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
Page Last Reviewed/Updated Tuesday, July 12, 2016