per assembly, and that's averaged over an
12 entire pool that maybe has a thousand fuel assemblies in it.
13 So we're dealing with the power of something like five
14 megawatts, and the pool is something in the range of eight
15 meters by seven meters, then the boil down rate, when the
16 water level is above the fuel is going to be somewhere in
17 the range of about five and half, 14 inches per hour, 14
18 centimeters per hour. Once it gets into the fuel, of
19 course, then we have a reduced cross sectional area, so it
20 roughly doubles. But it only takes about 35 gallons a
21 minute being added to obviously turn that around and stop
22 the level decrease. And that's always, of course, the thing
23 to be remembered. As I say, the -- there's the element
24 where time comes into play, and that's going to clearly give
25 you the most protection against any kind of release.
. 90
1 If we look at those conditions where the water is
2 partially boiled down in the pool, and we make some very
3 standard assumptions. First that it's quasi steady, making
4 it easy to analyze. And we're only dealing with steam and
5 water in the core. The inlet water is at saturation
6 temperature, TSAT. Decay and heat is QD, and we'll assume
7 for the time being it's constant over the fuel pin length.
8 The collapsed water level is used to represent the covered
9 part of the fuel assemblies, and we're looking for
10 temperatures might remain low enough before steam Zirc
11 oxidation begins. This results in a very simple expression
12 that the outlet temperature minus the inlet temperature has
13 nothing to do with power anymore. And that's a nice
14 convenient way -- I'll flip back and forth between these two
15 -- but what we're really looking at is -- that as we start
16 uncovering part of the top of the fuel, the amount of decay
17 heat, the steaming rate locally for higher fuel assembly, of
18 course, is higher, so it's cooling more of the assembly up
19 here. And, as I said, the power drops out of the equation,
20 so we can look at the peak outlet temperature as a function
21 of how much of the core or how much of the fuel assemblies
22 would be uncovered. And that is again this expression here.
23 And if we look at the results of that--
24 MR. WALLIS: This is a steady state. All the heat
25 goes into the steam, is that right?
. 91
1 CHAIRMAN POWERS: Yeah.
2 MR. HENRY: Steady state. All the heat below the
3 water level goes into creating the steam, and everything
4 goes into the steam thereafter, right.
5 So you come to this conclusion that if you set the
6 escalation to be somewhere in the range of about a thousand
7 centigrade, or if you want to make it slightly less, it
8 doesn't matter that much. We're talking about something in
9 the range of 60 or 70 percent of the pool needs to be
10 covered to keep that temperature at that level.
11 Of course, as Graham just said, this is a steady
12 state evaluation, so as you deviate from this, the rate at
13 which you get to that is fairly long for these very low
14 power conditions. But, again, as was discussed earlier, if
15 you say that nothing can done, sooner or later you would get
16 to this kind of condition. So as long as this much of the
17 pool is covered, then, in essence, the steam is sufficient
18 to keep everything cool. And this gets us to something
19 we'll come back to in a minute. If there's any part of the
20 pool that's covered, that the bundle has an easier time, or
21 the bundles have an easier time cooling themselves than if
22 it's just all air. So the issue of having any partial
23 blockage due to water in this estimation says that's really
24 a bogus issue. Any water, just like going back to accident
25 management days, water's good, right. So however much of
. 92
1 the bundle is covered, that additional amount of steam will
2 take some heat away because the entire bundle is uncovered,
3 and because steam is lighter than air, this accentuates the
4 natural convection. So that was just an issue that was
5 mentioned earlier, I don't think that there's--
6 MR. WALLIS: The steam velocities are very
7 moderate, aren't they?
8 MR. HENRY: Very moderate, yes.
9 MR. WALLIS: What was -- such as?
10 MR. HENRY: They're the order of a meter per
11 second, but to specific more like half a meter per second.
12 Reynolds numbers you're typically down in the range of a
13 couple hundred.
14 And what this gets me then to is again, a very
15 simple -- I always like to try and put some things in very
16 simple terms, but make sure we understand what it is that
17 drives the bus. If we just look at natural convection
18 cooling by air, and air alone, so that whatever we say the
19 flow path is, we can have a simple relationship between the
20 velocity and the operating delta P, and that's just due to
21 the density difference coming from the temperature.
22 And put two different lengths in here, as we'll
23 get to in a second, this is the height over which this delta
24 row applies. And this is the length over which the flow is
25 going through the bundle. So if there's some kind of
. 93
1 relatively open down comer if you will on the outside, then,
2 in essence, this length and this length are the same and
3 cancel out. If you want to consider the fact that there may
4 not be any significant down comer on the outside, but in
5 essence you have to go down one assembly and up the other
6 assembly. So now the length is twice what the height is.
7 And we could at least get some kind of perspective of how
8 much cooling we could get.
9 As I said earlier, the resistance is well
10 characterized. The delta row, which is operating as an
11 average point of view, the maximum delta row or T-out minus
12 T-in driving the bus, divided by two. The decay heat that
13 has to be removed -- and we're just talking about decay heat
14 here. Of course, it's coming from a straightforward energy
15 balance, it gives us the maximum temperature difference.
16 Assembling all these we get an expression for the maximum
17 temperature difference, the outlet minus the inlet. The
18 average temperature, and then a bunch of things -- they're
19 both to the one-half power, and the only reason I grouped
20 these together is these really are governed by how you
21 defined your boundary conditions. These aren't going to
22 change. Kinematic viscosity. The pressure. Gas constant,
23 et cetera.
24 As I mentioned earlier, this is the length over
25 which the flow paths [sic] has to go down and back up again.
. 94
1 You have basically twice the length of the bundle. The
2 height is the bundle. So if that's the flow path you chose,
3 this is two. But if you do that, then this is also half the
4 flow area. So when you make this assumptions going down one
5 set and up the other, then that changes this whole term by a
6 factor of four, or it can take the square root of a factor
7 of two.
8 So when you put in some reasonable numbers, as I
9 had in the pool before, since it was five kilowatts per
10 assembly, then these kinds of temperatures if you assume
11 there's no down comer, so it's going down one set of
12 assemblies and up the other. Again, when you finally get
13 down to the bottom of these low probability events, this
14 kind of temperature increase is more than a thousand degrees
15 centigrade. So when you get to that level, you know that
16 the oxidation is very strong. So it's going to drive the
17 bus. This power no longer is just limited to decay heat.
18 So when you take it down to the end of those kinds of
19 assumptions, which got you to this uncovered situation in
20 the first place, then this starts looking like what you
21 always analyzed for the at power conditions following scram
22 and eventual boil down of the core. Chemical energy is
23 driving the bus. You typically find then that you're now
24 limited by how much air you can get into the assembly. So
25 now you're oxygen starved by the air flow. And when you
. 95
1 take that back into this equation, you find that this kind
2 of fire is, as you were asking earlier, probably last
3 somewhere -- if you could get the air flow to it, which
4 means that the surface area doesn't change, it would take
5 you a few hours for it to burn, to oxidize. But you clearly
6 get to the temperatures well in advance of that that the
7 geometry begins to change. So now it begins to liquify
8 because of the interaction, the Uteknicator interaction
9 between molten Zirc and EO2, relocate, melt. So now the
10 surface area for heat transfer degrades roughly two orders
11 of magnitude. So when you take this through, it says you
12 can't get -- you really can't get much natural cooling by
13 air. But it also tells you that at this level, you're not
14 going to be able to get oxygen through this thing to oxidize
15 everything without the whole thing basically turning into a
16 TMI looking like configuration well before you oxidize all
17 the Zirc. So what this also looks like then to me is using
18 the kind of Ruthenium source terms we talked this morning
19 may be useable regulatory space from our perspective.
20 They're certainly look like they are very, very
21 conservative, because the Zirc is there to gather the
22 oxygen. I don't think you could ever find this to oxidize
23 all the Zirc. So.
24 CHAIRMAN POWERS: Dr. Henry, with this kind of
25 intense oxidation and given that the fuel is relatively old
. 96
1 in the sense that it probably has oxide layers both on the
2 inside and the outside, do you think that when you get to
3 liquefaction of the remaining metal, that it homogenizes
4 with the fuel or just drains down and leaves a -- perhaps a
5 zirconium skull clad fuel stack?
6 MR. HENRY: No, I think, again, we would see
7 pretty much what we've seen in TMI and what we've seen in a
8 lot of the core tests that the zirc is going to principally
9 take the -- most of the O2 with it. It would be very
10 unusual conditions to find unclad pellets still sitting
11 there. But whenever it goes out, it's also going to
12 preclude the -- start shutting off the circulation paths for
13 the air coming in.
14 CHAIRMAN POWERS: What I was thinking of is that
15 the zirconium metal presumably is oxygen saturated, maybe
16 from the start, but certainly very quickly gets oxygen
17 saturated.
18 MR. HENRY: At it's neutral temperature, yeah.
19 CHAIRMAN POWERS: And it's got an oxide film
20 between it and the fuel itself; that how you dissolve that
21 oxide to get to the fuel to do the homogenization?
22 MR. HENRY: Dana, the place I would start with
23 this, of course, and just like we always do, this isn't
24 enough to address the whole issue, but this is why I've
25 tried to put it in my own perspective. I'd say, now, let's
. 97
1 go to technical basis. Let's assemble the whole technical
2 basis. A key part of that scaling, as we know. So that the
3 place I always start, because I like to work backwards from,
4 is what I understood from TMI, and work back to the rest of
5 it. And it's too long to discussions today, but certainly
6 looking at these things related to -- as an example, the
7 CODEX test, and I see people are going to also be looking at
8 some PHEBUS tests. That's one of the best places. They're
9 not going to exactly address perhaps the fuel
10 characterization actually, but they certainly went to great
11 pains to try and take care of the zirc characterizations you
12 just mentioned. And certainly the CODEX tests, this is from
13 what I could gather from the relatively short write up --
14 they principally saw things which were oxidation -- a lot of
15 oxidation, but they made the statement from their
16 perspective, it was also limited by the supply. And once
17 that's the case, then that pretty much tells me that we've
18 changed the available area for oxidation substantially. But
19 I understand your question. I think there's where you would
20 really build the technical basis to look at these things.
21 DR. KRESS: Your main point here is that the
22 zirconium as it burns is there gathering the oxygen, and
23 oxygen can't get to the ruthenium to volatilize it. And
24 then when the zirconium is gone, then there's not heat
25 source to drive the volatilization so that these high
. 98
1 ruthenium releases that were observed by the Canadians how
2 did they come about?
3 MR. HENRY: Again, that's part of the technical
4 basis I would include here. But that really goes back, Tom,
5 to, you know, the things that everybody's talked about in
6 the past. That's a scaling issue; that if you start -- if
7 you give me something that's fairly small to begin with--
8 DR. KRESS: A small piece.
9 MR. HENRY: I can make sure I oxidize everything.
10 CHAIRMAN POWERS: I think you have to be very
11 careful. With the Canadian tests, they were done with
12 Canadian CANDU fuel that has clad that's only about a third
13 the thickness of U.S. clad. So they oxidized the cladding
14 probably before you can get any of this homogenization or
15 any kind of relocation. And it's a minor perturbation on
16 the test.
17 Now, Dr. Henry mentioned the PHEBUS test. I think
18 that's the crucial question that they're trying to address
19 is if your zirconium is there getting the oxygen, you don't
20 get the oxidative release. If the zirconium drains away,
21 the way they've seen in some of the tests and probably have
22 heard some to NTMI, then you're exposing -- I hate to call
23 it stacks, but piles of unclad fuel to oxygen sort of
24 bypassing this, and that's what he's saying is that once
25 you've done this analysis, now you've got to go look at a
. 99
1 more detailed and complicated question that's not so easy to
2 do, to see if you're getting this trade off.
3 DR. KRESS: I worry a little, Bob, about fuel is
4 about 12 feet long, and I don't know how homogeneous this
5 process is; that you may do what you say along one portion
6 of it, and expose fuel to the oxygen and still have
7 zirconium there with parts of it to drive the energy. So,
8 you know, it's a question of homogeneity of that whole
9 process. If you had a major earthquake, which initiated
10 this whole event, these stacks are going to be rattled with
11 aftershocks.
12 @@ DR. KRESS: Yeah, that may be -- that would be
13 another issue I think, yeah, as far as the geometry I think.
14 MR. HENRY: I think the -- I agree with what you
15 said before, Tom, exactly -- except that there was one other
16 point you should put in this. As part of the zirc
17 oxidation, you substantially change the geometry to limit
18 the surface area. And that's the case whether you start off
19 with all these rods intact and vertical or if they have been
20 shaken down into a bunch of whatever, you still come to the
21 same conclusion. You're going to start -- you can't undergo
22 all this oxidation and keep that surface area intact, and
23 particularly if you go to a strong seismic event, and you
24 say it's a pile of rubble to begin with, it's going to be an
25 even tighter pile of rubble pretty quickly thereafter. If
. 100
1 it's not a pile of rubble, then, again, you need more than
2 just these simple equations, but if we start telling you as
3 it starts burning from the top down, as the power -- the
4 power distribution doesn't matter, both in the fuel and
5 within the pool, and again, it also tells you, as I
6 mentioned earlier, you have time to get water on it, and
7 water is the savior. And if it's partially covered by
8 water, it's better than not having any at all. It's not the
9 other way around. So that -- the last part is if you take
10 me down to that end of the spectrum condition, then I think
11 you have a hard time getting the ruthenium out of it. And
12 that's -- the observation the CODEC tests have nothing to do
13 with ruthenium, but they did not see any additional state of
14 oxidation, EO2, as mentioned by the experimenters. It's
15 just one simple statement. I think it would be nice to --
16 if some people talked to those experimenters we'd be able to
17 expand the technical basis there. But they certainly had
18 major relocations, much like we see in all the end reactor
19 experiments at TMI Coral. So that's the -- Lynette, you
20 have some piece you'd like to offer, too.
21 LYNETTE: Not until the discussion on this piece--
22 MR. HENRY: Okay.
23 MS. LYNETTE: Then I want to shift gears a little
24 to more of the regulatory interface of this study.
25 CHAIRMAN POWERS: Do any of the members have
. 101
1 additional questions for Dr. Henry? I think you're up. You
2 can't avoid it. You got to talk to it.
3 MS. LYNETTE: I think I can play catch up here.
4 CHAIRMAN POWERS: But don't feel pressured to do
5 so.
6 MS. LYNETTE: Thank you. This one's just a test
7 to see if I can get them on, right. Looks like I can.
8 To revisit sort of the overall perspective of what
9 this initiative is about, the Commission in late '99,
10 through an SRM, said that they wanted the staff to develop
11 an integrated risk-informed rule making to address EP and
12 financial protection and security. Those weren't pulled out
13 for other than the fact that they were the ones that were in
14 some stage of rule making already, because they were the
15 first exemption requests that come up when you shut down for
16 decommissioning. My point in parentheses here is that there
17 are many other operating plant requirements that can benefit
18 from the results of the risk study so that we shouldn't get
19 too tied in with answering the risk-informed question in the
20 context of only EP or only financial protection.
21 I think to a certain extent the Commission is
22 going to have some challenges, if you will, in making some
23 of these decisions about the continued applicability of
24 operating plants, and bigger I think than just the
25 challenges of, you know, is EP even effective. I think
. 102
1 fundamentally they're applying the risk principles of the
2 agency to different circumstances. You have different types
3 of consequences. The risk is dominated by a single, very
4 low probability event, or two if you want to throw the cask
5 drop in.
6 Defense and depth considerations are likely to be
7 very different. You have a very short risk period, and you
8 have very few plants at risk during any given time period.
9 CHAIRMAN POWERS: I've always been puzzled --
10 struggled -- if I live next to a plant that's
11 decommissioning, do I care if there are plants over, say,
12 500 miles, from me that are also decommissioning?
13 MS. LYNETTE: I think as an individual you don't,
14 but perhaps as Commission policy and fundamental decisions,
15 for example, financial protection, it may make -- may make a
16 difference. But you're correct, as an individual, it
17 wouldn't matter that you were the only person in the country
18 living next to the decommissioning plant at that stage in
19 the risk.
20 So I think the Commission's going to have some
21 interesting things to deliberate on certainly. If you look
22 at what is needed or most valuable to the Commission to make
23 what are essentially going to be informed judgements,
24 there's no magic formula. We don't have any risk-based
25 rules to begin with, any, that you can just apply the
. 103
1 formula that the risk has gone down and certain amount and
2 ergo the regulation is no longer applicable.
3 DR. APOSTOLAKIS: So regulatory guide 1.174 is
4 magic, is that what you're saying?
5 MS. LYNETTE: No, it's not. It's not.
6 DR. APOSTOLAKIS: It's not. Not for this
7 application, but--
8 MS. LYNETTE: No.
9 DR. APOSTOLAKIS: That's a magic formula that's
10 missing here?
11 MS. LYNETTE: No, I think you have to -- to a
12 certain extent, you have the same challenge in operating
13 plants. I mean, Reg. Guide 1.174 was sort of guidance the
14 Commission uses going forward saying we have some
15 deterministic assurances, and we have defense in depth, and
16 ergo, you know, magic maybe, but certainly not a
17 quantitative formula. We have a sense that these sort of
18 deltas are acceptable.
19 MR. WALLIS: Now the word magic is gratuitous and
20 really is irrelevant. There's no formula.
21 MS. LYNETTE: Right. Right. I was being
22 dramatic. You caught me.
23 MR. WALLIS: Well, you might have a good formula
24 that wasn't magic, you know.
25 MS. LYNETTE: That would be wonderful, yeah. And
. 104
1 I guess I note here that the informed job chart requires
2 best estimates of risk using realistic scenarios. I think
3 this is really important. I'm a little disconcerted to
4 constantly be hearing, well, it's less than the QHOs were
5 done. For the type of fundamental decisions the Commission
6 may want to make, they have to have a best estimate. And
7 also I think the ACRS has gone so far as to recommend
8 revisiting this issue in the context of operating plants.
9 Certainly, there you have to have an apples to apples
10 comparison, or this risk, if it's -- based on bounding
11 estimates and unrealistic scenarios could end up garnering
12 regulatory and industry resources that should more rightly
13 be applied to other areas that are done on a best estimate
14 risk basis.
15 DR. APOSTOLAKIS: Let me understand this. You
16 were here when the staff made the presentation. You don't
17 consider that a best estimate of risk using realistic
18 scenarios.
19 MS. LYNETTE: No.
20 DR. APOSTOLAKIS: Because it was bounding?
21 MS. LYNETTE: It's bounding. I think it gets to
22 the point Dr. Powers made. It's the phenomenology of the
23 whole event, and also to some of the points that Dr. Henry
24 made. It hasn't been looked at in a way that say -- that
25 will give you real risk insights, such as what is the real
. 105
1 time frame of evolution. Are there mitigating actions
2 available? Get more water on it. I don't think it's been
3 flushed out in terms of what are the realistic scenarios,
4 and then often times I hear, well, it's bounding a little,
5 but I hear that a lot, and so I think when you look at all
6 the little bounds here, you're fooling about the time
7 frames. I mean, if you want to look at the risk at 60 days,
8 you know, between zero, shut down 60 days, the risk of your
9 earthquake is not going to be the same risk that it is per
10 year. It's going to be approximately.
11 DR. APOSTOLAKIS: But how would that change the
12 conclusion?
13 MS. LYNETTE: Well, I think you'd have a better
14 sense of the orders of magnitude difference between this and
15 your risk for an operating plant, and ultimately that's what
16 the Commission has to work with. The risk is different, and
17 it's lower, and is two orders of magnitude lower enough to
18 say, sometimes, you know, these things have to drop off.
19 You don't need financial protection anymore. It's just --
20 even though there's no number inherent in applying financial
21 protection, I think there's a sense that at some point, it
22 ain't good policy to continue to apply that--
23 DR. APOSTOLAKIS: So in essence what you're saying
24 is that the purpose of all this is not just to compare it
25 with the QHOs?
. 106
1 MS. LYNETTE: Correct.
2 DR. APOSTOLAKIS: There are other decisions that
3 may be affected by that.
4 MS. LYNETTE: Don't stop there. I mean, that's
5 certainly a very good comparison, but don't say, well, we
6 don't have to worry beyond this point about how realistic or
7 how bounding, because we're there. That's kind of what I've
8 heard a lot. That's -- but, you know, I think maybe that's
9 not the right.
10 DR. APOSTOLAKIS: Well, as long as there are
11 decisions that would be affected by this, I agree with you,
12 because I was a little confused. I mean, if we're -- if I'm
13 below the QHO and that's my only concern, by doing a
14 realistic analysis, I go even lower. But now you're saying,
15 no, there are other decisions that may be effective. Okay,
16 fine.
17 MS. LYNETTE: Or if you make that assumption,
18 you've essentially bought into the fact that this pool poses
19 the same amount of risk as a plant, and I think intuitively
20 we'd started out saying that that's probably not the case.
21 You don't have the complexity. You don't have the, you
22 know, high pressure system, on and on and on.
23 DR. APOSTOLAKIS: It's the reputation of PRA you
24 worry about.
25 MS. LYNETTE: Right. Right. And then again,
. 107
1 returning to the other part of the application, which is the
2 defense and depth consideration. I mean, this is obvious
3 stuff, but I want to keep saying it because when the
4 Commission has all this stuff put before them, I don't -- if
5 it's just a numbers game, and certainly if the number is
6 conservative by, you know, constraint of resources, you
7 certainly want them to appreciate--
8 DR. APOSTOLAKIS: Lynette, I have to disagree with
9 you here. I don't think this is defense and depth. And I
10 don't think you should even worry about defense and depth.
11 I mean, there is always an event, take an earthquake, that I
12 can make so strong that I just can't argue that I have any
13 defense and depth anymore. I mean, that's why I have
14 residual risk. So to stretch it and say just because
15 regulatory guide 1.174 asks me to consider defense and
16 depth, I'll do it no matter what. I never heard simplicity
17 of operation being part of defense and depth. Slow
18 evolution, that's part of the risk assessment. So I think
19 you're trying to stretch the requirement of 1.174 here. I
20 mean, even for plants at power, if you -- if their seismic
21 risks dominates, it's because it wipes out your defense in
22 depth. You don't have anything else. But, I mean, that's a
23 fact of life. So wouldn't worry about it here. I mean, if
24 the seismic event dominates, then it's natural that I don't
25 have anything else.
. 108
1 MS. LYNETTE: But I guess I don't disagree with
2 you at all, and my -- the point I'd like to make is the
3 Commission would benefit from that observation.
4 DR. APOSTOLAKIS: The thing that worries me is
5 that the integrated decision making process is so nebulous
6 in 1.174 that if we start, you know, taking literally
7 everything that is there, and try desperately to say, well,
8 Joe, I did look at defense and depth, I think that leads to
9 not a very happy occasion.
10 MS. LYNETTE: Silly conclusions, perhaps, sir.
11 DR. APOSTOLAKIS: Well, I didn't use the word
12 silly.
13 MS. LYNETTE: Which kind of brings me to the next
14 point, which is I think perspective is needed for when you
15 have risk driven by a rare seismic event. And to put all
16 this forward to the Commission and say, well, it's -- you
17 know, we're at the QHOs. Without all these qualifications
18 is not going to put them in the best position to make the
19 time of somewhat fundamental decisions they're going to have
20 to make. Extremely large seismic events that are background
21 risk factors for operating the plants that are dominating
22 the risk profile. I thought NUREG 1150 had some interesting
23 observations that where they attempted to put these huge
24 seismic events in perspective. And what they said was right
25 in the front of the document, where they talk about external
. 109
1 events as -- we've avoided even including off-site dose
2 consequences rom seismic events.
3 They did look at the CDF, but they didn't look at
4 the dose consequences, and they went further and recommended
5 that when those sort of seismic events are looked at that
6 they're put in the perspective of the total loss, monetary
7 loss, loss of life, in the area so that you don't get a
8 sense that -- you have a seismic event, and then you have an
9 impact on this nuclear plant. And, you know, the rest of
10 the region isn't very severely impacted.
11 DR. APOSTOLAKIS: The problem, Lynette, is that it
12 is a Nuclear Regulatory Commission. It's not the health and
13 safety executive of the U.K. This issue was raised more
14 than 20 years ago when design PRA was done in the Indian
15 Point, and there were, in fact, some studies that showed
16 that Chicago would be devastate because before you had a
17 nuclear accident. But it was decided not to use them because
18 the NRC, by its charter, cannot take that into account. I
19 think, Bob, you were there.
20 But, you know, as a general statement, what you're
21 saying is very true.
22 MS. LYNETTE: Yeah, and I agree--
23 DR. APOSTOLAKIS: I mean, you're not going to have
24 a city or people around.
25 MS. LYNETTE: Right. I agree that it -- again, it
. 110
1 doesn't lead you to a formula.
2 DR. APOSTOLAKIS: Right.
3 MS. LYNETTE: But I think if you keep emphasizing
4 that, you get the perspective of--
5 DR. APOSTOLAKIS: Yeah. That's for the people,
6 for the public. I get the feeling that you're trying to
7 address something that I don't know what that is. Do you
8 think the Commission does not have the perspective that you
9 want them to have?
10 MS. LYNETTE: I think they are going to be
11 challenged. They're going to be given a report that has a
12 bunch of numbers. The point Dr. Kress raised earlier,
13 they're going to be given the Livermore and the EPRI seismic
14 results. I mean, its going to be difficult. They're going
15 to be given an estimate that's very bounding, and if
16 you're--
17 DR. APOSTOLAKIS: Oh, I see.
18 MS. LYNETTE: Driven by the numbers, you may come
19 to the conclusion that this event is more significant than
20 it is, and warrants more resources than it might.
21 DR. APOSTOLAKIS: Okay.
22 MS. LYNETTE: And that sort of describes the top
23 conclusion. The bottom conclusion is a little more
24 esoteric, and I want to take a second to explain that. In
25 working with the staff in the evolution, if you will, of
. 111
1 this risk study, we've made some commitments across the way
2 to apply some practical risk insights. Some of the staff's
3 concerns early on were -- you have a change in configuration
4 then you do when you're operating. You have different
5 systems to supply water, and to supply cooling. They may be
6 qualified in a different way, and you certainly want to pay
7 attention to heavy loads. And many of these requirements,
8 we felt, were already in the regulations but they weren't in
9 any one place, so we went -- we put these down in a letter,
10 called them commitments, and we even committed to some
11 things that are going to cost extra money based on the risk
12 insights--instrumentation, additional instrumentation in the
13 pool once you shut down. We had originally committed to a
14 seismic check list that would ensure that the capacity of
15 the pool exceeded two to three times the safe shutdown from
16 earthquake. These are practical risk insights that could
17 have an effect on doing what is best to do when you're
18 trying to risk inform something. Ironically, if the event
19 comes out in an overly bounding way, you focus more on the
20 fantastic. You retain EP. You retain financial protection.
21 You treat it just like an operating plant, and then when you
22 go to do the rule making, because you still have all these
23 operating plant requirements focused on just the pool, and
24 just that one event, sequences that really in some cases
25 can't be mitigated, it will be very difficult in the rule
. 112
1 making context to justify the cost of these additional risk
2 insights even though, you know, they may not cost -- in some
3 cases, they wouldn't cost all that much. In some cases, the
4 cost would be significant.
5 So it's in sort of a weird unintended consequence,
6 if you will, to a certain extent if you overstate the risk
7 and it's comparable to the QHOs and that's the bottom line,
8 you actually forego your opportunity to apply some of these
9 more practical risk insights.
10 CHAIRMAN POWERS: Do members have any additional
11 questions they'd like to pose on this? I think that the
12 take-home lesson I get is there's more we need to understand
13 about this if we're going to go and make extensive decisions
14 on the risk significance of these meltdown accidents. Thank
15 you.
16 DR. KRESS: At this time, we have a perspective on
17 issues from the Institute for Resources and Security
18 Studies, Mr. Gordon Thompson.
19 MS. LYNETTE: I'd like to thank the Committee for
20 the opportunity.
21 DR. KRESS: Thank you very much for -- both of you
22 for coming and giving us the view.
23 MR. THOMPSON: Good morning. My name is Gordon
24 Thompson. I'm with the Institute for Resource and Security
25 Studies in Cambridge, Massachusetts. I'd like to thank the
. 113
1 Committee for this opportunity.
2 My first illustration shows you what has brought
3 us to this situation where pools can catch fire. The rack
4 on the left is an early PWR rack. These had center to
5 center distances of maybe 20 inches. The rack on the right
6 is a high density rack. These are being built with center
7 to center distances as small as nine inches to PWR fuel.
8 And the rack on the right provides a blanket that
9 creates a very hot fuel assembly in the event of partial or
10 total loss of water.
11 Some general observations. The potential for a
12 pool fire, and I'm going to use that phrase throughout in
13 lieu of--
14 DR. KRESS: Mr. Thompson, can I ask you -- what
15 are these racks made of? Are they stainless steel or is
16 there--
17 MR. THOMPSON: Generally, stainless steel, but the
18 neutron absorbing material has been either boraplex or
19 boral--
20 DR. KRESS: Right. Right.
21 MR. THOMPSON: Boral is now the preferred
22 material.
23 DR. KRESS: How are they held in? Are they just
24 lowered down and?
25 MR. THOMPSON: How are the racks held in the pool?
. 114
1 The racks are held by gravity. In many high density pools,
2 these racks are essentially wall to wall, with maybe an inch
3 or two inches at the side. Very constricted flow paths for
4 air cooling.
5 So I used the phrase pool fire for simplicity.
6 This could be an air oxidation reaction or a steam oxidation
7 reaction. This potential has been known since the late
8 1970s, and the Chairman of this Committee was a member of a
9 study team at Sandia that wrote about this subject,
10 published in 1978. It's been a neglected issue I believe
11 for most of the time since. The potential for a pool fire
12 can exist at any high density pool, but can be especially
13 significant for operating plants for two reasons.
14 One the presence of recently discharged fuel that
15 has a high decay heat, and two the potential for a reactor
16 accident to initiate a pool accident.
17 Pool fires have not been studied to anything
18 approaching the same depth as reactor accidents as
19 illustrated by NUREG 1150 or IPEs. There remain major gaps
20 in our knowledge about the probability of pool fires. The
21 associated phenomenology and the consequences. And I'll
22 return to that point later.
23 Pool fires deserve our attention primarily because
24 they could contaminate very large areas of land with
25 long-lived radioisotopes. The impacts from this
. 115
1 contamination could be highly significant in terms of
2 health, economic, social and political considerations.
3 DR. KRESS: Do I infer from that that you think we
4 shouldn't just compare to the individual top of the tower?
5 MR. THOMPSON: Absolutely. The analysis that we
6 heard today from the staff is dealing with the wrong issue,
7 or it's misdirected. Pools generally have a low inventory
8 of short-lived radioisotopes, and you, therefore, expect a
9 comparatively low potential for causing early fatalities.
10 Anyone familiar with consequence calculation knows this
11 without having to do a single analysis.
12 CHAIRMAN POWERS: You don't have to perform a
13 calculation to know this.
14 MR. THOMPSON: Finally, the potential for pool
15 fires could be almost completely eliminated by storing spent
16 fuel in a combination of low density storage and dry
17 storage. Dry storage technology, as you know, being already
18 approved pre-licensed. So this is an avoidable problem.
19 CHAIRMAN POWERS: Could we come back to that
20 statement. There must be more to it than this -- than just
21 using the low-density racks. Presumably, if I lost water on
22 a low density rack, I would have a potential for an
23 oxidation event, just as I would if I lost water in a
24 high-density rack?
25 MR. THOMPSON: In the rack on the left, with a
. 116
1 recently off loaded assembly, there is the possibility of
2 initiation of a fire. That's never been studied to my
3 knowledge. We know that at some time less than a year, the
4 rack on the left would be safe in the event of partial or
5 total water loss. What the precise time is should be
6 analyzed, but has not, to my knowledge.
7 Now, here's an indication of the amount of cesium
8 we're dealing with. Core inventory around five million
9 curies. Illustrative pool inventory here at 35 million
10 curies. And that's not the top end, as I'll be showing you
11 in a couple slides from now.
12 The release fractions concluded in NUREG 1150 are
13 generally small fractions or the core inventory of cesium.
14 I show here an inventory of point two as the release
15 fraction for containment bypass at Surrey. That's near the
16 upper end of what NUREG 1150 finds. For most sequences,
17 they're estimating lower release fractions.
18 The Chernobyl release fraction, as shown here,
19 this estimate is a 40 percent release.
20 The release fraction that's often used for a pool
21 fire is a hundred percent, as shown at the bottom. I think
22 that's probably not realistic, but it could certainly be
23 point five or above.
24 The conclusion here is that the release of cesium
25 from a pool fire could substantially exceed the release from
. 117
1 a reactor accident by at least an order of magnitude, and
2 that has obvious implications for risk.
3 And this is an indication of why we care about
4 cesium. This is the pattern of cesium deposition around
5 Chernobyl. There's another similar contaminated area up to
6 the north in Byelarus, a further distance from the plant.
7 And--
8 MR. WALLIS: This doesn't look like a Gaussian
9 plume to me?
10 MR. THOMPSON: No, that's probably characteristic
11 of any real release. A Gaussian plume may be the best you
12 can do for analytic purposes. There have been some efforts
13 to compare total contaminated area using Gaussian plume
14 models in actual situations. A colleague of mine did that
15 for the Windscale fire of 1957, where you had a similarly
16 erratic looking distribution, but found that the total area
17 was reasonably close to the Gaussian model.
18 CHAIRMAN POWERS: I might just interject, Graham
19 -- the -- what you see is a -- the result of a plume that
20 changed directions as the wind changed over the course of
21 the accident.
22 DR. KRESS: Lasting for 10 days as opposed to two
23 hours.
24 MR. WALLIS: Yeah, that's clear. Okay. So it's
25 important also that when we asked questions earlier about
. 118
1 how long the fire lasts and how big it is, comparisons with
2 Chernobyl can perhaps be made. I mean, it's a very
3 different kind of a fire we're talking about with a spent
4 fuel pool I think in terms of its loft and its duration.
5 CHAIRMAN POWERS: But, I mean, it seems to me that
6 the issue of how long the fire lasts was something the
7 previous speaker spoke to when he said you need to look at
8 these interactions fairly carefully, because things may
9 change as you progress across the pool, and you get some
10 water cooling and non water cooling.
11 MR. THOMPSON: Yes, I would expect the release to
12 evolve over time. The plume energy would change over time.
13 The composition would change over time.
14 What sort of areas might we be talking about?
15 This is from an analysis by BA that was actually done for
16 pool fires, presented to a German hearing in 1979. I might
17 add that as a result of that hearing that that government
18 ruled that they would not tolerate high density for that
19 pools for that facility.
20 The three releases that are shown here. Two in
21 curies is about the Chernobyl release. This was prepared in
22 the context of the Harris nuclear plant in North Carolina.
23 The projected inventory of spent fuel with an age of up to
24 three years would have 20 million curies of cesium. And
25 with an age up to nine years would have 70 million curies of
. 119
1 cesium. And if the release fraction approaches one, it's
2 obvious that the area of contaminated land is very, very
3 large. With a 70 megacurie release, we're at the land area
4 of North Carolina itself.
5 MR. WALLIS: When you say land contamination, do
6 you mean by a certain amount of curies per square meter or
7 kilometers or?
8 MR. THOMPSON: This is computed at 10 rem per
9 30-year threshold with a shielding factor of point 25. That
10 is a person unshielded would receive 40 rem.
11 MR. WALLIS: So it's a measure, but in effect
12 there are other areas with a lower dose which--
13 MR. THOMPSON: This is the area within which the
14 dose would exceed 10 rem per 30 years. And this was done
15 with a Gaussian straight line model by Janvier, then of
16 Princeton University.
17 Just as an indication of health effects. This is
18 the BEIR V estimate for lifetime risk at a tenth of a rem
19 per year. And 10 rem in 30 years will show up as -- if
20 sustained at that rate of about three-tenths of a rem per
21 year -- would show up as about a 10 percent increase in the
22 normally expected instance of fatal cancer. That's
23 certainly something that I think the public would feel
24 concerned about.
25 DR. KRESS: It's a lot more than 0.1 percent,
. 120
1 isn't it?
2 MR. THOMPSON: Excuse me?
3 DR. KRESS: It's a lot more than 0.1 percent,
4 isn't it?
5 MR. THOMPSON: I think if you -- this raises an
6 interesting question. If you tell people that the
7 relocation threshold is 0.3 rem per year, and all the people
8 in the range of 0.1 to 0.3 who are not being relocated see a
9 chart like this, they're not going to be very happy. In
10 fact, a big cesium release such as we're talking about would
11 be viewed by the public as an immense disaster of historic
12 proportions. It would lead to immense political effects,
13 huge litigation, and would be an historic event in the
14 history of the country.
15 Okay, safety goals. We've had a discussion about
16 safety goals. The primary goals as articulated by the
17 Commission are qualitative and they are as stated at the
18 top. The Commission did not specifically address the issue
19 of land contamination. I believe that they should do so.
20 The staff's analysis does not address land contamination,
21 which is the most important indicator of pool risk for
22 either decommissioning or operating plants. And, therefore,
23 this analysis does not provide a credible basis for decision
24 making.
25 So what are the next steps? I would argue that
. 121
1 there should be a moratorium on any regulatory action or
2 decision that could increase the risk of radioactive release
3 from any spent fuel pool pending the completion of new
4 studies. Those studies should be done to at least the depth
5 of NUREG 1150 on probability, phenomenology, and
6 consequences. On the probability side, it's particularly
7 important that for operating plants the work should address
8 potential interactions between reactor accidents and pool
9 fires.
10 DR. THOMPSON: When the methodology for this is
11 developed, I would argue that licensees should be required
12 to extend their IPEs to encompass pool fires.
13 DR. KRESS: Let me ask you a question. If the
14 major, one of the major concerns were this land
15 contamination and the residual cancers that you might get
16 from that, and the idea of this particular study was to see
17 if you could relax emergency preparedness, and there were
18 other things, but let's focus on the emergency preparedness.
19 Would the emergency preparedness plan that you
20 have make any difference to this concern?
21 DR. THOMPSON: Emergency preparedness would have
22 relatively small effect on the health consequences arising
23 from land contamination.
24 That is certainly true.
25 That by no means closes the issue in my mind as an
. 122
1 issue of protecting the public.
2 DR. KRESS: Do you think other measures need to be
3 taken to ignore the risk of spent fuel pool --
4 DR. THOMPSON: Well, I think the first step should
5 be to understand the risk properly so that the Staff needs
6 to be sent away and come back with a proper analysis of
7 consequences. What we have heard is just irrelevant. They
8 need to come back with a proper analysis of land
9 contamination.
10 DR. KRESS: Now the MACCS code will produce that.
11 They have the numbers. They just didn't show them to us.
12 DR. THOMPSON: They show doses --
13 DR. KRESS: Yes.
14 DR. THOMPSON: I think you'll find when you dig
15 into the MACCS code that that was assuming relocation of
16 populations.
17 DR. KRESS: Yes, but it will also tell you lead
18 contamination, square miles and essentially the numbers
19 you --
20 DR. THOMPSON: And I believe that should be
21 done --
22 DR. KRESS: That should be part of the study --
23 DR. THOMPSON: -- before this study is regarded as
24 having any relevance for decision-making.
25 I think there are important phenomenological
. 123
1 issues that need to be addressed and for the operating
2 plants, as I said, the relationship to reactor accidents.
3 DR. APOSTOLAKIS: What would the Staff do with
4 land contamination incidents?
5 DR. KRESS: What would they do with them?
6 DR. APOSTOLAKIS: Yes.
7 DR. KRESS: Well, like he said, I don't think it
8 would impact your decision-making on emergency response, but
9 it might give you a different perspective on both operating
10 plants and decommissioned plants as to how much you need to
11 protect the pool from having a spent fuel fire.
12 DR. APOSTOLAKIS: But if I don't have any guidance
13 from the Commission regarding land contamination how can I
14 make any decisions?
15 DR. KRESS: I would do it on an equivalent dollar
16 basis because that is the only metric in common, and what
17 you will find out is the dollars lost due to this land
18 contamination will far exceed the prompt fatalities, for
19 example, in this case, and so it ought to be more of a
20 concern.
21 The only common metric I can use -- I hate to
22 sound crass and say put it on a dollar basis, but you have
23 to have a common metric sometimes.
24 DR. APOSTOLAKIS: Yes, but I mean that is one
25 possible way of doing it.
. 124
1 DR. KRESS: I don't know of another and that is
2 the problem.
3 DR. APOSTOLAKIS: We don't know. I can
4 understand --
5 DR. KRESS: You could arbitrarily -- you know, the
6 .1 percent is an arbitrary choice. You could arbitrarily
7 say we choose a number also.
8 DR. APOSTOLAKIS: But I want to ask Mr. Thompson,
9 okay, the Staff analysis has not addressed land
10 contamination.
11 Now if they address it, then what should they do?
12 DR. THOMPSON: In addressing it, they will be
13 providing the decision-maker, the Commissions, with relevant
14 information.
15 The articulation of the safety goal to date does
16 not address land contamination --
17 DR. APOSTOLAKIS: Right.
18 DR. THOMPSON: -- and therefore I believe that the
19 qualitative safety goals, which I show on this slide, do not
20 have any appropriate metric developed in order to address
21 land contamination and therefore have no appropriate metric
22 to address the risk of pool fires and the Commission should
23 be apprised of that fact.
24 DR. APOSTOLAKIS: So what you are recommending
25 then is that the Commission also revise the safety goals to
. 125
1 include land contamination?
2 DR. THOMPSON: I think that's -- I would clearly
3 recommend that. What the Commission itself chooses to do,
4 of course, is beyond the control of the Staff, but I think
5 it is the Staff's responsibility to provide the relevant
6 information. I think it is this committee's responsibility
7 to make sure that they do.
8 DR. WALLIS: What sort of thing are you thinking
9 about within their mention of interactions between reactor
10 accidents and pool fires? What sort of scenarios were you
11 thinking of there?
12 DR. THOMPSON: I should, first of all, state that
13 this issues is under litigation at present. I am involved
14 in this litigation so that's going to limit what I say here
15 today.
16 There is a licensing case at the Harris facility
17 where this very issue is being addressed, so you will
18 understand that I am not going to say very much.
19 It is a complex issue and it requires the
20 extension of IPEs and PRAs in order to address it properly
21 and one of the issues is the contamination of the plant with
22 radioactive material to the extent that access is precluded
23 for purposes of operating the pools safely.
24 CHAIRMAN POWERS: Any additional questions you
25 would like to direct to the speaker?
. 126
1 Again, I think we hear your comment that there
2 needs to be a better phenomenological understanding. That
3 seems remarkably consistent with previous speaker.
4 Now you go on and indicate in more comprehensive
5 consequence analysis and certainly this committee has felt
6 that consequences were broader than just prompt and delayed
7 fatalities, some other communications in the past, and so I
8 think it's a welcome message to us.
9 DR. THOMPSON: Could I just add one --
10 CHAIRMAN POWERS: Sure.
11 DR. THOMPSON: -- thing about Dr. Henry's
12 presentation.
13 I think his analysis showed something quite
14 contrary to what he concluded from it, namely that water in
15 a pool is not always a good thing.
16 When the levels of water are low, actually water
17 can be a bad thing.
18 It can make the cladding harder than it would have
19 been otherwise, and this is very important when you are
20 talking about makeup and emergency response. If the level
21 is low, do you start doing makeup or not, and there are
22 times when actually making up water into a dried out pool
23 might be a bad thing.
24 DR. WALLIS: Can you tell me something about that?
25 How does air get to the elements if water is, say, 30
. 127
1 percent of the way up? There are access paths for air to
2 get in?
3 DR. THOMPSON: Could I show you a little, simple
4 picture?
5 DR. WALLIS: Does the water block the air flow or
6 not is the question really.
7 DR. THOMPSON: I have a nice, simple little
8 picture, if I can find it.
9 [Pause.]
10 CHAIRMAN POWERS: A direct answer to the
11 question -- I think some of the pictures we have seen show
12 you that you can change the convolution of the paths but
13 there is still access.
14 DR. WALLIS: Yes. The question is how much does
15 it block --
16 DR. THOMPSON: Okay. In some cases is the total
17 and instantaneous drainage case and if the racks are wall to
18 wall with an inch or two at the side, obviously that is a
19 very constricted flow path for air.
20 The bottom case you have no opportunity for air
21 convention.
22 As Dr. Henry pointed out, the only convective
23 cooling mechanisms is the rise of steam that's generated --
24 DR. WALLIS: So you are saying there is no other
25 air path.
. 128
1 My question was is there another air path which
2 lets air in from the side of something or not?
3 DR. THOMPSON: There are holes often at the very
4 bottom of the racks but at around 10 percent and greater
5 covery level there is no other path then.
6 The only possible cooling mechanisms for fuel
7 assembly are (1) longitudinal conduction; (2) longitudinal
8 radiation with multiple reflections; and (3) the forced
9 convention of steam rising from the submerged part of the
10 assembly, and for the equation that Dr. Henry has produced,
11 it shows that at low submerged fractions you get very high
12 cladding temperatures.
13 CHAIRMAN POWERS: Any additional comments?
14 [No response.]
15 CHAIRMAN POWERS: I am going to allow the members
16 to take about a 15-minute break and then we will resume with
17 the schedule.
18 My intention is to get through that portion of the
19 schedule that we had on the agenda before we break for
20 lunch.
21 DR. KRESS: I would like to thank the speaker for
22 sharing his views.
23 [Recess.]
24 CHAIRMAN POWERS: Let's come back into session.
25 Our next presentation deals with risk-informed
. 129
1 regulation plan. In my understanding this is a presentation
2 largely for our information in anticipation of a lot more
3 intensive activity in the future. Professor Apostolakis?
4 DR. APOSTOLAKIS: Yes. This represents a fairly
5 significant revision of the PRA implementation plan or
6 risk-informed implementation plan. Maybe we can just go
7 ahead with the presentation. The principal objectives, all
8 sorts of things.
9 MR. KING: For the record, my name is Tom King
10 from NRC's Office of Research.
11 You are right, what I'd like to do is keep this at
12 a fairly overview high level in terms of talking about the
13 purpose, the structure, the uses of this document. We're
14 not asking for a letter at this time. When we get to the
15 end, we'll talk about the schedule.
16 We view the next six months as sort of a
17 comment-gathering period both internally and externally, and
18 I think it would be appropriate at some time in the future
19 maybe a subcommittee could be put together and spend some
20 more time to talk about the details of this and the loose
21 ends that need to be cleaned up on it.
22 Let me start with just a little bit of background.
23 Back in 1995 the Commission issued their PRA policy
24 statement. It basically encouraged the use of risk
25 information in all regulatory matters where it was supported
. 130
1 by state of the art and where it was done in a fashion that
2 complemented our traditional deterministic
3 defense-in-depth-type philosophy.
4 After that, the staff put together what was called
5 a PRA implementation plan that basically was a catalogue of
6 what risk-informed activities were going on in the various
7 offices. It was organized by office.
8 Back in 1999 we got some criticism from GAO that
9 this really wasn't a strategy that defined where the Agency
10 wants to go and how it's going to get there for its
11 performance implementation. The Chairman committed to
12 develop such a document. We now call that document the
13 Risk-Informed Regulation Implementation Plan, and the SECY
14 paper that went up a couple of weeks ago, actually probably
15 just a week ago, SECY-000213, is the first attempt to put
16 together a complete package as best we could at this time.
17 There was an earlier paper back in March that sort
18 of outlined the structure and content and had a couple of
19 example sections, but this paper is the one where we've now
20 tried to fill in all the holes as best we could at this
21 time.
22 Also parallel with this the Agency's been
23 developing a strategic plan which just got issued a couple
24 of weeks ago, and as we'll go through, this document is tied
25 to the strategic plan in the sense that it's really the
. 131
1 document that gives the details of how these high-level
2 strategies in the strategic plan that talk about
3 risk-informed regulation are going to be implemented.
4 And then we said basically you can consider this a
5 road map to risk-informed regulation. It's going to define
6 where are we going -- in other words, what should be
7 risk-informed -- and then what needs to be done to get
8 there. That would include schedules, rulemakings, guidance
9 documents, infrastructure needs, training, communication.
10 Communication is a key part of this, in that the
11 Commission was very -- emphasized in their SRM that they
12 sent us on the March paper that this implementation plan
13 really ought to have a good communications element. They
14 emphasize the internal communications, but it also needs to
15 have an external communications piece, and we'll talk about
16 that later on.
17 CHAIRMAN POWERS: When you formulate a road map
18 like this, do you say the technology for driving risk
19 information is frozen, or do you prognosticate evolution in
20 that technology?
21 MR. KING: We prognosticate evolution in that
22 technology. You'll see it in the screening criteria that we
23 use to try and decide where does it make sense to go risk
24 inform something. If we think it's reasonable to go develop
25 the technology in that area, even though it isn't developed
. 132
1 yet, we consider that as a positive thing.
2 I do want to acknowledge that the development of
3 this plan has been an interoffice effort. It involves
4 heavily NMSS, NRR, Research, and Human Resources in the
5 training aspect. It also has some loose ends to clean up,
6 as I mentioned. So consider it work in progress, and like I
7 said, over the next six months we hope to polish it up.
8 Organization -- you've got three main parts. Part
9 1 is what we call general guidance. That's really the part
10 that I'm going to focus on in the presentation today. The
11 idea is to provide some consistency and important
12 considerations that whatever activity you're risk informing,
13 people ought to think about when they go do that. Part 2
14 are the arena-specific activities as broken up by the three
15 major arenas -- reactors, materials, waste -- that has the
16 details of what's going on, we're not going to focus on that
17 today. And part 3 is training and communications, and we'll
18 talk about the communications part.
19 The scope, we had to ask ourselves what goes into
20 this document. I mean, in theory if you wanted to stretch
21 it you could put everything the Agency's doing and tie it to
22 risk-informed regulation and put it in here. That wouldn't
23 make sense. So basically what we did was we said things
24 that were initiated subsequent to issuance of the 1995
25 policy statement to specifically implement the, you know,
. 133
1 the Commission's expectations would go in here, and any
2 other important things that we felt were important to the
3 transition to risk-informed regulation like, for example,
4 the IPEEE reviews. Even though they were started before the
5 policy statement, that body of knowledge is important to
6 transitioning to risk-informed implementation.
7 Things that are not in here are things like the
8 thermal hydraulic code development, severe accident code
9 development. They are important to risk-informed
10 regulation, but they were activities that would probably go
11 on if there were not risk-informed regulation, so we just
12 sort of give them a, you know, an acknowledgement in words
13 in here, but we're not putting all of that work in this
14 plan.
15 CHAIRMAN POWERS: You must be presuming that not
16 only would they go on plan or no plan, but nothing about
17 their mission changes, plan or no plan.
18 MR. KING: Yes.
19 CHAIRMAN POWERS: I think it's all about the same,
20 isn't it. I mean, we want good, fast-running, user-friendly
21 tools, and as best we can understand the uncertainties that
22 come out of those, and that's important for risk-informed or
23 non-risk-informed.
24 DR. APOSTOLAKIS: But quantifying the model
25 uncertainty associated with those codes would not be there
. 134
1 if you didn't have a risk-informed regulatory system in
2 mind.
3 MR. KING: Well, it's already there in the thermal
4 hydraulics area in that, you know, Appendix K has a
5 best-estimate option that requires you to quantify the
6 uncertainty in the codes.
7 DR. WALLIS: But, Tom, I think there's an
8 interface between the codes and risk informing regulations
9 which needs to be examined more. What output do you
10 actually need from the codes in order to really risk inform?
11 I don't think that question has been fully addressed. Just
12 saying that you need best estimates in uncertainty doesn't
13 tell you enough about how good an estimate it needs to be
14 and how you measure how good it is and all that sort of
15 thing. You know, it doesn't -- I think that you need
16 something more in the risk-informed world than you had in
17 the past.
18 MR. KING: I agree with your comment. Whether
19 that level of detail would go in the plan or not is not
20 clear to me, but I agree with your comment.
21 DR. SEALE: I'd be very careful to not recognize
22 that need.
23 MR. KING: In the plan.
24 DR. SEALE: Yes.
25 MR. KING: Okay.
. 135
1 DR. SEALE: Down the road I think you need to have
2 that to hang your hat on.
3 DR. WALLIS: I'm trying to get clear what your
4 syntax implies. You think he should not recognize the need,
5 or he should.
6 DR. SEALE: He should.
7 MR. KING: In the plan, specifically. Not
8 separate from the plan. Okay.
9 All right, let me talk now about part 1, which we
10 call general guidance. We've talked a lot about a vision
11 statement for risk-informed regulation, and really what we
12 ended up going back to was the Commission's 1995 policy
13 statement that said use risk information in all regulatory
14 matters supported by state of the art and consistent with
15 defense-in-depth philosophy, so forth. We think that's a
16 good vision statement.
17 CHAIRMAN POWERS: It is and it isn't. I mean,
18 it's fine as a vision statement. I never know exactly what
19 to do with these vision statements. But it does raise this
20 question of state of the art. I'm sure that I can find
21 somewhere someone who has a model for just about anything.
22 But I'm not sure that I can get agreement among specialists
23 or a consensus among specialists that a model for anything
24 is widely accepted. So I guess what I'm asking is how do
25 you know whether there's a good state of the art or not? Is
. 136
1 it central specific? I mean, it's state of the art within
2 your staff, or is it a broader state of the art?
3 MR. KING: You know, there's no formula that's
4 going to give you that answer.
5 CHAIRMAN POWERS: Yes. I'm just asking for kind
6 of intuition at this point.
7 MR. KING: I think it's expert judgment. It's not
8 limited to just people on the staff getting together. I
9 think you'd certainly have the other stakeholders, the
10 licensees, you know, the public, if it's an NMSS side of the
11 house, I mean, the States I think will be very interested in
12 this. And I think you listen to all points of view, and
13 ultimately the staff has to make a decision is that state of
14 the art exists or can exist with some development and --
15 CHAIRMAN POWERS: Well, I think you told me what
16 you wanted. It's something that you have to go look at.
17 You can't sit in your office and sit down and say I decree
18 that there's good state of the art here and not good state
19 of the art over here.
20 MR. KING: Yes.
21 CHAIRMAN POWERS: Until you've gotten a lot of
22 input.
23 MR. KING: Yes.
24 CHAIRMAN POWERS: But then sooner or later you
25 have to make that decision.
. 137
1 MR. KING: Certainly you have to make that
2 decision.
3 DR. APOSTOLAKIS: I would expect the answer to the
4 question where do we want to go to be a little bit more
5 specific than the policy statement. For example, maybe
6 where we want to go is a situation where we make decisions
7 according to some guidance that is risk informed. We have
8 quantified the important inputs to the decision, you know,
9 the uncertainties and so on, and we have computerized tools
10 for implementing our decision making process. That would be
11 in my mind a little bit more specific, and it would give,
12 you know, other people a better idea as to where we want to
13 go. Just to say we want to be in a situation where risk
14 information will complement or supplement defense in depth
15 doesn't say much to people. So maybe --
16 MR. KING: That's why there's two subbullets here.
17 You know, the question is where do we want to go. We've got
18 the general statement in the policy that doesn't say really
19 where you want to go. It just says we want to go there.
20 But the second bullet is really directed toward how do we
21 make the decisions to decide where do we want to go, and in
22 the plan you'll see a set of criteria that at least in our
23 view would be a set of criteria you could apply across the
24 Agency to regulations, to staff practices, and if the
25 answers to those questions are positive, then those would be
. 138
1 candidates to go make risk-informed changes.
2 Let me jump to slide 8.
3 DR. APOSTOLAKIS: I know. But what I mean, Tom,
4 is not just the criteria for selecting the activities.
5 After I have selected an activity, where do I want to go?
6 And I think that's the question I addressed. I mean, let's
7 say, you know, I've selected special treatment requirements.
8 What is my vision as to what would be needed to, in a good
9 risk-informed regulatory system?
10 Well, I will have to make a decision or a number
11 of decisions regarding special treatment, so can I help that
12 process in some way? I have to have good understanding of
13 the inputs to that decision and associated uncertainties,
14 and I may want to have computerized aids to make sure that
15 these things are implemented throughout the agency in an
16 efficient way, because that then will also give you a
17 high-level guidance to develop more detailed research plans,
18 it seems to me.
19 MR. KING: The way the plan's set up, those kinds
20 of issues and goals would be in part 2. In other words, you
21 apply these criteria -- let me put them up for a minute --
22 this is slide 8. If you go through and you apply these
23 criteria and you apply them to part 50, for example, and you
24 say I want to risk inform part 50 special treatment and you
25 apply these criteria and the answer is yes, that makes sense
. 139
1 to go do that.
2 DR. WALLIS: It seems to me, Tom, you're assuming
3 you have the tools to do it.
4 MR. KING: Well, you're not assuming, you're
5 asking yourself the question. As you walk through these
6 criteria, you're saying if I did risk inform special
7 treatment, would it resolve a safety concern? Would it make
8 things more effective and efficient, reduce unnecessary
9 burden?
10 DR. WALLIS: Yes, but there's an earlier question,
11 which is do I know how to risk inform, do I have the tools
12 to do it?
13 MR. KING: Then you come down to --
14 DR. WALLIS: You can't -- a program to do
15 something unless you have the wherewithal to do it.
16 MR. KING: Well, these are basically broken up
17 into two general categories. The first four bullets really
18 talk about the value of risk informing, is it worth doing,
19 and the last three bullets talk about the practicality of
20 doing it. If you look at the fifth bullet, I think it gets
21 to your question, you know, do we have existing, you know,
22 risk data and analytical models, or could they be developed
23 to go do this? So when you get to special treatment, you
24 ask yourself okay, it makes sense, there's a value in doing
25 that, now can I do it? And the answer would be yes, I've
. 140
1 got risk assessments, I've got, you know, importance
2 measures, I've got the tools I need to do that.
3 Now I'm going to have to do training, I'm going to
4 have to do a rule change, I'm going to have to develop
5 guidance documents, I may have to have some automated things
6 to help, you know, the reviewers or help the inspectors, but
7 I view that level of detail as being in part 2, not in part
8 1. Part 1 is a step above that that helps you make the
9 decision do I want to go in and even embark on this.
10 To continue on, given that you've got -- you apply
11 those criteria on Slide 8 and you identify things that you
12 want to go risk-inform, Part I then says, okay, there are
13 certain things that you have to keep in mind if you're going
14 to go risk-inform a set of requirements or an activity or a
15 staff practice.
16 And what we did in Part I was try to list those
17 things and list some questions associated with each of those
18 things. We just start defense-in-depth, you know, the PRA
19 policy statement says we want to risk-inform things
20 consistent with defense-in-depth philosophy.
21 That doesn't mean defense-in-depth has to be the
22 same in every activity you regulate. But I think what
23 should be the same are the questions the Staff asks itself
24 when it goes through and makes decisions on risk-informing
25 whether it's a reactor regulation or a material regulation,
. 141
1 or an inspection, plant oversight process.
2 DR. WALLIS: Well, I have trouble with this. When
3 defense-in-depth -- risk-informing is trying to get more
4 quantitative about decisionmaking, and safety margins can be
5 made quantitative.
6 Are there quantitative measures of
7 defense-in-depth that are agreed upon, or is there always an
8 argument?
9 MR. KING: There are no -- there is no
10 quantitative measure of defense-in-depth that applies across
11 the Agency.
12 DR. WALLIS: So it's all argument?
13 MR. KING: No.
14 DR. WALLIS: Sort of persuasiveness?
15 DR. APOSTOLAKIS: For multiple barriers, yes, the
16 PRA --
17 DR. WALLIS: For multiple barriers, the PRA picks
18 that up. So, the defense-in-depth is an aspect of PRA in
19 that case.
20 DR. APOSTOLAKIS: Right.
21 DR. WALLIS: But I think defense-in-depth gets
22 invoked in a much more general way than that.
23 DR. APOSTOLAKIS: Yes, I mean, it comes back to
24 our paper and what we're going to hear later about Option 3.
25 MR. KING: Option 3 took an attempt to quantify,
. 142
1 basically quantify the level, the various levels of
2 defense-in-depth that we'd like to see in the reactor
3 regulations.
4 Now, when you go to waste disposal, it may be
5 different, maybe three barriers doesn't make sense, given
6 the consequences and the likelihood of things going wrong.
7 We didn't want to put in this plan that, you know,
8 one size fits all on defense-in-depth. What we wanted to do
9 was say that a defense-in-depth has been a fundamental
10 underpinning of what this Agency does across the board.
11 The way it's applied can be different, but you at
12 least ought to ask yourself a consistent set of questions
13 about before you go and apply defense-in-depth. You ought
14 to ask yourself a consistent set of questions so that you're
15 addressing the right issues.
16 And basically what we did was, we took the
17 criterion Reg Guide 1.174, and we took some thoughts we had
18 gotten from ACNW on defense-in-depth, and we tried to fold
19 them in.
20 DR. APOSTOLAKIS: That was a joint letter?
21 MR. KING: Yes.
22 DR. APOSTOLAKIS: And please correct the report as
23 well. In Appendix I, you refer to it as an ACNW letter.
24 MR. KING: All right.
25 DR. APOSTOLAKIS: It was not.
. 143
1 MR. KING: We'll fix that.
2 DR. APOSTOLAKIS: Okay.
3 CHAIRMAN POWERS: When it comes to this
4 defense-in-depth business, one of the things that has always
5 been curious to me, is if I am working on, say, an
6 instrumentation and control system, they have a lot of
7 electric circuits coming in, something to put those out, so
8 I can read them; something to cause systems to actuate.
9 Do I set myself a defense-in-depth for that
10 activity, or is that one part of a larger defense-in-depth,
11 and I don't specifically address defense-in-depth when I'm
12 working on that particular activity?
13 MR. KING: In the reactor area, when you're
14 approaching it, it would be -- that would not be a separate
15 activity where you would look at it for defense-in-depth.
16 Defense-in-depth would be at a higher level, prevention,
17 mitigation, barrier integrity, no over-reliance on human
18 actions, that kind of thing.
19 CHAIRMAN POWERS: But when I look at the guidance
20 the Staff gives to the people preparing digital electronic
21 systems, they specifically call out defense-in-depth.
22 DR. SEALE: The example I can think of where you
23 might have defense-in-depth apply to a particular situation
24 -- and it may not apply in power reactors -- but in research
25 reactors -- is the idea that the control system fails safe.
. 144
1 If you have a failure, you shut the system down.
2 Now, that's defense-in-depth at that level.
3 MR. KING: Well, is that defense-in-depth, or is
4 that just good engineering practice?
5 DR. SEALE: It's defense-in-depth. It's also good
6 practice, but we don't always use it.
7 MR. KING: It's like negative void coefficients;
8 is that defense-in-depth, or is that just good engineering?
9 I mean, we've had this debate on the Option 3 framework
10 document. Should defense-in-depth include a general
11 statement of good engineering practices?
12 DR. KRESS: If you start including all those
13 things in defense-in-depth, in my mind, you just overwhelm
14 defense-in-depth and make it so nebulous that you can't
15 really deal with it.
16 I think you need limits on your definition of
17 defense-in-depth, and I wouldn't include things like that in
18 it.
19 I think your approach that was in the document,
20 actually, of looking at the cornerstones and not over
21 reliance on things in -- and I think I would go so far as to
22 redundancy and diversity on key safety functions as being
23 part of defense-in-depth.
24 MR. KING: In Option 3 it's at that level; it's
25 not within a system looking at defense-in-depth.
. 145
1 DR. APOSTOLAKIS: Well, a comment on this slide:
2 Since communication is important, in the report itself,
3 there is an implication that the Agency's safety goals are a
4 principle, and I don't think that they are a principle.
5 I mean, I can see safety margins,
6 defense-in-depth, and ALARA and radiation protection being
7 principles, but I wouldn't call the safety goals a
8 principle.
9 So, on page 1-3 and a few other pages, maybe you
10 can clean up the language a little bit.
11 DR. KRESS: Having safety goals is a principle of
12 good regulation.
13 DR. APOSTOLAKIS: Well, yes, but I wouldn't say
14 it's a principle like, you know, making sure you have
15 defense-in-depth.
16 DR. SEALE: For ALARA and the safety goal, as
17 stated, is a limitation.
18 DR. APOSTOLAKIS: Yes, well, guidance, I think is
19 the right word that you use on the viewgraph. But in the
20 text, it says principle there, and I'm not sure.
21 Again, referring to the text, you use the word
22 commensurate many times. Is defense-in-depth commensurate
23 with the risk and uncertainty associated with the risk;
24 commensurate with the expected frequency and so on.
25 Are there any plans to go beyond just that word?
. 146
1 MR. KING: At this point, we're looking for
2 feedback as to whether it would make sense to go beyond
3 that. When we first started writing this, I tried to write
4 this to be more rigid in terms of what that meant.
5 And the more we got into the way defense-in-depth
6 is applied across various ways, it became clear that we
7 needed to allow a lot of flexibility in how you would answer
8 these questions.
9 DR. APOSTOLAKIS: At this time.
10 MR. KING: At this time.
11 DR. APOSTOLAKIS: But when you say what I want it
12 to be, wouldn't -- why don't you try to do something about
13 it, so you don't have to use words like commensurate,
14 reasonable, or maybe a little bit more quantitative.
15 In fact, a related question is, in the Research
16 report discussion, I didn't -- I sort of expected that this
17 plan would have been a guide as to what needs to be done.
18 But yesterday, we hardly referred to it. Is it
19 because of the timing?
20 MR. KING: Yes. I think it's mostly a timing
21 issue.
22 DR. APOSTOLAKIS: I would say, you know, I have a
23 research plan; what do I need to address? Well, I have to
24 use these qualitative words so much in my plan, so here is
25 what I will try to address. I mean, that would make sense;
. 147
1 wouldn't it?
2 MR. KING: Yes. I view -- I mean, this is the
3 first time we've got all of the pieces put together. And I
4 think we now need to go through and see where are there
5 gaps, you know, where do we need to fold things into the
6 research program or fold things into other Office activities
7 to plug the holes that are probably going to be pretty
8 evident when we go through this systematically.
9 But again, at this point, it's our best shot in
10 putting together what we think are the important
11 considerations and the things that need to be done and what
12 our plans are.
13 DR. APOSTOLAKIS: Okay, so eventually, there will
14 be a connection between the report and --
15 MR. KING: Eventually there will be a connection
16 between what's identified in here, particularly under
17 infrastructure needs, and a Research program. And there
18 should be, you know, a connection between what's identified
19 in here for training needs, and the Training program.
20 DR. APOSTOLAKIS: Okay.
21 MR. KING: And communications activities should be
22 connected with the Agency communication goals and so forth.
23 I think you can find some of that now, but I
24 wouldn't claim that it's been a systematic search at this
25 point.
. 148
1 DR. APOSTOLAKIS: Okay.
2 MR. KING: Anyway, I wasn't going to go through
3 each of these things. You know, we tried to put down, as I
4 said, what we felt were important questions related to each
5 of these principles.
6 When we get to safety goals, we had put in an
7 appendix that said if you're going to establish safety
8 goals, there's a bunch of things you need to think about to
9 do that; you know, what's the population at risk, and are
10 you -- is the goal cancer incidence or fatality, and there
11 is a whole laundry list of things in there.
12 So your goals may be different, but the questions
13 you ask yourself are the same.
14 Then after these things we call principles, we
15 have some that we call implementation issues. And these are
16 really based upon our experience so far, that in going in
17 and risk-informing something, you know, we have a
18 performance-based initiative that basically is committed to
19 the Commission, that if we're going to make a risk-informed
20 change, we'll look and see if we can do that in a
21 performance-based fashion, so that connection is made in the
22 plan.
23 There is the issue of voluntary versus mandatory,
24 which on the reactor side, we've addressed it for the
25 reactor regulations. You know, for the plant oversight
. 149
1 process, it's not voluntary. Plants are all under the new
2 oversight process, so, you know, the answer can vary.
3 Selective implementation: That can be different,
4 depending on the activity you're looking at. And regulatory
5 oversight, you know, whatever you're risk-informing, you've
6 got to think about how are you going to factor this into
7 your oversight program, because it's not going to be
8 business as usual anymore.
9 So it's to get people to think about those things.
10 We can skip over to Slide 10. When you get to
11 Part II, which is arranged according to the three arenas,
12 rectors, materials, waste, you go to the strategic plan and
13 each of those three arenas has basically four performance
14 goals.
15 They are basically the same for each arena:
16 Maintain safety; increase public confidence; make things
17 more efficient, effective, realistic; reduce unnecessary
18 regulatory burden.
19 DR. APOSTOLAKIS: Isn't the fourth bullet included
20 in the third?
21 MR. KING: It could be, but if you go to the
22 strategic plan, it's not. I mean, this is right out of the
23 strategic plan.
24 And what we've tried to do in putting this
25 document together, was tie it to the strategic plan.
. 150
1 And the way we've done that under each arena, is,
2 we've said, okay, we've got a performance goal, maintain
3 safety, for example, and if you go to the strategic plan,
4 there's a number of strategies under that that talk about
5 how are we going to maintain safety?
6 Some of those strategies talk about using risk
7 information. So those are the ones that show up in the
8 document, and then under each of those strategies, we've
9 listed the activities that we believe either are underway or
10 should be underway to implement that strategy.
11 And then as part of that and part of things we
12 need to take a harder look at is, okay, for each of these
13 activities, you know, what are the infrastructure needs, the
14 training needs, the communication needs?
15 Now, I think one of the important things that you
16 can get out of doing it this way is, you can go back and
17 say, okay, I've got all these strategies from the strategic
18 plan; do I really have sufficient activities underway to
19 make those strategies come true? You know, it's a good way
20 to identify gaps.
21 And we've take a first cut at this, at this plan,
22 but I think there is a more systematic look that needs to be
23 taken over the next six months, and that's certainly where
24 we want feedback from all the stakeholders. Does it do the
25 job?
. 151
1 Now, Slides -- what I did on Slides 12, 13, and
2 14, was just list the strategies under the various goals
3 that pertain to a risk-informed activity. And these are the
4 things that show up in Part II of the plan.
5 I'm not going to repeat them here. You can read
6 them for yourself.
7 Maybe I'll just go to Slide 15, communications.
8 I don't know if you are aware but the Agency has
9 embarked upon developing what they call communication plans
10 in a number of key areas -- risk-informed regulation is one
11 of those.
12 There are probably two pieces to that. There's
13 the communication that goes along with this plan and I view
14 that as really, you know, having a twofold purpose. One is
15 to tell people what is risk-informed regulation -- what is
16 it all about? -- because we have been criticized in the past
17 from some stakeholders that risk-informed just equals burden
18 reduction and that is not really the case.
19 I mean the idea is to focus on the most important
20 aspects related to safety and that safety is our first
21 priority and I think by laying out a document like this that
22 clearly states that, hey, we're not throwing away
23 Defense-in-Depth safety margins, ALARA, those kinds of
24 things -- they all have to be factored in here -- it can
25 present a more comprehensive picture of what do we mean by
. 152
1 risk-informed regulation and what are the important things
2 that go into making a risk-informed change.
3 DR. WALLIS: On this business of not throwing away
4 a safety margin, you have now, say, a very conservative
5 estimate of LOCA, thus and so on, and Appendix K. You don't
6 really know how conservative it is now, how much margin
7 there really is now, do we?
8 So if you say we are not going to throw away any
9 margin, we don't know quite where we are now, so what are we
10 going to throw away?
11 DR. KING: We are not going to throw away the
12 principle of safety margins.
13 We may get rid of excess margins in the way we do
14 business today.
15 DR. WALLIS: You need to know more about what the
16 margins really are.
17 DR. KING: Yes, you do. Now that is not going to
18 come out in the general guidance. That is going to come out
19 in the details of implementing, you know, a change to 5046,
20 for example.
21 DR. APOSTOLAKIS: When you say "communication" and
22 the word "stakeholder," I think you usually mean
23 Intervenors, don't you?
24 DR. KING: No, I think the industry is a big
25 stakeholder.
. 153
1 DR. APOSTOLAKIS: Yes, the industry too.
2 DR. KING: The Staff is big stakeholder. You
3 know, there is education that needs to be done within NRC to
4 make people be aware of what is this all about, are we still
5 putting safety first?
6 There are still some people that think we are not
7 when we go into this regime.
8 DR. APOSTOLAKIS: Would a sub-bullet that says
9 something to the effect that we will use technically sound
10 methods or whatever, would that serve any purpose, that we
11 are --
12 DR. KING: I think the one that says "basis for
13 change is well-grounded" is intended to -- includes that.
14 We are not -- the intent is not to go off and use
15 untried methods or --
16 DR. SEALE: No flights of fancy.
17 DR. KING: Flights of fancy or fly-by-night
18 information. I mean that gets to the heart of the
19 importance of PRA quality -- people concerned about how good
20 are these risk estimates, is the information publicly
21 available, those kinds of questions really contribute to
22 convincing people that, yeah, we are doing things in a
23 fashion that does have a good basis --
24 DR. APOSTOLAKIS: Okay.
25 DR. KING: -- that does focus on safety and
. 154
1 hopefully when you get that kind of message across you get
2 better buy-in. That's buy-in across all the stakeholders.
3 The other key message to get across is what is the
4 Agency in risk-informed, where are we trying to go? So that
5 is really the part two of the plan.
6 Part one is what is this all about and that is a
7 key piece of communication, and part two is where are we
8 going and that is also a key piece of communication, and so
9 we have to go out and get this message across. That's
10 something we'll be trying to do over the next six months.
11 You know, you flip the page and we'll go over to page 17.
12 What we have planned to do is issue a yellow
13 announcement -- this is for internal consumption mainly, but
14 external folks could get it as well, about the availability
15 of this plan, the purpose of the plan, solicit comments on
16 the plan.
17 We are going to put it on the NRC website. We are
18 going to arrange some stakeholder meetings, both internal
19 and external, to sit down and talk about this thing.
20 Other things we have kicked around internally you
21 won't see in the plan right now but we want to talk about
22 are should we make some changes to management directives,
23 for example or the rulemaking handbook or maybe the
24 regulatory analysis guidelines, to bring in some of these
25 principles and some of these thoughts on risk-informed
. 155
1 regulation.
2 DR. WALLIS: I have a stakeholder for you. When
3 you have some good examples of risk-informed regulation,
4 which is thoroughly explainable and rational and logical and
5 persuasive, go out and give lectures to a class of
6 sophomores.
7 DR. KING: Are you inviting us up to Dartmouth?
8 DR. WALLIS: Sure.
9 DR. KING: Okay.
10 DR. APOSTOLAKIS: And tell them you will issue a
11 yellow announcement.
12 DR. KING: Yes, and give them the website.
13 [Laughter.]
14 DR. SEALE: You may not get it back.
15 DR. WALLIS: It's already been invaded.
16 DR. KING: Okay. Key challenges. We talked
17 about --
18 DR. LEITCH: Just a question here. I am looking
19 at the plan, part three, chapter 2, page 1, and it talks
20 about the audience for this communication strategy.
21 It says, "Although it is indeed vital to
22 communicate effectively with external stakeholders, the
23 primary focus of this communication plan is" -- and it
24 speaks about internal communication.
25 I don't see anything else in the plan about
. 156
1 external stakeholders.
2 Am I missing it here someplace? I thought
3 everything that was related to communications was in this
4 Part 3. That is the only thing I see, and it is just a very
5 tangential reference to external stakeholders.
6 DR. KING: I think the focus of what is in Part 3
7 now is internal stakeholders. You're right.
8 We mentioned it in the cover memo that sent it to
9 the Commission that we want to go externally to -- part of
10 what we -- you know, a lot of what is going to take place in
11 communication externally is going to be the details of what
12 we are doing, which would show up in Part 2, but I think you
13 have got a good point.
14 We ought to say more about the external in Part 3
15 as well.
16 DR. BONACA: I have a question, Tom. I was
17 looking back at page number 12, reactor arena, maintain
18 safety, and you have a number of strategies and when I look
19 at a lot of the plants we have here, you know, you have a
20 plant look at 50.46 and that essentially will provide a
21 better understanding, maybe a best estimate of what margin
22 there is available there.
23 A natural outcome of that may be have even higher
24 power uprates than what already are being planned -- just an
25 example.
. 157
1 We are looking at the PS rule and again where the
2 objective there is to make probably more margin available
3 for the life of the vessel and so many of these initiatives
4 individually, their burden I understand we're still shooting
5 for maintaining safety.
6 You know, one concern we debated here within the
7 ACRS is what kind of risk information are you trying to
8 develop to assess the implications of all this, let's say,
9 capturing of margins by licensees to operate the plants in a
10 different regime, that, you know, we already talked about
11 the synergistic effects that may result from a plant for
12 example going to power uprate, extending life, having cycles
13 with the more aggressive fuel designs, so on and so forth.
14 I mean is there any plan to be more capable of
15 understanding that? One example we have discussed is the
16 ability of modeling aging into the PRAs.
17 DR. KING: You won't find in this plan anything
18 directed specifically toward your question. We talked about
19 this yesterday as probably something that certainly needs
20 attention and I would expect when we decide what that
21 attention should be, it would show up in the plan, but right
22 now it is not in there.
23 DR. BONACA: I think it will be an important
24 element at some point. This is risk information regulation.
25 In many ways it will trade margins, provide additional
. 158
1 margins for further use by the licensees and there is an
2 implication at large by all this being used, being made and
3 so you are going to look at this issue, give some
4 consideration to that?
5 DR. KING: As Ashok mentioned yesterday, we took a
6 commitment to go look at that issue. Exactly what we are
7 going to do we have to figure out yet.
8 DR. BONACA: Because I think if you really scratch
9 the current regulation and you started with a blank sheet,
10 you probably would start with the first step of
11 understanding comprehensively what margins you have and how
12 you are going to divide them.
13 What I am trying to say is that it seems to me the
14 effort goes so much in the direction of looking at existing
15 rules and for each one of them to have a better
16 understanding of the basis and each one of them seems to me
17 like it's ending up providing additional margin to licensees
18 for those initiatives I mentioned, and so that begs the
19 question of what else is out there and can we do all these
20 things together.
21 DR. KING: Yes, and it is a question whether it is
22 a risk-informed change or a non-risk-informed change. It is
23 the same question because you have got plants doing power
24 uprates that aren't necessarily tied to any risk-informed
25 new rule or guidance document, but you have still got to ask
. 159
1 yourself a question what are the synergistic effects.
2 I think also in terms of doing Option 3 for
3 example, when we are changing 50.46 and maybe changing the
4 reactivity insertion accident and that kind of stuff, what
5 are the synergistic effects is a fair question to -- like I
6 said, you won't see that in the plan right now but I think
7 that is a good comment.
8 Okay. The Commission had asked for -- they called
9 it "impediments," we call it "challenges" -- what we think
10 are the most important challenges to implementing
11 risk-informed regulation. What I have listed on the
12 slide -- none of these are new to the committee but I think
13 these are the kinds of things when we start looking at
14 critical path activities are real things that could be
15 show-stoppers.
16 We are going to have to look carefully at how the
17 plan deals with these in terms of schedule and approach and
18 I am sure there's other things we could add to the list.
19 Where are we going with this? As I said, we are
20 now going into the solicit internal and external feedback
21 stage.
22 CHAIRMAN POWERS: Go back to your challenges. You
23 don't need to put the slide up. Just a question.
24 You spoke to the quality of the PRA. Don't you
25 have a challenge on quality and Appendix B? Isn't that a
. 160
1 major challenge that you need to address here as well?
2 DR. KING: I'm not sure I understand. Appendix B
3 in terms of when someone does a PRA, how do they make sure
4 they've got configuration control and all of that?
5 CHAIRMAN POWERS: It is the perennial problem of
6 is there any risk reduction work to quality activities. It
7 is something that the risk assessment technologies don't
8 address but it is a major part of Chapter 50.
9 I am just wondering if that isn't a challenge that
10 sits there simply because it is probably, if you can
11 risk-inform Appendix B we may have the biggest unnecessary
12 burden reduction that I can think of.
13 DR. KING: One of the things on our plate under
14 Option 3 is to risk-inform the special treatment
15 requirements themselves, not the scope of what things get
16 special treatment -- that is Option 2 -- but to go in and
17 look at the special treatment requirements themselves and
18 say which of these things have some risk implications and
19 which don't? Should we make changes in Appendix B, EQ?
20 Mary has got the whole list but we are looking at those
21 under Option 3 right now.
22 It is rough to sort out, you know, which are the
23 important things, which ones contribute and which ones
24 don't.
25 As I mentioned, you know, we would like probably
. 161
1 to schedule a Subcommittee meeting at some point in the
2 future to get into this in more detail. And we're going to
3 have some workshops to get some feedback.
4 We have a Commission briefing scheduled for
5 November 17th, to get their preliminary views on the plan,
6 and any direction they want to give us regarding the plan.
7 Then, as I said, we recognize there are some
8 followup activities. We need to apply the criteria. We're
9 on Slide 8. We need to take a look at what are our critical
10 path items, and develop some better integrated schedules.
11 We need to take a look at under each of the
12 strategies, do we have activities sufficient to implement
13 those, and what are our particular infrastructure and
14 training needs? Also, success measures was one of the
15 things that GAO had talked about, and I'll have you know
16 we've achieved success in this. We need to think about that
17 as well.
18 And then just in summary, we're not sitting still
19 while all of this takes place. We're continuing to do a lot
20 of things in risk-informed regulation, you know, Option 2,
21 Option 3, and NMSS is embarking on case studies to look at
22 materials and waste areas that are candidates for
23 risk-informing.
24 And we do plan to update this every six months.
25 So, in May or so you'll see another version. So that's my
. 162
1 overview, unless you have any questions.
2 DR. LEITSCH: Just one comment from kind of taking
3 a cold-read through this, and particularly as this applies
4 to external stakeholders: The theme that seems to come
5 across many places is reduce unnecessary regulatory burden.
6 But there is a statement in Part I-1, there's a
7 sentence there that I think perhaps needs to be emphasized
8 more uniformly throughout. It says: Risk-informed
9 regulatory approach can also be used to identify
10 insufficient conservatism and provide a basis for additional
11 requirements and regulatory actions.
12 I think there is so much emphasis in the plan here
13 that risk-informed is becoming synonymous with reducing
14 regulatory burden, and I think we -- I think a balanced
15 approach here, it's important to communicate that. And as I
16 say, it comes out in that one sentence, but I don't see much
17 -- I don't see any other discussion of that throughout the
18 report.
19 MR. KING: Okay. That's a fair comment. I think
20 this has and risk-informed has certainly been perceived as
21 an unbalanced approach. And I think if you look a the first
22 rule we recommended to be risk-informed, the combustible gas
23 regulation, you know, one of the recommendations in there
24 was to increase the requirement on two types of
25 containments, which was, we felt, plugging a hole currently
. 163
1 in the regulation.
2 So it's a fair comment.
3 DR. APOSTOLAKIS: I wonder whether the words,
4 insufficient conservatism are the right ones, in what Mr.
5 Leitsch just read? Maybe insufficient conservatism is not
6 the right terminology in a risk-informed system.
7 Just think about it. I don't have an answer, but
8 --
9 MR. KING: Okay.
10 DR. APOSTOLAKIS: So you've summarized your
11 summary?
12 MR. KING: I've summarized my summary.
13 DR. APOSTOLAKIS: Okay, any other questions?
14 [No response.]
15 DR. APOSTOLAKIS: Back to you, Mr. Chairman.
16 Thank you, Tom.
17 CHAIRMAN POWERS: I'm just thinking that what
18 they're doing right now is, they're soliciting feedback, and
19 I'm wondering if this is one of those things that ACRS
20 collectively ought to get together and give some thought to,
21 perhaps at a retreat or something like that, and be in a
22 position so that we could give them some feedback on issues
23 just like the one Mr. Leitsch just brought up.
24 They're putting a substantial amount of effort,
25 and maybe we can, outside of the Subcommittee structure -- I
. 164
1 mean, in preparation for a Subcommittee meeting, give some
2 abstract thought to this, and then come back in a
3 Subcommittee meeting, say, sometime after February or
4 something like that, after you've had a chance to get some
5 feedback and we've had a chance to think about this
6 collectively, might be a strategy to adopt here.
7 I think to keep having this hope that good plans
8 make subsequent work easier -- maybe they don't, but I keep
9 hoping that.
10 Okay, on that note, I will recess this until 20
11 minutes of 2:00, I guess.
12 [Whereupon, at 12:40 p.m., the meeting was
13 recessed, to reconvene at 1:40 p.m., this same day.]
14
15
16
17
18
19
20
21
22
23
24
25
. 165
1 A F T E R N O O N S E S S I O N
2 [1:40 p.m.]
3 CHAIRMAN POWERS: Let's go back into session. Our
4 next topic is proposed framework for risk-informed changes
5 to the technical requirements of 10 CFR Part 50. And
6 looking at the cast of dubious characters before us, I think
7 I probably have some modest organizational conflict of
8 interest. So though I may well pose questions that the
9 speakers will have to answer, the Committee need not pay any
10 attention to any comment I make on this subject.
11 @@ DR. KRESS: There's no such thing as a modest
12 conflict. It's like being pregnant -- you either are or
13 you're not.
14 CHAIRMAN POWERS: I am or I aren't?
15 In this case I have a feeling I have a severe
16 organizational conflict of interest.
17 MS. DROUIN: But that's not fair. If they don't
18 have to listen to it, why should we?
19 CHAIRMAN POWERS: Because I have to live with
20 them, and you only show up episodically.
21 DR. APOSTOLAKIS: But it's very --
22 CHAIRMAN POWERS: Hey, when General Drouin brings
23 her troops in here -- the Desert Fox, I can see this now.
24 With that introduction, I'll turn to Professor
25 Apostolakis and ask him --
. 166
1 DR. APOSTOLAKIS: Are we back in session?
2 CHAIRMAN POWERS: I put us back in session.
3 DR. APOSTOLAKIS: Okay.
4 CHAIRMAN POWERS: And see if you can give us a
5 proper introduction to this subject.
6 DR. APOSTOLAKIS: Well, we've seen this before, at
7 least twice that I recall. We were trying to schedule a
8 subcommittee meeting, which we never managed to do. Now
9 this is something that has already gone up to the
10 Commission, right? The SECY.
11 MS. DROUIN: Yes, 198.
12 DR. APOSTOLAKIS: Yes, which is what we're
13 discussing today.
14 MS. DROUIN: The framework part of 198.
15 DR. APOSTOLAKIS: Yes, the framework part, option
16 3. The Commission has not voted yet?
17 MS. DROUIN: The framework part, we did not ask
18 the Commission to vote on the framework. That was just for
19 information purposes only.
20 DR. APOSTOLAKIS: Ah.
21 MS. DROUIN: And as in the past SECYs, as in this
22 SECY, we've indicated that still this is a work in progress.
23 As we move forward, particularly with 5045, that's going to
24 be much more challenging. The framework could still undergo
25 revisions and changes.
. 167
1 DR. APOSTOLAKIS: And you would like this
2 Committee to write a letter?
3 MS. DROUIN: Not at this point.
4 DR. APOSTOLAKIS: Oh, you're not requesting a
5 letter?
6 MR. KING: I thought you did write a letter at one
7 time. I mean, we're not -- let me put it this way. We're
8 not requesting a letter. If you --
9 DR. APOSTOLAKIS: On the framework.
10 MR. KING: On the framework. If you have comments
11 and want to send them to us formally, we'll certainly, you
12 know, be glad to receive your letter, but we're not, you
13 know, we're not expecting one, we're not requesting one, and
14 if we don't get one, it's not going to hold us up.
15 DR. APOSTOLAKIS: Okay. So why don't we -- now we
16 have to finish this in an hour, so I'll leave it up to you,
17 Mary, to organize the presentation so we can finish in an
18 hour, including the questions.
19 MS. DROUIN: Well, that depends on your questions.
20 I have no control over that.
21 DR. APOSTOLAKIS: It depends on what you say.
22 MS. DROUIN: Okay. We're ready.
23 Okay, before we get started, just some
24 introductions. My name is Mary Drouin, with the Office of
25 Research. Also at the table with me is Alan Camp from
. 168
1 Sandia National Lab, Eric Haskin from ERI Consulting, Alan
2 Kuritsky from also the Office of Research, and Trevor Pratt,
3 from Brookhaven National Labs.
4 DR. APOSTOLAKIS: Okay.
5 MS. DROUIN: So we have the -- pretty much almost
6 the entire team that has worked diligently in putting this
7 framework together.
8 DR. APOSTOLAKIS: Now you are usually very good at
9 going to the heart of the matter. So can you exercise this
10 expertise today and skip the background and all that stuff?
11 MS. DROUIN: Go for the jugular?
12 DR. APOSTOLAKIS: Yes.
13 MS. DROUIN: Okay. Can I spend ten seconds on the
14 background?
15 DR. APOSTOLAKIS: We know it.
16 MS. DROUIN: Just that there's been SECYs, SRMs
17 directing us to do this work.
18 DR. APOSTOLAKIS: We never worried that you didn't
19 have enough to do.
20 MS. DROUIN: I won't even bother to put up the
21 objective slide.
22 DR. APOSTOLAKIS: That's fine.
23 MS. DROUIN: The purpose of the framework I think
24 as everyone knows is to help us formulate -- review, and
25 formulate the risk-informed alternatives to the technical
. 169
1 requirements of 10 CFR part 50.
2 DR. APOSTOLAKIS: I guess if you go to slide 8 you
3 will be making a lot of progress.
4 CHAIRMAN POWERS: Mary, could I ask a question
5 just for personal information?
6 When you formulated the various options that you
7 had to pursue this activity, what was your rationale for not
8 going -- taking a clean-sheet-of-paper approach?
9 MS. DROUIN: That is part of our approach. In the
10 implementation of the framework, we take two paths in using
11 the framework itself. One path is to look at the current
12 technical requirements and risk-inform those and come up
13 with options. Simultaneously in parallel with that we also
14 start with what we -- looking at the concern. If I look at
15 5044, for example, where the concern was hydrogen combustion
16 and challenging the containment integrity, if you just
17 started with that, you know, and you started with a blank
18 piece of paper, what risk-informed options?
19 CHAIRMAN POWERS: What I was wondering about is
20 suppose there were no nuclear reactors in this country and
21 somebody came forward and said I want to build a nuclear
22 reactor, and you were given the task of okay, what should I
23 require on this, and you had at your disposal all these risk
24 tools?
25 MS. DROUIN: All these risk what?
. 170
1 CHAIRMAN POWERS: Risk-analysis tools. Could you
2 write a set of regulations that would ensure that those
3 reactors were adequately safe in this country from scratch
4 essentially?
5 MS. DROUIN: I think so, yes.
6 CHAIRMAN POWERS: And I wonder why -- maybe you
7 are pursuing that. Is there anyone pursuing that kind of an
8 idea? The reason for asking about it is that we have a
9 variety of people suggesting that maybe they would come
10 forward with a gas-cooled reactor, look completely different
11 than any water-cooled reactor, probably have completely
12 different risk characteristics to it and what not, and so
13 even risk-informing the current regulations probably leaves
14 them inapplicable in some sense to this completely novel
15 design.
16 MS. DROUIN: That's probably very true what you
17 say. We had to stay within the scope of what we were asked
18 to do by the Commission, and we were asked to start with the
19 current set of the technical requirements in 10 CFR part 50.
20 So that right there set in some sense our approach where
21 that, you know, we couldn't throw away that whole body of
22 regulations and say how would I rewrite the entire set from
23 scratch. But since there was some benefit as we look at
24 each individual regulation, we are looking, as I said, at
25 the concern and looking at it from a fresh approach given --
. 171
1 pretending that there were not any technical requirements.
2 CHAIRMAN POWERS: Let me put my two cents in on
3 that. I think if one were starting with a clean sheet of
4 paper, given the safety goals we have today for reactors, I
5 think if you were developing a new set of regulations, you
6 probably would start with something pretty close to what
7 this framework has right now in terms of laying out
8 objectives for partitioning between prevention and
9 mitigation --
10 MS. DROUIN: Yes.
11 CHAIRMAN POWERS: How you would set up selection
12 of accidents, rare events, you know, more likely events and
13 so forth. The fact that you've got to deal with safety
14 margins, the defense-in-depth concept, and I think nobody
15 today is working on a clean sheet of paper thinking about,
16 you know, pebble bed reactors or anything. But I think if
17 we get turned on to do that, I think this framework document
18 is the place I would start.
19 MS. DROUIN: And I would go on to add that that
20 thought in creating the actual framework itself we went more
21 with that line of thinking, even though it's being applied
22 on a regulation basis, that wasn't -- the thinking process
23 that we went through in developing the framework, we stood
24 back more and looked at it more holistically.
25 CHAIRMAN POWERS: One of the items that comes to
. 172
1 my mind on this is how do I use risk information to keep me
2 from having reactors with positive void coefficients when I
3 know your cut sets don't take elaborate account of
4 neutronics.
5 DR. APOSTOLAKIS: I think the event trees would be
6 very different, Dana.
7 MS. DROUIN: I'm sorry, George?
8 DR. APOSTOLAKIS: The event trees would be very
9 different if you had a positive reactivity coefficient.
10 Right now they're built with the assumption that you have a
11 negative reactivity coefficient, right?
12 CHAIRMAN POWERS: Exactly so. And so if I am
13 creating a risk-informed regulatory structure that presumes
14 the reactor has a negative void coefficient, how do I apply
15 that to a completely novel design? It has not had the
16 benefit of the review of the FSAR and the design-basis
17 accident analyses, et cetera.
18 DR. APOSTOLAKIS: I can give two answers to that.
19 The first one is if you had a positive reactivity
20 coefficient, it would be very hard to meet the safety goal,
21 and the safety goal is there.
22 CHAIRMAN POWERS: The CANDUs have positives.
23 DR. APOSTOLAKIS: Huh?
24 CHAIRMAN POWERS: CANDU 6 has a positive void
25 coefficient.
. 173
1 DR. APOSTOLAKIS: All right. And the second is I
2 don't think, you know, for a new reactor, what is very true
3 is that what is the review that the NRC will do in the new
4 framework has not been determined. See, this is really like
5 water reactors, starting with the existing regulations, as
6 Mary said.
7 DR. SEALE: If you go back and remember what
8 happened in the LMFBR program back when the skies were clear
9 and the gods were tall and so on, there was an awful lot of
10 effort that went into trying to cope with the effect of a
11 positive void coefficient and ultimately the spoiled core
12 that essentially enhanced the leakage from the part of the
13 void was the response to that. So that in effect you got a
14 negative void coefficient from that effort. Now I don't
15 think there was an a priori requirement that there be a
16 "negative void coefficient," but there was an awful lot of
17 agony that went into looking at alternatives to in effect
18 meet that requirement.
19 DR. APOSTOLAKIS: I think I agree with Tom King's
20 and Mary's position that if you wanted to develop a
21 framework for an entirely new generation of reactors, you
22 would have to rethink the whole thing from scratch, but this
23 would be a good place to start.
24 DR. KRESS: You would have to --
25 DR. APOSTOLAKIS: We should not stretch the
. 174
1 applicability of this particular framework.
2 DR. KRESS: I think you would have to sharpen your
3 pencil a little bit, because the framework has things in it
4 like the balance between the conditional core damage
5 probability versus CDF, and if you -- but it has no
6 flexibility in that. That depends on how good your PRA is
7 and what are the uncertainties in both of those numbers, and
8 I think -- and what is the actual risk level you have. And,
9 for example, in the pebble-bed-type reactor, your CDF may be
10 10 to the minus 9 or something, and then the question comes
11 about, do you really need this .1 C conditional containment
12 failure probability under those circumstances. And what's
13 the uncertainty in that number, that 10 to the minus 9?
14 DR. APOSTOLAKIS: That's why this is --
15 DR. KRESS: So, yes, a good starting point, but
16 you'd have to sharpen up things like that.
17 DR. APOSTOLAKIS: Sure. Sure.
18 DR. KRESS: You couldn't.
19 CHAIRMAN POWERS: You'd still be looking at a
20 balance between convention and mitigation.
21 DR. APOSTOLAKIS: That's right.
22 CHAIRMAN POWERS: You wouldn't have it in those
23 particular numbers admittedly, but I think the notion that
24 you consider both of those is what they really mean by the
25 framework.
. 175
1 DR. APOSTOLAKIS: You would look at cornerstones,
2 but you would probably have to redefine them if necessary,
3 strategies. I mean, that's what I understand by a starting
4 point.
5 DR. SHACK: Mary, I have a question on your slide
6 number 5, to frustrate your going to 8. Are there special
7 therapeutic values that make this appropriate for staff use
8 but not for licensees?
9 MS. DROUIN: What this meant is that this is a
10 framework and it has quantitative guidelines, and when we
11 say it's for staff use, these are guidelines to help us
12 formulate a risk-informed regulation. These are not
13 guidelines or criteria or requirements that an individual
14 licensee is going to have to meet.
15 DR. SHACK: Okay.
16 MS. DROUIN: I'll just spend literally 30 seconds
17 on slide 6. I want to go through these at this point, but
18 the thing that I want to point out is that when you see the
19 framework, you will see aspects from each of these that were
20 built directly into the framework, and if you go to each of
21 these documents, you know, there's a lot of discussion in
22 there on what is meant by risk-informed regulation, what is
23 meant by defense in depth, and how these two complement each
24 other, and it was from these documents that we built the
25 framework. And that's all.
. 176
1 I will go ahead and get to 8. I think I can
2 capture -- everything that I was going to say on 4 I'll try
3 to say as I look at this slide, which shows the actual
4 framework.
5 In building the framework, we wanted what we
6 called a risk-informed, defense-in-depth approach, and
7 starting with that we want to protect the public health and
8 safety, and that is ultimately our goal with the
9 regulations, is to build a framework that tied from the
10 regulations down here, to the goal and how do you achieve
11 that.
12 So we developed this thing where we called this
13 risk-informed, defense-in-depth, but it's a hierarchical
14 approach where you start with the goal of protecting the
15 public health and safety, and in order to achieve that,
16 going to the reactor oversight program, looking at their
17 cornerstones for the safe operation of nuclear power plants.
18 They had their three major cornerstones of reactor
19 safety, radiation, and security. And then there are four
20 sub-cornerstones for reactor safety.
21 And then that led us to developing the four
22 strategies that you see there for accident prevention and
23 mitigation. Where we then -- I don't know if you --
24 DR. KRESS: Mary, with a steam generator tube
25 rupture accident, that challenges 10 CFR 100 because it's
. 177
1 just got the iodine spike and the iodine in the normal
2 coolant, didn't fit in this thing. Would that be under
3 radiation safety?
4 I mean, where would the regulation -- where would
5 you consider a regulation like that in there?
6 MS. DROUIN: I think it could fit in several
7 locations. I mean, if you have the frequency, you're going
8 to have your core damage, the frequency associated with it,
9 you're going to fit under Strategy 3, of having releases.
10 To me, this is one that cuts across all the
11 strategies.
12 DR. KRESS: Barrier integrity.
13 MS. DROUIN: Barrier integrity.
14 DR. KRESS: I'm just talking about the rule set up
15 with the design basis accident, and what you have to meet as
16 an acceptance criteria is a 10 CFR 100 dose. You don't ever
17 go to core melt. It does --
18 There is some challenge to core melt, but it's
19 down in the -- and the question is, you still have a rule on
20 the books that limits this dose, and it's a regulatory
21 objective to limit this dose.
22 Where would it fit under this framework, that kind
23 of rule; where would it fit under this framework?
24 If it doesn't challenge the core damage, you'd
25 just say we don't need it, because it's not a contributor to
. 178
1 CDF that we worry about. But it does release radioactivity,
2 and so, you know, I worry about that kind of rule in this
3 framework.
4 MS. DROUIN: Ultimately, every rule will fit under
5 one of the reactor -- I'm sorry, one of the plant oversight
6 cornerstones. Right now, our scope is limited to the
7 reactor safety cornerstone.
8 DR. KRESS: I understand that. I was tempted to
9 put it under radiation safety, but I'm not sure what all
10 radiation safety encompasses?
11 MR. KURITSKY: Part 20.
12 DR. KRESS: Oh, Part 20.
13 DR. BONACA: The leakage rate that you will have
14 will come after -- because you will have assumption of
15 barrier degradation, mitigation systems coming in, and then
16 you have a limit, radiation limit that comes from the
17 leakage plus the amount of activity. It's actually not
18 measured; it is really what is assumed to be, which is one
19 percent.
20 And so I think you will come on the left here.
21 MS. DROUIN: Right, if you go to --
22 DR. BONACA: I think, but I don't know.
23 MS. DROUIN: -- the NUREG on the cornerstones, and
24 you look at those four -- I have the words here -- the
25 initiating event is what they're talking about. There is
. 179
1 minimizing the events that could lead to an accident; the
2 next one is assure the ability of safety systems to respond
3 to and lessen the severity of an accident; maintain barriers
4 to the release of radioactivity in accidents, and plans by
5 the utility and government agencies to shelter or evacuate
6 people in the community in the event of a severe accident.
7 To me, clearly, it can fit under that cornerstone.
8 MR. SIEBER: It would have to fit there, because
9 prior to its actual occurrence, you wouldn't know whether it
10 was going to lead to CDF or not. You would just have a
11 probability that it might.
12 And if it doesn't lead to CDF, Part 100 is never
13 even approached.
14 DR. KRESS: I hope, prior to it happening, I have
15 some idea of whether it's going to lead to CDF.
16 MR. SIEBER: You know that it's going to --
17 DR. KRESS: He's going to tell me --
18 MR. SIEBER: How do you know that there is a
19 probability that it could? And so that's what make it fit
20 under the reactor safety here.
21 DR. APOSTOLAKIS: Let's take another set of
22 sequences. I remember we discussed this in one of the
23 earlier meetings, the role of the so-called external events.
24 My impression is that this particular diagram has
25 been developed really for internal events. Right now, the
. 180
1 language -- well, let me take an example.
2 Let's say that we have a fire analysis. When you
3 are asking me to limit the frequency of accident initiating
4 events, is that the frequency of the fire? Or is that the
5 frequency of the PRA initiating events and nuclear
6 initiating events that the fire induces?
7 MS. DROUIN: We are talking about limiting the
8 frequency when you go to the strategies there of an
9 accident. We are talking about a full-scope PRA.
10 DR. APOSTOLAKIS: Right, so the initiating event
11 is what, the fire?
12 MS. DROUIN: They would include both internal and
13 external events.
14 DR. APOSTOLAKIS: No, but usually in a PRA when we
15 say initiating event, we don't mean the fire; we mean LOCAs,
16 transients, that kind of stuff. The fire is sort of a
17 common cause that affects other things.
18 MS. DROUIN: True, yes.
19 DR. APOSTOLAKIS: So, I remember you had a box at
20 one time somewhere there for external events which now is
21 here. I vaguely remember that, but what is the -- yes,
22 there was some discussion at least.
23 I don't recall any discussion of these big
24 dependencies in this particular document. So, I mean, I can
25 follow the same line of question from Tom.
. 181
1 If I take an earthquake or a fire and I read the
2 document where it says limit the frequency of accident
3 initiating events, what am I supposed to limit? The
4 frequency of the fire or the frequency of the various
5 initiating events of the PRA that are induced by the fire?
6 And in the second case, I'm having a problem,
7 because these are conditional frequencies now. And, of
8 course, if I take it to the extreme, I can have an
9 earthquake that wipes out prevention, mitigation, and, you
10 know, the whole thing.
11 But then you guys take care of that in the
12 initiating events by saying ten to the minus six initiators
13 are rare and forget about them.
14 So in that case, the earthquake is the initiator.
15 MS. DROUIN: See, you have to remember you're
16 applying the framework. It's the regulation that is applied
17 to the framework.
18 DR. APOSTOLAKIS: Right, so if I want to regulate
19 fires, I have a risk-informed approach, and I take this
20 framework, and I think I'm going to need more guidance. I'm
21 not saying that it's useless, but I'm going to need a little
22 bit more guidance as to how to apply it.
23 CHAIRMAN POWERS: I think that it seems to me that
24 fire in 50.48, is more attuned toward risk-informing than
25 most of the regulations, because it comes along and says,
. 182
1 okay, we will define defense-in-depth as limiting the
2 frequency of fires.
3 Barrier two is mitigating -- detecting them and
4 trying to suppress then; and the third one, the third
5 element of defense-in-depth is prevent the fire from doing
6 these other things that you're talking about.
7 So I think the answer that they would give you, if
8 they were to follow the current practice in fire risk
9 assessment, would be to say, yes, that they're doing both;
10 they're limiting the frequency of fire, and they're limiting
11 the possibility that that fire would cause something that
12 would lead to core damage.
13 DR. APOSTOLAKIS: But that would be a statement
14 from the fire protection guys or the fire risk assessors,
15 which I think is a correct statement.
16 But the question is, this framework, how would it
17 guide me to do these things for a fire, if I hadn't done
18 them?
19 See, when you say limit frequency of accident
20 initiating events, you somehow have to explain. And I think
21 that in the report it says we usually group them into LOCAs,
22 transients, and so on.
23 So the report reads as if the person who wrote it
24 had in mind, an internal event PRA. I would take care of
25 that.
. 183
1 One way to take care of that is -- well, first of
2 all, you have to address it. And a related question that I
3 have is, why isn't defense-in-depth mentioned in the box,
4 Tactics?
5 Because it will have to apply defense-in-depth in
6 some places when the uncertainties are very large, even at
7 this lower level.
8 And if you put it there, then you can say what
9 Dana just said, well, I'm applying that to this particular
10 initiator, because now defense-in-depth means making sure
11 the fire doesn't start, or if it starts, that it's detected
12 and suppressed and so on.
13 But that's a different application of the concept
14 of defense-in-depth.
15 There are too many people who want to say
16 something. I think, Allen was first.
17 MR. CAMP: I think that certainly we can clarify
18 the initiating event issue. I think that from one
19 regulation to the next, the way we go at it may be one way
20 or the other as limiting the fire or limiting the transient.
21 When we look at 50.46 as an example -- and I know
22 that's not why we're here today -- you'll see later one when
23 we talk about whether or not we can get rid of the Large
24 Break LOCA, part of that question is going to be the
25 seismically-induced Large Break LOCA.
. 184
1 And so that's going to be addressed implicitly as
2 part of the new 50.46 that comes forward. So it's going to
3 be in there, and in that case, we're going to be looking at
4 some combination of earthquake frequencies, the hazard curve
5 and where it cuts off at ten to the minus six or whatever;
6 also how the pipe responds and do you get a LOCA with what
7 probability?
8 And so it won't be totally one or the other that
9 we're looking at.
10 DR. APOSTOLAKIS: I didn't expect you not to do
11 that. I'm not arguing that you do not have to approach it.
12 MR. CAMP: We just need to clarify it in the
13 writeup.
14 DR. APOSTOLAKIS: But this particular part of the
15 document should reflect this kind of thinking, because this
16 is really a higher level document.
17 DR. BONACA: The only thing that I wanted to
18 mention was that the original thinking when the
19 deterministic approach was used, all the initiators were
20 proposed and evaluated with the intent of determining what
21 kind of reactor protection actions you had to have,
22 mitigating systems and so on and so forth.
23 So you really developed your accident mitigation
24 through that process. And they were two separate things.
25 And the way you looked at, for example, fire, was
. 185
1 what's the likelihood that the fire will start an initiator?
2 What are the initiators that challenge the RPS or the -- and
3 so on and so forth, and also defeat the mitigating features
4 for that particular accident.
5 DR. APOSTOLAKIS: Let's not misunderstand what I'm
6 saying.
7 DR. BONACA: I understand what you're saying, I
8 think I do. I was just trying to reason through the
9 process, because I think it plays, again, into the
10 risk-informed approach.
11 And you have a case in point. I'm always trying
12 to see if that fits the mold there or not. You may be
13 right; that maybe --
14 DR. APOSTOLAKIS: I believe that with a minor
15 change, this thing could reflect that. You know what you
16 need to do. It's not that, my god, I never thought of it;
17 it's just that this particular thing was developed having
18 really the internal events in mind.
19 And then you will see, as you start thinking about
20 it, that now you have to explain what you mean by limiting
21 the frequency of accident initiating events. Do I take the
22 traditional events or do I say that for these big external
23 events, I want you to limit this, okay, which is the
24 frequency of the initiator itself, even though we don't call
25 it initiating event, which brings you now to what Dana said.
. 186
1 Now you are in a different regime where you apply
2 your defense-in-depth. But that's one approach. I don't
3 know, there may be others, but I remember vaguely that at
4 some point we had discussed it. But somehow it didn't come
5 to --
6 Now, why isn't defense-in-depth in the tactics?
7 MS. DROUIN: Defense-in-depth is on the whole
8 structure. I mean --
9 DR. APOSTOLAKIS: What's the difference between
10 defense-in-depth and safety margins?
11 MS. DROUIN: You get defense-in-depth through the
12 safety margins. You get defense-in-depth by the fact that
13 you have these strategies and you're balancing across the
14 strategies.
15 DR. APOSTOLAKIS: I agree; I fully agree. But
16 there will be also some need for defense-in-depth as part of
17 the tactics. Well, I guess you call it redundancy,
18 diversity, and independence.
19 DR. KRESS: That's defense-in-depth.
20 DR. APOSTOLAKIS: That's defense-in-depth.
21 MS. DROUIN: So, to me, the defense-in-depth theme
22 is inherent through the entire framework.
23 DR. APOSTOLAKIS: Right. So it's not just the
24 fact that you separate prevention from mitigation?
25 MS. DROUIN: Right.
. 187
1 DR. APOSTOLAKIS: Right, okay.
2 CHAIRMAN POWERS: And we examine control systems
3 in nuclear power plants. The Staff tells us one of the
4 requirements they place on their licensee coming forward
5 with a new system --
6 MS. DROUIN: I can't hear you over here.
7 CHAIRMAN POWERS: One of the requirements they
8 place on licensees who come forward with new systems for
9 instrumentation and control in the nuclear reactor is
10 defense-in-depth.
11 Within your scheme, do you apply defense-in-depth
12 at low-level systems?
13 MS. DROUIN: Yes.
14 DR. APOSTOLAKIS: You said yes a minute ago.
15 MS. DROUIN: The hesitation of the yes is not so
16 much the hesitation, but I guess I'm not understanding the
17 basis of your question or where the concern is.
18 CHAIRMAN POWERS: The concern is that every time
19 we impose defense-in-depth, we introduce complexity into a
20 system.
21 Why not reserve our defense-in-depth measures to
22 far more -- higher integration of systems together, rather
23 than doing it on control systems, on piping systems, on
24 valves, on all of these little subsystems of an overall
25 system?
. 188
1 Why not say defense-in-depth really is the set of
2 barriers that come from emergency preparedness, the
3 containment, the reactor coolant system, pressure boundary?
4 Why have defense-in-depth on HVAC systems, on control room
5 systems?
6 We're just introducing complexity.
7 MR. KURITSKY: Excuse me. I think that we do
8 mention that our defense-in-depth, as we see it, the
9 overarching defense-in-depth is applied on those four
10 strategies.
11 The defense-in-depth at a system level, which I
12 think you were referring to, is as I think Dr. Apostolakis
13 pointed out, when we mentioned redundancy and diversity,
14 that's where you're getting down to the more detailed level.
15 But the defense-in-depth main focus that we have
16 and the framework is on that strategy level. I think the
17 framework document does spell out that that is.
18 In fact, on page 2-3, we say defense-in-depth is
19 the approach taken to protect the public by applying the
20 following strategies in a risk-informed manner, and it lists
21 those four strategies.
22 DR. APOSTOLAKIS: Let me put the question in a
23 different way: I have defense-in-depth now because I have
24 four cornerstones and I have for each one, I suppose, a
25 strategy of prevention and mitigation, although that's not
. 189
1 really --
2 So that's defense-in-depth at a very high level.
3 Why are you asking me to be redundant when I design my high
4 pressure injection system?
5 MR. CAMP: George, if I could answer that, I see
6 the redundancy, diversity, all those things down at the
7 bottom as options that you could use if you need to for
8 particular system or regulation. And they're not things
9 we're aspiring to.
10 DR. APOSTOLAKIS: Then why do I need it?
11 MR. CAMP: Well, you need it when you need
12 something to help you meet those quantitative objectives
13 that we're setting.
14 DR. APOSTOLAKIS: That's right.
15 MR. CAMP: And an example is, if you look at
16 containment heat removal, we haven't touched it yet, so what
17 I'm saying now may or may not ever happen. But if you look
18 at the requirements now that those meet single-failure
19 criteria, we may decide when we get there that that doesn't
20 -- that that isn't necessary.
21 So, we would look at redundancy and single-failure
22 criteria as a tool that we would examine when we got to
23 containment heat removal, and see if we still needed it.
24 DR. BONACA: In fact, the driving force behind the
25 way this has been developed, the deterministic way, has been
. 190
1 the single-failure criterion.
2 I mean, that drove not only the redundancy; in
3 some cases drove the diversity because of concerns with the
4 specific function you had to perform. And now that you go
5 to a more insightful system by which you can reevaluate
6 whether or not you really need to assume a single-failure,
7 given, for example, that it would be extremely unlikely,
8 then you will have a lot more flexibility now to --
9 CHAIRMAN POWERS: It seems to me that that's what
10 I want to maximize, is that when people design subsystems,
11 they should have the freedom to use the best engineering
12 imagination they have, and make it as reliable as possible,
13 and not be obligated to introduce defense-in-depth in that
14 design which creates complexity.
15 DR. BONACA: In fact, you could set a goal that
16 says I will apply single-failure criterion, okay, and decide
17 that I'm not going to have a redundant system if the
18 reliability of the system is above a certain threshold. We
19 could make certain ones of that type.
20 DR. APOSTOLAKIS: I think it is all a matter of
21 uncertainty.
22 DR. BONACA: Of course.
23 MS. DROUIN: Yes.
24 DR. APOSTOLAKIS: Your uncertainties are very
25 large and, as Alan says, as you move up you realize that you
. 191
1 are going to have difficulty demonstrating that you met the
2 various goals you are going to show us in a minute.
3 Then you add extra protection measures, which is
4 the ideal of structuralism, right?
5 CHAIRMAN POWERS: Rationalist.
6 DR. APOSTOLAKIS: No. Because it is unquantified.
7 CHAIRMAN POWERS: Oh.
8 DR. APOSTOLAKIS: This is the preliminary proposal
9 that mixes the two?
10 DR. SEALE: You don't have to listen to that --
11 [Laughter.]
12 DR. APOSTOLAKIS: If you have already quantified
13 it, and again the uncertainties are very large, then you are
14 adding -- but that I would not call those Defense-in-Depth.
15 It seems to me that the Defense-in-Depth idea is
16 really the structuralist idea. I am uncomfortable. I am
17 putting things there to protect me. Otherwise it is just
18 engineering. You know, I have a common metric, an
19 availability. I am doing tradeoffs.
20 DR. BONACA: I gave you an example already. I
21 mean if I do not assume a failure probability for a
22 component, assume that it could fail, period, then that
23 would force me into a situation of saying, well, I need to
24 have some ways to cope with the situation.
25 Conversely, if you have --
. 192
1 DR. APOSTOLAKIS: Yes, but they are not asking you
2 to assume things are down and then see whether you like it.
3 See, that is the whole idea of a risk-informed approach.
4 I mean you have to take into account the
5 probability that the thing will go down, right?
6 DR. BONACA: That's right.
7 DR. APOSTOLAKIS: That's the whole idea so --
8 DR. BONACA: Sure.
9 MS. DROUIN: I am not real sure what slide to go
10 to.
11 DR. APOSTOLAKIS: Moving along, I would like to
12 congratulate you -- no, don't move yet -- I would like to
13 congratulate you on the excellent Section 4.2 on Safety
14 Margins where you clearly call upon people to accept the
15 typical probabilistic definition -- safety margin is the
16 probability or level of confidence -- somebody who didn't
17 understand probability wrote that -- that the design or
18 process will perform an intended function.
19 That is on page 4-2 on the left. This is great
20 and I hope the Research Program will take that into account
21 in the next round.
22 MS. DROUIN: Are you talking about there on the
23 top?
24 DR. APOSTOLAKIS: Yes, yes, yes -- believe me,
25 it's there -- and I am praising you. You don't have
. 193
1 question what I am saying, right? I am saying this is
2 good --
3 MS. DROUIN: I am thanking you.
4 CHAIRMAN POWERS: Go ahead and question him.
5 [Laughter.]
6 CHAIRMAN POWERS: If you don't, I will.
7 DR. APOSTOLAKIS: The thing that bothers me though
8 is that under treatment of uncertainty I don't see anything
9 on Defense-in-Depth -- and it seems to me that the safety
10 margin combined with the notion of redundancy, diversity and
11 so on is a package that is ways of handling uncertainty.
12 Under 4.0 I would expect to see something like the
13 discussion we just had regarding when do you involve
14 Defense-in-Depth at this level, because now you have gone
15 way down, okay, which again you bring back this what you
16 call Preliminary Proposal.
17 MS. DROUIN: I will acknowledge right upfront that
18 after being away from the framework document for awhile and
19 rereading it, you know, going through, of course, you find
20 lots of places where you would like to fix it, particularly
21 Chapter 4.
22 DR. APOSTOLAKIS: Okay, fine.
23 MS. DROUIN: We felt needed a lot more
24 clarification to explain some of the stuff.
25 DR. APOSTOLAKIS: So you will get the transcript
. 194
1 then and -- you will consider my comment?
2 MS. DROUIN: Absolutely.
3 DR. APOSTOLAKIS: And take appropriate action and
4 then reject it. No, no, you are not going to do that.
5 No, but let me make it clear. I think under
6 treatment of uncertainties there ought to be some discussion
7 of the role of the structuralist interpretation of
8 Defense-in-Depth when the uncertainties are large and
9 unquantified and also couple the safety margin idea with the
10 PRA.
11 MS. DROUIN: Yes. Just to go through --
12 DR. WALLIS: Mary, before you take that away --
13 MS. DROUIN: Okay.
14 DR. WALLIS: -- could you remind me what the
15 framework is for? Is this an intellectual exercise or is
16 someone going to use it for a specific purpose?
17 MS. DROUIN: No, we use the framework --
18 remembering that the framework is not just this figure,
19 there's two parts to the framework, there's the structure of
20 it here but associated with the framework are the
21 guidelines, the quantitative guidelines and this is used to
22 help us prioritize which regulations to risk-inform and then
23 once we identify a candidate regulation as we go through and
24 risk-inform it --
25 DR. WALLIS: So it is a procedure which says at
. 195
1 this time when you are making your decisions check the
2 framework and see where you are or something? It's all --
3 it's worked into actually what you do?
4 DR. APOSTOLAKIS: Yes, Chapter 5 gives that
5 guidance, how to implement it and select the regulations and
6 so on.
7 MS. DROUIN: Yes.
8 DR. APOSTOLAKIS: I am ready to move on to 11
9 unless a member has a question.
10 MS. DROUIN: Go ahead.
11 MR. CAMP: I would like to I guess ask a question
12 of you, George, and that is that one of the problems as we
13 write these things about safety margin and Defense-in-Depth
14 is as we go through each regulation is to decide how much is
15 enough.
16 DR. APOSTOLAKIS: yes.
17 MR. CAMP: Is this committee at any time going to
18 offer up any opinions on that such as I mean obviously we
19 found the paper that you wrote on Defense-in-Depth to be
20 very helpful. We have incorporated a lot of that.
21 Is there any chance of anything on the area of
22 safety margin and amount of acceptable uncertainty likely to
23 be forthcoming?
24 DR. APOSTOLAKIS: At this time, I doubt it, unless
25 my colleagues have brilliant ideas that they have not
. 196
1 expressed yet.
2 DR. WALLIS: Are you asking us to provide that? I
3 mean -- he's asking the Staff.
4 DR. APOSTOLAKIS: He is not asking us to do it.
5 He is asking whether we plan to do it.
6 DR. WALLIS: Who is going to do it?
7 DR. APOSTOLAKIS: There is an important
8 difference.
9 MR. CAMP: I mean obviously we are going to be
10 wrestling with that.
11 DR. APOSTOLAKIS: I understand that.
12 MR. CAMP: The extent to which there is guidance
13 from this body or anybody else --
14 DR. APOSTOLAKIS: I'll tell --
15 MR. CAMP: -- as to how to do that is helpful.
16 DR. APOSTOLAKIS: -- you, I believe that it is
17 premature to think about that and the reason is that I am
18 not sure that I have ever seen a quantification of the
19 margins in the sense that you guys describe it here, okay,
20 except in the containment. The containment I have seen
21 things like that, where they take their damage point and the
22 stressors and so on, the severe accidents, so I don't know
23 that I can offer any advice on that.
24 Right now I would imagine that those kinds of
25 "what is enough" are already in the numbers you have up
. 197
1 there in the next slide but at what degree of confidence? I
2 think that is what you are asking.
3 I don't know that we are ready to say anything
4 about that.
5 What do you think? -- although you say something I
6 don't like in the report, related.
7 DR. KRESS: But I must say I liked the start with
8 the safety margins.
9 DR. APOSTOLAKIS: I like it.
10 DR. KRESS: Yes, I think that is a good way to
11 start.
12 DR. APOSTOLAKIS: If you could quantify it for a
13 sequence then I think we will have a better basis for
14 addressing the question you raised, but on page 4-2, you say
15 a couple of things I am not sure we are ready to say.
16 You are saying that excessive margins may lead to
17 safety concerns. Can someone explain that to me, that if I
18 have excessive margins I may in fact compromise safety?
19 MS. DROUIN: I will give you an example.
20 Back when I first was working at an architectural
21 engineering firm, designing a piping system, a lot of
22 uncertainty with it and as you went through each part from
23 going to the source, talking about a fluid system to
24 wherever you are discharging, you would tend to add in 10
25 percent margins everywhere, and you're sitting here trying
. 198
1 to design the system in the pump head.
2 If you put so much margin in it, all of a sudden
3 you were designing a system that wasn't going to work.
4 DR. APOSTOLAKIS: That is not a safety concern.
5 MS. DROUIN: Well, it could be a safety concern if
6 that was a system you were needing for safety purposes, and
7 all of a sudden you have put in a design there and you have
8 put so much margin into it.
9 DR. APOSTOLAKIS: I guess that is what I don't
10 understand. How can something be so strong and so good that
11 it becomes useless?
12 MR. SIEBER: You may have too much capacity or too
13 much pressure.
14 CHAIRMAN POWERS: It's a little like a cast iron
15 airplane.
16 [Laughter.]
17 CHAIRMAN POWERS: You might be able to get it off
18 the ground and what not but I really, really do not want to
19 be in that airplane in a thunderstorm.
20 DR. SEALE: Or when it lands.
21 DR. APOSTOLAKIS: But what kind of margin am I
22 adding when I do that? I don't understand that.
23 DR. SEALE: There's all kinds of margin.
24 SPEAKER: Hailstones --
25 DR. APOSTOLAKIS: We are talking about margin with
. 199
1 respect to one particular issue which may hurt you with
2 respect to another issue, but that is not the way I read it.
3 I read it as the total margin, and I am confused
4 now. How can the total -- an excessive total margin lead to
5 a problem? I can't see that.
6 What you are saying, the example you are giving me
7 is that there may be five different things and I am
8 excessively conservative on Item 3 and that affects Item 4,
9 but that is not the way I read it. I thought it was the
10 total margin we are talking about.
11 I mean if it is clear to everyone else --
12 DR. WALLIS: I thought this was a very good point,
13 though, this business of what is a margin and how do you
14 decide what is enough. It's going to be one of the sticking
15 points and you can have all the frameworks in the world you
16 like and if you can't answer that question you may not be
17 able to proceed, so for me it is a very important issue.
18 You can't talk it away.
19 DR. APOSTOLAKIS: I understand, but on page 4-2
20 they say, "Excessive safety margins benefit neither the NRC
21 nor the nuclear industry. Excessively conservative
22 requirements can in fact lead to incorrect safety
23 conclusions."
24 DR. SHACK: I think that is on an item by item
25 basis.
. 200
1 DR. APOSTOLAKIS: Oh, item by item.
2 DR. SHACK: The safety margin here is really
3 discussed in kind of a limited context. It is not, you
4 know, can your CDF be too low.
5 [Laughter.]
6 DR. APOSTOLAKIS: All right.
7 DR. SEALE: Takes all the fun out of it.
8 MR. KURITSKY: I think the example -- Eric, you
9 may want to correct this -- we were discussing about the
10 capacity of the low pressure injection systems. Wasn't
11 there an issue that Westinghouse put up where right now
12 their design for large break LOCA and actually if you were
13 to reduce the capacity of those pumps it would be more
14 effective at some other types of accidents? Do you remember
15 something?
16 MR. HASKIN: Yes. I am not sure that is totally
17 relevant. The comment was that if we eliminate the large
18 break LOCA of the design basis accident it would allow them
19 to optimize the design of their systems to deal with small
20 LOCAs that are more risk dominant, and so in that sense
21 that's true.
22 I am not sure that is directly relevant to
23 George's question.
24 MR. KURITSKY: But if the flow capacity was based
25 on margin and you are making the capacity greater because
. 201
1 you are adding margin, it is kind of going the opposite
2 direction of what would be necessary.
3 DR. WALLIS: But you are saying margin is measured
4 by flow capacity. To me, margin should be measured by risk,
5 and it may or may not be that flow capacity is a good
6 measure of margin.
7 MR. HASKIN: I really think that is part of the
8 problem because traditionally safety margin has been
9 measured one dimensionally and that dimension has not been
10 risk, so what we are talking about here is where possible we
11 would like to measure margin based on risk.
12 DR. APOSTOLAKIS: And I agree with that.
13 Anyway, I thought it was an odd statement.
14 DR. SHACK: I took a completely different reading
15 here. I thought from reading this section safety margin was
16 in terms of things like capacity and that, and risk would be
17 at a higher level.
18 DR. APOSTOLAKIS: Yes, that is what I thought too.
19 MR. HASKIN: That's true. The statement that we
20 are talking about now is back to the old one dimensional
21 where people put on margin based on some parameter, be it
22 flow, be it stress and that needs to be clarified in the
23 writeup.
24 DR. SEALE: Could I ask another question about 11?
25 DR. APOSTOLAKIS: Yes, let's go to 11. We haven't
. 202
1 been there yet and now we are going to 11.
2 DR. SEALE: I thought you were. I'm sorry.
3 DR. APOSTOLAKIS: No, you can ask your question.
4 DR. SEALE: I realize you are starting over, if
5 you will, in this effort, but there are some people before
6 you who started over once, and those were pretty good people
7 too. Those were the people that put together the advanced
8 design utilities requirements document several years ago.
9 I think in that document and in every other case I
10 have heard people speak to any detail, the idea has been
11 that for the next generation of power plants which would
12 presumably proliferate to a greater degree than the present
13 lightwater reactor types that in order to keep the overall
14 fleet risk at approximately the same level, which is
15 unspoken gain, that the core damage frequency, the large
16 early release, if you will, the containment failure
17 probability, those numbers would be correspondingly lower
18 than they are generally expected to be for this generation
19 of plants and I see here that essentially the numbers we
20 sort of claim we'd like to have right now. I'm intrigued.
21 MR. CAMP: Well, first of all, these are for the
22 current fleet. It's my recollection that, in fact, in the
23 past, the Staff proposed lowering the safety goals for new
24 reactors, and the Commission came back with guidance not to
25 do that; that they were giving guidance to us saying that
. 203
1 the safety goals were safe enough and new plants didn't need
2 to be safer.
3 And that thinking, I think, has kind of guided
4 where we've gone with these numbers.
5 DR. APOSTOLAKIS: They stated that they expected
6 the new reactors to be safer, but they didn't want to change
7 the goals.
8 MR. KING: That's right, that's right, and the
9 utility requirements document went out and imposed
10 self-imposed more restrictive requirements on themselves
11 for, I think, partly for public acceptance reasons, partly
12 for economic protection reasons, and partly because they
13 wanted to please the Commission, to show that they were, you
14 know, doing better than today's generation.
15 MR. CAMP: I do think that if we were looking at a
16 next-generation plant, possibly the pebble bed or something
17 like that, based on the information I have seen, there is
18 the opportunity to relax the way that this is implemented in
19 terms of defense-in-depth requirements.
20 I mean, if somebody said it's got a really low
21 core damage frequency, and you can walk away from it and
22 won't melt, then maybe you don't need some of that stuff on
23 the right-hand side of the figure, the containment and
24 emergency planning.
25 So there are tradeoffs like that that we might be
. 204
1 able to make for new plants that we haven't thought about
2 here.
3 DR. KRESS: Looking at the three classes of
4 initiators on there, what do I do with the initiator
5 frequency that's between year and ten to the minus two per
6 year?
7 MS. DROUIN: You're on the infrequent?
8 DR. APOSTOLAKIS: No, it's greater than ten to the
9 minus two, but less than one, Mary.
10 MS. DROUIN: I'm sorry.
11 DR. KRESS: It seems to be left out of --
12 DR. APOSTOLAKIS: The range is left out between
13 one and ten to the minus two.
14 MS. DROUIN: Between one and ten to the minus two.
15 DR. APOSTOLAKIS: Yes, you are saying what to do
16 about greater than one and less than ten to the minus two.
17 What happens in between?
18 MR. KURITSKY: It's an artifact of when we
19 switched from being less than -- and we went from one to
20 greater than -- so it now leaves a gap.
21 DR. APOSTOLAKIS: May the inequality is wrong, the
22 first inequality. Do you mean less than one in the first
23 one? You probably mean less than one, right?
24 DR. KRESS: Up to one, maybe.
25 MR. CAMP: Well, it's up to as many as you have,
. 205
1 which is on the order of one. So, the --
2 DR. WALLIS: It means greater than ten to the
3 minus two, really.
4 MR. CAMP: Greater than's and less than's are a
5 little --
6 DR. WALLIS: It means greater than ten to the
7 minus two.
8 MR. CAMP: Yes, that's really what it means here.
9 DR. KRESS: Okay.
10 MS. DROUIN: We've got to fix that.
11 DR. APOSTOLAKIS: I have a couple of comments:
12 The first one is, again, the earlier comment
13 regarding the fires and so on, that somehow you have to
14 explain where these fit in here.
15 The second is, as you know, there was a confusion
16 when Rich Farmer published his line criteria way back, as to
17 whether one should evaluate individual accident sequences
18 using the plane, the acceptable/unacceptable region, or do
19 it in a cumulative way.
20 And, finally, people realized that it has to be
21 cumulative, otherwise you could artificially license a plant
22 by subdividing the sequences.
23 So, this cumulative effect, I wonder whether it
24 creates a problem here. When you say infrequent initiators,
25 let's say I have initiators that, you know, a class of
. 206
1 initiators that as a class they fall into the category
2 between ten to the minus five and ten to the minus two.
3 And I realize that I have to have a conditional
4 core damage probability of one in 100, and then I realize to
5 my horror that I cannot meet that.
6 Then I take the class, divide it into 20 different
7 subclasses, the frequency now goes below ten to the minus
8 five, and, boy, all they have is an asterisk there, so I
9 don't have to do anything about the conditional core damage
10 probability.
11 DR. SEALE: Creative report writing.
12 DR. APOSTOLAKIS: This is the cumulative.
13 DR. KRESS: It's the same question, yes.
14 DR. APOSTOLAKIS: I think you have to say
15 something. I think it's a matter of writing something to
16 avoid that.
17 MS. DROUIN: Yes.
18 DR. APOSTOLAKIS: You are already alluding to it
19 by saying, now, typically we consider the LOCAs and the
20 transients. If I go that way, then I can't do that.
21 I can't consider, you know, 20 different LOCAs.
22 DR. KRESS: No.
23 DR. APOSTOLAKIS: And push them all below ten to
24 the minus five, right?
25 MS. DROUIN: Yes.
. 207
1 DR. APOSTOLAKIS: But I think you have to
2 recognize that.
3 And the last comment is, I was intrigued, but I
4 think you should believe it from this present document, that
5 on page 4-2, which it may be a good idea for later, but it's
6 premature.
7 DR. KRESS: Four-two?
8 DR. APOSTOLAKIS: Four-two. On the right-hand
9 side column, last paragraph. You use best estimate code
10 calculations with uncertainty propagation to demonstrate
11 compliance based on the computed 95th percentile.
12 I think it's premature to set that at the 95th
13 percentile. That's the margin, right?
14 MS. DROUIN: Yes.
15 DR. APOSTOLAKIS: So you would have a probability
16 -- first of all, I don't know whether this is epistemic or
17 whatever it is, but this is probably the probability that
18 the stress will exceed the -- will not exceed the strength,
19 95 percent.
20 So the probability of failure is .05, right?
21 That's what this says; the probability of failure is .05.
22 MR. CAMP: It's also consistent with the Appendix
23 K option to do best estimate plus 95.
24 DR. APOSTOLAKIS: I understand that.
25 MR. CAMP: Okay.
. 208
1 DR. APOSTOLAKIS: You're saying that we're getting
2 a .05 probability, just from the margin. So if I go now to
3 this table, what happens?
4 I mean, the sequences now are affected
5 dramatically, the conditional core damage probability,
6 right?
7 I mean, if I get .05 from the margins, I already
8 have got two orders of magnitude of the ten to the minus
9 four.
10 MR. CAMP: I think we need to clean this up a
11 little bit.
12 DR. APOSTOLAKIS: I think so.
13 MR. CAMP: We're trying to illustrate our way of
14 thinking of the problems.
15 DR. APOSTOLAKIS: I understand.
16 MR. CAMP: As opposed to the rule.
17 DR. APOSTOLAKIS: That doesn't mean that for this
18 document, maybe --
19 DR. SHACK: He's also supposed to set his
20 acceptance criteria. He's going to build his margin into
21 his acceptance criteria.
22 DR. APOSTOLAKIS: Right.
23 DR. SHACK: Even though the 95th percentile was
24 there, he's got additional margin in the acceptance
25 criteria, presumably.
. 209
1 DR. APOSTOLAKIS: No. This is the margin for
2 safety.
3 DR. SHACK: No, no, no, it's not. It's the margin
4 -- it's between the acceptance criteria and compliance with
5 the acceptance criteria. It's only --
6 DR. APOSTOLAKIS: So the 95th percentile is the
7 acceptance criteria?
8 DR. SHACK: No, no, it's to show that he's met his
9 acceptance criteria for that particular topic.
10 MS. DROUIN: Right.
11 DR. APOSTOLAKIS: It says safety margins imposed
12 to account for uncertainties in data and models by
13 conservatism placed in acceptance criteria.
14 So one way of doing it is to pick the 95th
15 percentile of the probability curve, right; that you will
16 exceed the strength; is that what you mean?
17 MR. CAMP: As that paragraph is written, it's
18 implying both conservatism in the acceptance criteria and
19 conservatism in the calculation.
20 DR. APOSTOLAKIS: So maybe we can clear it up.
21 MS. DROUIN: The conservatism is in the acceptance
22 criteria only.
23 DR. APOSTOLAKIS: Well, does it say that? No.
24 Anyway, it confused me, and maybe other people.
25 MR. CAMP: I think your previous point about
. 210
1 accident classes, LOCAs, transients and whatnot, is an
2 important one that we haven't fully come to grips with,
3 because loss of offsite power is a good example that falls
4 into that top row.
5 And if you have a plant, for example, if their
6 containment is, on average, a really great containment, but
7 maybe it's a really lousy one for station blackout, then do
8 we need to do something?
9 Depending on how you group sequences, you'll get a
10 different answer to that question, and we're still
11 struggling with that; we're planning to add some examples
12 into the document to kind of clarify that that is.
13 But it's still an issue for you to think about.
14 DR. APOSTOLAKIS: You understand my comments, so
15 far?
16 MS. DROUIN: Yes.
17 DR. APOSTOLAKIS: Okay, because we have to finish
18 in a minute.
19 MS. DROUIN: One of the basic ground rules that we
20 had in here is that we wanted the technical analyses behind
21 things to be based on your best available calculations.
22 DR. APOSTOLAKIS: I understand that.
23 MS. DROUIN: And to be realistic, and to put the
24 conservatisms or the safety margin into the acceptance
25 criteria, not both places.
. 211
1 DR. APOSTOLAKIS: Okay, Tom?
2 DR. KRESS: On page 2-3, underneath the bullet, on
3 the right-hand column, you have this what I think is a
4 wonderful statement that says more stringent requirements
5 may be imposed in the presence of large uncertainties
6 regarding the effectiveness of one of the strategies.
7 I think it's a wonderful statement, and then I go
8 to this --
9 MS. DROUIN: I'm sorry, Tom, where were you
10 reading from?
11 DR. KRESS: Page 2-3.
12 MS. DROUIN: Page 2-3.
13 DR. KRESS: The second column underneath the
14 bullet, the very last sentence in that first paragraph
15 underneath the bullet. It says more stringent requirements
16 may be imposed in the presence of large uncertainties
17 regarding the effectiveness of one of the strategies.
18 I think that's a great statement, and then I go to
19 this table and I don't see anything in here that talks about
20 uncertainties or how they might affect these ranges or the
21 numbers.
22 MS. DROUIN: You're absolutely right, and you'll
23 see I have it circled here. We've already noted it; that
24 that is missing from Chapter 3 and needs to be added.
25 DR. KRESS: Okay, that was my point. It doesn't
. 212
1 show up in Chapter 3 or in this table.
2 MS. DROUIN: We've already caught that one and are
3 fixing it.
4 DR. APOSTOLAKIS: There are a couple more things
5 that confuse me. The definition of CDF on page 3-6, it says
6 a typical PRA criterion for core damage requires the water
7 level to be below a certain level with no imminent
8 restoration of coolant to the core regions, core melt,
9 release of fission products from the fuel is assured.
10 Is that a satisfactory definition?
11 And then it says this corresponds roughly to the
12 point where computer analysis becomes complicated.
13 Can that serve as a definition of core melt, that
14 the computer analysis becomes complicated.
15 DR. KRESS: That's where you start getting more
16 uncertainty.
17 DR. APOSTOLAKIS: I think you need to clean up the
18 language a little bit, because I'm sure you don't mean that.
19 DR. KRESS: On this table, again, Footnote 5 on
20 page 3-2, you've --
21 MS. DROUIN: What page?
22 DR. KRESS: Page 3-2 in the document, Footnote
23 Number 5 under this particular table, talks about
24 conditional core damage.
25 MS. DROUIN: Right, and large late release.
. 213
1 DR. KRESS: Yes, for large late releases. It
2 seemed to me that what you're concerned with with large late
3 releases is cancers and land contamination.
4 MS. DROUIN: Yes.
5 DR. KRESS: And I just wondered if this .1 was
6 back-derived from the latent safety goal like the .1 here is
7 kind of back-derived from the early fatalities. Where did
8 the .1 come from, or does it just look like a good choice?
9 DR. APOSTOLAKIS: And one last comment.
10 MS. DROUIN: Yes.
11 DR. APOSTOLAKIS: On page 5-3, there is sort of an
12 intriguing statement: Further investigation is necessary in
13 order to identify whether there are major risk contributors
14 associated with these accident types.
15 And you have SBO, ATWS, LOCAs. Are you telling
16 the world that you're not sure?
17 MS. DROUIN: What page are you on?
18 DR. APOSTOLAKIS: Five-three.
19 MS. DROUIN: Five-three.
20 DR. APOSTOLAKIS: Where the Table 5-1 is.
21 MS. DROUIN: Yes.
22 DR. APOSTOLAKIS: And there is a section that
23 starts by saying "Safety concerns not addressed in 10 CFR 50
24 ..."
25 MS. DROUIN: Right.
. 214
1 DR. APOSTOLAKIS: Look at the last sentence of
2 this little paragraph.
3 MS. DROUIN: Where it starts, "Further
4 investigation..."?
5 DR. APOSTOLAKIS: Yes. I don't understand that
6 statement. Am I misreading it?
7 You're saying that for station blackout, we have
8 50.63, 50.34 --
9 CHAIRMAN POWERS: It seems to me that if we're
10 just working on editorial things, that there's a more
11 efficient use of the Committee's time.
12 DR. APOSTOLAKIS: I think so.
13 DR. SHACK: But this is a major section here. I
14 mean, it's saying that there are major accidents --
15 DR. APOSTOLAKIS: This is not editorial work.
16 DR. SHACK: I was going to ask you if you were
17 going to work on a sample regulation to address some of
18 these. I mean, you make the point that you've got important
19 classes of things here addressed only by a regulation that's
20 not in effect for any plant.
21 MR. CAMP: I think part of what we mean to say is
22 that as we go through these things, we expect to find that
23 since the regulations were not originally risk-informed,
24 they obviously in many cases have a lot of conservatisms,
25 but they also may not have done things as effectively as we
. 215
1 would have liked.
2 For example, we have the station blackout rule; we
3 have Appendix R, but we still know that those are -- that
4 station blackout and fire are contributors to risk.
5 So, all of the changes that we make may not all be
6 in one direction. There may be some things we want to add
7 in as part of risk-informing.
8 I don't know of any huge gaping holes in the
9 regulation where there's an adequate protection issue.
10 DR. APOSTOLAKIS: You need to make that clear.
11 Are there any other comments from the Committee?
12 DR. KRESS: I think your quantitative measures is
13 a very bold stroke, and much needed.
14 I worry about the coarseness of your dividing up
15 initiators into the frequency range. It looked like an
16 awfully coarse frequency range.
17 For example, one that was below ten to the minus
18 two, but just below it, you treat it just like it was the
19 same as one just above ten to the minus five in frequency.
20 That's a big range of frequencies.
21 MS. DROUIN: Yes.
22 DR. KRESS: And to lump all of those together at
23 ten to the minus two, for example -- of course, conditional
24 core damage frequency, seems a little strange to me. It may
25 -- I worry about the coarseness.
. 216
1 I mean, the idea is good --
2 MS. DROUIN: You know, at different times in very
3 early versions of this framework, we didn't use the word,
4 guideline, and we finally came to the word, guideline,
5 because we were trying to soften; that these were not --
6 DR. KRESS: These are not hard and fast.
7 MS. DROUIN: These are not hard and fast numbers.
8 And I do envision, over time, that this framework document
9 will expand considerably, because these are guidelines.
10 And as we learn from each rule, and each one might
11 be applied with its own subtleties, that those subtleties
12 and nuances will be factored into the framework document.
13 DR. KRESS: I had in mind, since we did have PRAs
14 that look at the whole spectrum of frequencies that -- the
15 guideline might be -- the product ought to be below
16 something, and one end of it ought to be below something.
17 DR. APOSTOLAKIS: They have given already, the CDF
18 goal.
19 DR. KRESS: Yes.
20 DR. APOSTOLAKIS: So the product should be less
21 than ten to the minus four.
22 DR. KRESS: Yes, but why do you have to divide it
23 up into this coarse of a frequency? Why don't you -- the
24 whole spectrum?
25 DR. APOSTOLAKIS: It's a policy issue.
. 217
1 DR. KRESS: Yes.
2 DR. APOSTOLAKIS: Do you want more
3 defense-in-depth or not? But already they have the
4 cornerstones that talk about initiating events and
5 mitigating systems, so they have to do something about it.
6 DR. KRESS: But they don't break down
7 quantitatively to that level.
8 DR. APOSTOLAKIS: Well, if you look a their
9 numbers, the -- is ten to the minus four.
10 MS. DROUIN: Well, the product across each row is
11 less than ten to the minus five.
12 DR. KRESS: Yes, ten to the minus five.
13 MS. DROUIN: Ten to the minus five.
14 DR. KRESS: My point was, why don't you use the
15 product as the guideline, and not --
16 MS. DROUIN: And what you will see in a more
17 up-to-date version of the framework, more discussion of how
18 that plays a role in here.
19 DR. APOSTOLAKIS: Yes, I'd like to also see the
20 future justification of the ten to the minus five, when the
21 goal is ten to the minus four.
22 MS. DROUIN: There's more discussion on that that
23 we're going to add.
24 DR. APOSTOLAKIS: Okay, good.
25 MS. DROUIN: I was just going to wrap up.
. 218
1 DR. APOSTOLAKIS: Okay.
2 MS. DROUIN: We are going to continue to modify
3 this framework, and I think probably five years from now,
4 even though the basic concept, I don't think is going to
5 change, but a lot of explanation on how these numbers are
6 used, I think you'll see quite a bit.
7 Over the next six to seven months, nine months,
8 till June -- I've got to sit down and count -- we're going
9 to continue with 50.46 and the special treatment and other
10 regulations as we see; continue to hold our public meetings
11 and workshops.
12 And, as I said, we're looking sort of in the June
13 2001 timeframe with the paper to the Commission that will
14 address at least 50.46. And I don't mean necessarily the
15 whole thing, but some parts of it.
16 DR. APOSTOLAKIS: So the framework, when you say
17 paper, you mean this framework? I don't understand.
18 Are you ever going to ask the Commission to
19 approve this?
20 MS. DROUIN: Probably at some point in time.
21 DR. APOSTOLAKIS: But not in June of 2001?
22 MS. DROUIN: I don't know. I can't say yes or not
23 at this point.
24 DR. APOSTOLAKIS: Okay.
25 MR. KING: I think at some point if there are some
. 219
1 policy issues, we've already raised some policy issues to
2 the Commission, and I think there are some other policy
3 issues that come out of this and we'll raise it to them.
4 I'm not sure we'd want to ask the Commission to
5 approve the detailed document like the framework. We might
6 pick and choose a few key things from it.
7 DR. APOSTOLAKIS: If you use numbers like this, it
8 seems to me they ought to say yes or no.
9 MR. KING: And we pointed that out to them on the
10 first time we sent it up.
11 MS. DROUIN: Back in a previous SECY.
12 MR. KING: Yes, we tried to point out the key
13 things in here, the key assumptions, the key policy issues.
14 DR. APOSTOLAKIS: Okay, anything else?
15 [No response.]
16 DR. APOSTOLAKIS: Back to you, Mr. Chairman.
17 MS. DROUIN: Thank you.
18 DR. APOSTOLAKIS: Thank you very much.
19 CHAIRMAN POWERS: Whoever developed my schedule
20 thought we had a stronger constitution than I think we do,
21 so I am going to go ahead and take a ten minute break here
22 so we can set up for the next set of presentations.
23 [Recess.]
24 CHAIRMAN POWERS: We are back into session.
25 What we are going to discuss now is the
. 220
1 undertaking to address the differing professional opinion on
2 the steam generator and tube integrity.
3 Dr. Shack, did you have something to say?
4 DR. SHACK: Yes. I have a conflict of interest on
5 this, since Argonne National Laboratory is a contractor
6 working on steam generator problems for the NRC and I will
7 not be participating in the DPO discussion.
8 CHAIRMAN POWERS: Okay. Well, we will pretend
9 like Dr. Shack has evaporated or been kidnapped by aliens --
10 [Laughter.]
11 CHAIRMAN POWERS: Well, as most of you are aware
12 back in July the Executive Director of Operations made a
13 request to the ACRS to examine this question of steam
14 generator tube integrity under the conditions of the
15 alternative repair criteria. He asked that we function as
16 an equivalent of an ad hoc panel under a management
17 directive to review the issues raised by the author of the
18 differing professional opinion and the Staff's response to
19 those and report back to him.
20 We accepted the EDO's request, sent a memorandum
21 to him and in that memorandum we indicated we would function
22 as an ad hoc subcommittee but we would have to function
23 under the provisions of the Federal Advisory Committee Act,
24 which is a little different than what is usually done for
25 the differing professional opinions.
. 221
1 This lists the subcommittee members. I'll talk
2 about this a little more in detail later. Go on to the next
3 viewgraph.
4 What the Planning and Procedures Committee did was
5 to establish an ad hoc subcommittee to do what our
6 subcommittees do, and that is to gather and systematize the
7 information in the areas of contention, the data analyses
8 that exist to support the various positions people have
9 taken, to assess the applicability of the data, the validity
10 of any models and their applicability to the issue, and also
11 to look at the risk significance.
12 Based on this, the ad hoc subcommittee was to
13 develop a report to the ACRS and to draft a letter for the
14 ACRS to consider in responding to the EDO's request.
15 The subcommittee is acting not as an ad hoc panel
16 for the DPO but rather as a subcommittee of the ACRS and on
17 the next slide I show you the subcommittee in its glory, to
18 show the subcommittee members and the various topical areas
19 they are taking responsibility for.
20 The one notable area on this is that in the area
21 of metallurgy we have asked Ron Ballinger of MIT to come
22 serve as a member of the subcommittee and help us with this
23 arcane field of stress corrosion cracking.
24 I hasten to comment that in June of this year I
25 attended a conference on corrosion and someone gave an
. 222
1 overview of the area of corrosion, what he thought the
2 status was. He went through the various types of corrosion
3 that exist. He came to the area of stress corrosion
4 cracking and he says that field is just a total mess and
5 went right on, so we are in the midst of the stress
6 corrosion cracking field.
7 The subcommittee members themselves have been
8 looking at this monumental quantity of documentation that I
9 think all the members have received, looked at it in some
10 detail, and they are preparing sections of the report that
11 will be given to the ACRS but in order to assure high
12 quality for that we have an internal review system and this
13 slide shows the reviewers for each of the appropriate
14 sections. They are members of the ACRS but that is part of
15 our quality control procedures.
16 Again in the area of metallurgy we had sought some
17 assistance from Dick Riker at NIST to do the peer review for
18 us in the metallurgical areas.
19 There are a large number of issues that arise in
20 connection with the alternate voltage repair criteria and
21 the condition monitoring program that the licensees are
22 committed to comply with. What I have tried to do here is
23 to categorize the various contentions that exist, beginning
24 first with the accidents and the analysis of those
25 accidents.
. 223
1 We are in the midst of a change in the way the
2 world looks at accidents, so in this area we need to look
3 not only at design basis accidents but also severe
4 accidents.
5 Design basis accidents that are of interest here
6 are the main steam line break and the steam generator tube
7 rupture accident.
8 With the alternate voltage repair criterion there
9 is the possibility that defects could be left in place in
10 the steam generator tubes during normal operations and as a
11 result a main steam line break can involve more than just
12 leakage from the primary side to the secondary side. It can
13 actually involve rupture of tubes, so you have kind of a
14 synthesis of the two classic design basis accidents.
15 Within severe accidents are also two classes of
16 accidents that are of interest, those that are initiated by
17 a steam generator tube rupture followed by some other
18 failure, either a human failure of a systems failure but
19 progress to core damage.
20 Those kinds of accidents I think members are
21 entirely familiar with since they have been part of PRAs for
22 some time now. They were examined in the course of looking
23 at NUREG-1150. They have a peculiarity. Because they
24 involve bypass of the containment, they have rather high
25 source terms to the environment.
. 224
1 Consequently, even though they are not particular
2 frequent accidents, they can be risk dominant accidents.
3 There is another class of severe accidents that
4 does not get examined too closely in the past and those are
5 accidents initiated by something else, say a station
6 blackout accident. The primary coolant system remains
7 pressurized and heat loads are placed on the steam generator
8 tubes. I suppose it is also possible that pressure loads
9 can be placed on them such that the tube rupture in it
10 evolves into looking a lot like a steam generator tube
11 rupture accident with bypass of the containment.
12 We have to, as a result, look at two classes of
13 severe accidents as well.
14 The issue also of course involves nondestructive
15 examination techniques for steam generator tubes and their
16 defects and the limitations of those methods, what they can
17 and cannot detect. They are reasonably good at detecting
18 flaws in steam generator tubes, though in fact the
19 probability of detection that is used in the analyses tends
20 to be on the low side, around .6.
21 They are not so good at sizing those defects, and
22 that of course is what has given rise to the alternate
23 repair criterion because the past regulatory position was to
24 remove or repair tubes that had lost more than 40 percent of
25 the wall thickness. Now you can't tell whether it is
. 225
1 exactly 40 percent or not, so you need some other criterion
2 for deciding when they should be removed and the alternate
3 repair criterion of course uses the voltage from a bobbin
4 coil detector.
5 What gives rise to the defects and flaws in the
6 tubes of course is predominantly stress corrosion and the
7 stress corrosion leading to cracks and the cracks
8 propagating so that you can either leak or burst the tubes.
9 `We're having to look at the details of how much
10 it is that we know about these corrosion processes and the
11 cracking processes, initiation of cracks and the growth of
12 cracks. Those phenomena then figure in correlating the
13 non-destructive examination results and how you predict how
14 the steam generator will operate over the next cycle.
15 One of the issues that is central in this whether
16 in fact that is a predictable thing or not. There are
17 detail cracking issues that are important here. Cracks
18 initiate. They interlink. They then propagate through the
19 material with highly nonlinear phenomena. That is, of
20 course small cracks grow slowly but once they leak up then
21 they can grow through suddenly very quickly and so you might
22 go from a crack so small that you couldn't even detect in an
23 examination and then over the course of the cycle at the end
24 of that cycle something will pop up and you have a high
25 voltage indication that seemed to come from nowhere.
. 226
1 Is that predictable or not? If that were it, that
2 would be a challenging issue, but there are other things
3 that merit consideration here.
4 The author of the DPO has suggested that in the
5 course of an accident there are ways of producing damage to
6 the tubes that would result in higher flows from the primary
7 to the secondary system than you would predict just based on
8 crack growth and I would say one of the most significant
9 revelations presented in our discussions has been the
10 discussion of what happens when you have a main steam line
11 break and you depressurize the reactor coolant system.
12 We often think of this in terms of the thermal
13 hydraulic codes we have been examining as a relatively
14 benign event of fluid flowing out a drawn pipe on a piece of
15 paper, but in fact it is a choked flow with sonic booms and
16 enormous vibrations to the system, not only the system but
17 support structures.
18 The DPO author has asked how the predictive
19 techniques would account for what additional damage this
20 might cause the steam generator tubes, how it might affect
21 the restraint provided by the tube support plates around the
22 steam generator tubes.
23 He suggested that this might result in cracks
24 opening up, cracks that were plugged with corrosion products
25 becoming unplugged. Most important of these are not so much
. 227
1 the unplugging of cracks themselves since stress corrosion
2 cracks tend to be fairly tight and have a lot of corrosion
3 material in them, but many of the cracks are formed in areas
4 of crevices. These are annual spaces between the tube and
5 some other material, usually the tube support plate, which
6 is typically a carbon steel plate.
7 You clearly have a galvanastic voltage between the
8 tubes and the plate, but you get corrosion there and the
9 corrosion products create crevices where the chemistry is
10 radically different from the bulk fluid and as we all know,
11 the predilection of these things is to always go to the
12 worst possible situation so you get a lot of cracks in these
13 areas of corrosion product buildup both from the chemistry
14 and the stress that is imposed on the tubes.
15 If those corrosion products get displaced by these
16 violent vibrations and shaking of the material now the crack
17 is no longer constrained. It can open up and you can get a
18 flow.
19 The key issue here is certainly one of
20 radionuclide release in going from the primary to the
21 secondary system and then out through whatever break you
22 have.
23 More importantly, does the operator have time to
24 get the system back under control and have long-term
25 cooling? If in fact so much flow from the primary to the
. 228
1 secondary occurs, the water inventory available to that
2 operator might be consumed and it would not be possible to
3 get back into a long-term steady state cooling.
4 The DPO author has also suggested that as we all
5 know high velocity fluid flows, be they gases or liquids,
6 especially if they are carrying particulate material, can
7 cut metal and he wonders if they open up a crack in a tube
8 might you not also get the adjacent tubes to fail because of
9 the streaming fluid and particulate laden fluid that is
10 coming out, impacting on adjacent tubes much like you get in
11 jet cutting of metals.
12 The final category of contentions have to do with
13 the source term and clearly in the severe accident world you
14 have the radioactive aerosols and their behavior both on the
15 primary side and the secondary side.
16 Within the design basis world, you have this
17 phenomena called steam spiking -- I mean iodine spiking.
18 The phenomena is an observed one, that any time there is a
19 sudden change in reactor power or a depressurization of the
20 primary system the iodine concentration of the coolant goes
21 up, and that is called an iodine spike.
22 Clearly in assessing the consequences of a
23 design-basis accident you need to know something about how
24 big that spike got to be and the contentions are over what
25 spiking factor to use.
. 229
1 CHAIRMAN POWERS: The subcommittee held a meeting
2 to gather additional information to have presentations by
3 the DPO author and the staff on their positions on the
4 various items. He held that meeting October 10 through 14.
5 Let me say that seldom have I been privileged to witness
6 three days of such superb presentations. All parties did a
7 wonderful job in presenting their points of view and the
8 information they had to support them. Just time after time
9 after time we were getting outstanding presentations.
10 The DPO author deserves a certain congratulations.
11 He with a little bit of assistance from Robert Spence spent
12 a whole day standing in front of us. It was a heroic effort
13 on his part, and he did an outstanding job.
14 Unfortunately, even with the 3-1/2 days that we
15 gathered this information, we did not allow what I would
16 call rebuttal comments in the presentations. There just
17 wasn't time. But we did indicate that as part of both the
18 full ACRS meeting today and the ACRS meeting in December
19 that if people felt that they had comments to make either in
20 rebuttal or to expand on what they already said, we'd make
21 time available for them, and we will be doing that today.
22 That's all. I'm done.
23 I'll just provide you I think this and the next --
24 provide you a thumbnail sketch of where we stand. The staff
25 has issued its alternate voltage repair criteria.
. 230
1 You will recall that the ACRS had reviewed this,
2 that we had said that as an interim measure it looked
3 useful. There was a variety of things that still needed to
4 be investigated at that time.
5 The DPO author has come back and contended that
6 there is insufficient technical support, and this has
7 allowed degradation of the protection provided for the
8 public health and safety.
9 The DPO's author has concluded that the Generic
10 Letter 95-05 should be rescinded, and they should return to
11 the 40 percent plugin criteria of the past.
12 Staff's basic position is that they have devoted
13 a lot of effort to this. I think it's safe to say that they
14 have devoted a lot of effort, and that they have given
15 complete attention to the issues that were raised in the DPO
16 and provided some documentation for that.
17 There are still issues that are outstanding, and
18 they have initiated research on them. Notable among these
19 issues that they're researching is this idea that the fluid
20 would -- streaming out of a crack in one tube the
21 pressurized fluid would cut another tube. So that work is I
22 would guess about halfway through their effort. There are
23 some preliminary results, but certainly not a final word on
24 that. They've been attacking it with a classic combination
25 of both computational fluid dynamics analyses of the problem
. 231
1 and experimental studies trying to make prototypic
2 experiments to the extent you can in nonreactor facilities.
3 One of the issues that is really quite remarkable
4 on this is the concept of the high-pressure blowdown of the
5 system and how dynamic it is, that it is not simply the flow
6 of fluid out of the hole, but there is a choking in it that
7 causes remarkable events to take place in the whole system.
8 There really has not been a main steam line break in an
9 operating fuel reactor, but there has been the equivalent of
10 a main steam line break in a facility that was not fueled --
11 two of them, I'm sorry -- Jack's quick, two of them -- and
12 we're fortunate to have on the staff a staff member who got
13 to endure those events, and so he can give us a firsthand
14 account of the kinds of things he saw, and it's really
15 worthwhile to hear the description. As he indicated, he
16 subsequently left these things. Maybe it was battle
17 fatigue. But they really are issues on --
18 DR. WALLIS: You said -- you said main steam line
19 break. You meant a tube rupture?
20 CHAIRMAN POWERS: No, main steam line break.
21 DR. WALLIS: Okay. Okay. Thanks for confirming
22 what you said.
23 CHAIRMAN POWERS: I mean, you understand what
24 you're talking about when you're breaking the main steam
25 line, breaking a big system, and you get a lot of flow, but
. 232
1 nevertheless the biggest hole you can have still goes choked
2 on you at these things. And so you get, to quote Mr.
3 Spence, sonic booms and sympathetic vibrations of the
4 system, and I don't do justice to the description he can
5 give.
6 DR. SEALE: World's biggest whistle.
7 CHAIRMAN POWERS: It is a huge whistle. But more
8 than that he describes the undulating of walking platforms
9 and what not. Then he brings a very important point. He
10 says in escaping this he goes to the control room and in
11 there not for a short period of time but for literally an
12 hour the noise is so high that he can 't talk to the other
13 operators.
14 Now one of the areas where there is good agreement
15 between the staff and the author of the DPO is that there
16 are key operator actions that have to take place in
17 addressing these things. And one of the pieces of
18 information that was presented to us were some tests of
19 operator performance on simulators at the Halden project,
20 and the conclusion that came out of that is that when they
21 had a manpower shortage, it was difficult to meet the kinds
22 of schedules for handling these events that you'd like to
23 see. To my mind that means there's a lot of teamwork among
24 the operators that has to take place, and of course if they
25 can't talk to each other and hear each other, that teamwork
. 233
1 must surely get impacted.
2 It says on the viewgraph here that lessons learned
3 from Indian Point 2 steam generator tube rupture event will
4 be evaluated and a decision made with regard to the
5 improvements that need to be done to ensure the steam
6 generator tube integrity. Quite frankly the subcommittee
7 has been steering clear of the Indian Point event simply
8 because it's too new for us to get any kind of complete
9 assessment. I suppose that as information becomes available
10 we'll try to take it into account, but we have certainly not
11 had any in-depth discussions of it to this point.
12 Which brings us to the schedule. We've had the
13 information-gathering. We're in the process of trying to
14 assemble a coherent document. The senior author, however,
15 is totally incompetent and has not been keeping his end of
16 the bargain up. We're going to try very hard to get this
17 draft document back out to the subcommittee members so they
18 can correct what grievous damage the senior author has done
19 to the work, with the idea of getting it to the peer
20 reviewers sometime in November so that they'll have a chance
21 to get back to us about corrections so that we can have this
22 for our committee, the ACRS full committee deliberations in
23 December with the objective of reporting back to the EDO in
24 our December meeting.
25 I guess all this says to you is that the burden of
. 234
1 making up for the sins of the senior authors falls on the
2 rest of the ACRS. So time schedules are liable to be tight
3 for doing your peer-review activities.
4 That's what our status is at this point. Are
5 there any questions or suggestions on what we ought to be
6 doing here?
7 MR. SIEBEL: I'd suggest that you send it to us on
8 electronic media.
9 CHAIRMAN POWERS: Yes, we undoubtedly will. We're
10 having a little trouble figuring out exactly how to do that
11 with the figures, because --
12 DR. SEALE: Just tell us what the figures are.
13 CHAIRMAN POWERS: Yes.
14 DR. SEALE: My calendar tells me that we will be
15 here, many of us, most of the week of the 13th of November,
16 and so we will be doing that review in our off hours of that
17 week while we're killing all of these other dragons.
18 CHAIRMAN POWERS: That's right. We don't -- I see
19 nothing in your contract with the NRC that specifies taking
20 time out to sleep. So --
21 MR. SIEBEL: The maximum is 8 hours per day.
22 CHAIRMAN POWERS: That's what you get paid, but
23 that didn't say that's what you need to work.
24 DR. SEALE: That's the part I was remembering.
25 CHAIRMAN POWERS: Well, just remember, that only
. 235
1 says the amount you get paid. It does not say how much you
2 have to work.
3 DR. WALLIS: Some of us have other employees and
4 we have similar obligations to them.
5 CHAIRMAN POWERS: That's right, and if you check
6 your contract with them, it says nothing about eating,
7 either. So --
8 DR. KRESS: Did you ever investigate cloning?
9 CHAIRMAN POWERS: Okay. Then I think we can move
10 on. We're not attempting to present to the ACRS anything
11 that approaches the comprehensive nature of the
12 presentations the subcommittee has gotten. You are really
13 at a disadvantage here. You're going to hear addenda,
14 rebuttals, and that like from people we'll be presenting
15 now, not a complete exposition on the issue.
16 So with that introduction I will call upon Dr.
17 Hopenfeld if there is anything he wanted to say at this
18 point.
19 DR. HOPENFELD: Thanks a lot, Dr. Powers. You
20 have identified the problem very well.
21 I'd like just to fill in a couple of things that
22 give you my perspective on this. What I would really like
23 today is just to give you -- to summarize as to where I
24 think we are, and how I look at the problem.
25 Obviously, there's a tremendous amount of data,
. 236
1 material, and you name it and it's there, and you can just
2 get lost in it.
3 What I'll try to do is just to give you a feel for
4 what I feel are the major problems here. I thought I would
5 read it to you.
6 In my introductory remarks to the ACRS Ad Hoc
7 Subcommittee on the DPO issues, I pointed out that the risk
8 to the public from not removing degraded steam generator
9 tubes from service is at least 100 times larger than has
10 been reported to the public.
11 This is the crux of the DPO. The entire DPO
12 focuses on this issue; that we have 100 or more than 100
13 times difference in predicting what the risk in leaving the
14 degraded tubes in service.
15 DR. APOSTOLAKIS: When you say, risk, what is
16 that?
17 DR. HOPENFELD: The risk in this case, I mean as
18 the core melt frequency with the containment bypass. The
19 predictions that -- my predictions are on the order of ten
20 to the minus four per reactor year of core melt per reactor
21 year with the containment bypass, and I believe that the
22 NRC's predictions are ten to the minus six. And the
23 guidelines are ten to the minus five with the bypass.
24 So we have 100 -- at least 100, factor of 100
25 difference. I'm ten to the minus four.
. 237
1 DR. APOSTOLAKIS: And you're a factor of ten
2 higher?
3 DR. HOPENFELD: Factor of ten higher than the NRC
4 guidelines.
5 DR. APOSTOLAKIS: Correct.
6 DR. HOPENFELD: This is the crux of the DPO:
7 Briefly, I would like to discuss this in terms of four key
8 factors:
9 One is instrument capabilities; two, primary to
10 secondary leakage predictions; three, operators' response to
11 main steam line break accidents; and, four, the NRC process.
12 The NRC, which Dr. Powers didn't allude to, is
13 also an important factor.
14 One, instrumentation capabilities: The ability to
15 detect defects which could lead to catastrophic tube failure
16 is based only on laboratory tests.
17 After more than ten years of research, large
18 cracks with small voltage readings are missed, even in the
19 laboratory. Actual plant experience such as the recent
20 Indian Point 2 event demonstrates that significant defects
21 will not be detected in the field.
22 The probability of detection of .6 allowed by the
23 NRC, is arbitrary, totally unfounded, and non-conservative.
24 And that's probably the most important thing in the
25 instrumentation capabilities.
. 238
1 The eddy current probe has some inherent
2 limitations to it. You're looking with a tiny little probe
3 with a limited view range, and you're looking at a whole
4 bunch of cracks, an network of cracks, which the probe
5 cannot -- which the probe gives you different readings,
6 depending on the distribution and the concentration of the
7 cracks.
8 So there's an inherent limitation to the ability
9 of eddy currents to define what you're really interested in.
10 What you're really interested in -- you're not interested in
11 the networks or volumes of cracks; you're really interested
12 in that crack or that threshold of a crack that will lead
13 you to a -- that will open up during the steam line break.
14 And that's what is not -- the instrumentation is
15 not capable to provide you that.
16 Two, the second factor, is the primary to
17 secondary leakage prediction. The NRC voltage methodology
18 for predicting leakage is grossly non-conservative. The
19 correlation between voltage and leakage are inconsistent
20 with basic physical laws governing the flow of fluids
21 through cracks.
22 After more than ten years of research, large flows
23 with small voltage readings are missed in both the
24 laboratory environment and in the field.
25 Cracks which exhibit a low signal-to-noise ratio
. 239
1 may result in tube failures with catastrophic consequences.
2 In other words, those cracks which give you this lowest
3 signal-to-noise ratio are difficult to detect, and those
4 could give you very high leakages.
5 The tube leakage and database used to correlate
6 the voltage with leakage was not obtained under realistic
7 and valid conditions.
8 The NRC predictions of the leakage are several
9 orders of magnitude lower than those that can be expected
10 during steam line break accidents.
11 Three, Operator Response: The NRC assigns 99.9
12 probability of success to an operator's ability to
13 depressurize and cool down the primary cooling system before
14 the reactor core is uncovered.
15 Operating experience of 10 steam generator tube
16 ruptures which were relatively mild in comparison to SLB
17 events, do not warrant such optimistic predictions.
18 Four, the NRC process: The NRC regulatory process
19 primarily protects the financial interests of the nuclear
20 industry. Public safety takes a back seat to necessary
21 corrective actions which would be costly.
22 The generic safety issue program and other
23 research activities create the appearance that the NRC is
24 concerned with plant safety, and is effectively resolving
25 safety issues.
. 240
1 In fact, these programs delay the implementation
2 of necessary and urgently-needed corrections.
3 The attached table shows that it takes up to 17
4 years to resolve high priority safety issues. Existing
5 design safety margins are being drastically and dangerously
6 reduced under the cover of risk-informed regulations.
7 DR. WALLIS: This is a nonspecific statement.
8 DR. HOPENFELD: This is both nonspecific and
9 specific. And --
10 DR. WALLIS: Well, it's not specific to the
11 particular issue.
12 DR. HOPENFELD: Yes, it is specific to the
13 particular issue, too. As part of the DPO, the initiation
14 of the DPO, there was at the same time, a generic safety
15 issue, GSI-163 that was supposed to be resolved in parallel.
16 And the ACRS requested NRR on several occasions to
17 resolve that issue. Now, Research and NRR, whoever manages
18 these Divisions, constantly postponed and delayed the
19 resolution of that issue, yet they proceeded to the
20 rulemaking and then through Generic Letters and now to an
21 agreement with NEI, to first resolve the issue before you go
22 on to a rulemaking, to find what -- the generic safety
23 issue, the procedure for resolving generic safety issues is
24 different than DPOs.
25 You're looking into cost/benefit studies; you're
. 241
1 looking to alternative engineering fixes. Those kind of
2 issues are not discussed in the DPO context.
3 So, first you -- what was done here, they
4 continuously have been sitting on it, delaying the
5 resolution and it's still not being resolved.
6 But if you'll look as a whole at all the generic
7 safety issues, you see -- and all of them are high priority
8 issues -- you see that it takes on the average like four and
9 a half years, and this is -- to resolve a high priority
10 issue, and usually that is a management problem, because
11 most of the technical issues here were resolved way before
12 the issue itself was considered to be close.
13 But it's an important part of the mindset that you
14 have in the Agency when you address safety issues. You
15 can't divorce that just form that just cutting tubes,
16 because that's not the only factor that affects what the
17 outcome is going to be.
18 The bottom line is, you just want to protect the
19 public, and the process is a very important part of it.
20 At the October meeting, Mr. Joe Donahue described
21 how the technical specification for the Braidwood and Byron
22 I were relaxed using a modified RELAP-5 code, even though
23 inappropriate data was used to modify the code.
24 Mr. Steve Long described how the Farley plant was
25 allowed to skip a scheduled inspection, ignoring the fact
. 242
1 that there was no data to show how small cracks could
2 rapidly and catastrophically propagate tube damage by jet
3 erosion.
4 The NRC IG recently documented that inexperienced
5 engineers, poorly supervised, and constrained from
6 conducting free discussions with licensees, are responsible
7 for major safety determinations.
8 The NRC has also been successful in preventing the
9 Staff from identifying safety issues through the DPO
10 process.
11 In summary, uncertainties in instrumentation,
12 leakage predictions, operator response, and inadequate NRC
13 oversights of license submittals substantiates the
14 conclusion that leaving degraded steam generator tubes in
15 service can easily lead to catastrophic consequences.
16 Even with the unrealistic optimistic assumption
17 that an operator will be 90-percent successful in
18 controlling the accident, the actual risk to the public is
19 100 times larger than predicted by the NRC.
20 The stress -- as was mentioned by Dr. Powers, the
21 stress corrosion problem has been around -- it wasn't
22 invented by the nuclear industry. It has been around for
23 200 years.
24 And the practice in the industry, in the oil
25 industry, in the chemical industry, is to design around it.
. 243
1 You select materials that are not going to corrode.
2 Once you find that you have stress corrosion
3 problems, you don't keep the material in service, unlike
4 fatigue cracks, because you can't predict that phenomenon.
5 But now what we are doing here, and what I would
6 like to impress on you is that we have a severe stress
7 corrosion problem that nobody argues about. At the same
8 time, we are leaving those tubes in service, and at the same
9 time, we don't have a management program that supervises
10 what is being removed or what's not being removed.
11 And that really is the whole picture that you have
12 to consider when you look at how is the public being
13 protected here?
14 In summary, uncertainty instrumentation, leakage
15 prediction, operator response, and inadequate NRC oversight
16 of license submittals substantiate the conclusion that
17 leaving degraded steam generator tubes in service can easily
18 lead to catastrophic consequences, and that's the bottom
19 line here.
20 At the October meeting, the NRC for the first time
21 admitted that GL 95-05 is not valid for leakages of more
22 than 30 gpm. Because existing prediction indicates that
23 leakage could exceed 1,000 gpm, the use of GL 95-05 cannot
24 be supported.
25 Now, I would like to say a few words about Gl
. 244
1 95-05. That's the vehicle which allows the NRC to leave
2 degraded tubes in service, which allows the NRC to not use
3 the 40-percent plugging criteria on which is the basis for
4 which these plants were originally licensed.
5 So, 95-05 is that vehicle; that if you kill that,
6 then you really resolve the DPO.
7 Because existing prediction indicates that leakage
8 could exist at 1,000 gpm, the GL 95-05 cannot be supported
9 -- and I'm repeating it.
10 Jet erosion, tube vibration, bending, and buckling
11 during steam line break events can lead to leakage of
12 thousands of gallons per minute.
13 We had Dr. Schrock from Argon describe some very
14 interesting studies about crack propagation, crack growth,
15 and leakages through cracks. But those all were tested
16 under conditions which really have nothing to do with the
17 real environment.
18 So unless you can take all that data that has been
19 -- as a basic study, it's beautiful and it's very important,
20 but it's more so that it fits into the NSF kind of study
21 that you would do. It really doesn't relate to the
22 real-life conditions in which a tube is going to be exposed
23 or not during a steam line break.
24 It's going to be exposed to buckling; it's going
25 to be exposed to vibration, bending, and all these forces
. 245
1 are coming in. And you cannot just ignore them.
2 DR. WALLIS: I need some clarification. This
3 thousands of gpm, is that your estimate, or is that somebody
4 else's?
5 DR. HOPENFELD: I'm glad, sir, that you brought it
6 up. The thousand gpm was estimated by NRC Research during
7 the Trojan -- when Trojan developed a leak. NRC had
8 request, I guess from NRR, to come up with the estimate of
9 the risk of getting Trojan back online, and they came up
10 with an estimate between 30 to 1100 gpm, with 145 gpm mean.
11 This was done for Trojan. There were other
12 predictions from outside consultants. I think it was like
13 600 gpm.
14 My own predictions are that the uncertainties are
15 so high that when you look at all the parameters that are
16 coming in here -- and it isn't one; it isn't a jet or it
17 isn't bending or one particular factor. When you take
18 everything as a whole, you cannot conclude that the leakage
19 could be so large that the operator will not be able to
20 control it.
21 So in that sense, I would say that my prediction
22 would be more than a thousand gpm.
23 CHAIRMAN POWERS: Maybe we should try to -- and
24 I'll try from memory, Joe, and maybe you can correct me --
25 but there have been steam generator tube rupture accidents
. 246
1 in the United States and in other countries.
2 And in a few of those cases, we've seen flows in
3 the steam generator tube rupture, and my recollection is
4 that the flows there range from some very low values, but
5 the higher ones, 175, 600 gallons per minute, stick in my
6 mind as some of the higher values.
7 DR. HOPENFELD: Right.
8 CHAIRMAN POWERS: That gives you some perspective.
9 Now, if one believes in damage propagation and additional
10 tubes flowing, you see that you can get up to big numbers
11 pretty quickly.
12 The NRC Staff itself has sponsored some studies
13 that have looked at the timing of operator actions for flows
14 up to, I think, one thousand gallons per minute.
15 DR. BONACA: The only thing to point out is that
16 the events we're referring to here are steam generator tube
17 ruptures with no blowdown or secondary side of the steam
18 line break.
19 CHAIRMAN POWERS: That's right.
20 DR. BONACA: Therefore, you have a driving
21 pressure, that pressure between primary and secondary side,
22 which depends on the injection rates and so on and so forth
23 on the primary side.
24 Now, in the events we're looking at here, you have
25 a blowdown of the secondary side, and that is followed
. 247
1 simultaneously by a cooldown of the primary side and
2 depressurization. Therefore, you have the driving pressure
3 down.
4 And then at some point, the question is, how you
5 recover pressure of the primary side, and that depends on
6 what kind of hole you've got.
7 So you have some --
8 DR. WALLIS: If it's -- of course, the flow rate
9 may be independent of the secondary side pressure.
10 DR. BONACA: I'm sorry?
11 DR. WALLIS: If it's choked.
12 CHAIRMAN POWERS: Choked where?
13 DR. WALLIS: At the steam generator.
14 DR. HOPENFELD: At the nozzle.
15 DR. BONACA: Yes.
16 DR. HOPENFELD: Yes, but if you choke the nozzle,
17 the flow is so high that it's over 1,000 gpm, and the
18 operator wouldn't be able to handle it anyway. If you're
19 talking about choked at the tube, if you have one tube, that
20 would be a different story, but we're not talking about one
21 tube; you're talking about the tubes with thousands of
22 cracks, and maybe each one of them may or may not be choked.
23 And because you have to realize that the time --
24 we're getting into details here, but the time of flow
25 through the crack with the delta-P of 2500, the velocity is
. 248
1 very high and the flow is metastable.
2 But you get -- if anything, you can get one-phase
3 flow through the crack, and you just get the spray before.
4 There are so many different conditions coming into play
5 that, you know, after awhile, you can develop several Ph.D.
6 theses out of this.
7 DR. BONACA: The only thing I wanted to say, that
8 it is a very important issue. However, to look at the
9 dynamics of this event because if you have a very large
10 leak, primary side also is depressurizing, driving delta
11 pressure is decreasing.
12 If you have a very large leak in fact it may go
13 down so fast that you have -- that is going to affect the
14 leak rate.
15 DR. HOPENFELD: Right.
16 DR. BONACA: If you have a very small leakage,
17 then it becomes closer to a steam generator tube rupture
18 insofar as the dynamics of the transient --
19 DR. HOPENFELD: Right.
20 DR. BONACA: -- and that also affects the operator
21 response, because he is more likely to be confused about
22 whether or not he has -- so it's quite, so it is not as
23 simple as --
24 DR. HOPENFELD: And you may have something in
25 between and then you have something that lasted for a long
. 249
1 time causing you vibration over a very long period of time
2 instead of a short period of time, and so it is a
3 complicated problem.
4 I would like to touch upon another subject which
5 wasn't brought up before, but I think it's relevant and was
6 brought to my attention yesterday and I would like to read
7 it to you just for your information, nothing else:
8 "The draft report on regulatory effectiveness of
9 the anticipated transients without scram rule made public on
10 October 18" -- that was very recently -- "addresses the high
11 peak pressure and up to 37 percent unfavorable exposure
12 times that lessen mitigative functions required by the ATWS
13 rule, especially if there is no diverse scram system in
14 Westinghouse reactors. This is scheduled to be addressed at
15 the coming February ACRS meeting. The percentage of fuel
16 cycle during which Westinghouse ATWS is unmitigated
17 proportionately increases the risk. High peak pressures
18 with less than 10 seconds from an unmitigated ATWS will also
19 have an effect on tube sheet cladding separation and tube
20 weld cracks similar to the Robinson cold hydro. This high
21 pressure rise in an additional mechanism for steam generator
22 leakage with containment bypass during the severe
23 accidents."
24 In other words, this is another accident when you
25 are going to have very high pressures going all the way --
. 250
1 exceeding 3200, and I remind you that that unit was only
2 designed for 1500, so now you are talking about 3200 psi,
3 which would -- which, as Mr. Roberts told you, he had
4 experienced leakage or he had evidence of leakage in
5 Robinson which was hydro to 3000.
6 I am bringing that because I am not too versed
7 with that accident but it was brought to my attention that
8 you get these high pressures there, and additionally those
9 high pressures are not pressure that all the database that
10 has been generated on leakages, that they went, I believe,
11 as high as 2600 pounds but nobody ever went to 3200 pounds
12 and generated data, so the data is not applicable for these
13 conditions.
14 But again we are into the severe accident scenario
15 and I just brought it for point of information purposes.
16 At the October meeting the NRC Division of
17 Research strongly advocated the need for additional
18 research. If research is still needed, and still needed
19 after 10 years in order to prove the validity of GL 9505 and
20 to provide tech spec relief for the 40 percent through wall
21 tube plugging, it shows that this document has no valid
22 technical basis.
23 The alternate repair criteria as specified in GL
24 9505 should be rescinded immediately without further delay.
25 Failure to do so will continue to mislead the public
. 251
1 concerning the safety of allowing plants to operate with
2 severely and unacceptable defective steam generator tubes.
3 This is really the bottom line as to rescind GL
4 9505 and go back to the 40 percent plugging rule that served
5 us well. I am not saying that this is ideal, but it did
6 serve us well, and there's some theoretical justification to
7 it, but you don't expose the public to something of that
8 magnitude without having the ability to handle it
9 analytically or any other way, so you can say that, you
10 know, I am too conservative but I am erring on the side of
11 caution. I think this does warrant that and that is
12 basically the bottom line.
13 I would like to say one thing though, that I came
14 across this recent application from a licensee which
15 requested permission under 9505 to go to higher voltages, to
16 go from one to three volts and the Staff very, very
17 eloquently raised a whole list of issues which some of them
18 related to the kind of thing that we have been talking about
19 so there is something positive coming out of this.
20 I know I have taken a lot of people's time and
21 really you can appreciate when you really go to start
22 looking into the guts of the problem that it is not a simple
23 problem.
24 CHAIRMAN POWERS: Any questions you would like to
25 pose to Dr. Hopenfeld?
. 252
1 DR. APOSTOLAKIS: You said that the 40 percent
2 plugging criteria has served us well.
3 DR. HOPENFELD: Yes.
4 DR. APOSTOLAKIS: What do you mean by that?
5 DR. HOPENFELD: Well, it's been around since the
6 first day when all the reactors have been operating. We
7 haven't had any major -- I think the frequency of tube
8 ruptures is something like 10 to the minus 2, and to some
9 degree I would think that it would be becuase of that.
10 DR. KRESS: If the detection probability is low,
11 lower than the .6 --
12 DR. HOPENFELD: .6, yes.
13 DR. KRESS: -- wouldn't you have a problem even
14 with this 40 percent plugging criteria becuase you are going
15 to miss those things there also?
16 DR. HOPENFELD: I think the answer is yes, you
17 would, and I think if you remember one justification for the
18 rule, for the rulemaking activities, they were going on
19 since was it '94 through '97, was that very point that 40
20 percent was obtained on the basis of wastage data, and now
21 we are dealing with a different -- and it's true, although
22 there was some theoretical basis that 40 percent it still is
23 the limited capability of the probe.
24 DR. KRESS: I guess the nature of my question
25 is --
. 253
1 MR. SIEBER: Maybe I could address that and if I
2 am incorrect --
3 DR. HOPENFELD: Yes, sir.
4 MR. SIEBER: -- maybe you can correct me, but in
5 order to come up with crack depths since bobbin coils are
6 volumetric probes, they need to use a rotating pancake coil
7 which operates at one 24th the speed of a bobbin coil when
8 you are making the examination and there are about 40,000
9 intersections where these crevices exist in a steam
10 generator and to take five or 10 minutes travel time and
11 analysis time for 40,000 intersections, most of the reactor
12 existence would be spent examining tubes.
13 Basically the 40 percent was based on really a
14 minimum wall thickness requirement of 40 percent plus two 10
15 percent allowances for inaccuracy and also for continued
16 erosion of the tube through the next cycle and so it has
17 some of its arbitrary nature associated with it anyway.
18 On the other hand, the first steam generator tube
19 rupture that I am aware of was caused on an eroded tube
20 where 40 percent was accurately measured by bobbin. It was
21 done in FREESPAN, where the measurement should have been
22 very accurate but nonetheless the rupture occurred and so
23 that is sort of the basis is time of examination and there
24 are some uncertainties associated with using the rotating
25 pancake coil.
. 254
1 No measurement of anything in the world according
2 to Pauley is totally certain.
3 DR. KRESS: Eisenberger.
4 MR. SIEBER: Eisenberger, right.
5 DR. APOSTOLAKIS: Pauley gave it a different spin.
6 [Laughter.]
7 DR. KRESS: Pauley was a regulator.
8 DR. BONACA: But the other question I had was on
9 the same issue. If you really have a violent dislocation of
10 the internals of the steam generator as a result of the
11 blowdown and I am sure there are -- you know, we heard, we
12 have witness account of Bowman, would you expect also
13 significant leakage from -- you know, with this plugging
14 criterion rather than just only with the alternate?
15 DR. HOPENFELD: I think one answer -- you probably
16 would, but because Mr. Spence went back to that affected
17 steam generator because these were good tubes so that would
18 answer that question, and he thought when he compared the
19 several units, the unit that was affected exhibited a larger
20 number of leaks, but he really hasn't gone beyond that, so I
21 don't know whether that means anything.
22 The point is that probably when you have such an
23 event, probably it may not make any difference but you are
24 just trying to again err on the side of safety and that is
25 what you do.
. 255
1 See, here what the problem is, the problem is we
2 are dealing with something that happens very, very
3 infrequently. We don't get steam line breaks all the time,
4 so you can get this feeling, well, everything is fine, you
5 know, why worry about it?
6 That is why I think the IP-2 accident is
7 significant, because it's a precursor, because the only
8 thing you can do in these kind of things is look at
9 precursors. That is the only thing you can do. There is
10 nothing else you can do, but then if you look at this thing,
11 the logic tells you you just don't operate with tubes that
12 you don't know what they look like and what they are, and
13 you just really don't know.
14 CHAIRMAN POWERS: Let me hasten to point out that
15 IP-2 involved, as I understand it, a crack in the bend
16 region.
17 MR. SIEBER: Yes.
18 CHAIRMAN POWERS: GL 95-05 does not address that
19 kind of crack.
20 MR. SIEBER: That's right.
21 CHAIRMAN POWERS: That's a different beast.
22 DR. HOPENFELD: Except with one respect is that
23 the implication in the GL 95-05 was that the tube is going
24 to be -- and your letter to the Commission indicated very
25 clearly that is what your impression was, that the tube was
. 256
1 going to be constrained within the support plate, and that
2 is not going to happen, even by -- and that is what my
3 problem with the ANL tests are, because they assumed that
4 the tube is going to sit there and be protected by the
5 support plate but given minimum calculation, just the boring
6 of the tube sheet tells you that it is not going to stay
7 there.
8 MR. SIEBER: My understanding of the ANL tests was
9 that the tube wasn't constrained during burst tests or
10 leakage tests.
11 DR. HOPENFELD: No, but the environment, that they
12 are testing it only under internal pressure, and what I am
13 saying, when that tube is -- when the tube sheet bursts you
14 are going to have, the tubes are going to be buckled. Some
15 of them are being buckled because especially those which are
16 frozen into the support plate, and they are not looking at
17 that kind of stuff, and they can't.
18 I am not blaming them, but they are looking for
19 simple things that you can describe and I have no problem.
20 I think that's fine but when you take that and now you use
21 those equations which come from these nice experiments and
22 you apply the thing to the regulatory arena that's where I
23 have a problem with it, and they have done that in IP-2, and
24 that is where the problem comes in.
25 It is fine to do those tests, and I hope that Dr.
. 257
1 Shack can keep on doing it, and the same thing with the jet
2 erosion. I mean there's no way, practical way, of
3 predicting what those particles are. I mean we heard some
4 predictions about the particles going through cracks. You
5 can have an infinite number of different crack geometries.
6 You are going to have different flows in each crack. Some
7 of them are going to be sonic. Some of them are going to be
8 hypersonic. Who knows what? You can't analyze that kind of
9 stuff and then go and run tests and say I am going to have
10 these parts because these parts are going to be --
11 especially to do this in a three month period, it is just
12 unrealistic.
13 CHAIRMAN POWERS: Okay. I think we need to move
14 on.
15 Mr. Strosnider, you are going to present us --
16 DR. HOPENFELD: Thank you very much.
17 CHAIRMAN POWERS: And again you for the
18 outstanding presentations during the October subcommittee
19 meeting.
20 DR. HOPENFELD: Thank you.
21 MR. STROSNIDER: I wanted to start off by, first
22 of all, thanking the ACRS, the main committee and also the
23 special panel looking at the steam generator issues for
24 taking this issue on.
25 I appreciate what you are talking about, putting
. 258
1 in extra time and stuff, because there is really a
2 tremendous amount of information. There's some very
3 interesting and detailed technical issues and it take a lot
4 of effort to pull it all together and I just want to
5 acknowledge your efforts in taking that on.
6 I didn't come today to really make any rebuttal.
7 I just wanted to hit some of the high points from the
8 October meeting. In fact, Dr. Powers had a viewgraph that
9 did it exceptionally well on one page. I wish I had that --
10 not quite as concise apparently -- but just to make a few
11 points.
12 This is the agenda we used back in the October
13 12th and 13th meeting and as I just mentioned and you can
14 see from this it is an extremely multidiscipline issue and
15 one of the challenges is making it all come together.
16 I think you can see from the discussions we had
17 back then that the NRC Staff really has had a lot of people
18 from all the right disciplines involved in this.
19 We have had NRR and Research, Systems, PRA,
20 metallurgists, and we were talking about the fact that we
21 have got metallurgists asking PRA questions and PRA experts
22 asking metallurgy.
23 CHAIRMAN POWERS: And neither one knows what the
24 other one is talking about, right?
25 MR. STROSNIDER: I think that's good. But the
. 259
1 other point I wanted to make here is that the material that
2 was presented I think goes beyond what was initially in the
3 different professional opinion but nonetheless I think that
4 is good for two reasons.
5 One is we tend to have some problems identifying
6 exactly what the DPO is, in terms of it has changed somewhat
7 with time but putting that aside, I think it is important
8 that we get all the issues on the table so that we make sure
9 we capture them and that we can look at it from this
10 integrated perspective. I think that was the right way to
11 go at this.
12 I wanted to make a few comments just about perhaps
13 some overall trends, maybe big picture sort of things here
14 with regard to some timelines.
15 If we start off on the bottom line here, it really
16 marks out some of the major milestones with regard to the
17 DPV being filed and the DPO.
18 I mentioned an ACRS meeting here and also this
19 refers to the Staff DPO considerations document.
20 In fact, and I didn't put them all on here, but
21 just more or less for the record I would like to mention
22 that there were 11 ACRS meetings between I think 1994 and
23 the present talking about steam generators.
24 Something like seven of them dealt specifically
25 with the DPO issue.
. 260
1 This was another conscious decision to look at
2 this stuff again, look at it from a fresh perspective, and I
3 think that is also a good thing to do because we have new
4 information to provide at this time and so I think that was
5 the right approach there.
6 I also wanted to make a comment with regard to --
7 if you look at the process, and I tried to put these in some
8 sort of logic, starting with the steam generator rule, and
9 one of the comments I made in October and I would like to
10 make it again because it always kind of bothers me to hear
11 people talk about the failed steam generator rule effort,
12 okay? -- and I think I understand the perspective because
13 there is not a new steam generator rule on the street today
14 so I do understand that.
15 On the other hand, what the Staff is we followed
16 the rulemaking process, and there is a very well-defined
17 process and it does involve a regulatory impact study which
18 requires doing some risk assessments to support cost and
19 benefits and that sort of thing, and we did that.
20 When we went through it we concluded that we could
21 not support implementing a new rule based on the work that
22 was done there.
23 There was some real good work that came out of
24 that and I will say something more about that in a second
25 with regard to NUREG 1570, but we also acknowledged coming
. 261
1 out of that work that there were deficiencies in some of the
2 regulatory framework that exists, that the technical
3 specifications that are in place today were developed back
4 before a lot of the types of degradation that we have.
5 Certainly it doesn't recognize the advanced
6 technology in terms of eddy current testing, et cetera, so
7 we knew that there was still some improvements that need to
8 be made. They could be made within the existing rule
9 framework, and we took that on with regard to the generic
10 letter.
11 Well, as we were moving forward in that, the
12 Commission was developing some directions setting
13 initiatives. One of those was the work to figure out the
14 role of the industry in industry initiatives and in fact the
15 industry had been very involved looking at what we were
16 doing here.
17 I like to think that they acknowledged some of the
18 technical issues and regulatory issues that we were trying
19 to address and they took the initiative to develop NEI 97-06
20 and this is a very good initiative on the industry's part.
21 I think we need to credit that.
22 They committed to follow the guidelines that are
23 in that document before we have ever come to a point where
24 they have been made a regulatory requirement and they have
25 some good things in there, like doing condition monitoring
. 262
1 and operational assessment of steam generator tubes that
2 they were not required to do before.
3 Some of the discussions we are having now are
4 because they doing those things that hadn't been done in the
5 past.
6 I also just wanted to mention then this top
7 timeline, and I mentioned NUREG 1570, and that is a very
8 important document, I think somewhat of a milestone in
9 looking at steam generators from a risk perspective.
10 If you go back to some of the earliest work back
11 in the '80s with NUREG 0844 and compare it to what was done
12 here, this was a much more sophisticated analysis but most
13 importantly is it addressed this issue of severe accident
14 induced tube failures, which had not been explicitly
15 addressed before.
16 The other thing I wanted to point out particularly
17 from the regulatory, the NRR perspective is Reg Guide 1.174,
18 which is where a lot of this risk informed thinking, where
19 the Staff actually got some guidance on how to apply this in
20 plant-specific reviews, so there's been a continual
21 evolution here with regard to risk-informed thinking. It
22 has been very beneficial.
23 I just bring this up because I think it is good to
24 keep that in perspective when you deal with some of these
25 issues and recognize that, yes, we are increasing our
. 263
1 knowledge all the time and we are factoring that new
2 information into our reviews of licensing applications and
3 what is going on at the plants.
4 Dr. Powers had this concept on his viewgraph, one
5 of the things I emphasized back in October. We do take the
6 DPO issues and steam generator issues in general very
7 seriously.
8 A couple of examples. There's been extensive
9 documentation -- and I think actually we had some
10 discussion -- one of the members mentioned, I think was like
11 89 pounds that must have been mailed to him. I don't know
12 how much that was wrapping material --
13 CHAIRMAN POWERS: None. That did not count
14 wrapping material.
15 MR. STROSNIDER: But in all seriousness, there is
16 a lot of information there and I offered this at the last
17 meeting and I want to make sure it is clear to the people
18 who are going to take on the peer review.
19 That is a lot of information to find your way
20 through, and if we can help in pointing to the right
21 reference and the right spot in it, please just let us know
22 and we will do that.
23 CHAIRMAN POWERS: Good. Tell me where your steam
24 spiking factors came from in the plot.
25 MR. STROSNIDER: And in fact we have a list of
. 264
1 questions that we received and we are looking to provide
2 that information.
3 DR. WALLIS: Weight isn't everything. It would
4 help to have some guidance that says these are the real
5 areas of contention and this is the real evidence.
6 CHAIRMAN POWERS: I think if you had a chance to
7 sit through the subcommittee meeting that all the speakers
8 did a really outstanding job of doing just exactly that.
9 They really held our hands and walked us right
10 through that on all sides of the issue. That roadmap was
11 the thrust of everybody's presentation. It was just a
12 superb set of presentations.
13 Again, I hope you have been able to pass on to
14 your team how much we appreciated their input.
15 MR. STROSNIDER: And that is greatly appreciated
16 by everyone, but if we can provide further assistance we
17 certainly will.
18 With regard to development of the regulatory
19 framework, some of these issues that have come up in terms
20 of risk-based like the cutting issues and ablation of tubes,
21 one of the things I wanted to point out here is that some of
22 these have to be dealt with from a process point of view,
23 and we have in fact done that.
24 When we talk about the NEI 97-06 framework, one of
25 the things that the industry wanted when they came in with
. 265
1 that was they wanted to change the technical specifications
2 such that they could establish their own repair criteria and
3 their own repair methods without coming to NRC for approval,
4 and we haven't allowed that at least in the discussions we
5 have had so far and in the package that has come in to us
6 for the proposed tech spec changes.
7 We have not allowed that, and the reason we
8 haven't is because we want to be able to take a look at
9 these alternate repair methods from a risk perspective.
10 Unfortunately, there's not the sort of guidance
11 out on the street where we could say the industry is
12 following that and we have confidence that using that sort
13 of guidance that they will come to the right conclusion.
14 We don't know what the next alternate repair
15 criteria is going to look like. The industry is very
16 innovative, very creative, and so it is hard for us to
17 analyze those things ahead of time, but the point I am
18 making is that in the framework we are developing where we
19 know there are risk possibilities of increasing risk
20 vulnerabilities, we plan on taking a look at those.
21 With regard to plant-specific evaluations, we had
22 two risk-informed steam generator related issues there,
23 applications rather. I would just make the point as far as
24 how we are implementing these that in one of them we found
25 the licensee's proposal acceptable and the other one we
. 266
1 didn't.
2 Arkansas shut down and did an inspection about two
3 months before replacing their generators because we couldn't
4 find some uncertainties that existed in the risk assessment
5 and in their deterministic licensing basis. We couldn't get
6 past those.
7 I would also mention -- Dr. Hopenfeld just brought
8 it up -- and I appreciate his acknowledging with regard to
9 we have an application that came in recently to use a 3 volt
10 repair criteria. That wouldn't be allowed. What they are
11 proposing would not be allowed under Generic Letter 95-05,
12 but they are using some similar methodology.
13 There was one plant in the past where we, or one
14 or two I guess, where we approved this where they had
15 actually turned some of the steam generator tubes into stay
16 rods, if you will. They expanded the tubes and the support
17 plates in order to ensure that the support plates would stay
18 in place and mask the degradation.
19 This particular application they are not proposing
20 to make that modification and they are relying on thermal
21 hydraulic analysis to show that the plates won't move, and
22 we have gone back to them with some I think fairly tough
23 questions regarding that sort of analysis and so that is
24 going through the process, but we are asking those kind of
25 questions.
. 267
1 Finally, I want to point out and really give a lot
2 of credit to the support we have been getting from the
3 Office of Research here. It's been very responsive in
4 helping us deal with these issues and their programs have
5 been responsive and flexible enough to deal with some of the
6 issues that come up as we go through doing these
7 risk-informed reviews.
8 MR. SIEBER: Could I ask a question, please?
9 MR. STROSNIDER: Sure.
10 MR. SIEBER: There was discussion during the ad
11 hoc committee meetings about some research associated with
12 finding out exactly what the vibrational modes of the tube
13 support plates was during blowdown conditions and I don't
14 think that was -- nobody said they were going to do that.
15 Is that in the plan someplace?
16 MR. STROSNIDER: Well, I think one of the
17 discussions we had, and I am not sure I remember that
18 specifically, but with regard to the blowdown loads and the
19 event that happened -- was it Turkey Point, I guess?
20 MR. SIEBER: Turkey Point.
21 MR. STROSNIDER: When that issue was raised, what
22 we did is we sent it to Research. That is, NRR sent it to
23 Research to put into the Generic Safety Issues process and I
24 think Jack Rosenthal talked about that at the last meeting
25 and the plan that has been put together.
. 268
1 MR. SIEBER: He said, yeah, they have a plan to do
2 it, but --
3 MR. STROSNIDER: Right. That may lead to
4 additional research. We have to see what that process leads
5 to.
6 We are also seeing what information we can find.
7 As was mentioned, it was quite a while ago, so we are
8 researching the files, et cetera, to see if we can find what
9 sort of examinations might have been done after the event.
10 CHAIRMAN POWERS: Let's be clear. They are only
11 looking at this thing to see if it's to be a generic issue
12 or not. That would -- I presume that it -- somehow I have a
13 safe bet that it's going to be a generic issue.
14 And they do have a different process now for
15 handling those generic issues than they did in the past.
16 But this does not look like one that gets explored
17 thoroughly very quickly.
18 It's a challenge for -- it's working on those
19 parts of the code that we don't have codes to work on.
20 MR. SIEBER: It's a difficult TH problem.
21 MR. STROSNIDER: But as it's worked through that
22 generic safety issues process, gets put into that process,
23 it can end up driving research in terms of the dynamic
24 response and tube response that we have to let that work
25 through the process and see where it goes.
. 269
1 Regarding the maintaining safety, I did just want
2 to point out -- you know, we spent a lot of time talking
3 about some tough issues, and I think we shouldn't lose sight
4 of perhaps what's actually been going on in the field.
5 I know it's somewhat difficult, right on the heels
6 of a tube rupture earlier this year, but, in fact, if you go
7 back and look at the history, it had been seven years since
8 we had a rupture, and prior to that they were occurring at a
9 somewhat higher frequency.
10 And it's not statistically significant. We looked
11 at the numbers, and --
12 CHAIRMAN POWERS: I tried to look at that myself,
13 because the first one, I think, was in 1975, as I recall.
14 And I came to the conclusion that this is a --
15 process, and that the rate is about the same.
16 MR. STROSNIDER: From '75 to '93, there were eight
17 domestic ruptures, and then there was Indian Point earlier
18 this year. And I acknowledge no statistical significance
19 there, however, I also have to look at it from the
20 perspective of understanding what's been going on in the
21 industry in terms of the much more extensive inspections,
22 the operational assessments, the condition monitoring that's
23 going on.
24 I'd at least like to think that that's contributed
25 to the last seven years.
. 270
1 CHAIRMAN POWERS: The other thing that changes the
2 statistics on you and makes it hard to understand, is that
3 they're replacing steam generators, too, with Alloy-690
4 material.
5 And as I understand it, the 690s have stainless
6 steel support plates?
7 MR. SIEBER: Yes.
8 MR. STROSNIDER: And that certainly does affect it
9 also. The bad news is that some of the steam generators
10 that are out there, which took longer to start developing
11 degradation, it is occurring.
12 And there is degradation now in some of the BNW
13 generators, which went for a long time before it showed up.
14 CHAIRMAN POWERS: One of the things that comes
15 very clear from the document -- what I didn't emphasize in
16 my presentation that I should make clear is that the Staff
17 has had a huge program going on in understanding this
18 cracking phenomenon.
19 And it isn't all black art; they actually have
20 some science to it. But what they find is that these tubes
21 get annealed in the manufacturing process, in a way that
22 gives them a variety of microstructures.
23 And some of those microstructures are more
24 susceptible to cracking. They're all susceptible to
25 cracking, but it's when you start seeing is a function of
. 271
1 time.
2 And so that there are some generators that haven't
3 shown a big problem now, that would just begin to show
4 problems, and some that were a problem from the day they
5 were installed.
6 MR. STROSNIDER: There are a lot of competing
7 variables there, including the stress state, the environment
8 and the materials, and it --
9 MR. SIEBER: Chemistry.
10 MR. STROSNIDER: -- that dictates the time. And
11 it is different for different units, depending upon how they
12 were fabricated.
13 DR. WALLIS: I wasn't with the Subcommittee, so
14 what you're telling us now in some way answers Dr.
15 Hopenfeld's issues; does it?
16 MR. STROSNIDER: Well, you'll have to make that
17 determination. I think actually --
18 DR. WALLIS: You're not going through sort of
19 saying he raises this question and this is how we resolved
20 it?
21 MR. STROSNIDER: No, I'm not doing that. And as I
22 indicated, I wasn't really planning on offering some sort of
23 rebuttal to that. I think we provided a lot of information.
24 DR. SEALE: Is it fair to say that the evidence of
25 the last -- of the frequency of these and so on -- suggests
. 272
1 that the increased attention you're giving to this problem
2 -- and clearly you are -- has not made the problem worse?
3 [Laughter.]
4 MR. STROSNIDER: There is no evidence that I'm
5 aware of that says anything is going to be worse, and I'm
6 actually suggesting that --
7 CHAIRMAN POWERS: One of my firm beliefs is that
8 there is no problem a little management attention can't make
9 worse.
10 [Laughter.]
11 MR. STROSNIDER: But the point I wanted to make is
12 that we spent a lot of time discussing some of these
13 difficult problems, and we get down into risk of extremely
14 low probability events, and I'm not suggesting we shouldn't
15 deal with them, right?
16 But I'm just trying to back up a little bit and
17 say let's look at what's going on in the field, and the news
18 isn't all that bad.
19 This graph shows the trend, if you will, in steam
20 generator forced outages due to tube leaks. Again, I'm not
21 presenting any statistical analysis here, but, you know,
22 there is a decreasing trend.
23 CHAIRMAN POWERS: What would really help on this
24 plot is to show it as normalized to the number of steam
25 generators that have the susceptible tubing in them.
. 273
1 Because you are going over the same period of time,
2 especially from about '88 on where you've got a significant
3 fraction of the generators that have been changed.
4 MR. STROSNIDER: That's true.
5 MR. SIEBER: There is another impact there, too,
6 you know. There was leakage caused by loose parts and
7 foreign objects stuck in the tubes.
8 CHAIRMAN POWERS: They pulled that stuff up in
9 '75.
10 MR. STROSNIDER: I could mention that none of
11 these leaks are doe to ODSEC at tube support plates under
12 the 95-05 alternate repair criteria.
13 CHAIRMAN POWERS: Now, the issue that you can't
14 help escape is this condition monitoring program where a
15 fairly elaborate set of analyses are going on. They go
16 through to look at the condition at the end of cycle, end of
17 the previous cycle, and based on that, make prognostications
18 abut the future.
19 You would hope that because in that process they
20 have to examine, not just the cracks covered by GL 95-05,
21 but all the cracks; is that correct?
22 MR. STROSNIDER: That's correct.
23 CHAIRMAN POWERS: And one would hope that that has
24 some positive impact.
25 MR. STROSNIDER: If you look at the --
. 274
1 CHAIRMAN POWERS: Even accepting the problem,
2 though, of non-predictability, just in looking at it.
3 MR. STROSNIDER: And I pointed out earlier, some
4 of the weaknesses in the existing regulatory framework and
5 technical specifications, and this is an example where
6 people implementing existing technical specifications would
7 shut down plug at 40 percent, but not really have to do any
8 assessment of how serious the degradation was or what it
9 meant with regard to the next cycle of operation. They are
10 doing that now.
11 And I also should go back and point out that we
12 recognize the deficiencies in the existing framework -- and
13 I forget the exact number, but there are over half a dozen
14 generic communications we put out in the 90s dealing with
15 issues like optimizing inspection methods, looking for
16 circumferential cracks, dealing with u-bend cracking.
17 So, we've been doing that, unfortunately on
18 somewhat of an ad hoc basis. I'll say something a little
19 bit later about wanting to move forward and get this
20 improved regulatory framework in place.
21 But again, maintaining safety is our main
22 objective here.
23 DR. WALLIS: What else is your purpose? Why do
24 you have to -- this is your job.
25 MR. STROSNIDER: Yes.
. 275
1 DR. WALLIS: You don't even have to say it.
2 DR. KRESS: It's your only priority.
3 DR. APOSTOLAKIS: Because he was accused that he
4 wasn't doing that.
5 DR. WALLIS: The test that you're doing your job
6 -- it shouldn't be necessary to even say it. You know that.
7 MR. STROSNIDER: Well, but I'm trying to explain,
8 you know, how we're doing that. And, in fact, we're taking
9 these risk-informed insights that we've been gaining over
10 the last several years. We're factoring those into our
11 evaluations.
12 We're monitoring what's going on with Generic
13 Letter 95-05. And I'm doing more than just saying I'm doing
14 my job; I'm trying to give the explanation as to how.
15 Approaches to resolution, I would point out that
16 when you look at the different approaches to resolving these
17 issues, in some cases, you'll see some specific technical
18 answers where we're looked at assumptions and concluded
19 that, yes, this is an acceptable assumption to go into this
20 analysis.
21 And, you know, we've provided our technical bases.
22 In other cases, though, it may be a process, and
23 there's two ways you can look at that. I gave the example
24 before of building processes into our regulatory framework
25 so that we're sure we have the opportunity to look, you
. 276
1 know, from a risk-informed perspective, as an example.
2 The other thing you can read into this is issues
3 like the steam generator blowdown loads being put into the
4 GSI process, so there are a couple of different ways that we
5 end up dealing with these issues, depending upon their
6 significance and our understanding of them.
7 DR. WALLIS: So when Dr. Hopenfeld says there's a
8 factor of 100 difference between what he predicts and what
9 NRC predicts, you have some specific answers for that?
10 MR. STROSNIDER: Yes, and actually, the leakage
11 calculations he was referring to were performed back around
12 1993 or so, '94, in NUREG 1477, which was an ad hoc task
13 group that looked at the voltage-based criteria.
14 There is some discussion in that, and I think
15 that's the first place I would go to look.
16 CHAIRMAN POWERS: There are a couple of documents
17 that are useful in addressing the risk issue. There is the
18 NUREG 1477; there's also an INEL study on looking at issues
19 of operator response under conditions of multiple tubes
20 failing. And there is different phenomenology.
21 MR. STROSNIDER: Yes. In fact, I was just
22 referring to the leakage model itself, which was a
23 voltage-based empirical correlation that's used in 95-05
24 versus some other analytic methods that were used in the
25 other leakage analysis.
. 277
1 But it does require, again, looking at it from an
2 integrated perspective with regard to operator action and
3 system performance.
4 So anyway, just then to get to some conclusions
5 here on our future actions, I have already mentioned that we
6 want to continue to support the effort that you have
7 underway right now.
8 And if we can help you find something in that vast
9 amount of information, let us know. We are -- there is
10 mention, or I'll mention that we've had, coming after the
11 Indian Point 2 steam generator tube failure, the Office of
12 Investigations took a hard look at that and they have
13 published an event report.
14 The Staff, the NRC Staff, we implemented and
15 initiated our own lessons learned effort, and I believe that
16 that lessons learned task force report was sent up to the
17 EDO within the last day or so, and so it should be made
18 public, I would hope, in the near future.
19 And we've been working for some time now on the
20 NEI 97-06 licensing change package. We actually -- when the
21 Indian Point 2 rupture occurred, we put this on hold; we
22 stopped review of this.
23 It was a conscious decision, because we wanted to
24 factor any lessons learned from that rupture into this
25 review before we moved forward with the new framework.
. 278
1 DR. SEALE: When do you think the ACRS is going to
2 have a chance to look at that latest bit of information,
3 that, and the Indian Point thing and the OIG report?
4 MR. STROSNIDER: The OIG report is actually out
5 now. And we'll have to -- I could let Dean know where you
6 could get to it. I'm not sure if it's out on the Web or
7 exactly where it's at. That's available.
8 This is with the EDO. I'm not sure of the exact
9 schedule, but I would hope it's a matter of weeks or so that
10 that should be out.
11 CHAIRMAN POWERS: The question I have always had,
12 Jack, on the Indian Point 2, I mean, it's a rupture in the
13 u-bend, a specialized location. It's always susceptible to
14 the special -- has special strains and whatnot in material.
15 Did they really learn much looking at Indian Point
16 2 with respect to resolving the DPO?
17 MR. STROSNIDER: I'd suggest that actually we
18 have. One of the things that's come out of it, in effect,
19 is that the industry acknowledges and is already working on
20 it, is dealing with the quality of the eddy current data in
21 terms of noise.
22 That particular location is a difficult issue
23 because of the geometry, and getting a probe to fit and ride
24 the surface properly.
25 But also, the deposits, the copper deposits that
. 279
1 were on the tube had a strong influence. That's not the
2 only location in a steam generator where you might have
3 those sorts of problems.
4 So the industry is taking a look at the EPRI
5 guidelines and trying to address that issue of the quality
6 and the noise issue, so that's one example.
7 CHAIRMAN POWERS: But are the kinds of things that
8 complicate your detection up there, aren't those already
9 recognized in your POD of 0.6?
10 MR. STROSNIDER: This is where you have to be
11 careful. The .6 is applied in the Generic Letter 95-05
12 context only, okay?
13 And since you bring it up, it's interesting
14 because that's talking about the probability of detecting a
15 certain voltage amplitude, all right? And it goes into a
16 calculation where you make an estimate of how many tubes are
17 left in service because you failed to detect them.
18 And I think we did point out that we have gone
19 back and looked, based on the reports we're getting from the
20 licensees on implementation of 95-05, and that .6 factor
21 appears to account reasonably well in terms of predicting
22 how many and the sort of voltage indications that you'll
23 have at the end of the cycle.
24 I will tell you that one of the issues there is
25 that you have to understand how many indications you missed
. 280
1 and left in service, and you have to understand how many new
2 ones show up, because they have reached an incubation
3 period.
4 It's hard to separate the two, but at least our
5 look back at it is that the .6 works reasonably well.
6 And a final comment: There's a lot going on here
7 in terms of the OIG reports, Indian Point 2 Lessons Learned,
8 97-06, and we've committed to the EDO that we will have put
9 together by, I believe it's November 17th, an integrated
10 plan that pulls together how we're going to address all
11 these issues.
12 So, that's basically what I wanted to present, I
13 guess I'd ask if there are any questions? Joe?
14 DR. HOPENFELD: Can I make just one comment with
15 respect to Indian Point 2? I think it is relevant to the
16 DPO in the following sense:
17 One of the assumption that I have been criticizing
18 in the whole concept of how you calculate leakages through
19 that GL 95-05, is that the assumption there is that you use
20 prior experience on crack growth in previous cycles and
21 assume that that's going to happen next. You project prior
22 experience into the future.
23 And in a sense, this was done at Indian Point 2.
24 They said, well, nothing happened for the last 23 years;
25 nothing's going to happen next cycle, and that's what
. 281
1 happened.
2 But that's not the main thing; the main thing is,
3 when you get -- and it's not relevant, really, whether the
4 leakage was in the free span or wherever it was. You have
5 that final probability that whatever the leakage was, that
6 you'll overfill the unit.
7 Once you overfill the unit, you've got the safety
8 valves that are going to be stuck, and that leads you back
9 -- the steam line break, I mean, a stuck-open valve is a
10 steam line break, and that leads you back to the accident
11 I'm talking about.
12 Now you're talking about operating with a unit
13 with all kind of cracks and cracks in the sludge pile or
14 everywhere that you didn't detect and there is the potential
15 for them opening up. So it's a precursor, and I think that
16 in that sense, it is relevant.
17 MR. STROSNIDER: And the one comment I'd make on
18 that is, Dr. Hopenfeld raises a point about utilizing past
19 cycle data to project what's happening in the next cycle.
20 And that is part of the operational assessment.
21 The Staff does look at that as part of our review, and I'd
22 just point out that with regard to Indian Point, it does not
23 appear that that affects growth rates, and that unexpected
24 crack growth rates are really the issue there. They missed
25 a very large indication.
. 282
1 If you believe the eddy current sizing, it was
2 perhaps on the order of 80 percent of so through-wall. What
3 goes along with this assumption is that you can use power
4 cycle operation, is that you've got some threshold of
5 detection that supports applying that.
6 And so when you start getting back to the root
7 cause of this problem, it would appear to be the fact that
8 they missed a very large indication.
9 Anyway, are there any questions?
10 CHAIRMAN POWERS: Any other questions to be posed
11 to Mr. Strosnider?
12 DR. WALLIS: Well, again, I wasn't on the
13 Subcommittee, so I'm not as knowledgeable as my colleagues
14 here, but I'm not quite sure why what you told me helps me
15 to form any opinion about whether or not Dr. Hopenfeld is
16 right.
17 CHAIRMAN POWERS: I don't think that was the
18 intention.
19 DR. WALLIS: That was not the intention?
20 CHAIRMAN POWERS: Again, because our Subcommittee
21 meeting only allowed us a certain amount of time --
22 DR. WALLIS: That wasn't the purpose.
23 CHAIRMAN POWERS: -- we didn't allow them -- we
24 didn't say now you've heard what everybody said; come back
25 and talk to us some more. So we just put some more time.
. 283
1 You're listening to a Subcommittee presentation right now.
2 So you're not getting the full-blown story; you
3 get that from us.
4 DR. KRESS: In the Subcommittee, it lasted three
5 days. You couldn't compress that.
6 MR. STROSNIDER: I think it was a 20-hour
7 presentation by the staff.
8 DR. KRESS: You couldn't compress that.
9 MR. STROSNIDER: And I'm not going to try to
10 summarize that.
11 CHAIRMAN POWERS: And I think it took every bit of
12 that to understand, from soup to nuts, both what the
13 contentions are, and what the rebuttals are. It takes -- I
14 mean, I don't think there was a lot of fluff in those
15 presentations.
16 Again, I thought everyone on your team, even the
17 guys from Argon, did a bang-up job.
18 MR. STROSNIDER: Once again, I want to thank the
19 Committee for your efforts in looking at these issues. And
20 if we can be any further assistance, please let us know.
21 Thank you.
22 CHAIRMAN POWERS: Thank you. Okay, at that point,
23 I think we can suspend the transcription at this point and
24 we can take 15 minutes break.
25 [Whereupon, at 4:38 p.m., the meeting recessed and
. 284
1 reconvened into a unrecorded session.]
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Page Last Reviewed/Updated Tuesday, July 12, 2016