Reliability and Probabilistic Risk Assessment - June 29, 2000
UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION ADVISORY COMMITTEE ON REACTOR SAFEGUARDS *** RELIABILITY AND PROBABILISTIC RISK ASSESSMENT Nuclear Regulatory Commission Room T-2B3 Two White Flint North 11545 Rockville Pike Rockville, Maryland Thursday, June 29, 2000 The committee met, pursuant to notice, at 8:30 a.m. MEMBERS PRESENT: GEORGE E. APOSTOLAKIS, Chairman THOMAS S. KRESS, ACRS Member JOHN D. SIEBER, ACRS Member MARIO V. BONACA, ACRS Member ROBERT E. UHRIG, ACRS Member WILLIAM J. SHACK, ACRS Member PARTICIPANTS: Paul A. Boehnert, ACRS Staff John T. Larkins, Executive Director, ACRS Michael T. Markley, ACRS Staff Howard J. Larson, ACRS/ACNW Staff Noel F. Dudley, ACRS Staff Cindi Carpenter, Office of Nuclear Reactor Regulation Tom Bergman, Office of Nuclear Reactor Regulation Mike Cheok, Office of Nuclear Reactor Regulation Mohammed Shwaibi, Office of Nuclear Reactor Regulation Tim Reed, Office of Nuclear Reactor Regulation Joe Williams, Office of Nuclear Reactor Regulation Mark Rubin, Office of Nuclear Reactor Regulation Bob Christie, Performance Technology, Incorporated Thomas King, Office of Nuclear Regulatory Research Mary Drouin, Office of Nuclear Regulatory Research John Lehner, Brookhaven National Laboratory Trevor Pratt, Brookhaven National Laboratory Allen Camp, Sandia National Laboratory Steve Floyd, Nuclear Energy Institute Adrian Heymer, Nuclear Energy Institute Biff Bradley, Nuclear Energy Institute. C O N T E N T S ATTACHMENT PAGE Proposed Schedule 210 Introductory Statement 210 Petition for Rulemaking Combustible Gas Control 211 PECO Energy Letter March 3, 2000 213 Risk-Informed 50.44 322 Memorandum to Ashok C. Thadani 344 Risk-Informed Part 50 Option 2 384 Risk-Informed Regulation 384 . P R O C E E D I N G S [8:30 a.m.] DR. APOSTOLAKIS: The meeting will now come to order. This is the second day of the meeting of the Advisory Committee on Reactor Safeguards, Subcommittee on Reliability and Probabilistic Risk Assessment. I am George Apostolakis, Chairman of the subcommittee. ACRS members in attendance are Mario Bonaca, Tom Kress, William Shack, Jack Sieber, and Robert Uhrig. The purpose of this meeting is to discuss the status of risk-informed revisions to 10 CFR Part 50, including proposed revision to 10 CFR 50.44 concerning combustible gas control systems, issues in the Nuclear Energy Institute letter dated January 19, 2000, option three, and public comments related to the advanced notice of proposed rulemaking on 10 CFR 50.69, and Appendix T, option two. The subcommittee will gather information, analyze relevant issues and facts, and formulate proposed positions and actions, as appropriate, for deliberation by the full committee. Michael T. Markley is the Cognizant ACRS Staff Engineer for this meeting. The rules for participation in today's meeting have been announced as part of the notice of this meeting previously published in the Federal Register on May 16, 2000. A transcript of the meeting is being kept and will be made available as stated in the Federal Register notice. It is requested that speakers first identify themselves and speak with sufficient clarity and volume so that they can be readily heard. We have received no written comments from members of the public. However, Mr. Bob Christie, of Performance Technology, Incorporated, has requested time to make a presentation concerning proposed revision to 10 CFR 50.44. I will remind the audience here that we wrote a letter to the Chairman, dated October 12, 1999, on primarily option two, but also option three. We will now proceed with the meeting and I call upon Ms. Cynthia Carpenter, NRR, to begin. MS. CARPENTER: I thank you very much. Tom Bergman, who is one of the co-leaders of the risk-informing Part 50 effort, is going to start off and then we'll have our speakers. DR. APOSTOLAKIS: Good. Thanks. MR. BERGMAN: Yes, just quickly. As you noted, we are here to present on option two, which is risk-informing the special treatment requirements. We gave the committee three handouts this morning, one which is the slides. We got a late comment letter, we've provided that to you, and we gave you a copy of our user needs memo, in addition to the material provided several weeks ago. The presentation, we have two new presenters for us. They've both been on the core team since its inception. The first is Mohammed. He will be presenting the section on the ANPR comments. Mohammed Shwaibi. He's been with the agency for eight years, about both Region III and six years here at headquarters of the eight. In terms of option two, Mohammed, he led the effort to develop the criteria and methodology for selecting the rules in option two. He drafted the ANPR and he is leading our evaluation of the comments we received. Joe Williams will be presenting preliminary staff views on treatment and he'll also cover the status and schedule. He has been with the agency for ten years in the Division of Licensing Project Management, mostly with Browns Ferry, and 11 years in the industry. In terms of option two, he is the PM for our pilot program. He's our principal interface with the South Texas exemption and he is leading the review of the NEI guideline on categorization and treatment, as well as the peer certification process. And you all know Mike Cheok. He'll be addressing the portions on categorization and PRA peer certification process. MR. SHWAIBI: Good morning. My name is Mohammed Shwaibi, and, as Tom indicated, I'll be talking about the ANPR comments today. As you know, the ANPR was published on March 3. There was a comment period of 75 days, which ended on May 17. In response to the ANPR, we received 11 comment letters, which included over 200 comments. We received six letters from licensees and industry groups, two letters from law firms, one letter from a consulting firm, one letter from a professional society, and one from a member of the public. The comments were divided into four major categories. I'll be talking about categories that related to the approach that we took. We described in the ANPR comments that were related to the categorization process in Appendix T, comments related to treatment, and comments related to the pilot program that was presented in Appendix T -- I'm sorry -- in the ANPR. With regard to approach, we received comments that were in general agreement with the list of rules that were identified, with a proposal that the risk-informing of the rules be taken in a phased approach. The proposal was to risk-inform the rules that were special treatment type rules, such as 50.49 EQ rules, seismic qualification, and those types of rules, and the administrative rules, such as reporting, documentation, be done in a later phase, and to break up the tech spec rule in Appendix R and fire protection as a separate and parallel effort. We were told to be performance-based in the approach that we take, make sure that the resulting rules are optional and allow for selective implementation for both the resulting rules and systems at the plant. We were told that the resulting rules should provide for limited NRC prior review and approval and that we should apply the backfit rule to option two. DR. APOSTOLAKIS: Now, when you say that you received these comments, you have here 11 different letters. MR. SHWAIBI: Yes. DR. APOSTOLAKIS: Were they unanimous in these recommendations? MR. SHWAIBI: I'm sorry? DR. APOSTOLAKIS: Were they unanimous in the recommendations? I mean, all 11 asked that the thing be selective, the subject of selective implementation? MR. SHWAIBI: Pretty much, yes. With regard to selective implementation, they were pretty unanimous. MR. BERGMAN: For those that commented on it, to qualify that. MR. SHWAIBI: Yes. MR. BERGMAN: Where there was a comment on selective implementation, most of the commenters didn't address all the questions in the ANPR. But for those that did, they were supportive of selective implementation. DR. UHRIG: By selective implementation, do you mean voluntary? That is, the utility has the option, or is this the NRC's selectivity? MR. SHWAIBI: No. It's the utility would have the option. DR. UHRIG: Have the option. MR. SHWAIBI: Yes. DR. UHRIG: Then why would you apply the backfit rule? MR. SHWAIBI: Well, one of the commenters did indicate that we ought to apply the backfit rule so that the Commission can have a full understanding of what we're requiring in the proposed rule or in the final rule. They said that we should not bypass the backfit rule just because it's voluntary. DR. UHRIG: Is that standard procedure? I thought it was not. MR. SHWAIBI: I don't believe. MR. BERGMAN: We have not decided how to address each individual comment. We'll be doing that as part of our August paper. DR. UHRIG: My comment was intended -- is this the procedure within NRC? I thought if a rule was voluntary, you did not apply the backfit rule. MR. BERGMAN: That is true. That is how we've done it in the past and I believe that's what was said. DR. UHRIG: Then why would this be different? MR. BERGMAN: We haven't decided how we will address it. We have a commenter who has proposed that we apply the backfit rule. Whether or not we'll accept that comment we have not decided. MR. SHWAIBI: This is just a summary of the comments that were provided. We don't have positions yet on them. DR. KRESS: By selective implementation, I thought you meant they could choose which of the rules they want to go by and which ones they didn't. MR. SHWAIBI: That's correct. The comments were provided on selective implementation with regard to the rules and selective implementation with regard to systems at the plants. That is, a licensee could select which rules they want to implement and then they could also select which systems at the plant they chose to apply the risk-informed approach to. Moving on to the categorization of Appendix T. We received comments that the Appendix T that was in ANPR was unduly detailed, prescriptive and burdensome. DR. APOSTOLAKIS: Can you give us an example of what people felt was burdensome? MR. CHEOK: I think one quick example of that, George, is that we had asked for all SSCs to be identified and what the current requirements for those SSCs are, and I think that one of the comments was that's asking too much for the SSCs, since there are thousands of them. DR. APOSTOLAKIS: But didn't South Texas do this? MR. CHEOK: In a sense, they did, yes. DR. APOSTOLAKIS: Yes. MR. CHEOK: I think when we wrote Appendix T, the idea was that there was going to be minimal staff review and approval. I think industry may be backing up a little bit instead of maybe there should be some kind of a submittal and, you know, when we make the submittal, we would like the rules to be less prescriptive, because then whenever they want to change the process, it's not allowed because it's not in the rules or the regulations. DR. APOSTOLAKIS: Is this related to a recommendation I saw somewhere that Appendix T be eliminated and replaced by a regulatory guide? MR. CHEOK: That's related to that recommendation, yes. DR. APOSTOLAKIS: And how does the staff feel about this? MR. CHEOK: We are open to that recommendation. The material in Appendix T can either reside there or it can reside in another document, like a reg guide or an industry document, which the staff endorses. DR. APOSTOLAKIS: So this is still to be decided. MR. CHEOK: It's still be decided. DR. APOSTOLAKIS: And the unduly detailed, can you give an example of a detail that shouldn't be there? MR. SHWAIBI: I could provide you some examples. We have a list of examples. DR. SHACK: Just coming back, while you're at it, I thought there was some discussion whether, if you put it in a reg guide, you had to do more review and approval, whereas if it was in the rule, you had less review and approval, and that was the justification for making it part of the rule in the first place. MR. CHEOK: That's correct. And like I said, I think industry is backing off a little bit and they feel that they would like the flexibility. DR. SHACK: Okay. So they have the flexibility. MR. CHEOK: And perhaps maybe make a limited form of review, what they call a template type review. DR. APOSTOLAKIS: A what kind of review? MR. CHEOK: A template; in other words, an agreed upon five-pager or something. DR. APOSTOLAKIS: Okay. MR. BERGMAN: The unduly detailed example would be the precise makeup of the IDP. It was spelled out in the draft Appendix T included in the ANPR. DR. APOSTOLAKIS: I'm really curious myself about the IDP, because we are performing unduly detailed reviews of the PRA part and then we are turning over the results to the IDP, and they are pretty much free to do whatever they please. I think we are spending our resources reviewing something that can't be reviewed because it has some quantitative elements in it, whereas the IDP is just the deliberative process. So at some point in the future, we really have to think about it very hard. Maybe there ought to be some general guidelines regarding the conduct of the deliberation, because we worry about little things here and there, the significance of other -- sensitivity of importance measures of various things, and then we're taking the results and saying now you guys do what you like with those. Well, it's not like that, but. MR. CHEOK: I think the staff is working with the industry on the guidance document and how the IDP should deliberate, and hopefully we'll come up with something that's acceptable to everybody. MR. SHWAIBI: I think the comments were mostly to take the details and put them in the guidance document. They want to be able to -- some of the things that were pointed out were as advances in technology occur, we want to be able to take advantage of those without having to go back to rulemaking. And if you lock them in in Appendix T or in the rule, that would make it hard to do. You'd have to go through rulemaking, and those were some of the comments on the level of detail that needs to be included in the appendix and in the rule versus the level of detail that could be included in a guidance document or a reg guide. DR. APOSTOLAKIS: I don't understand the second bullet. What is the consensus PRA standard, the ASME and the ANS? MR. SHWAIBI: In the proposed -- DR. APOSTOLAKIS: Can you move up the microphone a little bit? MR. SHWAIBI: I'm sorry. DR. APOSTOLAKIS: Good. MR. SHWAIBI: In the proposed Appendix T, I believe it was identified that the PRAs would have to meet the consensus standard, ASME, ANS, whichever one would be available at the time. It was expected, I guess, that they would be available at the time. DR. APOSTOLAKIS: I wonder the people who wrote this had seen the standard. MR. CHEOK: The people who wrote this, and you're looking at him, has not seen the standard at the time. DR. APOSTOLAKIS: Wait a minute. You are not the public. MR. CHEOK: You mean the comments. DR. APOSTOLAKIS: Yes. MR. CHEOK: All right. DR. APOSTOLAKIS: You're not public. I mean, it's such a high level document, that I don't understand what it means, that it should not be accepted as the only method. It's not even a method. MR. CHEOK: I think the intent of the thing is maybe we should look at other avenues; for example, the PRA certification that's being forwarded to the staff to look at and that the ASME standards should not be the only way. That is a public comment that we have to address. DR. APOSTOLAKIS: It's really such a high level document, that I don't know that it cannot be the only acceptable method. Basically, it tells you to do fault trees and event trees. Why is it important to minimize the levels of the risk significance? MR. SHWAIBI: I'll give you what was provided in the comments. It was suggested that with the different levels of risk significance, you would need to come up with different treatment requirements for those levels and that would just be too complex. DR. APOSTOLAKIS: South Texas didn't. Essentially, they have four categories or something. MR. WILLIAMS: South Texas has four levels of risk significance, but really only two types of treatment. DR. APOSTOLAKIS: That's right. MR. WILLIAMS: Because the high safety significant and medium safety significant are treated essentially the same. DR. APOSTOLAKIS: Exactly. So in essence, they have two categories. MR. WILLIAMS: Yes. DR. APOSTOLAKIS: So I don't understand that comment either, or there is a concern that there might be different -- MR. SHWAIBI: Different levels of treatment. MR. RUBIN: I'm Mark Rubin, from the PRA Branch. Just one quick data point. South Texas, originally, in the GQA submittal, you may remember, had an intermediate category. So perhaps the public comment was reflecting that. Allowing too many categories gets too complex, and I think South Texas concluded that and they went back to a single cut point. DR. APOSTOLAKIS: And what's a functional categorization? MR. SHWAIBI: Yes. In the ANPR, the way that categorization was discussed was to bin structures, systems or components, and that comment was saying that we ought to look at the functions that these components will be providing and those functions. So in other words, you're dividing -- you're taking the different functions of a component and binning them into the different bins. You will be taking the different functions that a component will be providing. Instead of binning all of the functions of the same component in one class, in one safety class or one risk class, you could take the different functions of that component and bin those functions into the different classes. So a component could have functions in different classes. DR. APOSTOLAKIS: And then you would go with the most stringent category or you would have it targeted? MR. BERGMAN: We'd target by function. MR. SHWAIBI: You will address the functions or the attributes that give you the function that ended up in the different bins. DR. APOSTOLAKIS: I wonder how you would do that? MR. BERGMAN: George, I think this is a lesson learned from the maintenance rule. The maintenance rule was written on an SSC basis, as well, but most licensees implemented it with a function-based approach, which, after we observed it, we concluded it was a very good way to tackle the problem. In South Texas, in a meeting a couple months ago, they pointed out that they've looked at something like 22,000 components, but those components only fill 500 functions. So by taking the function-based approach, it does appear to simplify the problem somewhat. They're just saying consider allowing us that flexibility to tackle the problem using either approach, component-based or function-based. DR. APOSTOLAKIS: I guess it's not clear to me. I thought it would make the process more complicated. So you're not taking -- the way you described it, Mohammed, you're taking one component that may have three functions. MR. SHWAIBI: That's correct. DR. APOSTOLAKIS: So you categorize -- now, instead of categorizing one component, now you have three functions to categorize. MR. SHWAIBI: That's correct. DR. APOSTOLAKIS: So that makes it more complex. MR. SHWAIBI: That's correct. DR. APOSTOLAKIS: We were just told that it's actually simpler. MR. BERGMAN: Because there only may be, take the example, 500 functions. You may have 50 components, all fulfill the same function. DR. APOSTOLAKIS: I see. MR. BERGMAN: So especially when you get into the monitoring aspects, you're monitoring functions, which is typically what you look at anyway, are you injecting enough water, are you maintaining the bus voltage, and when you fail to meet that condition, then you track down to figure out what's causing the inability to fulfill the function. But when you look at your plan, it's easier to look at it from a functional perspective. DR. APOSTOLAKIS: But eventually you would have to go to the hardware. MR. BERGMAN: You always have to end up going to the hardware. DR. APOSTOLAKIS: Because there may be an intermediate step that makes it a bit easier. MR. BERGMAN: Right. As long as stuff works, it's easier from a function-based. When you have a lot of stuff breaking, you're ultimately back at components anyway. MR. SHWAIBI: I think if you go back to our four-box diagram, one of the comments said that if you have a component that ends up in box one or RISC-1, ends up as a RISC-1 category, yet it does not need to be environmentally qualified because of the functions that it provides, it should not have to meet the 50.49 environmental qualification criteria. You could take the environmental qualification parts, and maybe that is lower safety significant. DR. APOSTOLAKIS: But, still, it's new to me, so bear with me for a minute. I'm categorizing the function of a particular component for a function that is needed for the safety of the plant. MR. BERGMAN: It's kind of both. I think the example that Mohammed brought up is very good. You're looking at the functions and you'll have RISC-1 functions and you say which components fulfill that function and the function, though, may only be applicable for, say, seismic reasons, but not the rest of EQ. So right now, if you do a component, if it has any -- if it's safety-related for any reason and it's safety-significant for any one reason, the entire SSC and all the functions its performed remain subject to all the special treatment rules. If you go on a function-based approach, you would only need to apply those special treatment rules for those functions that are both safety-related and safety-significant. So it does allow a little bit more flexibility. DR. APOSTOLAKIS: There could be a downside, though. I remember a complaint from South Texas was that if a function is in the box RISC-1, then automatically all the components supporting that function were RISC-1. Do we have the same problem here? MR. BERGMAN: Oh, you definitely could. There's a balancing between -- it makes your approach to the component much more complicated and I think South Texas has concluded if it's high for any reason, just throw the whole thing in high and just keep dealing with it the way you deal with it today rather than trying to pare away at it. So it's a comment we've got. Like you said, how it plays out, we don't know. They're just asking for the flexibility. I don't think even utilities will know till they actually get implementation which approach is necessarily better. MR. REED: Utilities still may want to do it on specific circumstances. The component that comes to mind for the examples, for me anyway, is the RHR pump. For ECCS, it may come out, for that low head pump, maybe come out low safety significant, let's say, but for mid-loop, it may come out mid or high or for achieving maintain safe shutdown condition, it may come out high or medium. You would want to treat it appropriately for mid-loop or achieving maintain safe shutdown condition, those functions as high or medium. ECCS now, which is a lot of what the best treatments are focused on on that particular component, while they're not doing anything for you, you know what I'm saying. So there may be a few where they want to split it apart, but like Tom said, it does get complex. It simplifies the categorization process, because you're categorizing, say, 500 things and then you're mapping components into that, if you will, but in the end, when you come down to the component and you've got this thing in different boxes, in a sense, based on its functions, that gets complex and I think that's where South Texas said we're simplifying this. So if it's got anything that's high, it's high for everything. DR. APOSTOLAKIS: The example you just gave considered two different modes of operation. MR. REED: Yes. DR. APOSTOLAKIS: In which case, the PRA is different, the measures are different, everything is different. That's not what I understood by function or categorization. I understood that for the same mode, the same importance measures and so on, instead of going directly to a component and asking what is its role, in general, you are looking at specific functions that it supports, and for some of these functions, it may be RISC-1 and for others it can be something else. But for the same mode, if you change the mode of operation, then you're changing the importance measure. So it doesn't surprise me that -- MR. REED: Yes. Actually, change mode and actually change internal and external events, too. I did the whole thing there. I think we've got to consider the whole gamut, actually. I don't know. DR. APOSTOLAKIS: They are talking about, in one mode, having different functions and categorizing those. Is that the intent of this bullet? MR. SHWAIBI: I'm not clear if the comment is actually talking about one mode. I think the comment was to allow the flexibility to do this. It's not specific as to for one mode, you would need to do it. But I would imagine that if you were doing -- DR. APOSTOLAKIS: Now, you changed the mode of projection. Why? MR. SHWAIBI: It just went off. I guess it was a screen saver. It went off. DR. APOSTOLAKIS: Okay. That's better. Why does it say random access? Shouldn't it be alliatory access? Okay, Mohammed, go on. You're kind of slow today. MR. SHWAIBI: I guess move on to the last bullet, then. DR. APOSTOLAKIS: Yes. MR. SHWAIBI: That we need to address the use of results from PRAs or tools with different levels of conservatism or uncertainty. DR. APOSTOLAKIS: That's a mystery to me what it means. Can you elaborate? MR. SHWAIBI: Yes. The intent of the comment here is that we have different tools that we're using for categorizing. For shutdown, we have certain tools, for seismic, for fire, and for internal events, and we don't want the conservatisms and uncertainties in one tool to mask significance that would come out from another tool. We don't want, for example, our shutdown tool to drive a component into the low bin when it should be high for the internal event scenarios. DR. APOSTOLAKIS: So how would you handle this? MR. SHWAIBI: I think that's already -- DR. APOSTOLAKIS: It sounds like a research project, to me. MR. CHEOK: I think in Appendix T, what we say is that if you do have a PRA model for fires, for, let's say, low power and shutdown, that you take the importance measures and you treat them both cumulatively and individually, and you need to look at the results, both sets of results, and categorize your SSCs based on both sets, not just the cumulative results. DR. APOSTOLAKIS: By cumulative, you mean PRA as one. MR. CHEOK: Consider all the cut sets into one. DR. APOSTOLAKIS: And then you separate out, say, the fire and see what happens. MR. CHEOK: Right, and low power and shutdown, because they do tend to be maybe more conservative and they might skew the results. DR. APOSTOLAKIS: Okay, sir. MR. SHWAIBI: Moving on to treatment. We have comments that suggested that any additional treatment for safety significant attributes should be determined by the licensees and that they should rely on existing licensee programs. Those would be in the RISC-1, RISC-2 boxes. DR. APOSTOLAKIS: What happened then to the declaration that risk-informing the results may, in fact, lead to additional requirements? MR. SHWAIBI: Again, this was the comment. We're still addressing it. We don't have a response to that yet. DR. APOSTOLAKIS: I think Mr. Riccio would like this. Okay. MR. SHWAIBI: For LSS SSCs, commercial programs provide sufficient treatment, was the comment on those. DR. APOSTOLAKIS: Somebody asked the question on Tuesday, which I will ask you now. MR. SHWAIBI: Yes. DR. APOSTOLAKIS: What is a commercial program? MR. SHWAIBI: We're in the process of trying to learn what that is, trying to understand what it is. MR. BERGMAN: We'll talk about that a little bit later. MR. SHWAIBI: The rulemaking should eliminate existing commitments for LSS SSCs, was another set of comments. Any existing commitments would be eliminated. DR. APOSTOLAKIS: Now, a matter of terminology again. Low safety significance is what used to be called low risk significance. MR. SHWAIBI: Low safety significant SSCs, here, what we're basically talking about is the RISC-3 box. Any commitment -- DR. APOSTOLAKIS: I understand that. I mean the terminology. Safety significance is the former risk significance. MR. CHEOK: Yes. DR. APOSTOLAKIS: Yes. And why the change in terminology? MR. CHEOK: I think for a while now we have been using low safety significance to be consistent throughout the whole agency. I think we have been using LSS and HHS for the last three years or so, since we were implementing the maintenance rule. DR. APOSTOLAKIS: So something can be safety-related and of low safety significance. MR. CHEOK: Right. MR. SHWAIBI: And, finally, that the risk-informed -- a risk-informed change process should be included in the new rule. This is where we recognize 50.59 may not be sufficient and we need to include something other than 50.59. DR. APOSTOLAKIS: Change process? MR. SHWAIBI: Change control process, yes. Finally, on pilot programs, we received comments that the final rule should not be backfit on pilot plants that have reviewed and accepted processes for categorization and treatment. In addition, we received comments that since STP has already demonstrated that they can categorize and provide treatment for different types of components, that there is no need for other pilot plants to do the same. This is where, in the ANPR, it was suggested that pilot plants would need to include passive components, active components, mechanical, electrical and all types of components, and the commenter was saying that there is really no need to do that, since South Texas would have demonstrated that. If there are no comments, I'll turn it over to Joe Williams. I think he's next. MR. WILLIAMS: I'm Joe Williams. I'm the project manager for the review of the NEI guideline documents. What we have here, first of all, is we have three guidance documents. Two of them are actually part of what will be a single document. First of all, NEI submitted NEI-0002 in April of this year. That provides their peer certification process. They have also provided their categorization guidance for option two application. That was submitted in March of this year. Then in June, they just submitted their draft treatment guideline. We'll start discussing the review of NEI-0002. We're working with the Office of Nuclear Regulatory Research in the review of this document. We have an overall outline that's presented the next couple of slides. Mike will be speaking to most of these points. I just want to point out that our intent is to provide comments to NEI. We have talked to them a couple of times at public meetings about the categorization, the treatment and the peer review. We're going to be providing formal comments to NEI over the next couple of months and then we'll be working with them to develop a final document that hopefully will be endorsed in the regulatory guidance that will go forward with the proposed rule. Mike will now talk about the particulars of the review of the NEI-0002. MR. CHEOK: This NEI-0002 is basically the industry peer review certification process. This is the process that was, I guess, adopted from the BWR Owner's Group process that was presented to the committee probably six months ago. To review this document, we came up with a task plan and like Joe was saying earlier, this is going to be an Office of Research and NRR effort. The task plan has four tasks in it. Task one basically calls for us to review the process itself, the overall process, to see if it meets the general staff expectations of what we think a peer review should look like. We also would like to look at the QA requirements that are being put on the PRA. In Reg Guide 1.174, we did say that parts of Appendix B should apply to the PRA, since it's used to change the licensing basis. We would like to see those parts of Appendix B being implemented. In task two, this is the task that our Office of Research, I guess led by Mary Drouin, is going to be looking at. This task basically looks at the technical elements of the peer review process. Again, we are writing a SECY paper, due to the Commission probably the middle of next month, that outlines our high level expectations of what we would like to see in the PRA. So the Office of Research is going to use these high level expectations as a guidance to look at the peer review certification process to see if they meet this guidance criterion. The next thing they're going to do is they're going to look at the sub-tier criteria that's provided by NEI-0002. We should note here that in submitting the document, NEI has asked that we do not look at sub-tier criteria -- I mean, do not review the sub-tier criteria. They would just provide it for information. The staff had come to the conclusion that we do need to look at the sub-tier criteria because this is the basis for which the peer reviewers are going to make the decisions whether something is graded with one, two, three or four. So we have written a letter back to NEI to state that, yes, we will look at your certification process, but we will have to look at them in concert with the sub-tier criteria. So part of task two here is for the Office of Research to look at this sub-tier criteria and, in a sense, compare it to existing and available documents, the ASME standards being one of them, to see how consistent they are. That's the first task in this sub-task. And then if there are inconsistencies, determine if these inconsistencies will affect applications in option two. Task three is then actually to look at what we actually want to do in option two. Remember, NEI has asked us to review the certification process in conjunction, in light of applications to option two. So what task three does is that we need to look at what the requirements of option two are. In other words, what's the role of PRA in option two and how are we going to use the results of a PRA in option two; how are things like defense-in-depth, safety margins, and the other expert panel type issues going to affect the results of the PRA. DR. SHACK: Mike, I as recall, I didn't think you assigned an overall grade to a PRA. Wasn't it sort of graded on an element by element basis? Do we take an average or something? MR. CHEOK: No. You are actually correct. We do give a grade to the elements themselves and NEI-0002 is very clear in saying that we do not assign an overall grade to a PRA. DR. SHACK: Then what does task two mean then? Somehow it sounds as though there's something like the grade three PRA, grade three PRA is the overall beast. MR. CHEOK: I think I'll go to that a little bit. Basically, what happens is NEI has asked us to look at this certification process with respect to option two and since they have four grades, the grade that corresponds to what the PRA should be to be good enough for option two is the grade three PRA. So basically you're looking at the sub-tier criteria for grade three to see if these sub-tier criteria are what we think is needed to be applied for each element for option two. DR. SHACK: Suppose he comes out with grade three in X attributes and grade two in Z attributes. MR. CHEOK: And I think we will discuss this in the next slide, but basically what happens is -- one of our comments to them is, look, you have to come up with a certain level of conformance in each element and then if you do not conform, we would like to know why you don't conform and document that, so that your expert panel will know why you don't conform and how you can get around this, what we call the tradeoffs to apply to option two. But this thing has to be documented well enough for your expert panel to know and for the staff reviewer to know that you have used these tradeoffs correctly. In a sense, I'm not -- I hear this committee discussing the grading levels yesterday for the ASME standards, and the more I listen to the discussion, the more I'm thinking the ASME is talking themselves out of having different grades, because, in a sense, they are defining a standard for grade four, for example. And you come in and say, hey, look, for this application, I may not meet the standards, but I do not meet them because, here's my reasons, and there could be more bullets for grade one and maybe less bullets for grade two. But I think when you apply something, you need to know those bullets anyway and you need to know how to get around those bullets or those bullets have to be fixed in an update of the PRA. So in essence, you're really comparing yourself to a standard, which is the grade four standard, and for each application, you need to know how you -- the tradeoffs are probably different for each application. So in that sense, I'm not sure. DR. SHACK: It's a useful tool for self-assessment. MR. CHEOK: That's right, that's right. DR. SHACK: I can see where my PRA has had shortcomings. It's not category one, two or three or grade -- DR. APOSTOLAKIS: Use your mic. DR. SHACK: He's going to have to come in, as you say, and defend it. MR. CHEOK: Right. DR. SHACK: He'd be better off to go to grade four or category three and be done with it. MR. CHEOK: Right. In essence, a grade two is I'm conforming to the standards, but to a less extent than grade three is. And how I don't conform to it, I have to document anyway, because I need to know why it doesn't conform, not just that it's a grade two. A grade two by itself doesn't mean anything to me and a grade three doesn't mean anything to me. I like to know why it's a grade three. So in that sense, putting the different grades out there doesn't really make that much sense. DR. SHACK: Except for a self-assessment tool or to tell you how to fix it up. MR. CHEOK: So basically, here in task three, like we were talking earlier, we would like to understand why something is not up to par and if there are any compensatory measures or tradeoffs that can be done; for example, sensitivity studies or more defense-in-depth to get around this one element that's not up to par, so to speak. And the last task, we are trying to define what elements of the peer review process are important enough that it needs to be submitted to the staff and what elements are important enough to be kept on-site for staff inspection, so to assure us that this process has been carried out correctly. Here, we met with NEI on Tuesday, two days ago, and our comments on NEI-0002, these are the high level initial comments, basically are summarized in these four bullets. We said that we need sub-tier criteria to review the certification document. MR. WILLIAMS: If I may. One thing I want to point out, too. It's not just that the sub-tier criteria are integral, but also that we received only the BWR sub-tier criteria. We also need the sub-tier criteria for the pressurized water reactors, as well. MR. CHEOK: We said that it's important for this process to document all the findings very well. It's in these findings that make the peer review process itself useful. For example, I'll give an example. For grade three, there's a lot of "should's" in there. The "shalls" go in grade four, and the "mays" go in grade two. If you look at the definition of a "should," it is you shall either have met the requirements or you have good documented reasons why you didn't meet the requirements. I think that the licensee, PRA analysts, the expert panel, as well as the staff, need to know what "should" means in each of these cases. Is it because they met the element or is it because they had something that -- some documentation that says they don't have to meet this element. This documentation or justification might actually affect the application and if that's the case, I think we should have it ready for the expert panel to review and act on. Keep in mind that the certification process could have been done two years ago with no application in mind. So I think good documentation of this process is essential. The third bullet here is how applicable are previous peer reviews. In other words, the majority of BWRs have already been peer-reviewed. A lot of the PWRs are being reviewed currently. How do we apply the results of past peer reviews? I mean, the sub-tier criteria were not put down on paper until probably recently. What do we do with cases like that? What happens if the staff finds some discrepancies and would like to add the staff guidance into the guidance document? What do you do with plants that have already been certified? So that's a topic that we need to discuss with industry. Independent decision-making panel, again, I think Dr. Apostolakis pointed this out earlier, we are saying that perhaps the PRA doesn't have to be as good as our standards. Our expert panel can take care of this. I think we need to have good guidance to the expert panel as to how they can take care of this. Categorization. NEI submitted this document to us in March and in the document they talked about PRA scope and quality. It was at that meeting that actually it was suggested that they submit to us NEI-0002 to address the quality issue. As far as the scope issue is concerned, they are proposing that an external events PRA and a low power shutdown PRA is not necessary for this process, and in Appendix T, the staff actually basically said the same thing. NEI has proposed processes where you can use 91-06 criteria that's for risk management and shutdown configurations and how you can use analyses and the seismic margin analyses in light of PRA and how we can categorize components using those kinds of analyses. The staff is looking at those proposals and, again, I think it comes down to the role of the expert panel. It's my personal feeling that if you are going to not do a PRA, you should be a little bit more conservative in your assessment of categories. You should encourage the use of PRAs and they should be able to tell you the more correct results. So, again, this comes to how we define the role of the expert panel. DR. APOSTOLAKIS: Mike, as I recall, the South Texas project categorized something like 21,000 SSCs, roughly. And a typical number of SSCs that appear in a good PRA is on the order of maybe 1,200. That's the number. MR. CHEOK: The number of events, the basic events in a database normally is 1,200, 2,000, something like that. That could be less SSCs even, because -- DR. APOSTOLAKIS: Less SSCs. MR. CHEOK: Yes. DR. APOSTOLAKIS: The point is that it's 20,000 versus 1,200, 1,500, 1,000, that kind of number. MR. CHEOK: Right. DR. APOSTOLAKIS: Which means that you had almost 20,000 SSCs that were categorized without the PRA. MR. CHEOK: Correct. DR. APOSTOLAKIS: Are you familiar with the process that they followed to do this? We had a short presentation here once. MR. CHEOK: I'm not totally familiar with what South Texas did, but what Appendix T calls for basically is to first map the SSCs on to the PRA, if you can; in other words, to implicitly model SSCs. So instead of the 2,000, you might actually have 6,000, the piping, the instrumentation that's dependent on the operators, the tanks. DR. APOSTOLAKIS: Yes. So this is a variation, if you will, of the risk-informed ISI approach, the Westinghouse surrogate component. MR. CHEOK: In a sense it is, yes. MR. WILLIAMS: If I may, with regard to the South Texas, I know that part of that process is that they have several critical questions that they ask about these non-PRA components, if you will, such as whether or not they're involved in the emergency operating procedures or not. They assign a numerical rank from essentially zero to five rank. DR. APOSTOLAKIS: Yes. MR. WILLIAMS: According to some criteria, and then they have some weighting. Then they have rules that they've applied that according to the sum of all those ranks for the individual questions and the weights, that they then will bin those SSCs according to where they fall out of that numerical system. DR. APOSTOLAKIS: So I think -- I mean, I fully agree with you that this sounds, and it is, a structured process, but I'm not sure that, as a community, we have really understood what the process is and what it means to do certain things. For example, do these different constructed scales they use, do they represent independent attributes, mutually exclusive attributes, do they have to be independent, is there a risk of double counting perhaps. You know, all these things -- we're getting now into the structure of deliberative processes, which is not a new field for some community, but that community is not part of our community. So I think since they are making these important decisions, I think we should look at this process a little bit more carefully. These may turn out to be great, that's fine, I'm not saying there are problems with it. In fact, I was very pleasantly surprised when I heard South Texas presented and showing that they actually tried very hard to put some structure into the process. MR. CHEOK: We do have some experience with South Texas and we do have experience with ISI and IST, with the pilots in there. So we do have some kind of experience with how expert panels do work, but I think we still need to nail down exactly how the process should go. DR. APOSTOLAKIS: Do you have any document that describes what was actually done? Because I haven't really seen it. Besides the presentation, I'm not familiar that we have anything else. MR. WILLIAMS: For South Texas? DR. APOSTOLAKIS: Yes, or anybody who has used -- MR. WILLIAMS: Certainly we have the submittals that South Texas has made to date. DR. APOSTOLAKIS: It's just the results of the process itself. MR. WILLIAMS: It's the process itself. They described the process in some -- DR. APOSTOLAKIS: Can you give an example or two? MR. WILLIAMS: I believe so, yes, because we have copies of some of the risk significance basis determination documents. DR. APOSTOLAKIS: Is that a huge document? MR. WILLIAMS: The risk significance determination document, that's pretty large. The documents involving the description of the processes, those are manageable. DR. APOSTOLAKIS: May you can coordinate it with Mr. Markley here. MR. WILLIAMS: We can take care of that. DR. APOSTOLAKIS: And see if part-time people can review it in a reasonable amount of time. MR. CHEOK: I'll keep going through this bullets quickly. The second bullet basically says the role of importance analysis can be used in sensitivity studies to bound an increase in risk, or do we just depend on the importance matrix, such as Fussel-Veseley and RAW. The next bullet there is, again, the expert panel, what role it plays, and we have discussed that quite a bit already. The fourth bullet there is how we treat low safety significant safety-related components, and I think Joe will talk about that in later slides. The last bullet there is what role should monitoring and feedback play in this whole process, how does this affect the PRA updates and how does this affect the whole process in general. DR. APOSTOLAKIS: Now, you know that we had this presentation from Palisades on top event prevention methodology. Is that left out because the authors of the NEI-0002 were not aware of it, as most people are not, or because they decided they would go with Fussel-Veseley and RAW? MR. CHEOK: I think the Fussel-Veseley and RAW is something that's known to everybody. Everyone has the capability of doing it. So that's the way that NEI chose to go. The top event prevention is another acceptable method to do this. I think they have at least one licensee, maybe two or three others might follow, but on a generic sense, they do not have enough -- I guess we can call it support for that methodology. DR. APOSTOLAKIS: But, again, though, they don't have enough support because people have studied it and they decided not to support it or because it's just brand new and they are not really familiar with it? MR. CHEOK: I believe they are probably somewhat familiar with it. They just chose to go with the method that they already know how to do and they think will work. DR. APOSTOLAKIS: Do they allow other methods to be used? MR. CHEOK: I think the documentation is such that they would be flexible to allow any other methodology, and the staff, in fact, say you can actually submit a top event prevention, if you like. There's nothing to stop them from doing that. We will just have to review it on its own basis. DR. APOSTOLAKIS: Well, at some point, though, we have to understand what the differences are. Are we getting more or less or are you doing one methodology versus the other? MR. CHEOK: We have an e-mail, I guess, we have communicated with Palisades on what they should be doing about the top event prevention methodology. They want to apply it to IST. We haven't gotten back to them yet, but basically what we're going to tell them is that, sure, go ahead and submit it, we would like to find out more about it, and find out how applicable it is to option two and the rest of the stuff we're doing. DR. APOSTOLAKIS: Okay. So you are in the process then of examining further that methodology. MR. CHEOK: That's correct. DR. APOSTOLAKIS: Okay. MR. WILLIAMS: We will now talk about some of the feedback we've given to the industry regarding treatment guidelines. Most of this discussion was provided to NEI and other industry stakeholders on Tuesday. The first bullet deals with the definition of commercial practices. One of the predominant issues here is that commercial practices covers a very wide range of activities. Consider, for example, the distinction between a Rolls Royce and Yugo. Both of those are commercial vehicles. They're presumably for the same end, but clearly much different in their application. The staff is interested in basically defining the set of commercial practices that provides an adequate assurance of functionality, both in the context of the preservation of the design basis for the components that are categorized in the RISC-3 area, and also when those commercial practices are applied for the, for lack of a better term, severe accident attributes in the RISC-1 and 2 areas. The NEI document, NEI-0002, has provided a useful outline of how they propose to proceed in this area, but the staff is going to need additional details before we can complete our review. The next bullet deals with the preservation of the design basis. Fundamentally, that's, under option two, the existing deterministic design basis is supposed to be preserved. It cannot be changed by the rule itself. If the licensee chooses to do that, they'd have to choose another regulatory mechanism. So the need to identify and provide adequate protection of those design basis attributes is an inherent part of the process. The next bullet deals with change control. The issue here is that 10 CFR 50.59 is focused exclusively on the preservation of the deterministic licensing basis. So it's not an adequate tool to address facility changes that might affect severe accident performance. For example, a pressurizer PORV could play an important role in a facility risk profile by providing the capability for feed-and-bleed, the once-through core cooling scenarios. If a licensee chose to somehow diminish that capability, the existing 50.59 might fully allow the licensee to proceed and make that change, because it is not a design basis event. However, it could have a very significant effect upon the facility's risk profile. NEI has indicated that they agree with the staff on the need to address these severe accident attributes and has mentioned that in their guideline documents on treatment. However, again, the staff is going to need some additional details. We have a lot of work to do before we can actually define the process in the way that it will be applied by the industry. The predominant concern, at least in my mind, is that there's some level at which, similar to 50.59, at which prior staff review would be required before a facility change could be made. We want to strike a balance between the facility's ability, the licensee's ability to make reasonable changes to their facility and to manage their own risk profile, but also recognize that there is some threshold that they could reach, hopefully rarely, that we would want to be engaged prior to those changes being made. The next bullet, dealing with the adequate assurance of RISC-2 capability. The fundamental issue here, in my mind, deals with performance monitoring. Basically, performance monitoring under normal operating conditions may not be relevant to the severe accident conditions that might be seen. So how will you actually be able to meaningfully monitor these components as they normally operate and say that they derive information that's actually meaningful for their usage in a severe accident environment. Also, I will point out that NEI has proposed that we divide the RISC-2 category into two subcategories. Basically, one is those non-safety-related SSCs that are subject to some other regulatory treatment, such as fire protection components or station blackout components, and also one category where there's no existing special treatment, no existing regulatory requirements for special treatment. So this has led to a distinction between the SSCs in the RISC-2 box. Finally, with regard to the adequate assurance of RISC-3 functionality, again, we have the other side of the coin with regard to performance monitoring. Performance monitoring, again, during normal operating conditions may not provide meaningful information regarding design basis capability. So a program that exclusively relies upon, say, for example, the maintenance rule, may not, in and of itself, be sufficient, may require some other attributes. You may have to rely upon other attributes of a commercial program or other regulatory controls to provide the adequate level of assurance for the protection of those design basis functions. Are there any questions? This slide outlines how we're proceeding with our review. Presently, the risk-informed Part 50 core team has been tasked with developing guidelines for the review of the South Texas exemption. We will be complete with that task within the next few weeks. Following on from that, we're going to develop acceptance criteria for the review of option two treatment. I anticipate that this will be an evolution from the South Texas guidance and will be the basis for our review of the generic industry guidance. This means that the staff needs to have a good understanding of how well the South Texas and NEI proposals conform to one another. We've had a presentation both by South Texas and by NEI addressing this topic. There seems to be, at least on first blush, a good level of agreement between the proposals. South Texas clearly did not apply the NEI guideline. It didn't exist at that time. But they have participated with NEI in development of their guidance and largely they conform to one another in terms of their processes. This last slide provides the risk-informed Part 50 option two schedule, as it stands at this point. I will point out this is a tentative schedule that we recently set up for management review, and some of the assumptions that are inherent here are still being refined. If you'll note, for example, we say that in January 2001, the pilot program would be initiated. However, in our meeting with NEI on Tuesday, we had some indication that pilot activities might begin much sooner than that, perhaps this September. So some of these schedule assumptions and the overall timeline here is subject to change as we proceed. At this point in time, however, we're projecting that we would complete the final rulemaking to the Commission by the end of 2002. DR. APOSTOLAKIS: Do you know who the pilot plants will be? MR. WILLIAMS: We don't have specific information at this point. We know that the first out-of-the-box will probably be some boiling water reactors. I could speculate, but that would be all it would be at this point. DR. APOSTOLAKIS: And where does the South Texas exemption request fit into this? MR. WILLIAMS: The South Texas review, right now, we're anticipating that that will be completely by April of next year. I'll point out we've characterized the South Texas review as a prototype versus a pilot, basically as a demonstration of the concept. Since they don't conform or did not develop their process in full conformance with the NEI guidance and hopefully what will ultimately go into the regulatory guidance, we couldn't really call them a pilot. They do play a significant, but distinct role in the effort. I believe that concludes our presentation. DR. APOSTOLAKIS: This concludes the whole thing? You want us to brief the Commission in September on what you've got here? MR. WILLIAMS: I guess so. DR. APOSTOLAKIS: Fine. Any comments or questions from the members? You will expect a letter from us at some point? MR. BERGMAN: We considered this informational briefing. Of course, if you want us to provide something, we're certainly happy to get it. We are coming back to you, though, at the -- I think it's the September meeting, but it's August 30th, maybe. DR. APOSTOLAKIS: Okay. MR. BERGMAN: Right. And at that point, we'll have a full Commission paper for you that addresses the ANPR comments and other related issues and we'll certainly want a letter from you at that point. DR. APOSTOLAKIS: In September. MR. BERGMAN: Right. I think our Commission briefing is going to be the week of the 18th of September. DR. APOSTOLAKIS: I don't think there's time to have another subcommittee meeting. August is a month. But we don't need a subcommittee meeting before the presentation to the full committee. MR. BERGMAN: I don't think so. I mean, that's up to you, but I don't think so. DR. APOSTOLAKIS: Okay. Any other comments from members of the public, NRC staff? Thank you very much. Very informative. MR. WILLIAMS: Thank you, sir. DR. APOSTOLAKIS: Appreciate it. Now, the next item on the agenda is Mr. Christie's presentation, which is scheduled to start at 10:30, and I'm advised that I have to stick to that schedule. So we will recess for 50 minutes. [Recess.] DR. APOSTOLAKIS: We are back in session. Mr. Bob Christie, the floor is yours. MR. CHRISTIE: My name is Bob Christie. I am the owner of a firm in Knoxville, Tennessee, called Performance Technology. I have been in the commercial electric power business, nuclear, for about 26 and a half years or so. The first 15 and a half years was as an employee of the Tennessee Valley Authority. The last 11 years, I've been a consultant, not only for the nuclear business, but also for other places, such as the railroads and that, basically doing risk and reliability evaluation. I am here today to talk about a petition for rulemaking which was filed last year and to describe it and hopefully answer your questions about it. I have been asked and requested by a staff member of the Nuclear Energy Institute to clarify completely that the views that I express today are not endorsed by the Nuclear Energy Institute. So with that, I'd like to start. DR. KRESS: Is there any particular reason that they haven't endorsed this? MR. CHRISTIE: I have no idea. You'd have to ask them. This petition is a petition that does not come from the Nuclear Energy Institute and my views and the views that I express today, which I explained to you the last time when I talked to you, and I think it was March 1st, whatever it was, are the views of myself and a bunch of other people who have been following this and working this area for many years. They asked me to make that statement, I made the statement. The agenda today, I'd like to start by talking about a letter I sent to Dr. Tom King on May the 30th. I think I'll skip the introduction and background. We'll go back to it if we have time at the end. So really, I'd kind of like to do A, C, D, E, F, and then go back and do B, if we have time, and then summarize at the end. DR. KRESS: What is the status of the petition? Has it been -- MR. CHRISTIE: We'll talk about that. DR. KRESS: You're going to talk about that. MR. CHRISTIE: We'll talk about that in full detail. DR. APOSTOLAKIS: I wouldn't skip the background completely, though. I think you should -- DR. KRESS: You're going to talk about SONGS and the background. MR. CHRISTIE: Okay. We'll go over the background then, quick. The thing that's really amazing to me is I believe that we're very, very close to having a rule that's acceptable to everyone. We've had many meetings. I guess the first public workshop was back last September. Then we had another public workshop in February, and then we had a -- I think it was called a public meeting, not a public workshop, we had a public meeting in May. We've also had other interactions and so on. But at the May meeting, there were six criteria that the Nuclear Regulatory Commission people were using to make decisions with respect to 10 CFR 50.44. And so in the meeting on May the 17th, I think we got pretty good agreement and pretty good definition of where we didn't have agreement. So I'd like to tell you that today. I think we're pretty close. And let's go over what we discussed on May the 17th. The first one has to do with the hydrogen monitoring, measuring the hydrogen monitoring concentration. That's their slide 23. And I think we have come to agreement that the hydrogen monitoring system can be commercial grade. It no longer has to be safety grade, safety-related, with all the bells and whistles. That was my interpretation of what happened. Where we disagreed is, or I think we disagreed, that the staff of the Nuclear Regulatory Commission still believes that there should be requirements, quote, for the long term for hydrogen monitoring and while these requirements would allow you to be commercial grade, there are still going to be requirements and you will still have to have inspection and so on. My position and the position of the others that I've worked with on this is basically this hydrogen monitoring is not safety-significant, it is not a primary indicator that is used for anything but to turn on the hydrogen control systems, which are the recombiners and the purge systems, and if we make the recombiners and the purge systems non-safety-related and don't have requirements, then fine, the hydrogen monitoring is also not safety-significant and it would be put in a non-safety-significant category, turned over to the utilities, and the utilities would be responsible for it. Not that they're going to immediately go out and dump or anything like that. They are just the ones that are now making the decisions. DR. KRESS: When you say not safety-significant, that means it has insignificant impact on CDF and LERF. Is that what you mean by that? MR. CHRISTIE: Insignificant impact on anything. DR. APOSTOLAKIS: Have you subjected these to the Ross' adoption two that he is proposing? MR. CHRISTIE: In option two, all this stuff isn't even in the PRA, doesn't get -- this is non-safety -- if you did an option two on this, this is all non-safety-significant. It's below the line. DR. BONACA: Bob, this is the question I have. I remember when we did review, first of all, an application by the Westinghouse Owner's Group for elimination of the PASS system or of certain portions of the PASS system. I remember that they put the burden on the hydrogen monitoring system to perform some function in support of the emergency actions level, because -- MR. CHRISTIE: Severe accident management guidelines. DR. BONACA: So I don't remember the exact details, but I remember that they committed to maintain the hydrogen monitoring for a specific function in the severe accident management guidelines. Now, when San Onofre came with the proposal to eliminate the hydrogen monitoring system, they took it back at some point, because they said that this had to go with the same option on severe accident management provided by Westinghouse and for that, they needed a monitoring system. Could you address that point? MR. CHRISTIE: Okay. Let me explain what happened with the severe accident -- well, let me explain what happened with San Onofre. San Onofre went in with a 50.12 exemption request, which you people finally approved, but in the course of negotiations with that, San Onofre basically withdrew the application for the hydrogen monitoring to be declared non-safety. They put it on hold. They didn't say it wasn't non-safety. They didn't say it was non-safety. They just put it on hold and they received the approval from the staff of the Nuclear Regulatory Commission for the recombiners and the purge to be non-safety and to pass onto the purview of the utility only. Their position still was it was not safety-significant, it was not safety-related, and they just put it on hold. The Westinghouse Owner's Group, and I haven't been following it for PASS very much, but to the best of my knowledge, what the Westinghouse Owner's Group have said, and I believe there was a change in the middle of their submittal. They started out with the hydrogen monitoring as being safety-related and they changed it to be non-safety-related. It's going to serve a function in the severe accident management guidelines which are not safety-related pieces of equipment and you don't have to have safety-related pieces of equipment to do it. DR. BONACA: I understand. I'm only saying that I know the reason why they left it in, however, they did not ask for exemption on that system, was because -- I asked the question specifically and they answered that they, yes, would want to go with the WOG for the SAMG and because of that, they withdrew the monitoring system from the exemption. So I think there was a logic behind the reason why it stayed there yet and I want you to keep it in mind as you go forth with this. MR. CHRISTIE: And I haven't followed the Westinghouse Owner's Group post-accident sampling system that closely, but I have been told that the post-accident sampling system WOG submittal changed the requirement for the hydrogen monitorings in the middle of the application to move the hydrogen monitoring from safety-related to non-safety-related and that's where it is today. It was approved on the basis of non-safety. MR. SNODDERLY: Excuse me. Dr. Bonaca, my name is Mike Snodderly. I did the San Onofre hydrogen exemption and I also did the review for the Westinghouse Owner's Group PASS sampling. From my perspective, I believe that if you look at the exemption request for the PASS, it did credit the continuous hydrogen monitors, the safety-related continuous hydrogen monitors, as the way to measure hydrogen and to support core damage assessment, the Westinghouse Owner's core damage assessment guidelines. Also, if you look at Regulation 50.47(b)(9) for an effective emergency plan, most licensees meet that with Reg Guide 1.101, Revision 3, which endorsed an NEI guideline. But that states that a general emergency is a loss of any two barriers and potential loss of a third barrier. Potential loss of a third barrier includes whether an explosive mixture exists inside containment and most licensees use the hydrogen monitors for that determination. So just to support and refresh what had happened. DR. BONACA: Okay. I didn't have enough detail, and you have provided that. Okay. But I remember there was a connection. It wasn't in this part of the rule. It was, however, in the severe accident maintenance. Now, whether your requirements may be to implement severe accident management steps and if they are part of the law, I don't know. MR. SNODDERLY: The severe accident management guidelines are a voluntary initiative and as Mr. Christie points out, yes, we agree that the hydrogen monitors are not needed to actuate any of the mitigative features in 50.44, the hydrogen recombiners, the igniters, those types of things. So that it could be eliminated from 50.44, the requirement of hydrogen monitoring. But there still is a requirement in NUREG-0737, the post-TMI requirements, that you have hydrogen monitoring, and, also, in how licensees have met this 50.47 requirement through the appropriate reg guides. So as we go through this process, we've got to keep that in mind and we'll have to go back and reconsider those types of situations, but I believe the staff still believes that hydrogen monitoring, although you may not have to do it with safety-related continuous monitors that are currently installed, they still would be needed for the core damage assessment and EP. DR. BONACA: Thank you. MR. CHRISTIE: And I think that's a pretty clear description of where the differences are between the position we have and what the staff has. DR. APOSTOLAKIS: Again, I asked you about option two and you said it's below. How could it be below? If there is a box there that says non-safety-related, non-safety-significant, so this must belong there. MR. CHRISTIE: In the option two, if you didn't change the rule, this would go into RISC-3. DR. APOSTOLAKIS: Which is? MR. CHRISTIE: Non-safety-significant, but safety-related. DR. APOSTOLAKIS: Okay. DR. KRESS: Because it's already designated. MR. CHRISTIE: It's already. But if the rule, the petition for rulemaking is approved, that's the whole purpose, first -- and, see, we're doing box two and -- we're doing option two and three all at once with hydrogen. Okay. We're moving it from safety-related, box one, to box four, by this rulemaking. DR. APOSTOLAKIS: Right. MR. CHRISTIE: So, again, that's the situation as far as measuring the hydrogen concentration. MR. SNODDERLY: Excuse me. Dr. Apostolakis, I just wanted to address one thing on your point. If you just use option two and the criteria of CDF and LERF, yes, I agree with Bob Christie, it's going to be in the group three. One thing I think we all need to keep in mind, though, is those other regulations, such as 50.47 and EP, aren't really addressed as one of the criteria for determining whether something is in box one or box three. So I think that that's -- and sometimes in the expert panel, that's considered and sometimes it's not. Currently, the way the hydrogen monitors and all the monitors that are required safety meet Appendix E, the emergency response data system. All those things currently would be in the low safety, no risk significance. I think that's one issue that we're -- I don't think we're disagreeing on the categorization, but how do we treat that, because there is this acknowledgement that it is needed for those things. DR. KRESS: Would that be considered a defense-in-depth requirement? MR. SNODDERLY: I believe so, Dr. Kress, but I think what we're -- the issue, I think, that's being brought up here is that, yes, if you just apply the criteria of CDF and LEF, yes, this stuff clearly is not needed to meet that, falls out, but do we maybe consider another criteria or are we adequately addressing it through the special -- you know, how is it going to be treated. DR. APOSTOLAKIS: Let's go on. MR. CHRISTIE: Okay. For the next, which is the mixed containment atmospheres, again, as far as we're concerned, basically, mixing the containment atmosphere comes because you have systems for containment heat removal, either fans and coolers or sprays, some plants have both, and these systems we're not proposing any change at all in. We're still going to be mixing atmosphere. The proposed rule doesn't change that. Nothing changes that, and we're in complete agreement on that. The next one is the control of the post-LOCA combustible gases. The Nuclear Regulatory Commission, on their slide 25, says remove post-LOCA hydrogen control from 50.44. Complete agreement, we remove it from 50.44. On the reactor coolant system high point vents, again, complete agreement. DR. APOSTOLAKIS: Either you don't use them at all or -- MR. CHRISTIE: All right. I guess this is slide 26, which is the reactor coolant system high point vents. Again, we got complete agreement. We're going to keep all the stuff in for the reactor coolant system high point vents. On the slide 27, which is the inert, the MARK 1's and MARK 2 containments, again, complete agreement. We're not going to change -- no change. We're going to keep them inert. The last one, and we should spend a little time on this one -- DR. KRESS: Where will I find that slide? MR. CHRISTIE: Slide 28. MR. MARKLEY: Page four. DR. KRESS: Page four. MR. CHRISTIE: Page four. It comes as an attachment -- DR. KRESS: No, I mean the real slide that he's talking about. MR. CHRISTIE: That comes as the attachment to the May 30th letter from me to Dr. King. DR. KRESS: Okay. I've got that here somewhere. DR. APOSTOLAKIS: No, no, no. He means this. I think what Dr. Kress means is the actual slide 28. MR. CHRISTIE: Right. It comes in the attachment. You got a letter from myself to Dr. King dated May the 30th and you have all the slides that we're talking about here. DR. APOSTOLAKIS: Right. MR. CHRISTIE: Okay. This is a very -- again, it's a tough slide to read. We had a very difficult time with it on May the 17th. It's unclear. And really, as far as we could tell, it busts into two pieces. One had a requirement for all plants, and that's to demonstrate the containment will withstand both short and long term a specified source, so and so, and then they had one for MARK 3's and the ice condensers, which is do something with the MARK 3's and the ice condensers for the igniters during station blackout. What we have said, and we'll talk about this in a minute, is for the large dry containments, we have a requirement now, we've added a requirement to address the containment capability during severe accidents. And I don't know whether the staff of the Nuclear Regulatory Commission agrees with that one. It's very unclear what this one was. The a second part, and this is the one that I think we need a lot of discussion on, because it got brought up again in the Commissioners' briefing last week. The staff of the Nuclear Regulatory Commission believed that the igniters should be operable during station blackout. Myself and the others do not understand this requirement at all, so let me explain what we see for station blackout and the igniters for the MARK 3's and the ice condensers, which are the ones that have the igniters. If you have a station blackout, you have a loss of all AC power, it becomes basically a timing problem. And let's take the boilers first. Let's take the MARK 3 boilers. If you lose all the AC power and your offset power is lost and your emergency diesel generators are lost, what you boil down to is you will have the turbine-driven reactor core isolation cooling. That will supply water to the reactor vessel. You will also have the control systems that are DC powered coming off the batteries and you will also have the control systems that are AC powered that come off the inverters, which are powered by the batteries. This is what you're down to. At this point in time, and this will go on, depending on whether it's a four-hour plant in battery life or an eight-hour plant in battery life, it just depends on how much battery you have, you will basically not see any core damage, we think. I mean, basically, your RCSI, reactor core isolation cooling system is supplying water, the core is covered. You're not removing any heat from the containment, which is kind of worrisome, but you can control it, and, if that works, everything is fine. DR. KRESS: You're saying the contribution of that station blackout sequence to core damage frequency is very low. MR. CHRISTIE: That's an automatic, yes. To get to the point where you have a loss of off-site power and the failure of all your emergency systems on-site. DR. KRESS: You've got to have a lot of things happen. So that particular sequence -- MR. CHRISTIE: Right, that sequence -- DR. KRESS: -- for those plants don't add much to core damage frequency. MR. CHRISTIE: To get to that point, we're talking a very low number on initiating events, plus those subsequent -- you know, the initiating event is loss of off-site power and the subsequent event is a loss of all your emergencies. That's a very low probability event. On top of that now, you have the situation where you're keeping the core cooled with the RCSI and you have control because you've got your DC and your AC power systems. And this goes on for approximately four hours. At the end of four hours, and let us ask yourself what role do the igniters play during this time. Why would you want to have them operable? DR. KRESS: If they were operable, they might preclude -- MR. CHRISTIE: First off, how would you put them operable? They require electrical power. DR. KRESS: That's another question. But if you did go into core damage, which you say -- MR. CHRISTIE: No, we're not into core damage yet. DR. KRESS: Then they have no purpose. MR. CHRISTIE: Right, exactly. DR. KRESS: But if they did go into core damage, then they may preclude an early failure of the containment. MR. CHRISTIE: Wait a minute. Wait a minute. Now we're going down -- and now we're four hours into it. Let's say we have a four-hour plant, and the batteries fail. When the batteries fail, you're no longer going to be able to control RCSI, even though at Browns Ferry we thought we could jam the steam emission valve wide open and just take our chances. But say we lose now the ability to have RCSI. Okay. The core -- the inventory in the core is going to start to deplete, depending on what codes you believe for -- you know, whether you use MAPP or MELCOR. Within a couple of hours, the core is going to be melted and probably going through the reactor vessel and laying on the floor. But if you don't have -- and during this period of time, if you were able to restore anything, electric power, what you -- well, in the first part, the first four hours, if you got electric power back, what would you restore? Well, the first thing you would restore is you'd take -- you'd stop using the batteries and you'd try and get your AC and DC systems on control systems working off your emergency power systems or off-site. If you've got off-site back, everything comes on and you don't even worry about it. But if you've got just one diesel, the first thing you're going to restore is the AC/DC. The second thing you're going to do is you're going to start going down to probably the suppression pool and start cooling the suppression pool, because HPSI and RCSI, if you don't -- well, you don't have HPSI. You've got RCSI. RCSI can't cool -- can't use the water from the suppression pool forever because it's too hot. So if you get an electrical diesel generator back in the first four hours, you're first going to make sure your batteries are taken care of and then you're going to go and start suppression pool cooling, because your HPSI is keeping you alive and you're not worried about it, and now you want to start removing heat from the containment. So that period. The next period of time is where the core is melting and when the core is melting, again, the first thing you're going to do, if it's after four hours and your batteries are gone, but then you get an emergency diesel generator back, you're going to go first for the control systems. You're going to want to know where you are, what's the status of my core, is it completely melted, is it on the floor or et cetera, et cetera. So that's the first thing. The second thing you're going to have, and it's going to be a hodgepodge with boiling water reactors. What you're probably going to want to do is even if you melt it, you're probably going to want to put water on the core. Probably. That's probably something that you've really got to consider. The other thing is you're going to have to start removing heat from the containment and you're going to either do it RHR through the suppression pool or sprays or whatever method you can do, you're going to do it. Again, we don't see that igniters -- having operable igniters during this period of time is the thing that would cause is -- we wouldn't take our emergency diesel generator and automatically start taking some of that power off to power up the igniters. It doesn't seem logical to us. DR. KRESS: If the rule says you need to have power to the igniters, all this other stuff is what ifs. But the question you're asking is should you have power to the igniters, should that be part of the rule. And if I were to ask why would I want power to those igniters, it's if my CDF is low enough in the sequence already, do I need a conditional containment failure probability that's very, very low, because I've already got the CDF low enough. And does having power to the hydrogen monitors do anything to my conditional containment failure probability? I don't know that it does or not. I would suspect it does. MR. CHRISTIE: I think, Tom, I'm not explaining myself well. You have X amount of dollars. With station blackout, you've got X amount of dollars. The staff is telling us they want to spend some additional money on station blackout to have the igniters operable during station blackout. It makes no sense to us. If I am spending money on having an additional source of power at the plant during station blackout, other than my off-site power and my emergency diesel generators, I'm going to have it powering my AC and DC control and my containment heat removal systems. The last thing on my mind, because I want to prevent core damage, and so my money is not going to go to igniters. Igniters don't do anything in the first four hours. Igniters probably don't do anything in the next four hours. Igniters may not even do anything in the long run. We don't see why anyone would spend money on igniters, on AC power sources for igniters, when the money would be better spent somewhere else, either preventing core damage or taking care of the heat removal. Ultimately, supposing you could never get heat removal back, you're failing. You're going to fail the containment. I mean, it's just if you don't have heat removal, you're going to fail the containment. I don't care if you've got igniters powered forever. So this is the -- in our mind, the most -- and that's why we wrote the rule the way we did or we proposed the rule the way we do. DR. KRESS: I wasn't looking at it as an either/or. I was looking at it as you need containment heat removal and maybe you need igniters, also. MR. CHRISTIE: Right. But from my standpoint, from a technical standpoint, if we are going to go and provide additional electrical power sources other than what we've already got, we're not going to put it on igniters. If it were cost-beneficial, we'd probably put it on AC and DC control systems, we'd probably put it on heat removal, on boilers, because we want to keep water in the core, and so on and so forth. Go over to a pressurized water reactor. Again, the same thing. You lose the off-site power, you lose your emergency AC. What are you down to? Well, you're down to the turbine-driven emergency feedwater systems and, once again, the DC and the AC systems. On a pressurized water reactor, this is going to go on for X amount of time. All right. Now, depending on what model you use for the reactor coolant pump seal LOCA, you're going to have a LOCA at the same time because your steam generators are cooling, but you're also losing water inventory off of the primary system and you're not replenishing it, because you have no AC power and there's no turbine-driven pump that supplies water to the reactor pressure vessel. So for the first four hours, again, you're in a situation where you're cooling with the steam generators and your AC and DC power control systems are working fine. You know what's going on, so on and so forth. Once again, if we restored anything during that period of time, we wouldn't be worrying about the igniters. We're going to go -- if we get something back, the first thing we're probably going to go do is start safety injection, the higher pressure safety injection pumps and restore the inventory we lost off of the reactor coolant pump seal. So that's -- again, it's this priority of things that doesn't make any sense to us. If I were going to have an additional power source on the plant, I'm not going to put it on igniters. It makes no sense to us. The same goes when you start to melt the core. Now, when you start to melt the core and if you get the core on the floor, et cetera, et cetera, you're, again, going to be faced with a choice, hey, do I turn on containment heat removal sources or do I try and get safety injection working so that I pump water on wherever it is, dropped in the sump at the bottom of the reactor vessel, wherever it is. But you're not going to worry about the igniters. Once again, it's containment heat removal. So from a technical standpoint, we don't understand the staff's belief that having an additional power source to make igniters operable at MARK 3's and ice condensers -- why? And what we've said to them is, if you've got a problem, identify it. We have still not seen any technical justification or even know what exactly the problem is. And then if you've got the problem, go put it in the context of the 51.09, the backfit, because then we'll all understand it, we'll understand what the problem is, we'll understand what the alternatives you're proposing are, and we'll understand what the cost-benefits are. So that's where we are on that. Okay. Now, if you want to talk background. Again, this work comes out of the Arkansas Task Zero and the San Onofre Task Zero, and the objective of those pilot programs, we wanted to have a more objective and efficient way of doing business. This is what we talked about before. We wanted to take the whole plant, we wanted to consider a whole plant package, which included cost generation and risk. DR. APOSTOLAKIS: What happened to the whole plant study, by the way? MR. CHRISTIE: It's gone. DR. APOSTOLAKIS: It's gone. MR. CHRISTIE: You can't get any -- it appears clear that the staff of the Nuclear Regulatory Commission are not interested and we can't interest them in the whole plant study, as far as I can tell. So what we're down to now is working off of specific pieces. Though I think that's a mistake and we'll talk about that a little bit later. And, again, the basis of what we said is that the primary responsibility for the public health and safety lies with the people who run the plant. Their regulatory process is only good for public health and safety. We've also got that the public health risk is different for each nuclear, and it changes with time. We have addressed, we thought, in the whole plant study, all the major problems that were placed in the Kemeny report, which is the President's report on Three Mile Island, and the best definition I saw of them was Dr. Thomas Pickford's separate opinion and we thought the whole plant addressed every one of those items. We also believe that we don't have that much time to do all this stuff. We've got to get more efficient and effective, because basically, in an economically deregulated power industry, if you're not effective and efficient, you're not going to be a producer of power very long. So the more effective and efficient we go, we believe the best definition of what's going to happen to the nuclear power plants in the future is contained in this paper by Mr. Shiffer, who is a retired executive out at Pacific Gas & Electric, someone who has lived through it on Diablo Canyon and, in this paper, told us what the future would look like. Now, just a quick -- and we've been all through this before with you. If you look at the San Onofre evaluation report, again, this we've known for 20 years. The overall public health risk is dominated by severe accidents where the core is damaged and containment is bypassed or breached. DR. KRESS: Before we take that one off, let me ask you a philosophical question about that. Should NRC just be concerned with the dominant accident sequences or should they be concerned with accidents that may be more frequent, but not as dominant in terms of risk, but have some consequences? That's the philosophical question. MR. CHRISTIE: Tom, you know, when I'm talking about the whole plant, you -- and I talked to you on March 1st. To my mind, what the nuclear industry should be shooting for is what I call the whole plant study. We should know from top to bottom what that plant represents with respect to public health risks. We should know it. I mean, we should know what the health effects are. We should know what I call source term. I look at it from the standpoint of a PRA. From the standpoint of a PRA, out of a level three, I get source terms to the people. We should know what those source terms are. We should -- they're in categories. We should know what the probabilities are, we should know how much tellurium and cesium. We should know that. We should have that information for every plant. To go back to the output of the level two, we should know what the plant damage states are and level one and level two in the containment event tree. We should know all of this material. We should know it all from top to bottom. We should know what the systems are responding, you know, that it's a high probability, we should know it all. When we get that information, we ought to make the -- the plant ought to be able to determine for themselves what level in there should we go to to make sure that we have provided the adequate protection of public health and safety. They're almost invariably going to go to lowest level possible. They're going to go down to system, because that's what they can control. But we have to know the whole picture, because I'm not smart enough to know the whole picture without doing the whole picture. So that's -- my philosophical thing is we ought to consider everything, every accident -- you know, risk assessment doesn't say, hey, I only consider the ones that have a frequency greater than so-and-so. They take every frequency. They don't consider just single failures. They consider every failure possible. They don't consider, hey, we can't have this plant damage. They consider all plant damage. That's the beauty of the risk assessment. So that's what I'm -- DR. KRESS: I agree with you. The question is what should the NRC concern itself with in terms of their activities and rules and so forth. MR. CHRISTIE: They start from the top. If they are absolutely convinced we meet the top, why do they go any further? If you meet the goals, why are you going any further? The rest of it belongs to the utility. See, I look at the thing, there's a certain amount of expertise in the Nuclear Regulatory Commission. They should focus on the things that they can do well and that they can add to value. The value that they can add is that they're nuclear. We don't have a coal regulatory commission. We have a Nuclear Regulatory Commission. If you wanted to protect against coal boiler explosions, would you create a coal regulatory -- no. You let the utility handle it. DR. KRESS: I guess where I may different a little from that is it seems to me like the goals of NRC are not just the two quantitative health objectives. They have goals that are different than those and they have to do with controlling things like worker exposure and controlling smaller releases of higher frequency. There are rules in there that deal with those kind of things, and I call those goals, also. And I don't know where NRC should -- MR. CHRISTIE: Do you believe -- you know, the nuclear business, in my time, has undergone tremendous revolution. It's phenomenal to me now today that you can have nuclear power plants with a total quantitative dose to all the workers at the plant is under 100 rem. I mean, we used to do 500 rem. Okay. Now, that happened in my lifetime and it didn't happen because the Nuclear Regulatory Commission dictated it. DR. KRESS: It was because the guys at the plant -- MR. CHRISTIE: The guys at the plant are not dumb people and they have -- and I'm not afraid to admit self-interest, man. We're going into a private enterprise system in which our survival is necessary. But the beauty of it all is, and as you heard me say this before, you can't have a safe plant that isn't a good economic plant. You can't have a good economic plant that isn't a safe plant. It doesn't work that way. You will do things right across the board. The guys in maintenance ops, engineering, et cetera, they don't -- no, it's a balance of plant, I'm going to do a shitty job, or it's a balance of plant, I'm going to do a great job, or it's safety-related, I'm going to do a great job, or it's safety-related, I'm going to do -- they don't do that. They do a good job, and so from my standpoint, there's an absolute role for the Nuclear Regulatory Commission. It's there. People have a fear of nuclear and they have created, through law, the Nuclear Regulatory Commission. I think you're a good thing. Okay. I mean, they haven't created a coal because people don't think that coal kills people to the same degree they do about nuclear. So the Nuclear Regulatory Commission has a role and they can fill it, great, but they ought to focus on the things that they can do the best and the part that's the best for them is the nuclear part. What's the health effects part? You know, we've ignored that for almost ten years now, as far as I can tell. DR. APOSTOLAKIS: So you would like every unit to have a level three PRA. MR. CHRISTIE: Absolutely. Absolutely. DR. APOSTOLAKIS: Well, let's see what -- MR. CHRISTIE: I don't know. I think we are derelict in our duty if we don't. When I was at the Tennessee Valley Authority, in charge of PRAs, every PRA we did was a level three, every one. DR. APOSTOLAKIS: As you know, the trend now is to work with LERF and CDF. MR. CHRISTIE: That's because you made it so. I didn't make it so and some of the people at the plants didn't make it so. Some of them still have level three and we still use them. Going back to -- again, the San Onofre safety evaluation report, it's not the design basis accidents, again. It's the severe accidents, et cetera. The stuff that we put in the plants for design basis accidents don't work in severe accidents. They're way under-sized, et cetera, et cetera. Paying attention to design basis accidents can be detrimental. In the case of San Onofre, it was the distraction of the operators because they're paying attention to a non-safety-significant piece of equipment, at the detriment of the safety-significant equipment, and we want that. So I guess the things we learned out of the San Onofre -- DR. APOSTOLAKIS: I guess I have to understand that a little better. MR. CHRISTIE: Sure. DR. APOSTOLAKIS: What is this distraction we're talking about? It seems that's a high level statement. They have to do an extra thing so that -- MR. CHRISTIE: If you have a design basis set of things and you do it, one of the things is that you have as part of the design basis, all the 50.44, et cetera, et cetera, and you will have to have hydrogen monitoring in place at X amount of time after an accident. If you follow the NUREG-0737, it was 30 minutes, but now, because of Arkansas Task Zero, a lot of the plants have moved to 90 minutes, et cetera. But some of the plants still have the 30-minute requirement. What it means is that one of the reactor operators in the horseshoe, and, generally, if you're at minimal crew, there's only two, at some point in the accident, as dictated by the rules, has to pick himself up off out of the main boards and go set up the secondary boards, because almost all these things are not on the main boards. It's not something -- you have to go to back boards and hook things -- you've got to coordinate. That's the other thing. You've got to start coordinating. Health physics has got to get in, chemistry has got to get in, ops has got to get in, maintenance has got -- it's a whole process thing. And you just pick somebody up out of the control room, paying attention to things like, you know, what are the thermocouples in the core telling us, what are the radiation monitors telling us. It goes away. DR. KRESS: Just to start up the hydrogen system. MR. CHRISTIE: Just to start up the hydrogen system, which we all agree doesn't work for anything that really counts. That's the distraction. DR. KRESS: And he could be using that time to do other things. MR. CHRISTIE: Right. Hey, monitoring what your thermocouple is telling you, what your heat removal, making sure your steam -- if you're a pressurized water reactor, making sure your steam generators are working correctly, watching your pressure/temperature curves, et cetera, et cetera. All these things that we put in since Three Mile Island, now we're pulling a guy up and away he goes. And it's not that one single operator can't do it, they can do it. But can they do it as well as two? Probably not. And when they get near drills and that, you know, they get -- well, anyway, it's just not right and we've -- you know, and I think the staff of the Nuclear Regulatory Commission agreed. They approved the Arkansas, they approved the San Onofre, and we're moving towards rulemaking and we're getting agreement on a lot of things. So anyway, my read is the important things that we've got to pay attention to with severe accidents, we've got to focus on containment integrity. The existing recombiners and purge don't work and the existing procedures. And this is what we got out of San Onofre. Following the February meeting with the Nuclear Regulatory Commission, it became clear to us that it wasn't -- the rulemaking was not going to be something that happened tomorrow. So what we did is we started an effort in the industry to go out and now we're going to produce exemption requests similar to the San Onofre exemption request and file it under the 50.12 process. So this is the results from the first six plants that we looked at under this process. So what we're interested in here is what are the action levels, what's your design pressure and your failure pressure, what kind of system do you use in the plant, and when do you use them and how permanent are they and all the rest of that kind of stuff. What came out of this study is that, man, there's a wide variation in the implementation. I mean, we were working off of San Onofre, which is basically recombiners with a backup purge that was never used. And then we found that at the plants that the recombiners were either off-site or on a warehouse on-site or somewhere. So if you wanted to do something with hydrogen in the short term, you were going to have to use the purge. Also, believe it or not, they have systems called repressurization systems, because if you do design basis work, what happens is you don't reach the action levels in hydrogen until days into the accidents and by that time, you have cooled the containment to such a point that you don't have any driving point, any driving force to move the stuff out, the hydrogen out of the containment. So you have a repressurization system. This is caused by design basis accident analysis. So theoretically, you've got a system in a plant, if you're using purge, the way you have to pump up the containment to move out the hydrogen. So we saw this and the recombiners, some of them were off-site, some of them were on-site, et cetera, et cetera. What we saw is the use of the pressurization purge and the moveable recombiners, that's dangerous. It's dangerous to both the workers on-site and the people off-site. If we have a severe accident and somebody opens up a purge valve, the calculations on the inner system LOCA and the steam generator tube ruptures are going to be dwarfed in comparison. That's just a fact of life. DR. KRESS: How big are those valves? MR. CHRISTIE: Okay. The six we got, we got a six-inch, a four-inch and 48-inch butterfly. DR. KRESS: Pretty good size lines, aren't they? MR. CHRISTIE: Well, 48-inch butterfly is fairly big. They're using the normal purge, but they got a stop in it at 15 degrees instead of 90, which they -- then the other thing is, what are these things going to do? Suppose you really did, in a severe accident, open up a purge valve. You're going to saturate the HEPA filters without even blinking. They're going to be ineffective to beat the band, and who knows what's going to happen to the valves. The junk that's flowing through those valve paths are is just phenomenal. I wouldn't -- again, I think everybody agrees it's not a good thing to do. No one would recommend opening a purge valve in a severe accident. Simple thing. No one would recommend hooking up a recombiner. We got recombiners that are stored off-site or stored on-site, where you have a pad right next to the reactor building containment, and theoretically, you're going to go hook up a recombiner in a severe accident. Now, the doses at Three Mile Island were measured outside the reactor building, if I remember correctly, were measured in the tens to hundred R per hour range. You believe the workers are going to go out there and hook anything up? It's not going to happen, and we know that. You see, it's not something we don't know. The other thing we checked out is that on all the large -- these are all large drives. What's the containment capability? Can you stand the burn without the recombiners, without the purge, without the monitors, without anything? Fine. As far as we can tell, everybody has got more than enough capability to withstand a burn. Three Mile Island did. It was designed for 50 psi gauge, it probably had an ultimate of about 150. It stood 28 without even blinking. And the hydrogen production rate at Three Mile Island was probably about as great as you can get, when you think of the zirc water we did. We drained it, we threw the water by it as we were draining it, then we filled it again, we drained and we threw the water by it again, then we drained it and threw the water by it. I mean, if you were talking about how much hydrogen can you get out of the zirc water, you'd run a Three Mile Island accident. So we had, what, 45 percent zirc water, somewhere in that neighborhood, and about an eight percent containment volume and it burned and got a 28 gauge. The large drives, they can stand it. It's not -- okay. Here's a personal belief of mine. I just don't like going to the plants and having the plant people understand that they're writing procedures for design basis, where the reality is they're not going to follow them, and if they did follow them, it was going to really be hurting people. Now, these are not dumb people. Okay. But their problem, again, is the application of probabilistic risk assessment at plants across the United States is not uniform. Not everybody does use PRA across the board. So you've got to be careful. Some plants are still living in design basis space to a much greater degree than others. So my problem is I don't like knowing that we got in the plants the potential for problem, knowing what the solution is, and we can't take immediate action. Now, the plants are going to turn in their exemption requests, because they don't believe the rulemaking is going to be anything that's going to be fast, and the exemption request will take care of it, because we'll get rid of the recombiners and the purgers and the hydrogen monitors, et cetera, and you'll never have to worry about it. It will become blanked off for purges, et cetera. But that takes a while. I mean, you know, we're thinking maybe about two in the month of July and a couple more and so on. And it just seems weird. Now, we've had some discussions. Mr. Mike Snodderly just told me today there may be another way. Maybe we can go in and change the emergency operating procedures generically type of thing. I suggested in my letter of May 30th an information notice to let people know that they shouldn't be doing things like opening purge valves and moving recombiners in a severe accident. So I think we're going to do something. It's just, myself personally, it takes a while. It just flat takes a while. Why do we have a system where we have a known problem with a known solution and it takes the bureaucracy an awful inordinate amount of time to get the solution solved? All right. Let's go. We're running out of time, but we'll see if we can get through. Let's go the proposed rulemaking. The first thing we wanted to do was change Appendix A, Part 50, Appendix A, criterion 41. If you look at criterion 41, what it says is you'll have a hydrogen control system and what it's going to do, it's going to reduce the concentration and quality of fission products released in the environment by postulating accidents and the control of the concentration of hydrogen or oxygen to the other substances in a containment atmosphere following postulated accidents. Okay. So this is what it says. It says we're going to reduce the concentrations that can get out and we're going to control the hydrogen in. All right. And it's going to be postulated accidents. Okay. Well, we know that the postulated accidents don't work. So when we were looking at a GDC and you were asking how are you going to risk-inform the GDC. Okay. The first thing we're going to do is we thought we'd take out the stuff about the postulated accidents. Providing things to reduce the concentration of fission products and to control hydrogen and oxygen for postulated accidents, that doesn't work in risk-informed space. So what we said is, okay, let's write it this way. We're going to have systems to control fission products, hydrogen, et cetera, et cetera, to assure that the reactor containment is maintained for accidents in which there is a high probability for fission products to be present. We moved it out of design basis space into severe accident space. Now, in some of the comments to the rulemaking, they said, well, you're going to have to define high probability. I agree, we're going to have to define high probability, but the reality is we already got the tool to do that. We got the risk assessments and every plant in the United States can define the sequences with high probability. Now, where they cut it off or where everybody cuts it off, we'll argue about that till the day I die. But the reality is the plants have the tool, and so we've got this -- we re-created a GDC or a criterion to move it from design to severe accident space. We think we've done a good job. So we knew we had to change the GDC, because if we had changed 50.44, but we didn't change the GDC, then we're still in the postulated accident space, et cetera, et cetera, et cetera. It wouldn't work. So we've got to change the GDC, if you want to go to severe accident space. Inerted containment, no problem. MARK 3 and ice condensers have to have the igniters, no problem. Leave it the same. We added one and this one is causing a lot of consternation. We thought it was pretty simple. What we said was if you're a large dry and you're depending on the containment capability, which is what you're doing now, all the plants in the United States with large drys depend upon containment capability to stand the burn. If that's what's going on, why the regulation that says withstand a burn? Okay. Regulation could correspond to risk. If the risk is the containment is going to withstand it or not, then write your regulation that says it. So that we wrote it, say, and we put in words and, man, these words caused trouble, based upon realistic calculations, can withstand, without any hydrogen control system, hydrogen burn for accidents, again, with a high probability of existing -- of causing severe -- you know. But to me, this is the words that you got to have in severe accident space. Now, the beauty of this is every large dry in the United States has already done this. We've all done mid-core. We all went back after Three Mile Island and all of us looked at it, the NRC looked at it, the industry looked at, and all of use evaluated the ability of the large drys to withstand the burns. We all evaluated it in the 51.09 backfit space. We all said they can withstand the burns, we don't need the igniters in the large dry. So what this proposed rule says, let's make our knowledge that came out of Three Mile Island and all this stuff from mid-core and so on and so forth, let's make it correspond to what we're doing today. We are basing it on the ability of large drys to withstand the burns. Write the regulation to say large drys withstand the burn. DR. KRESS: When you talk about hydrogen control system, does that include something to be sure the hydrogen is mixed and doesn't reach -- MR. CHRISTIE: Yes. That's the other part. We'll go back in the 50.44. We didn't change any of that mixing. DR. KRESS: That's still part of it. MR. CHRISTIE: Right. This is a whole package. DR. KRESS: Then why did you choose 75 percent? MR. CHRISTIE: Because that's what is in there, that's the igniter. DR. KRESS: That's already in there. MR. CHRISTIE: If you had a large dry that couldn't withstand a burn, then you go through the backfit process to see whether you put the igniters in. If the backfit process says put the igniters in, you put the igniters in. That's all this proposal will say. DR. KRESS: Okay. MR. CHRISTIE: It's, we think, pretty simple and it says, hey, this is what we're doing out there in the world, write the regulation to say what you're doing in the world. We wanted to make the regulations and the risk comparable. I mean, that's the whole purpose of risk-informed performance-based regulation. And we think we've done it. That's what we think. And this is just the same thing for the high point vents. DR. KRESS: Let me ask about large drys. Clearly, in my mind, they can withstand the hydrogen burn, but a hydrogen burn with igniters spreads the pressure out over a long period. MR. CHRISTIE: I'm not sure that that's true. DR. KRESS: It burns the hydrogen as it goes in, supposedly. MR. CHRISTIE: It depends on when they're turned on and how fast the stuff goes. DR. KRESS: Yes, but that's the general idea. Whereas if you didn't have igniters in those, then you could build up the hydrogen to a pretty high level. They're designed so you don't get up to detonation levels, generally, but you can get up to a pretty high level and then ignite them by some ignition source and you get a much different kind of pressure spike. MR. CHRISTIE: Three Mile Island is proof of that. DR. KRESS: Yes. Would there be any difference in the nature -- I'm also assuming if you've got that much hydrogen in there, you're in a severe accident and you also have fission products present at the same time. MR. CHRISTIE: Sure. DR. KRESS: Would there be much difference in the release of fission products and their subsequent consequences between those two scenarios? MR. CHRISTIE: No, because in order to get the fission product releases, if you got that point, just like we did at Three Mile Island, you're going to have to have heat removal. Containment heat removal has got to fail. Whether you burn it fast, slow, it doesn't matter. What's going to happen is the containment is going to stand the burn, whether it's slow or fast doesn't matter. DR. KRESS: I'm assuming that you still have leakage, normal leakage path, and -- MR. CHRISTIE: But the amount of material that's getting out in the normal leakage path is so infinitesimal and such a small -- I mean, what -- we're not going to -- DR. KRESS: You say it's not worth worrying about that difference, is what you're saying. MR. CHRISTIE: Not worth worrying about. Not worth worrying about. Not that I can see, Tom. I mean, you know, we -- I think. And Lord only knows that we've spent gobs of money. DR. KRESS: Suppose the difference between those types of burns meant you -- in the one burn with the igniters, you were below 10 CFR 100, and the other burn, you were just a little above it. Is that something that -- MR. CHRISTIE: I'll tell you this, flat. If you're in severe accident space -- DR. KRESS: You don't worry about 10 CFR 100. MR. CHRISTIE: 10 CFR 100 is long gone, in my mind. If the tech support center is worried about 100, when we're in a severe accident, especially Three Mile Island, we -- DR. KRESS: They're focusing on the wrong thing. MR. CHRISTIE: Yes. We ought to start replacing that tech support center fast. I'm almost close to being on time. To summarize, I have and others have -- a sufficient knowledge exists to change the regulation of combustible gas control. We don't need anymore studies, we don't need anymore work. We don't - you know, we can spend hundreds of millions of dollars more if we want to, but we don't need to. We're done. I mean, enough is enough. I absolutely believe that if you're going to rewrite the regulations, you've got to focus them on the severe accident. I mean, that's just the way you've got to write the regulations. I know that's a major cultural change with the staff of the Nuclear Regulatory Commission and they are also included as a major regulatory change for the staffs at the nuclear power plants, but we've got to get through that. I mean, we've got to. In order to have more effective and efficient regulations, we've got to get through that. DR. APOSTOLAKIS: Actually -- go ahead. DR. KRESS: Go ahead. I've asked enough. You go ahead. DR. APOSTOLAKIS: Isn't the notion, the very name severe accident tied to design basis accident, the concept of design basis? MR. CHRISTIE: I don't think so. DR. APOSTOLAKIS: Beyond design basis is severe. DR. KRESS: In a sense, that's an arbitrary thing, but to say the regulations ought to focus on severe accidents, which is what I heard right there, seems a little bit problematic to me, because the reason the severe accidents are the risk dominant ones is because the regulations have gotten rid of all of the high frequency -- MR. CHRISTIE: I disagree with that completely. I disagree with that completely. DR. APOSTOLAKIS: But I think you can rephrase it. Let's say we start with a clean slate. Instead of using terminology that is really tied to the existing system, what you're really saying there is focus on risk. DR. KRESS: Yes. MR. CHRISTIE: Okay. That would be appropriate. I accept that. DR. APOSTOLAKIS: Which now will lead you -- what does it mean to focus on risk? Well, you are going to the sequences of events that dominate that risk. MR. CHRISTIE: That's right. DR. APOSTOLAKIS: And you are free now from design basis or whatever. You look at the sequence and say what's the best job I can do to make sure the risk is low. DR. KRESS: I think that would be a good approach. DR. BONACA: I think there is space for both needs. When you design ECCS systems, you are using some requirements of the existing law that make sense. You are designing how much water do you need to deal with some limiting accidents of some type. Now, when you evaluate the response of other systems, you may need to focus the way you're saying. What I'm trying to say, you cannot exclude it. It's the same way in which, after TMI, we discovered that using the analysis made to design the plants, to train the operator was wrong, because life, things don't move the way that the analysis for designing the plant moved. So in designing the plant, you are looking at limiting events to build equipment that will deal also with the limiting event. In training the operator, you develop analysis that gives you what will really happen in most cases, so that the operators are able to deal with those issues. So I'm only saying that to say we only should focus on severe accident is somewhat limiting insofar as regulation is concerned, because you still have some design objectives that you want to maintain there. Who knows? Maybe we'll beat another plant ten years from now. DR. APOSTOLAKIS: Yes, but that's not in the present. Anyway, we understand the spirit of this. MR. CHRISTIE: Yes. We've been through this many times. Okay. The next thing is when we put together the proposed rulemaking, again, it's a first of a kind and we were doing the best job we could. But what we thought -- and, see, this is our framework. Our framework starts on public health risk is dominated by severe accidents, with containment bypass and breach, probabilistic risk assessment is the best tool to measure it. You're going to focus on the sequences that are most significant, and so on and so forth. So our framework just said, okay, let's start with that kind of a philosophical approach to the whole problem and then let's just go ahead -- and what we said is let's retain whatever is in there that's effective and efficient. So if you got high point vents, if you got inerting, if you got igniters for the MARK 3, and so and so, fine, retain it. Just keep it. No sweat. Add where necessary. We added the section that had to do with the large drys, checking their containment capability. Now, we would believe that that would be an appropriate addition to the rulemaking process to cover the large dry plants. In essence, we've already done it. Let's just put it in the regulations. To me, it passed the backfit rule. DR. KRESS: Because you've already done it. MR. CHRISTIE: Right. DR. KRESS: It's not going to cost you very much, is it? MR. CHRISTIE: Right. You've already done it, it doesn't cost you anything, and it meets the backfit rule. And then delete whatever is not effective and efficient. We wiped out, and I think we got agreement now, we wiped out all the post-LOCA design basis hydrogen requirements. I mean, to me, it's a simple problem in just -- I don't need option three, I don't need option two. I can sit down and write a rule to address the specific thing just doing exactly what I said. Go through, identify the things that are important in probabilistic risk assessment space, retain what's effective and efficient, add where it's necessary, and delete whatever is left. I absolutely believe this is -- this petition for rulemaking is a risk-positive thing. By that, I mean the potential for health effects on people surrounding the plant will be less if this petition for rulemaking is approved than if it's not. We have identified certain problems with the existing regulations, which have to do with distracting the operators, having the potential for purge valves open, having the potential for guys wandering around the yards after severe -- you know, et cetera, et cetera. We can eliminate those. We can have not only a safer plant, but a more cost-effective. I think it's -- why aren't we doing this tomorrow? I mean, that's the thing that's driving me the nuttiest. It's everybody comes out ahead. I sat there in that meeting with the Commissioners on the 20th and had people come up and say the industry is only interested in the economics and we'll -- you know, that's all we're -- we're not. We're interested in doing things that make the plants better, both from a safety standpoint and an economic. Anything that makes it better, and both is our first priority. DR. KRESS: What is the status of your petition? MR. CHRISTIE: You'll find out this afternoon. I hope you ask, and ask well. DR. BONACA: I have a question. On your slide C, where you're talking about basing the removal of -- for dry containment based on the reactor containment capability. MR. CHRISTIE: Right. DR. BONACA: For the first time really, you are tying a requirement to a severe accident assumption of performance in the containment. What I mean is that if you look at the containments, the way they are today, they are designed to meet whatever the design requirement, assuming the design basis, say 50 psi capacity. And then in your presentation, you're saying 140, 130 psi capability. We made these estimates for the PRAs. They were based on the original design of the containments. Now, we know containments with aging go through relaxations. There are relaxation of tendons and there are requirements that by the time they get down to certain values, these tendons be retensioned and things of that kind. But that's an interesting concept, because you are tying a specific requirement to a performance, which right now is not being regulated. It's just an estimate. And the question I have is, for example, would it be a problem for a containment? MR. CHRISTIE: To the best of my knowledge, it's not. We've already been through this. It's not a problem for the containment. DR. BONACA: No, no, no. We haven't gone through this, I'm sorry. We have used the original design of the containments to perform an estimate of their ultimate capability. But the fact is when you monitor containment tendons performance, you allow them to relax, until they get to a certain limit, where the capability for design basis used is challenged and then you are going to replace and retension and so and so forth. You have programs to do that. So I'm saying that the age of a containment, its ultimate capability may not always be at the level we use for the IPEs or the PRAs. MR. CHRISTIE: Okay. And, again, I'm going back to the -- we've had a test case for large drys and hydrogen combustion. It's called Three Mile Island Unit 2. In there, this was a containment designed for 50, probably had an ultimate of 150. We produced hydrogen in Three Mile Island in bushels, probably more than we're ever going to ever produce again. The spike was less than the design. Okay. We can argue, hey, it might be 30 psi if we had done something different, et cetera, et cetera, but the difference between 30 and 150 is -- fine, maybe 20 years from now, the containment ultimate is 135. I don't care. I mean, as long as we've got this kind of tool that allows us to put these things in perspective and to make decisions to spend money, we've got to use it. If we don't use it, then we won't be as effective and efficient. And I'm the last person in the world to say it's the perfect tool, I'm the last person to say that it's going to be fixed in concrete and we're never going to change our knowledge base and what we know and understand. I'm the opposite. I think it's going to change constantly and we've got to be -- and I'm with you. We've got to be aware of those. But what I want to focus on is the things that are safety significant. I want to stop spending the money on the things that aren't safety significant. DR. BONACA: I agree with you 100 percent. MR. CHRISTIE: Okay. DR. BONACA: But there has to be some defensible technical basis for everything that we are going to discuss and accept. I'm saying that I see the reference here to the ultimate capability of containments. There has been a criterion in here proposed that bases itself on that capability and I'm saying the capability, however, is not one that is guaranteed. In fact, it varies because all the monitoring systems for the capability are not focusing on the ultimate capability, they are focusing on the design pressure of the containment, which is well below that. MR. CHRISTIE: I do not read -- DR. BONACA: That is what's being preserved. What is being preserved is a design capability of the containment. That's all I said. MR. CHRISTIE: I don't see anywhere in that proposed rulemaking that says that we're using the ultimate capability. We're just saying that the large drys are going to check their containment capability. We don't use ultimate capacity. We just say for high probability events, check your containment capability. Whether you use design, whether you use ultimate, whatever you do, that's for the people at the plants to decide how to do it, and they've done it. Now, the last thing is, and this goes to the schedule and where we stand. You've got to remember that this rulemaking didn't come from me as a rulemaking effort. It came from me as a letter to the Commissioners saying based upon what I read in the San Onofre safety evaluation report, we've got a problem and here is my solution to the problem. And the Commissioners sent it down to the Office of Nuclear Reactor Regulation for fix. The fix that Nuclear Reactor Regulation and I have worked out is the rulemaking. This is a rulemaking. This is not part of option three, it doesn't depend on option there. When we agreed to rulemaking, it had nothing to do with option three. The letter I sent back in to NRR -- and I was dealing with NRR and I'm still dealing with NRR on the rulemaking -- it says fine, if you want to take this over to the option three people in Research and treat it as one of the things that they should be using as the first set of things to do, I have no problems with that. But this is a rulemaking. Okay. The status of it is today, the people who are doing option three tell me -- well, first off, in the February meeting, they told us they'd have recommendations to the Commission in June. Now, as I understand it, the recommendations go to the Commission in August, but they'll talk to you about that I hope this afternoon. But it's not a rulemaking. The recommendations that they're going to make to the Commissioners in August are not -- they were explicit that it's not going to be a rulemaking. What they're going to send is something to the Commissioners that's going to recommend something and then they're going to wait for the Commissioners to make a decision as to what to do and then they're going to start a rulemaking process. And I'm looking at the thing and I'm saying to myself, wait a minute, we sent it in, we didn't send it in as a rulemaking, but then we agreed that the rulemaking was the best way to handle the thing, and you're going to send a recommendation up that doesn't say rulemake. It doesn't make any sense to me. Now, I don't understand the bureaucracy at the Nuclear Regulatory Commission any more than I'm sure they do, but it just doesn't seem to me to be efficient to go back and start a whole new affair. We've got a rulemaking, we've got a petition. We've been through the public comment period. If you wanted a -- if the staff of the Nuclear Regulatory Commission wanted to push that through, all they got to do is say, boom, final rulemaking, and away it goes. Now, if the staff of the Nuclear Regulatory Commission doesn't want to have the petition approved, fine, then they say they don't want to have it approved and here's the following reasons and they send me a letter back saying we've considered it, but we don't want to do it, and too bad. DR. KRESS: Well, it seems to me like if NRC has embarked upon option three, which is risk-informing a great deal of Part 50, and that they have a priority system on which regulation they're going to do first and that if this one fits into that priority system, so that it gets early attention, and that all they're doing is trying to figure out what they mean by risk-informing the rule, it seems like it's appropriate to include it under that process. MR. CHRISTIE: I have no problem at all. DR. KRESS: That's the way I read what's going on. MR. CHRISTIE: The way I read it is they're going back to square one. DR. APOSTOLAKIS: We'll ask them this afternoon. MR. CHRISTIE: Right. Ask them this afternoon. The things that I have a problem with right now are the bit about station blackout for the MARK 3's and the ice condensers. It doesn't make any technical sense to me. And then where in the world are we with respect to getting a rulemaking gone? I think we've got a good rulemaking. I think we're very close to having complete agreement on a rulemaking, as far as I can tell. DR. APOSTOLAKIS: Okay. Thank you very much, Bob. We will recess until 12:45. [Whereupon, at 11:47 a.m., the meeting was recessed, to reconvene at 12:45 p.m., this same day.]. A F T E R N O O N S E S S I O N [12:47 p.m.] DR. APOSTOLAKIS: We are back in session. This afternoon we will discuss the risk-informed revision of 10 CFR Part 50 option three and 10 CFR 50.44. I understand most of the presentation is on 50.44. I thought it was to be on option three. I don't know. I have a number of comments on SECY-00-0086, which I don't know now how we can transmit to you and when the committee will have a chance to review that, because this is a very important document. At some point, we will have to write a letter, I suppose. That would be sometime in December, Tom? MR. KING: I think we're talking about two letters, probably. We owe the Commission a paper in August, which is primarily to present our recommendations on 50.44, but, also, as part of that paper, we're going to provide an updated framework document, recognizing that this is all work in progress and things are evolving as we go. We're also going to present the Commission any policy issues they have to deal with and we're going to talk about those today. I think these are policy issues that aren't limited to 50.44. These are broader policy issues. It was 50.44 that brought them out. DR. APOSTOLAKIS: So in July we will have to write a letter. MR. KING: So I would like a letter either in July or, at the latest, I know you don't have an August meeting, but you have a meeting starting August 30th, I think, your September meeting, that would probably be the latest for that August paper. Then in December, we owe another report to the Commission and we would probably be asking for a letter on that one, as well. DR. APOSTOLAKIS: And that would be on option three, independent of 50.44. MR. KING: Well, 50.44 is part of option three, remember. It's the first test case that we're running through the framework. DR. APOSTOLAKIS: But we haven't discussed the framework, that's my concern. MR. KING: We had discussed it back in March or so. We went through the framework. You had given us a number of comments. We didn't ask for a letter. DR. APOSTOLAKIS: Without the benefit of a document, though. We had the viewgraphs, I remember, but not the document. I mean, this is the first time I see this document and it's being transmitted to the Commissioners April 12. MS. DROUIN: It was transmitted then, but it was out for public comment. The first of February, we sent it to the public document room. It was on the web site. So it's been out there since February 1. DR. APOSTOLAKIS: Well, sure, but -- MR. KING: Recognize, the April 12 package is not the final framework. It was a status report and we gave the Commission what we called a draft framework. It represented work that was done up to that time and the framework is evolving as we proceed and we get into these test cases. So your comments on the framework are certainly -- it's not too late. We certainly are interested in those. I mean, if you want to give those to us today, fine. If you want to talk about them, if you want to schedule another meeting to talk about them, whatever. DR. APOSTOLAKIS: I don't know how the members feel. DR. KRESS: I think it's an extremely important document and we ought to schedule a meeting and get the full benefit of comments from as much of the committee as we can, the subcommittee anyway. MR. KING: We got comments -- not recently -- a couple months from NEI on the framework. We're having a meeting with them tomorrow to talk about their comments on the framework. So they continue to come in as time goes on. DR. KRESS: This is a seminal document. We really ought to -- DR. APOSTOLAKIS: It's a big thing. We can decide. I don't know. Mario, do you have any thoughts? Bill? Bob? Jack? Any comments? MR. KING: It would be useful to do that, though, before this August paper goes up. The sooner the better. DR. KRESS: We'd have to do it with subcommittee then? MR. KING: But when? DR. KRESS: Just have a subcommittee of the whole. Get all the members here. DR. APOSTOLAKIS: The problem with full committee meetings is that you can't really get into detail. DR. KRESS: You can't get into enough detail for this document at the full committee. I'd rather have a subcommittee of the whole and have a full day of it, at least, maybe even more, if they would agree to it. MR. KING: We're having our own internal retreat on the framework document and 50.44 on July 14, as you know. Even internally we're getting some comments on it. DR. BONACA: Will you be here the morning of the 11th? DR. APOSTOLAKIS: I will be here in the morning, yes, Tuesday morning, the 11th. In fact, the P&P meeting can be pushed a little bit into the afternoon, or another way would be to maybe do the P&P in the morning, finish with it. DR. BONACA: We can adjust that. MR. KING: That would be much better for us, because we have a meeting with NEI in the morning of July 11th. DR. KRESS: We could cover it pretty well in half a day, I think. MR. KING: Half a day would probably be -- DR. APOSTOLAKIS: I'm sorry. What? MR. KING: A half a day meeting would probably be enough. DR. APOSTOLAKIS: Half a day, you know, four to five hours. I don't think we need more to that. MR. KING: The afternoon of July 11th is okay. DR. APOSTOLAKIS: We have to coordinate it with the Chairman of the committee so that the P&P subcommittee meeting -- MR. KING: We come to you again on the 12th as a full committee, but we only have two hours. DR. APOSTOLAKIS: It's out of the question. We have to get into it in more detail. I mean, this is one of the major changes in the regulation. So let's say until we settle it with Dr. Powers, that the 11th, which is a week from this coming Tuesday, right? MR. KING: Right. DR. APOSTOLAKIS: We'll try to find a block of four hours or so. DR. SHACK: That's Tuesday afternoon. DR. APOSTOLAKIS: Tuesday afternoon, so that members can fly down in the morning. The ones who are members of P&P will have to come Monday night anyway. Does Noel know that? MR. MARKLEY: Noel is going to check with Dana now. DR. APOSTOLAKIS: Yes, but he should propose that the P&P be in the morning. MS. DROUIN: At this meeting, George, are you looking for more a round tabletop discussion where you would just be walking through and giving us -- DR. APOSTOLAKIS: I'm sorry, Mary. What did he just tell us? 1:00 for us. MR. MARKLEY: 1:00 for us. DR. APOSTOLAKIS: I'm sorry. Go ahead. MS. DROUIN: Are you more interested in just sitting down and going through the report and having more of a roundtable discussion or us preparing a formal presentation? DR. APOSTOLAKIS: Well, you must have presented this to other people. I don't want you to spend extra time preparing. I mean, it's always helpful if you guide the discussion by a set of viewgraphs, but if you have to start from scratch, no. MR. KING: We could take previous viewgraphs and hit the high points. DR. APOSTOLAKIS: The things that you have used already, because we can always go page by page. DR. BONACA: Actually, you have a number of charts in the report. MR. KING: Yes. DR. BONACA: That you can push that out and we can -- DR. APOSTOLAKIS: But a lot of this stuff I see for the first time, like the prioritization decision tree. In fact, we will ask the whole committee to come. This will be a subcommittee with the full committee present. MS. DROUIN: I think we ought to give them the more updated version. MR. KING: There obviously isn't much time. MS. DROUIN: George, as Tom had pointed out, this has been a living document. DR. APOSTOLAKIS: Exactly. MS. DROUIN: And we have been receiving comments and as we have received them, we have been updating it. I think it would probably help for you to not -- we can get it to you this afternoon, right away, the more updated version. DR. APOSTOLAKIS: If I can get it this afternoon, that's fine with me, because I'm going out of the country this weekend. MS. DROUIN: We can give it to you. In fact, I have one right here, I can get it copied right away. DR. APOSTOLAKIS: Are there dramatic changes? Because I've marked this up. MS. DROUIN: In terms of the basic framework, those figures, no, but chapter five, which talks about the implementation, the prioritization and everything, that has changed quite a bit. DR. APOSTOLAKIS: So the members who have not read it have an advantage here. DR. BONACA: Well, we'll need to see it sometime next week. DR. APOSTOLAKIS: Yes. So I think it's a good idea. Tuesday. MS. DROUIN: I can give it to you right now. I have it right here in front of me. DR. APOSTOLAKIS: Tuesday, I hope Dana will agree. MR. KING: So we'll plan July 11th, in the afternoon then. DR. APOSTOLAKIS: Let's plan on that, until we hear from Dr. Powers. MR. KING: And you'll probably get -- maybe touch on some of the framework issues today as we go through 50.44. DR. APOSTOLAKIS: Of course. MR. KING: All right. For the record, we have put on the front slide the primary participants who have been working on the 50.44 issue. They're not all at the table. With me at the table, for the record, my name is Tom King, from the Office of Research; John Lehner, from Brookhaven National Laboratory; Mary Drouin, from the Research staff; Trevor Pratt, from Brookhaven; and, Allen Camp, from Sandia. As I said, we're not asking for a letter at this point. This is a status report. We're going to focus on 50.44 and we're going to touch on some of the potential issues that will probably be in the paper that goes up to the Commission in August. Some are policy, some are technical, but even the technical ones even have some policy nature. So any thoughts you folks have on issues would be very useful to get today, so we can start to formulate them for the paper. Stakeholder input. One question we had gotten was that you were interested in talking about the January 19 letter we got from Joe Colvin of NEI. It basically came in and reported the results of a survey that NEI had taken of the industry. It was an attempt to give us some information on their priorities and potential cost savings associated with risk-informed changes they'd like to see, and that would help us prioritize our activities. We did get the letter. We basically found it matched up pretty well with the priorities we had already come to from our own thinking and previous discussions among ourselves and with industry folks. We had also gotten a letter dated April 18 from Steve Floyd to myself, with a lot of comments on the framework document. If you don't have that, I've got a copy here you can have. You do have it? DR. APOSTOLAKIS: Yes. MR. KING: Okay. And as I said, tomorrow we're meeting with NEI on the comments in that letter. We had responded to the Colvin letter. It was just a short response thanking them for their input and recognizing that the main points in their letter we were accommodating as part of our framework and option three work. That's all I really intended to say about the January 19 letter. I don't know if you have any other specific points you want to get into. DR. KRESS: I had, I guess, a question. The 19th letter had some associated cost-benefit savings that NEI had gotten from their survey. MR. KING: Yes. DR. KRESS: Do you have any plans for using those numbers in any way in this process? MR. KING: I think those numbers will certainly help us in trying to prioritize our work. If we see regulations that look like they would have a large cost-benefit impact associated with them, they may go more toward the top of the list of things we work on first. DR. KRESS: But they won't be used -- I think prioritization -- MR. KING: That's not the only factor. DR. KRESS: Prioritization is a good thing to use them for, but I was concerned that you may have some sort of an inverse backfit part of your system that has as cost-benefit type of -- MR. KING: We're going to get to that issue when we get to the slides. DR. KRESS: You wouldn't use those numbers for that, though. MR. KING: If we do some sort of reverse backfit test, those numbers could be useful. But whether we do one or not is an issue at this point. I was going to talk about that when we get to another slide. DR. KRESS: I'll wait. MR. KING: In terms of prioritization, I mean, clearly safety is the first factor. DR. KRESS: Of course. MR. KING: If we go through a risk-informed look at something and there's a safety issue that really goes to the to of the list. But then there's other factors to sort out the other things, and cost is one of them. The other thing I want to mention, maybe you're not aware of, is some members of the Research staff had a meeting with Commonwealth Edison a week or so ago, June 14th, actually, where Commonwealth Edison gave out quite a detailed list of changes that they feel could be made, risk-informed changes to the regulations, and cost savings associated with them. They went into more detail than the Joe Colvin January 19 letter. If you are interested in that list, I've got a copy of it here. You might find it interesting to see what they have to say. DR. UHRIG: I don't think we have it. MR. MARKLEY: What we have is a June 7 NEI letter. DR. BONACA: That's right. MR. MARKLEY: That's the latest thing you have. MR. KING: This is the results of the June 14 meeting, but it's publicly available. This is an extra copy, if you want to take it. With that, I'm going to turn it over to Mary to talk about our approach and the alternatives we've come up with on 50.44. MS. DROUIN: We weren't planning on getting into any of the details of the framework, but we wanted to talk a little bit about the framework, because it does establish our approach and how we came up with our options. We have developed a framework and in developing that framework, what we tried to do was blend the principles from Regulatory Guide 1.174 with the cornerstones that are outlined in the oversight program for safe nuclear power plant operation. So what we tried to do there was balance an accident prevention and accident mitigation to achieve a defense-in-depth approach in building the framework. The way we went about trying to achieve this defense-in-depth balance is that we defined four strategies in the framework and these four strategies were looking at limiting your initiating events, or limiting your core damage accidents, or limiting your radionuclide releases, or limiting your public health effects. And taking these strategies and this accident prevention and mitigation, we then came up with quantitative guidelines for these. We only developed these quantitative guidelines to help us frame the options, such that when we come up with a risk-informed option, we wanted to make sure the option was neither too conservative, such that we were imposing undue conservatism or undue burden; at the same time, we wanted to make sure an option is not too relaxed. So whereas we don't expect to ever see this quantitative guidelines in a regulation, they're just helping us put it in the ballpark, such that we have achieved these strategies in maintaining our defense-in-depth. DR. KRESS: Might those numbers appear in a reg guide or somewhere else? MS. DROUIN: They could appear in a reg guide for -- and I'm going to get to that in one of our options on 50.44. DR. KRESS: At the moment, those numbers are intended to be means. MS. DROUIN: Yes. DR. APOSTOLAKIS: I have two comments already on this. I don't think that calling it balanced high level defense-in-depth really represents what you're doing and you say that also in the document, the approach recommended by the ACRS and adopted herein is a structure of this high level defense-in-depth approach. Well, the ACRS recommended the preliminary framework that also addressed what to do at lower levels, and you are very silent on this in this chapter two, to the point where one wonders have you rejected this or what do you plan to do at lower levels. But then you go to chapter four and on page 4-3, it says decisions regarding lower level defense-in-depth in the form of redundancy or diversity are generally well suited to this type of analysis, which is a PRA. I mean, you just said it was a PRA, that's what it refers to. So I was wondering why you focus so much on high level defense-in-depth and you don't actually say this is defense-in-depth at the high level, rationalist at the lower level, except when the uncertainties are such that defense-in-depth must be invoked at lower levels, as well, as you strongly imply later that you're going to do. I mean, if I read chapter two, the only thing you are telling me is that at the high level, you are structuralist, and nothing on anything else. But later on in the implementation, you say, no, no, no, we're going to use PRA at lower levels. I'm just wondering why that was -- MS. DROUIN: Are you disagreeing with what's later on in the -- DR. APOSTOLAKIS: I am not disagreeing. I'm just wondering why you focus so much on the high level defense-in-depth and even in your title here says balanced high level. What happens at lower levels? MS. DROUIN: I think that is one of the comments from other people and from ourselves that we have noticed in the document is this inconsistency of things written in one part of the report and the other. That's not meant to be. DR. APOSTOLAKIS: No, I understand. I'm just making the comment. When I read it, I was wondering, because you talk about structuralist and rationalist, and they say we're going to apply structuralist at the high level. So I said, well, gee, on the next page they're going to tell me why rationalist at lower levels is a crazy idea, because they don't want to use it. But you are completely silent. But then when it comes to chapter four, you are speaking about it matter of factly, yes, and we're going to do PRA evaluation. So I think it needs tightening up. MS. DROUIN: Absolutely. DR. APOSTOLAKIS: I don't disagree with what you say. DR. KRESS: You're not opposed to this high level. DR. APOSTOLAKIS: I just don't think this is what they're doing. This is part of what they're doing. DR. KRESS: It's part of what they're doing, yes. MS. DROUIN: Yes. DR. KRESS: But I would not have called this the structuralist approach. DR. APOSTOLAKIS: No. It's the preliminary proposal. DR. KRESS: I would call it the rationalist. But I would have called that a rationalist approach there. DR. APOSTOLAKIS: Well, because at the high level we impose the structuralist, without any reason. What they call four strategies, that's what we do, too. That's why it's preliminary. DR. KRESS: I still disagree. This is rationalist. DR. APOSTOLAKIS: Because of the uncertainties at that level, but I think that's too subtle for most people. Anyway, your name is on the paper. You can't disagree. DR. KRESS: It's too late to disagree. DR. APOSTOLAKIS: It's too late to disagree. And the other thing is, which is related to what you said, I'm just limiting my comments on what you said, which brought -- you really go out of your way in this document to tell people that one quantitative risk numbers will not appear in the regulations and you say that so many times, that there must be a reason behind it. So what is that reason? MR. KING: That makes it risk-based. MS. DROUIN: It makes it risk-based. DR. APOSTOLAKIS: If you use one number, it makes it risk-based? MR. KING: Put the number in the regulation. DR. APOSTOLAKIS: The way you put it here, Tom, is that nothing that smells of frequencies will appear anywhere. It's one thing to base everything on the QHOs and quite another to have such sweeping statements that the regulations will be deterministic. And the way I understood it, then, PRA will be used to rationalize why you go this way and that way. But it seems to me that in some instances, you may have to use unavailabilities or something. You are saying no, no -- that's a no-no. Would that make it risk-based if you use the little number here and there? MR. KING: If we start putting risk numbers, whether they're core damage frequencies or reliability numbers or something in the regulations, our view is, yes, that makes it risk-based, because then that becomes the sole factor on which you make your decision. DR. APOSTOLAKIS: No, because the oversight process uses performance indicators and explains to us that -- MR. KING: But they're not in the regulations. We're talking about the regulations. Reg guides are a different story. SRPs are a different story. DR. APOSTOLAKIS: Reg guides are not regulations. MR. KING: Reg guides are not regulation. Reg guides are one way to meet the regulations, one acceptable way. DR. APOSTOLAKIS: It's just that you say it so many times in such a definitive way, that I thought there was something behind it. MR. KING: Maybe we over-emphasized a bit. DR. APOSTOLAKIS: Yes, I think you did. MR. KING: But the idea is -- DR. APOSTOLAKIS: Heaven forbid if we become risk-based, a quarter of a page. DR. SHACK: They keep getting advice from advisory committees to put numbers in, George. DR. APOSTOLAKIS: Say it again. DR. SHACK: They keep getting advice from some of their advisory committees to put numbers in, letters. DR. KRESS: They continue to reject it. DR. APOSTOLAKIS: Another thing is I don't know if -- it looks like the risk-based is really something that the agency must avoid at all costs. I thought risk-informed was a good terminology that really reflected reality, but now it has come to the point where people say we're not going to do this because if we do, it's risk-based. In other words, it has become something to avoid at all costs. DR. KRESS: I guess I would disagree that having actual risk numbers as targets makes it risk-based. If you do all the other things, like defense-in-depth and deal with uncertainties and appropriate margins, also, then I think you can have numbers in there and call it risk-informed. DR. APOSTOLAKIS: I think so, too. But anyway, if you tone down the document here. It appears like you're so sensitive to that. MR. KING: I mean, we don't want it so slanted that -- we're sensitive. DR. APOSTOLAKIS: I mean, we are sensitive people. Allen, you wanted to say something a half an hour ago. MR. CAMP: I forgot what it was. DR. KRESS: We talked you out of it. DR. SHACK: One comment we made before was that these were PWR numbers and they're not really BWR numbers. You end up with LERF about the same range, but at least looking at the IPEs, these look like sort of unrealistic expectations for conditional containment probabilities for BWRs. MR. KING: They are really numbers that come out of the Commission's 1990 SRM that told us how to implement the safety goal, that laid out the ten-to-the-minus-four CDF and the LERF we derived working backwards from the QHO. You're right, we haven't tried to distinguish between B's and P's and that's certainly an issue in implementing it that we're going to have to deal with. But they're intended to be a generic set for both. DR. APOSTOLAKIS: One last comment, because I think it's relevant to what's coming. The document 00-0086, well, gee, you know, today we have 00-02, now we have 00-0086. DR. KRESS: 007. DR. APOSTOLAKIS: Yes. A lot of zeroes. This document proposes a framework and then it goes on to propose some target numbers for containment, the right frequencies, and so on. Now, this is intended to be used by the regulator, correct? MS. DROUIN: Yes. DR. APOSTOLAKIS: And I was wondering -- and, again, I don't have a position on this, but I think it's something that needs to be aired. The QHOs were always intended to be goals, not measures of adequate protection or inadequate protection. In other words, you can exceed them and you can still provide adequate protection. Is it really appropriate for such fundamental documents to use the QHOs, something that is a goal and we know already 19 PWR units are above it? Not the QHOs, but I mean the lower CDF. Is it appropriate to do that, Tom? I don't know. Or should we looking for something else? DR. KRESS: George, I would prefer we not even discuss that subject anymore, because I like the idea of using those goals as guidance to guide the regulations, even though they may or may not represent adequate protection, as we know it, in terms of real numbers. Those are pretty good goals and I see very little wrong with them. They're accepted. And I think as a guidance, you're not going to -- you're not going to say you must meet these numbers. That's why they're saying they're not going to put the numbers, partly why they're there, but they're going to use it to guide how they get the regulations and the regulations are going to be crafted in such a way that they won't exactly meet them. Some plants will do better and some plants may not do as good. So I think it's a nice target to set when they're going to want to gauge how to write the regulations. It avoids this issue of quantifying adequate protection and trying to put numbers in the thing. So I'm comfortable with it now and I kind of want to avoid even mentioning it anymore. Go on and let's do it, because it does raise issues of are we ratcheting and are we doing -- DR. APOSTOLAKIS: Well, that's exactly my concern. But if we start using them in the regulations, before you know it, they will become measures of adequate protection. DR. KRESS: That's what I would really -- I have a hidden agenda. That's what I'd really like to have happen. DR. APOSTOLAKIS: Well, my agenda is out in the open and I really don't think that would be appropriate. For example, I don't understand -- you just said that we can use those to write regulations and then we'll say, you know, gee, some of you guys will not meet it. I don't understand that. I mean, if we write the regulations, then -- DR. KRESS: They'll meet the regulations, but they may not exactly meet those target values for the frequencies and the CDFs and things like that. They'll be close enough -- if you do the regulations, it's good enough, but there won't be a direct one-to-one link between the -- DR. APOSTOLAKIS: There won't be a direct one-to-one link, but if we follow the framework and we have numbers for core damage frequency, conditional probability of containment failure and so on, and we write the regulations that way, then eventually they will meet those, because the regulations will be a coherent code. DR. KRESS: I say good. DR. APOSTOLAKIS: I realize that, but the question is whether you want this to be a measure of adequate protection. MR. KING: No. No. The issue you're talking about is going to have to be discussed in the August paper and the Commission is going to have to buy into using the safety goals as the measure of how far we want to go in risk-informing things. DR. APOSTOLAKIS: I think that we are using the safety goals because that is what's available, correct? It's a number that's available. MR. KING: It's what is available. It avoids having to deal with the question of defining adequate protection, which, in my view, isn't necessary for risk-informing. DR. KRESS: And it gives you an idea of what a proper balance might be based on tradition. MR. KING: And it reflects the Commission's -- remember, the safety goal policy statement is a statement of the Commission's expectations of the safety they'd like to see achieved. DR. KRESS: It has a lot to -- MR. KING: If that's what they want to see achieved, it would seem reasonable to me that we proceed down a path that lays out a framework to achieve that. And maybe it won't pan out in all cases for cost-benefit reasons or something, but why not aspire to that? DR. APOSTOLAKIS: I can see this ratcheting up the regulations down the line. MR. KING: That's another issue. DR. APOSTOLAKIS: So I really think -- and I'm glad you said that you will raise the flag to the Commission to think about it. DR. KRESS: But risk-informed regulations are going to be voluntary, right? DR. APOSTOLAKIS: Yes, that's right. Okay. DR. SHACK: And the proposed rule will have to meet a backfit rule. DR. KRESS: I think it's entirely appropriate that they use these. DR. APOSTOLAKIS: This is something that I think we should discuss at the subcommittee meeting. DR. KRESS: Yes, we ought to. DR. APOSTOLAKIS: Which I was informed will be at 1:00 on Tuesday, July 11th. DR. KRESS: Okay. And you put that on the calendar. DR. APOSTOLAKIS: That will certainly be on the calendar and see how we can approach it. DR. KRESS: Sure. DR. APOSTOLAKIS: Because it may be that you put a few qualifiers there and then the concern goes away. But I don't think we should just matter of factly pick up those numbers and just run with them and starting writing the regulations. I just don't think so. MR. KING: It's clearly a policy issue, in my view. DR. APOSTOLAKIS: Yes. It's a policy issue and before the decision is made, I think the decision-makers should be informed what the concerns are. MR. KING: Okay. DR. APOSTOLAKIS: Of the various agendas we have. Maybe a three-vision approach may be useful. MR. KING: Shades of gray, maybe. DR. APOSTOLAKIS: Shades of gray are working very well. See, if you look at the ASME thing, I think now the categories make much more sense to me. MS. DROUIN: Are you ready? DR. SHACK: Category one plants and category two plants. DR. APOSTOLAKIS: Yes. MS. DROUIN: In looking at our approach, there's three major steps. The first one, of course, is selecting what regulation to risk-inform out of 10 CFR 50. What we have done to date is that we have performed a core screening and we're going through our prioritization. The core screening, first, just looked at whether or not the regulation dealt with accident prevention and accident mitigation and if it did, then it was a candidate for risk-informing regulation, and then we started going through the prioritization. And in going through the prioritization, the first thing we would look at once it was a candidate, is it even warranted to risk-inform the regulation. What we mean by that is in looking at the regulation, if you risk-informed it, were you going to be able to gain any safety benefit, was there any conservatism in that regulation, was there any excess burden. And if you weren't going to gain any safety benefit and there wasn't any conservatism and there wasn't any excess burden, then it would be eliminated further. Then in further prioritization, the next thing that we would look at is what was the safety significance of that regulation, and then looking at what the resources it would take to risk-inform and then also looking at the potential for reducing any unnecessary burden. And taking these factors into account, 50.44 fell out as a very high priority and that was the one that we identified as the first in what we're using as our test case against the framework. DR. APOSTOLAKIS: You didn't know that, Bob? MR. CHRISTIE: Excuse me? DR. APOSTOLAKIS: This morning you said you didn't know what the staff was doing. They are telling us this is the test case. Is that consistent with what you said this morning? MR. CHRISTIE: They're testing against the framework. MR. MARKLEY: Microphone, please. MR. CHRISTIE: This is Bob Christie, Performance Technology. The staff, in option three, is testing the framework with respect to 50.44. In my comments to you on March 1st, we vehemently disagreed with the framework that has been outlined in the February meeting to us. The framework has changed. We don't see that the changes have been any for the better and they can go on having this very complicated framework document that they can test any regulation about to infinity and that will be their objective. Our objective is to make the regulations more effective and efficient and we don't need this framework and testing out 50.44, in the petition we submitted or I submitted or any submits, against this framework seems to us to be kind of an academic exercise that doesn't move anything towards decisions to make the regulations more effective and efficient. Just flat, hey, we're now shooting for how safe is safe enough, is adequate protection, just blows my mind. It just absolutely is crazy. If that's what we're fighting for, you know, you talk about this being voluntary, you want to say how safe is safe enough is now adequate protection, you just carry on with voluntary efforts, if that's the framework and that's the decision criteria. MR. KING: That's not what we said. We're not equating how safe is safe enough to adequate protection. MR. CHRISTIE: Again, this is Bob Christie. I heard that we are going to aspire at writing the regulations to achieve the levels as set forth in the safety goals, which are, as far as I can tell, in the June 11 or whatever, the June 1990 memorandum was, from the Commissioners to the staff, was very clear that the safety goals were how safe is safe enough, the point at which even backfit was not to be considered. Aspiring to them and using those as the guides and the criteria for the regulation, to me, look like you're moving the rules to how safe is safe enough. DR. APOSTOLAKIS: And this is certainly something that we'll discuss on July 11th. Anyway, thanks for the comment. MS. DROUIN: In developing our options, we took two different paths to come up with the options. The first was to look at the current requirements and to risk-inform the requirements. The other path to come up with options was to start with a blank piece of paper, look at the threat or the concern and if you started with a blank piece of paper, what requirements would you come up with, and then that would give us the pool of options and then based on those options, we would come up with alternatives for a risk-informed 50.44. In going down the path of looking at the current requirements, what we would do and what we have been doing on 50.44 is to go back and look at those requirements and look at what was the technical bases of those requirements, look to see how those requirements were implemented by the licensees, identify any related regulations in the implementing documents, because one of the biggest things that we have learned is you can't take these things in isolation. A lot of the problems that you have with a particular regulation is not so much in the regulation itself, but it comes out perhaps in a related regulation or in the implementing documents. And then start doing a risk evaluation of that and coming up with options. We'll get through this more as we get into 50.44. Once we have identified the alternatives, then I do an evaluation to try and prioritize those. In this particular prioritization, what we would do is look at both the impact to the NRC and the impact to the licensees. In looking at the NRC, we would look at such things as is it going to require a rule change, what would be the impact of doing this on other regulations and implementing documents, would we have to modify them, would we need to create perhaps some new regulatory guides, what would be the extent of regulatory analysis that we would need to do. If the alternative would require some submittal from the licensee, what would be the extent of NRC review, what would be the impact on inspection. These are just some of the factors that we would take into account in trying to prioritize the regulation. Also, of course, goes in there what the safety benefit would be. In looking at the licensee and the impact on the licensee, what would be the need for perhaps new or make modifications to the current design in place, looking at the need for analysis on the licensee's part, the impact on their maintenance and inspection activities, impact on technical specifications, impact on procedures and training. And here, again, are just some examples of the things that we would be looking at. Moving on. I'm just going to quickly go through it, because hopefully we'll get, after this slide, right into 50.44, how this ties explicitly back to the framework and how we've used the framework to come up with our risk-informed options. First, starting with what is the concern that we're dealing with and how does this concern tie into the four strategies that are outlined in the framework. Is it dealing with accident prevention and mitigation and within those two things, is it dealing with just limiting the initiating events or is it dealing with limiting the accident sequences or is it dealing with limiting your radionuclide releases or is it dealing with the health effects. For example, on 50.44, the fact that we're dealing with the combustible gases, the strategy that is of concern, that it's tying directly to into the framework, is our third strategy of limiting your radionuclide releases, trying to contain your conditional containment failure probability. And then identifying that concern, the relative importance of that concern against the quantitative guidelines. If, for example, and this is just hypothetical, if you look at this and you see from your risk insights that your conditional containment failure probability from this concern is -- say it's extremely high, close to unity, that's identifying that this is a concern, this is something that the regulations ought to be dealing with. If you are on the other extreme and you're getting something that's, making up a fictitious number, 1E-to-the-minus-four for your conditional containment failure probability, then you would question why are we regulating this. DR. KRESS: Do you mean the contribution of that particular rule to the conditional containment failure probability? MS. DROUIN: The contribution of that concern. DR. KRESS: That concern, I mean. The concern you said here was containment failure. MS. DROUIN: The concern is hydrogen combustion. DR. KRESS: The concern is hydrogen combustion. MS. DROUIN: So if I look at, if I'm going through and I'm looking at my risk insights -- DR. KRESS: Now, will you do this on a plant type by plant type basis? Like you will look at it for -- I mean, will the rule apply to all plants across the board or would you look at the ice condenser separately from -- MS. DROUIN: Both, both. DR. KRESS: -- BWRs? MS. DROUIN: Both. MR. KING: You will see the alternatives. Some are broader that apply generally and some are the specific like the current rule, but modified to reflect risk insights. DR. KRESS: So you might have this rule written differently for ice condensers as opposed to large drys. MS. DROUIN: Absolutely. DR. KRESS: Okay. MS. DROUIN: So you would look at the specific containment types, but you would also look at it generically. And then once you've identified whether or not this is a concern that should be in the regulations, then in developing the options, we would look and make sure that a single class doesn't contribute more than ten percent to that guideline. DR. KRESS: Now, that's a very interesting number. DR. APOSTOLAKIS: That's a Sizewell guide. DR. KRESS: It's the rule of ten, which is a nice rule to use when you have to pick something out of the air, which is basically what you have to do. It's a policy type thing. But I guess I wanted to ask, could you have an additional line there and say "or" more than ten percent to the uncertainty? MS. DROUIN: In doing that, you would have to account for the uncertainty in there. DR. APOSTOLAKIS: To the standard deviation. DR. KRESS: Or to the standard deviation or something like that. MS. DROUIN: Yes. DR. KRESS: Could you have an additional -- that would help me a little, because it ties into my concept of what defense-in-depth is related to uncertainties. DR. APOSTOLAKIS: I have a number of comments on this. First of all, I don't know that you can separate the variance from the mean, because if you reduce -- you see, these are wide distributors. If you reduce the high tail, the mean also moves. DR. KRESS: Sometimes. DR. APOSTOLAKIS: So clearly it's something we have to understand. DR. KRESS: But you need to understand the point with the uncertainties, too. DR. APOSTOLAKIS: The other point I want to make is this is really a defense-in-depth structuralist kind of approach, because you're just declaring it ten percent. I'm wondering why. I agree with what you said earlier that you will have defense-in-depth at the top level, but what is it -- or maybe it's rationalist view here. DR. KRESS: It's a rationalist view, in my opinion. DR. APOSTOLAKIS: Realize that the uncertainties are very large at the accident sequence level and you want to take a defense-in-depth measure to handle those, and that is that no single class should contribute more than ten percent. You could rationalize it that way. DR. KRESS: I'm not sure I know exactly what they mean by accident class here. DR. APOSTOLAKIS: I think they mean the initiator. DR. KRESS: The frequency range of initiators. DR. APOSTOLAKIS: Yes. In other words, I think they define it here. DR. KRESS: It could be different kinds of sequences. DR. APOSTOLAKIS: There is a definition here that these must have -- DR. KRESS: But what you do when you do that is -- DR. APOSTOLAKIS: -- similar physical characteristics. DR. KRESS: -- you lump into an accident class different kinds of sequences, which bothers me, to some extent. DR. APOSTOLAKIS: Like small LOCA will be one class. DR. KRESS: No, I don't think so. Those initiators of frequencies within certain ranges are a class. DR. APOSTOLAKIS: No. That's not -- DR. KRESS: That's not what you mean. DR. APOSTOLAKIS: No. MR. KING: No. DR. KRESS: You actually mean the traditional sense of -- DR. APOSTOLAKIS: Small LOCA. DR. KRESS: Okay. Then I'm not bothered by it. DR. APOSTOLAKIS: Physical and chemical characteristics. DR. KRESS: Okay. I'm not bothered by it then. DR. APOSTOLAKIS: These are not the categories that they have in the framework. DR. KRESS: They're not related to these categories. Okay. DR. APOSTOLAKIS: Infrequent events. MS. DROUIN: That's right. DR. APOSTOLAKIS: But, again, maybe you want to think a little bit about it and the rationalization here of the framework. MR. KING: Yes. DR. APOSTOLAKIS: Also, it appears -- I think it should be given more prominence. I think all the defense-in-depth measures you are taking in this framework should be collected in one place. DR. KRESS: I do, too. DR. APOSTOLAKIS: And justified. DR. KRESS: And this is, in my mind, a defense-in-depth -- DR. APOSTOLAKIS: I have an agenda, too, by the way. DR. KRESS: -- issue there. DR. APOSTOLAKIS: If you justify them in terms of uncertainty, the whole document is rationalist. DR. KRESS: I agree. DR. APOSTOLAKIS: And I think this clearly will be justified on the basis of uncertainty. DR. KRESS: Absolutely. DR. APOSTOLAKIS: You're talking about things that are so uncertain. MR. KING: The Commission asked us to come back and present to them what is our definition of defense-in-depth and collecting all of this in one place, we are going to have to do that. You'll see it show up on the issues slide as something we're going to have to lay out so that they don't have to search for it here and there in the document. DR. APOSTOLAKIS: Also, you really don't -- this is too important, it seems to me, to talk about it matter of factly somewhere there. But I think you can clearly justify just about all the defense-in-depth measures you have here using the arguments based on uncertainty, even the high level. DR. KRESS: I would attempt to do that. DR. APOSTOLAKIS: And that would be a significant step over the rational -- DR. KRESS: It would be a wonderful step, George. DR. APOSTOLAKIS: I knew you would agree. DR. KRESS: Yes. Okay. MS. DROUIN: Okay. Now, we're going to get explicitly into 50.44. When you just look at 50.44 itself and not any related document, any related regulation or related regulatory guide or something and just look at the specific requirements, what I have here is that there are both analytical requirements and what I call physical requirements in 50.44. When you look at the analytical requirements, it's telling you to deal with a postulated LOCA. For some containment types, you go to a degraded core accident. It specifies the combustible gas. It's just hydrogen, and the source is from the fuel cladding oxidation. The amount, there's two different amounts you have to deal with, depending on the containment type. I've just quickly abbreviated here. One is a five percent versus a 75 percent metal-water reaction. When you get into the physical requirements, what I call physical requirements, there's six of them. You have to measure your hydrogen concentration, ensure your mixed atmosphere, control your combustible gas concentrations, and those are applying across all your LWRs. The last three, inerting your MARK 1 and 2 containments, installing the high point vents, that's for all light water reactors, and installing your hydrogen control system, your igniters, for your MARK 3's and your ice condensers. So at a high level, those are the requirements, the specific technical requirements. DR. APOSTOLAKIS: When was this rule passed? MS. DROUIN: There were three stages. You had the original, then you had a 1981 amendment, and then you had a 1985 amendment. DR. APOSTOLAKIS: The original is from? MS. DROUIN: 1979. That's right, 1979. DR. APOSTOLAKIS: So it was a TMI type. MS. DROUIN: Right. The 1981 -- DR. APOSTOLAKIS: Oh, just before. MS. DROUIN: Right. And then the 1981 was a TMI update, response to TMI. DR. KRESS: Are the requirements, either in this particular regulation or in some other? MS. DROUIN: No. These are -- DR. KRESS: These are in .44. MS. DROUIN: -- specific in 50.44. DR. KRESS: Are there requirements elsewhere -- MS. DROUIN: I'm going to get to that. DR. KRESS: -- that ask for purge and vent capability? MS. DROUIN: I'm sorry. Purge and vent is part of the control of combustible gas concentration. You need to do the recombiners or the purge and vent. DR. KRESS: Or the purge and vent. MS. DROUIN: Yes. MR. KING: The rule doesn't say have recombiners. It just says be able to control post-LOCA control and some plants have chosen to do the purge event. DR. KRESS: Okay. MS. DROUIN: Okay. DR. KRESS: But that's part of this rule here. MR. KING: Yes. MS. DROUIN: Yes. When you look at how the licensees have complied with 50.44, what I've noted on the left side are just the physical requirements, and when you look at the predominant means of compliance, you see some difference between what you saw in the previous slide and this slide. For example, the physical requirement was just to measure hydrogen concentration. That's all 50.44 required you to do. The way licensees have complied with that is to install safety-grade continuous hydrogen monitors. And as you go down, you'll see the same type of thing, against some of those requirements, that it doesn't match up directly to 50.44, and there's reasons for that and that's where we get into why you cannot look at a regulation in isolation. You have to see all the tentacles and look below the surface. I haven't listed them all here. It's quite extensive. I've only pulled out three just for an example. One regulation that's tightly tied to 50.44 is Appendix E to Part 50, the emergency planning and preparedness for production utilization facilities. This one is requiring your continuous hydrogen monitoring for your emergency response data system. When you go to 50.46(b), it's specifying your maximum hydrogen generation and postulated LOCAs for the purpose of complying with your ECCS acceptance criteria. That's getting into the five percent. When you look at Reg Guide 1.97, that's what is imposing the safety grade. So, again, I have just listed three of them here, but there's a lot of other regulations that are tied to this, 50.34 is another one. There's other regulatory guides. So as we talk about risk-informing 50.44, we're also talking about perhaps having to change other regulations and other documents if you want to receive the complete benefit. DR. APOSTOLAKIS: That's a purely administrative issue. It's not that these other regulations bring some other aspect of the accident or some other accident depends on this that we have to deal with. It's just administrative. I mean, somebody decided that hydrogen control is important and then that thought is reflected in a number of regulations. Is that a correct statement? MS. DROUIN: In a simplistic answer, yes, but not always necessarily so, because it might be in another regulation for something totally different. DR. APOSTOLAKIS: In a different context. MS. DROUIN: Yes. DR. APOSTOLAKIS: But in this context, the three examples you gave us, it seems to me, they all stem from the fact that somebody decided this is important and now it's reflected in emergency planning. MS. DROUIN: Right. DR. APOSTOLAKIS: Core cooling design. MR. SIEBER: Well, 50.46(b), that's an input assumption to the calculation that establishes the final acceptance criteria. MS. DROUIN: Right. MR. SIEBER: So that one is different than the other tow. The other two, to me, seem to be administrative. Somebody here in White Flint wants to know what your hydrogen is during an accident. That's why it's on ERDS. And it's in Reg Guide 1.97 because it's in ERDS, part of emergency planning, and because the rule existed in the first place, 50.44. So to me, 50.46(b) is different because then you have to recalculate to see if you continue to comply with the final acceptance criteria under Appendix K. DR. APOSTOLAKIS: There is a technical reason for 50.46(b) in terms of the cladding integrity, coolable geometry, actually. DR. BONACA: I think what you're trying to say is that in some cases, it may not be only a threat of issue. There is some other type of technical reasons for having the requirement. Is it true? MS. DROUIN: Yes. DR. APOSTOLAKIS: Yes. MS. DROUIN: Now, when you start looking at going back and trying to pull your insights, your risk insights out to see what risk-informed 50.44 or how would you create a risk-informed regulation that deals with combustible gases. When you start looking at your PRAs and you start looking at your core damage and your core melt accidents, the first thing that comes out is that the combustible gases that can challenge your containment integrity are not limited to hydrogen. You have both hydrogen and your carbon monoxide that you need to deal with and also the sources for these are not just limited to your fuel cladding oxidation. You also need to take into account your core-crete interaction. One thing I might point out here is that in terms of your core-crete interaction, that is highly a function of what concrete you use. If you're using basalt concrete, then you aren't getting your carbon monoxide generation. Also, you're generating these combustible gases, both in the early stages of the accident and in the late stages. So in dealing with trying to control your gases, you have to look at the whole timeframe. You can't just limit yourself to the early parts of the accident. When we start going through the history of our PRAs and the insights coming from the PRAs, starting with WASH-1400, is the biggest insight that came out of there, of course, is that LOCAs are not the only thing you have to deal with. You have other accidents that are contributing and you're getting hydrogen generation, you're getting combustible gases from these others, and when you particularly look at WASH-1400, it did predict a fairly high conditional containment failure probability from hydrogen combustion. Even though it wasn't the dominant contributor, it was a dominant one. You then had the Three Mile Island accident, which generated the start of the severe accident research program. Primarily what this program did was confirm the importance of your hydrogen, your hydrogen combustion, and the way that, when you look at the first -- the 1981 amendment and the 1985 amendment and the things that were put into place to control these, that these were the right measures in dealing with it. When we move on to -- DR. APOSTOLAKIS: Now, has anyone questioned these findings from the industry? This is all NUREGs and NRC-sponsored projects, but I'm sure the industry has done work. MS. DROUIN: We're going to get into some industry stuff here, too, on the next one. DR. APOSTOLAKIS: You will. Okay. MS. DROUIN: I couldn't fit it all on one slide. DR. APOSTOLAKIS: So it just happened randomly there. MS. DROUIN: Now I'm going chronologically. The next big risk study that came out was NUREG-1150. It pretty much still confirmed the same thing. You had other accidents that were contributing to your core damage. Hydrogen combustion was a dominant contributor and that the new thing was that hydrogen combustion was not a challenge to your large volume containments. MR. SIEBER: That's based on the assumption that you use ultimate strength as opposed to design strength. MS. DROUIN: Yes. MR. SIEBER: And the ultimate strength, as I picture it, has tremendous uncertainty associated with it. MS. DROUIN: That's correct. MR. SIEBER: And that's why you don't design bridges to the ultimate strength. MS. DROUIN: That is accurate. Let me move on to the next set of PRAs that came out, which are our IPEs, and you look at the insights that were documented in NUREG-1560, again, you still see a wide range of accident initiators are contributing. A thing that came out of here which we also saw in NUREG-1150 was that your hydrogen combustion or station blackout accident sequences was a significant contributor to containment failure for your MARK 3's and your ice condensers. Those two little -- those last words got left off the slide, but that's very important, because these refer to your MARK 3's and your ice condensers. Some more recent research that was going on related to the resolution of your DCH and this is a program that was looking at your challenge to containment integrity for just large drys and ice condenser containments for station blackout, and the information that's coming out of there is that hydrogen combustion can be a challenge to your containment integrity for ice condensers during SBO. DR. KRESS: How is that related to DCH? It's because it's putting the hydrogen combustion pressure on top of the DCH pressure or is it because you get additional hydrogen out of the DCH event? I'm trying to figure out why that particular bullet is under the DCH issue resolution. MR. CAMP: What happened is basically this study was done to look at DCH in ice condensers. That was the driving purpose of the study. When they did the study, they concluded that the ice bed was going to be pretty effective in mitigating the DCH part of the event. But the calculations that they did accompanying this with the hydrogen source terms that they got from RELAP/SCDAP calculations and CONTAIN calculations to look at the containment load almost always produced containment failure from hydrogen. DR. KRESS: So it was an ancillary result. MR. CAMP: It was an ancillary result to the DCH study. MR. KING: And remember in DCH there's a bunch of hydrogen that accumulates in the containment before the actual DCH event and it's that preexisting hydrogen plus the DCH -- DR. KRESS: Plus the pressurization, plus the ignition. MR. KING: Yes. MS. DROUIN: And the last point that we want to make is that as you go through and look at all these insights that have come from WASH-1400, 1150, the IPEs, they've all dealt with just internal events and when you start taking the fire and seismic into control, the accidents associated with those tend to have characteristics very similar to station blackout. So whereas you're getting station blackout to be a concern from your internal, it's even more important when you start taking into account fire and seismic. DR. APOSTOLAKIS: Is this a good time to take a ten-minute break? MS. DROUIN: Sure. DR. APOSTOLAKIS: Back at two. [Recess.] MR. KING: What time do you want us to finish up? DR. APOSTOLAKIS: Well, NEI needs about half an hour, they told me, and I'm leaving at three. So can you finish in 15 minutes or is that too much to ask? MS. DROUIN: Yes, we can. MR. KING: We can. I think most of the issues we've talked about, so the last two slides ought to go pretty quickly. But we ought to just spend two minutes on the last two. DR. APOSTOLAKIS: I'm sorry. We're back in session. Are you disagreeing with Mr. Christie's evaluation or you are not even thinking at all in those terms? MS. DROUIN: We're doing our own evaluation. DR. APOSTOLAKIS: I didn't say your own. Mr. Christie's. He presented to us his arguments why he thinks things should go a certain way in terms of the first four hours of the accident and after that and so on that it will distract the operators from other important things. So all this is in agreement with what he said or is there a disagreement? MS. DROUIN: I think we're in very close agreement. What we'll do, instead of going through each of the containment types, why don't we jump to slide 16. DR. APOSTOLAKIS: Okay. Good. I like experienced presenters. MS. DROUIN: The previous slides, it just goes into more detail, containment type, what we learned from the various PRAs and research. If you jump to slide 16 and you summarize what we've learned and what conclusions did those insights lead us to, what we're saying here is that when you look at the results from these PRAs and from research, and you look at that relative to combustible gases and the potential for combustion and its challenge to containment integrity, what we have learned from that is that a risk-informed 50.44 needs to have these kind of ingredients. In looking at the accident types, it needs to focus on what we would call the risk significant core melt accidents. It shouldn't just deal with LOCAs or just the degraded part of the accident. It should deal with the full accident and it should focus on those that are the risk significant ones. DR. APOSTOLAKIS: Risk significant with respect to public health and safety? What's risk significant for core damage is not necessarily significant for public health. MS. DROUIN: Public health. DR. APOSTOLAKIS: And Mr. Christie emphasizes the fact that we should focus on public health and safety. So when you say risk significant, you refer to those, as well. MS. DROUIN: I'm talking about those that can challenge containment and that would lead to a large release. DR. APOSTOLAKIS: So you're in agreement. MS. DROUIN: Yes. DR. KRESS: Is there any concept in here of containments leak and you put a hydrogen burn into it, that may increase that leak? For two reasons; it may make the hole bigger and it may increase the pressure to the driving force. You don't calculate a containment failure, but there's some probability that you're increasing leak and you've got fission products that you're blowing out at a faster rate. It's generally a regulatory objective to limit that kind of release, although it's not a LERF, because it's not the kind of large early release. It's a small early release. But is there any thinking in that, in risk-informing this, that you need to deal with those kinds of things, also? MS. DROUIN: I wouldn't say directly, but indirectly it gets covered. Bear with us and hopefully we'll answer it. The next major ingredient is what combustible gases do you concern yourself with, and you shouldn't just be limited to hydrogen, but also take into account your carbon monoxide and also looking at the sources, not limiting yourself to the fuel cladding, but also looking at core-concrete interaction, and the source terms should be based on realistic calculations. When you do the realistic calculation, you may end up at a 75 percent metal-water reaction. It might be less than that or it might be more, but we would perform and come up with a specified term here. DR. KRESS: Are there any core-concrete interaction sequences that contribute to LERF? I'll ask Trevor, maybe, or you. I don't know. Is that a non-LERF phenomenon? MR. PRATT: Maybe one or two in the boilers. DR. KRESS: One or two in the boilers. MR. PRATT: Not in the large volume containments. DR. KRESS: But this sub-bullet is intended to look at it in context of LERF or just to look at it in general? MR. PRATT: In terms of early and late. It's a dominant contributor to late. DR. KRESS: So you're not limiting this risk-informing just to early failures. MS. DROUIN: No, and that's where we get into -- DR. KRESS: You're including late. MR. PRATT: Yes. MS. DROUIN: That's where we get into the next bullet, is that in controlling the combustible gases, you need to look at both the early phases and the late phases of the accident. DR. KRESS: So it's not just -- so you can't just look at LERF. MS. DROUIN: Correct. DR. KRESS: Glad to hear you say that. MS. DROUIN: Okay. If you remember, earlier, we had said that 50.44 dealt with both what we call analytical requirements and physical requirements. So when we look at the analytical requirements, we are saying that the analytical requirements, in going from what is currently in 50.44, the changes that we think ought to be done, it should, again, account for your risk significant core melt core melt accidents. You need to account for your combustible gases generations from both your fuel cladding oxidation and your core-concrete, and specify the amount and rate of your combustible gas generation based on realistic calculations. So that matches up exactly to the previous slide. These were things that were specifically called out in 50.44 and these are how we would propose the alternative for the analytical requirements. When we get to the physical requirements, we've identified three alternatives. The first alternative is taking the specific requirements from 50.44 and modifying them. For example, eliminating the requirement for safety grade continuous monitors, add the capability to measure your hydrogen concentration under degraded core conditions, such that whatever instrumentation you had there, it would survive under high pressure, high temperature, whatever those conditions were. Insuring your mixed atmosphere for your risk significant accidents, whichever they are for your particular plant; eliminate your post-LOCA hydrogen control system, that would be your recombiners. It would also be your vent and purge system. Adding a long-term hydrogen control for your risk significant core melt accidents and ensuring your hydrogen control for your risk significant core melt for your MARK 3's and ice condensers. These are lining up one to one to your six physical requirements. DR. KRESS: Let me ask you about that first sub-bullet, eliminate the requirements for safety grade continuous monitors. Is that based on some criteria that says that the risk achievement worth of being safety grade is not worth having them safety grade or is this an SSC that's low on your priority list for some criteria? MR. KING: It's based on the fact that these would be used in an accident management situation and traditionally we have not required accident management type stuff to be the full pedigree safety grade. It will be consistent with that approach. DR. KRESS: Even though it might be a significant contributor to changing the conditional containment failure probability, you wouldn't. If it had real significance there, you would still eliminate it. MR. KING: See, you're not eliminating the requirement. You're just saying it doesn't have to be safety grade in terms of -- DR. KRESS: It doesn't have to be safety grade. MS. DROUIN: It doesn't have to be safety grade. MR. KING: -- all the pedigree requirements that go with it. MS. DROUIN: We still want you to measure, but we aren't going to be prescriptive that the way you go about measuring is that -- DR. KRESS: We were looking at what are the potential requirements for things to be safety grade or not safety grade not based on tradition, but if you had to do it again, and the concept was, well, if it had a pretty big contribution to RAW or Fussel-Veseley for either CDF or LERF, then maybe it ought to be called safety grade. But this is not the kind of criteria you used here at all. MR. KING: No. MR. PRATT: I was just going to add that the emphasis was moving away from it needed the continuous monitoring early to an accident management strategy later. So that gets you away from having to have the information available for immediate responses, which we don't envisage there would be any. So it would be a long-term accident management. DR. KRESS: Yes, but isn't it still needed or used in some of the emergency response criteria? You use it there. Wouldn't that call for sort of an early measure? I thought it was one of the indicators that you use to decide on emergency response measures. MR. PRATT: In that context, it certainly may be necessary. In the context of what we're looking at here, in terms of hydrogen control, there's no immediate measures that we would need, we believe, that requires this information. That's the point. But you certainly would probably want that information further, but for this particular -- MR. SNODDERLY: Excuse me, Dr. Kress. This is Mike Snodderly, Probabilistic Safety Assessment Branch. The other thing is early I think we would be relying more on the indicators of containment high range rad monitor and the core exit thermocouples and then supplemented by the hydrogen monitors as a confirmatory. As far as the emergency planning goes, right now, the way I see it, the way -- the biggest thing it's used for is distinguishing between a general emergency and a site area and is there a threat to the containment due to high hydrogen concentration. DR. KRESS: You could almost infer that from the fission. MR. SNODDERLY: Right. And I think we've adequately addressed that, as Bob Christie pointed out, in the ANO amendment, where we went from a -- we went to a performance-based requirement for hydrogen monitoring. DR. KRESS: Thank you. MS. DROVIN: The next alternative is one where we would eliminate all the physical requirements and replace it with a high level type of statement to control your combustible gases for all light water reactors for your risk significant accidents. Under this alternative, we would leave it to the licensee to demonstrate how they would meet this requirement. This is much more performance-based type of requirement. The licensee could come in and, say the way they meet that requirement is using the current physical things that are there in the plant now, or they could come in and do some type of analysis to show -- DR. KRESS: But when you looked at what they brought to you, you would probably take these bullets up above and say how did you deal with this, this. MS. DROVIN: That could be one way. DR. KRESS: It still would be the same consideration. MS. DROVIN: Or they could come back in and show that they aren't challenging the containment through some type of analysis. So, this would have a lot of flexibility in how they would go about doing it. They could be very prescriptive, or they could come back more in an analytical. MR. KING: This one has an attractive advantage for future plants. It would give them a lot more flexibility. MS. DROVIN: Yes. MR. KING: So, again, it depends on how much weight the Commission wants to put on current versus future plants. It could drive, you know, some selection of one of these alternative over another. DR. KRESS: Would you put an Appendix K like thing in there that says we want 95 percent confidence that you don't fail the containment? MR. KING: Well, we'd have to specify what the performance is that we're looking for. DR. KRESS: Yes. MS. DROVIN: The third alternative, actually what it is is what you see, alternative two is kind of a mixture of alternative one and alternative three. Alternative three is going to the strategies of the framework and demonstrating that you're meeting those strategies so that you would go through and demonstrate that your containment integrity is not challenged from your combustible gases or seeing an order of preference by limiting your radionuclide releases or your core damage accidents or the initiating events, or you come in and insure your emergency preparedness. You would go in now into your regulatory guide, and this is where you might, for example, see numbers. DR. APOSTOLAKIS: I was about to say that. MS. DROVIN: Yeah. DR. APOSTOLAKIS: But then wow, you would be risk based. MS. DROVIN: No, it's not in the regulation. It's in the regulatory guide. Then the last of this is that all of these options may require, you know, confirming changes in other regulations in order to get the full benefit. So, that's an issue we would have to look at. On that note, I'm going to get Tom wrap up. MR. KING: Let me just take the last two minutes and talk about slide 20, some of the potential issues we're going to have to address in the August paper. Selected implementation had been raised a couple of years ago when we were talking about risk informing Part 50. I think it really boils down to all of these things are going to be voluntary alternatives, but does that mean you can pick, say, 5044, for example. You can pick the current 5044, or you take a risk informed 5044, the total package, or should licensees be able to within the risk informed 5044 pick and choose the pieces they like that's going to effect -- DR. KRESS: How -- what are you going to use as guidance in deciding on this policy issue? Is this one that you just allow the commissioners to say yes or no, or do you have some thoughts as to why one might want to allow selective implementation or why not? What I had in mind there is is there some principle that if you allow selective implementation that you can pick this one, this one, and this one, and therefore, your risk status has changed to the point that it's unacceptable, or your uncertainty in that is unacceptable. Is there some thinking about what you would use to, as a criteria, for deciding on selective implementation or not? MS. DROVIN: I'm giving my personal thoughts here. I think that if you allow selective implementation within a regulation, you're not risk informed because the risk inform is supposed to focus in on the risk significant. In some cases, there may be cases where to be risk informed, it's just reducing things. There may be cases, as you see in 5044, to be risk informed, you're going to take some things away, but you're going to have to add some things. There were things missing. DR. KRESS: So, you're thinking it's either a whole package or not at all? MS. DROVIN: And I think that if you allow selective implementation, then you would just go after those things that are going to reduce, but the things that were safety significant -- MR. SIEBER: Compensatory. MS. DROVIN: -- may not be selected. I don't think that's a risk informed regulation. MR. KING: We have already said risk informed is a two-edged sword. When you go in and, well, the risk information can lead you to -- DR. KRESS: I was assuming both edges of the sword would show up in each individual regulation, and I may be wrong there, so that I could choose this regulation, but I have to choose all of that particular regulation. I get both ends of the sword there. MR. KING: Yeah, yeah. DR. KRESS: And then I would allow selective implementation because -- MR. KING: On a regulation by regulation basis. DR. KRESS: Right, regulation by regulation basis. MR. KING: Yeah, I think that's Mary's personal opinion is what she likes, and I think we're going to have to go to the Commission with a recommendation and talk about the criteria to why we picked that one, whatever it ends up being. MR. SNODDERLY: Sorry, Tom. I think, Dr. Kress, a concrete example of that would be the Mark 3 containments. So, right now, we had igniters for Mark 3 containments, but from a risk study, we knew that 96 percent of the risk is from station black-out sequences. So, the igniters aren't available during -- they're not there. So, it doesn't make sense to say okay, we're going to risk inform this regulation, and the Mark 3 owners are going to say hey, you know what? I like that I don't need recombiners, and that makes sense, but does it also make sense, as you've gone through this process to say, shall we continue to allow this anomaly to exist? Now, there's a lot of other things to consider, but I think like that's a hard example of what, you know, that we -- one example. MR. KING: Okay, and it's related to the backfit question. If you're going to do a backfit analysis, are you doing it piece by piece, or are you doing -- taking the whole rule as a package and doing the backfit and looking at the aggregate of how things come out. Also, the backfit, one of the stakeholders raised the question of the reverse backfit test. If we are, you know, ending up net burden reduction, should there be some, you know, tangible benefit in doing that? Should there be some criteria for which to do that? So, that's part of the issue when we talk about backfit. The risk informed guidelines we talked about, that's the issue of using the safety goals and the subsidiary objectives as the guidelines for making the changes. I mean, that has to be a Commission decision versus future plants, how much weight do you give from one to the other. Technical issues, long-term containment performance -- one of the NEI comments was they didn't think that ought to be a consideration. We're going to talk to them about that tomorrow. Defense in depth, it's the issue, George, you talked about, rolling those considerations up as something the Commission can deal with is an issue. Safety margins is the same thing, and treatment of uncertainties is tied in with that. So, a number of things that we're going to have to lay in front of the Commission that aren't just unique to 5044. I mean, these things are generic to this whole risk informed process. With that, I think we've talked about the schedule already, so that's it. DR. APOSTOLAKIS: Any questions? Any comments from the members? DR. KRESS: Well, I'd like to offer my encouragement to this framework. I think they're going in the right direction. You know, I may argue with some of the details right now, but yeah, I think it shows some good thinking and some good progress. I like the approach, at least. DR. APOSTOLAKIS: Speaking of the approach, the three alternatives that you put up there, you will evaluate those in some way at some point, or are there preliminary alternatives? Is that the idea? MR. KING: Right now they're preliminary alternatives. I think there's an open question. We go to the Commission in August. Do we recommend one of those, or do we recommend to the Commission that each of these has some merit, and maybe the next step ought to be an advance notice of proposed rulemaking to get feedback on those three alternatives. DR. APOSTOLAKIS: Once you do any sort of analysis to -- MR. KING: Enough analysis to say that they're feasible. We don't want to give the Commission unfeasible alternatives. DR. APOSTOLAKIS: No, they are feasible, but I mean, if I go with alternative two versus alternative three, what am I doing to the condition containment failure probability, for example? MR. KING: Pros and cons kinds of things, yeah. DR. APOSTOLAKIS: You will do that kind of thing? MR. KING: Yeah. DR. APOSTOLAKIS: Quantitatively. MR. KING: As best we can. DR. APOSTOLAKIS: As best as possible. I knew that was coming. Okay, thank you very much. Any other comments? No? DR. KRESS: I also wanted to mention, I would hope at some point along the line that uncertainties get more explicitly built into the criteria. DR. APOSTOLAKIS: I think in the July subcommittee meeting, we'll probably have a lot to say about that. DR. KRESS: Yeah. MS. DROVIN: We skipped over that slide. DR. KRESS: Yeah, okay. Maybe so. I'll look through your slides. DR. APOSTOLAKIS: Well, I don't know whether we want to ask how this is affecting the rulemaking, but Mr. Christie requested. DR. KRESS: Well, that wouldn't hurt to ask. DR. APOSTOLAKIS: Whom do we ask? Do we ask you or the NRR? DR. KRESS: I think ask these people over here. MS. DROVIN: Ask Cindi. MS. CARPENTER: This is Cindi Carpenter from NRR. Basically, we have the petition for rulemaking. It was noticed, and we're waiting -- it basically is being incorporated into what they're doing in Option 3. Option 3 in 5044 is going to provide the technical basis for how we resolve that petition for rulemaking, so they're very well aware of what Mr. Christie has proposed, and we're waiting on the outcome of that right now. DR. APOSTOLAKIS: So, the technical basis you are referring to is the August document? MS. CARPENTER: The August and probably the December document also, I'm thinking. MR. KING: The August document -- MS. CARPENTER: The August, okay. MR. KING: -- would be the one that deals with this issue. DR. APOSTOLAKIS: Okay, thank you. DR. KRESS: I'm glad to see the National Laboratories so well represented. DR. APOSTOLAKIS: So, the next presentation is from NEI, and it's on a different subject, 10 C.F.R. 50.69 and Appendix T. DR. SHACK: Option 2, Option 3, we cover everything. DR. APOSTOLAKIS: Boy, this is a dynamite subcommittee meeting. Mr. Floyd, would you introduce your colleagues, because we have not seen them before. MR. FLOYD: Yeah, right. To my left is Adrian Haymer, and to my right is Biff Bradley, all from NEI. DR. APOSTOLAKIS: Wonderful. MR. FLOYD: We're actually going to talk about two topics today. We have some brief comments on Option 2, and we also have some brief comments on Option 3 to cover. DR. APOSTOLAKIS: Very good. MR. FLOYD: Okay, with respect to Option 2, just wanted to advise the committee here that we have submitted the PRA certification industry peer certification process to the Commission formally for review. It's in a document called NEI-00-02. All plants will be peer reviewed by the end of 2001. Our status right now is that all the boiling water reactors have completed the peer certification process, and I don't know, maybe perhaps as many as 25 percent of the PWR's at this stage, but they all are scheduled to be completed with their review by the end of 2001. DR. KRESS: Is your intention to say that peer review would correspond to Category 2 or 3 in the ASME guide? Say that you would like to say go through this peer review. Then we correspond to Category 2 or 3, or are those completely separate? MR. FLOYD: They're somewhat separate because we don't know what the final form of the ASME standard will be. What we're saying and what we suggested in our letter is that the peer certification process is adequate and appropriate for Option 2 of risk informing Part 50. DR. KRESS: Okay. MR. FLOYD: It may comport with what the final ASME standard comes out with in one of the categories, but we don't know that at this point. MR. BRADLEY: It's an application specific review. MR. FLOYD: Yeah, and we're saying that only for this Option 2 application. DR. KRESS: Okay. DR. SHACK: Just, I got confused about this this morning. Since you're not grading the whole PRA, you're doing it element by element, do I then get a list of elements that have to be grade three or above to support Option 2, or do I get an overall road map that really tells me now? MR. BRADLEY: I think those are issues that we're going to get into when we get into the review process with NRC. I think our initial thinking is, you know, there is no overall grade. There's a grade on each sub-element, and generally speaking, I think we're talking about something, you know, for most of the sub-elements, a grade level three in the certification process to support Option 2. One of the things we've got to do is coordinate or correlate our Part 50 guideline document in that whole process with the certification process so that as we do that, it will become more clear which elements are most important. I think the general thinking of what I saw from NRC, and probably we don't disagree, is that we're looking at a level three for most, if not all, of the sub-elements to support that. DR. KRESS: My first thought, Bill, is I was going to find out, look and see how much -- which elements are more important, put a weighting factor on and take the grades and get an overall grade. Anyway, they're probably doing something like that but not putting numbers on it. MR. FLOYD: And I think the third bullet really goes to what we've been talking about, is at the meeting that we had this week, the NRC did lay out at least the outline of what the review process will be for the peer certification process. The other item we had down here, as you know, the South Texas nuclear project has an exemption request in, and of course, one of the concerns the staff has is how does that relate to the overall Option 2 generic rulemaking issue. South Texas has prepared a matrix comparing their risk categorization and treatment process to the NEI guideline process approach. We haven't had a chance to fully review that, but in general, it looks to be -- looks like they captured the essence of it. The conclusion we think we would agree with is that, although when you get down at the level of detail, their approach is a little bit different. When you get at the how-to level, the concepts and the philosophy that are embedded in the guideline document are consistent with the approach that they had. That wasn't by happenstance. Of course, we were cognizant of what their approach was when we were developing the guideline document, as well as being cognizant of the ANPR and the SECY papers. Our guideline document has been evolutionary to accommodate those changes as they've gone along. The other thing I wanted to say in general about Option 2 here is that, just give you an update on the pilot status. NEI has been coordinating with the owners' groups a pilot project effort to test out a number of systems at lead plants in each of the owners' groups. So far, the Westinghouse, BWR owners' group and CE owners' group have all now voted to fund a project. The scope of the projects are very similar in scope and schedule. They will be starting roughly in the September time frame, late August, early September time frame, and having the preliminary bulk of the work done by around the end of the year and then writing up an evaluation report. All three owners' groups will be using the NEI guideline as their template for conducting their review to see what enhancements and revisions need to be made to the guideline document as we move forward. Some other Option 2 issues, we've identified a couple of issues, I think, the we would like to get an early read from a legal perspective. The first one is the differentiation of design basis from special treatment. Let me see if I can explain what we mean by that. Take, for example, the environmental qualification rule. We view that as a special treatment requirement. We hear some in the staff say that no, part of the design basis is the specific special treatment requirements that are embedded in the 5049 regulation. That makes a big difference on how we write our guideline and what special treatment provisions you put in place, and also it makes a big difference as to what the bottom line benefit of Option 2 is, depending upon the resolution of that issue. So, we'd like to get a fairly timely read from the agency on what is their position with respect to the relationship between special treatment requirements and the design basis requirements. Are they one and the same, or are your design basis requirements -- I mean, our view is that, for example, in the EQ area, if you have a pump that has to deliver 100 gallons per minute to each steam generator under a harsh environment, that's the design basis functional requirement. The special treatment regulation is the how do I provide that assurance that it can do that. There may be other ways of providing them, and we're really talking about the RISC-3 category here, by and large, which is the low safety significant but safety related category. Do you need that same degree of pedigree that's embodied in 5049 to provide a reasonable assurance that that design basis function is agreeing under those high energy line break conditions can be met. DR. KRESS: It seems like if you're going to be risk informed, that that's the appropriate way to view it. MR. FLOYD: Well, that's our view. In fact, it's almost an oxymoron to say that if the intent of the Option 2 approach is to grade the pedigree of treatment, but you can't do that because the pedigree of treatment is indeed part of the design basis, and you can't change the design basis, you, in essence, don't have an Option 2. We just want to get it nailed down and get agreement on that. We've heard conflicting opinions on that. The second issue is related to Part 21, and its applicability to really RISC-2 and RISC-3 boxes. The position that we've taken in the guideline is that Part 21 is certainly applicable to the RISC-1 box. It's not applicable to the RISC-2 box because the RISC-2 box doesn't contain components that pass the definition of what is a basic component. Our position in the RISC-3 box is that Part 21 also would not apply because the test in Part 21 is you have to have a substantial safety hazard as a result of the deficiency. We're talking about items that if we have agreement on the risk categorization process, we've defined that if they were to have a problem, they don't create a substantial safety hazard. So, that's our position, and we, again, would like to have an early-on legal read of that from the agency, again, because it has great bearing on how we develop our guideline document and the overall benefits of Option 2. Next issue we have is commercial treatment for the RISC-3 category. This is where, obviously, all of the discussion or the bulk of the discussion has gone on to date. We have provided in our early draft of our guideline to the staff what is known as an Appendix A, which they characterized earlier this week as a good outline of the key elements that should be in a commercial program. The feedback we got was that they would like to see some additional detail in the document, not so much on the how to you provide, the assurance that the design basis function will be maintained, but what are the desired results and outcomes that you'd like to see from that. So, we're looking now to go back and see if we can add some more detail to the Appendix A that focuses on desired results outcomes, without getting into the how to. This is the difficult road to walk down because we could very easily wind up replacing 18-point criteria in Appendix B with a very detailed and prescriptive commercial program if we go too far in one direction. Then you've eliminated, again, the benefit of trying to reduce emphasis on low safety significant SSC's. DR. KRESS: The parts that you are going to -- the components that you're going to get from the commercial program, it seems to me like there ought to be some thinking about what is the actual reliability. There ought to be enough data out there to say that in general, the reliability of commercial parts for this particular pump or whatever it is has this mean and this variance, and therefore, it very well meets the requirements I have on my PRA that shows this is not a safety significant thing. You wouldn't have to go into the -- you know, it's sort of a Bayesian approach. You wouldn't have to go into the commercial process and what they do and what their quality controls are. You just use the database. Is there any thinking along that line? MR. FLOYD: Yeah, there is. We actually see two elements in it. That is certainly one element, and we think that's appropriate for those design basis functions which can be routinely monitored through normal plan operation. That's an appropriate way to treat those. We acknowledge that there are other certain design basis functional requirements in harsh environments, for example, or seismic conditions that you're not going to be able to do that monitoring for. So, then you need to have some design control and procurement control aspects put in place to assure that the equipment is being ordered with those considerations in mind. MR. BRADLEY: There may be uses in other industrial applications where the equipment, it spurns routinely to a harsher environment, and you might be able to get some benefit out of that. DR. KRESS: Yeah, that's what I had in mind, other industrial applications. MR. BRADLEY: Right. We will be looking at that. MR. FLOYD: Right, looking at those. DR. APOSTOLAKIS: You mentioned the benefits. Do you have any idea what -- I mean, if the industry implements this, what the benefits will be to them in terms of dollars? MR. FLOYD: The only plant that's really looked at in much detail is south Texas, and they're telling us that this could have benefits of, you know, in the $2 million, $3 million per year potential in terms of, largely in the procurement related area, procurement and inventory control. DR. APOSTOLAKIS: So you think that will be more or less the ball park? MR. FLOYD: It's hard to say. That's really what the -- because they're a little bit unique because of their late vintage of their licensing basis for their plant. They have a lot more safety related SSC's identified in their licensing basis than a typical plant has. That's really what the pilots, I think, are going to tell us, which is why the owners' groups are very interested in having a scoped pilot effort that will give them some of those answers. DR. APOSTOLAKIS: It might be less for others? MR. FLOYD: It could be less for others, and it may turn out that that is representative. We just don't know yet. DR. APOSTOLAKIS: And after these savings are realized, then people will spend a billion dollars to have a good level three PRA perhaps? MR. FLOYD: Well, you know, we are going to talk about that a little bit in just a few more minutes, but no, I think -- DR. APOSTOLAKIS: Category three. MR. FLOYD: This is a market-drive, I think, application in general, and I think if people start seeing the benefits of applying risk technology, and the benefits are real intangible, you're going to see people embrace them more strongly and be willing to make a greater investment in it. DR. APOSTOLAKIS: And after they do that, we'll need a new ASME standard because everyone will be category three. DR. KRESS: We'll have to have a category four? MR. BRADLEY: Don't get too far down the road. We've got to make the baby steps first here. MR. FLOYD: The other issue we have on Option 2 is treatment of prior commitments. As we know, and I think as we talked before at this committee, a typical plant has somewhere around eight to 10,000 commitments. These are things that are beyond strict regulatory requirements. Many of them are the way the licensee has chosen to meet a specific regulatory requirement, but nonetheless, that's about a ball park number in the order of eight to 10,000 per plant. What we would hope could happen under this Option 2 approach would be a replacement, a blanket replacement, of those current commitments, which by and large, went to special treatment considerations and, you know, pedigree of equipment, with an acknowledgement that the new risk categorization process and the new defined treatment process for the various boxes is indeed a replacement for the previous commitments, without making a licensee go through the commitment management guideline, which has been endorsed by the agency, and write up a piece of paper for each of those 8,000 commitments to disposition it. We'd like to see a blanket replacement. Otherwise, it is a fairly extensive bureaucratic process to go through to officially eliminate those commitments from the docket. DR. BONACA: Just before you leave this overhead, I would like to ask you about Part 21. The benefit there is that probably in the procurement components which are not subjected to the requirement. MR. BRADLEY: Yes. MR. FLOYD: Yes. MR. BRADLEY: Well, Part 21, it's imposing a significant liability on the vendor or the dedicator, so even if we can have -- let's say relaxation of other requirements, if Part 21 is still there, much of the benefit is not going to be gained. DR. BONACA: Certainly, however, there is still a benefit to the industry of communicating problems of the components, and maybe they could be communicated in other vehicles than Part 21. MR. BRADLEY: Yes, we've thought about that as well, that maybe, you know, there's ways to meet the intent without imposing a type of liability that drives the parts prices up. DR. BONACA: Or eliminate suppliers, actually, because I mean, some of them don't want to supply any more. MR. BRADLEY: Right, and this is an issue. On RISC-2 where you may say, you know, in a BWR, for instance, at the condensate system or feedwater system is risk significant, you know, trying to impose Part 21 on those types of vendors would probably be very problematic for the industry. MR. FLOYD: As a close-out for Option 2, I think I'd be remiss if I didn't say that I think we've been having some very good meetings with the staff on Option 2. Obviously, the industry has some more gelling to do. The staff has some more gelling to do with respect to positions, but philosophically, I don't think we heard anything in our last meeting this past week that is greatly out of sync with what the industry thinks is the necessary things to do with respect to treatment for the SSC's under Option 2. So, I think we're on the right pathway. I'd like to switch now to Option 3. We do think that there has been a tremendous amount of thought and effort put in by the NRC staff in establishing the Option 3 framework. Some of the comments specifically that we have, and we provided these in our public comments, we tend to see that -- we think this is leaning a little bit more towards risk base than risk informed. At least our impression. It may be a wrong impression, but there seems to be more emphasis on the bottom line number results and less on the integration of the thought process that's embodied in Option 2 approach with the integrated decision making process. DR. APOSTOLAKIS: I didn't get that impression from the representation. MR. FLOYD: Okay. DR. APOSTOLAKIS: But again, risk based is something that should be avoided at all costs, right? MR. FLOYD: It comes across more in looking at the framework than I think it does on the individual rulemaking efforts, you know. This is more a comment on the framework document, I think, than -- DR. APOSTOLAKIS: Well, they go out of their way, as I pointed out, than earlier, to state in several places that risk numbers will not appear in the regulations. I thought that was excessive. So, I guess it's in the eye of the beholder. An excessive number of times. I mean, they can say it once. Not that they went too far, although that might be the case, too. I don't know. We're going to have a subcommittee meeting July 11 at 1:00 to discuss the framework here, Option 3. It was decided today. MR. BRADLEY: That's one none of us can support. We have a -- MR. FLOYD: Right, we have a risk informed regulation working group meeting on the 11th. MR. HEYMER: And a date with senior management. MR. FLOYD: And senior management, PRA steering committee meeting with the staff on Monday morning. MR. BRADLEY: One thing, since you mentioned this, it might be helpful is we could -- there's so much going on in risk informed right now. One of the reasons we couldn't be here all day today is we have many meetings going on within the industry. If possible, if we could coordinate our schedules in advance, we might be able to support these meetings. DR. APOSTOLAKIS: I believe that is an excellent suggestion. Let me tell you about today's meeting. I mean, it was scheduled, really, this afternoon when we realized that the emphasis on the staff's presentation today was on 5044. You know, the committee really has not been given the opportunity to discuss the details. So, it was something that it was decided at the last moment, but I think this is a great idea because, you know, I at least learned a lot by hearing different viewpoints. MR. FLOYD: Right. DR. APOSTOLAKIS: It's unfortunate. Maybe you can go back and look at your commitment and maybe one of you can show up for awhile. DR. KRESS: But that seemed to be about the only date we could come up with. DR. APOSTOLAKIS: Yeah, this is the only day, you see, because it's just before the full committee meeting. MR. BRADLEY: Well maybe just make an effort to share our schedules and try to do that. DR. APOSTOLAKIS: No, we will. Are you in town that week? MR. BRADLEY: They are in town that week. I'm on vacation. MR. FLOYD: We're partly in town that week. DR. APOSTOLAKIS: I mean, if you can show up around 4:00. Do your meetings go late? MR. FLOYD: They go -- this one is from one to 5:00 in the afternoon, since we're meeting with the PRA steering committee in the morning from nine to 11:00. DR. APOSTOLAKIS: Anyway, that's -- MR. BRADLEY: Your full committee meeting is the 12th? DR. APOSTOLAKIS: Yes. You plan to be there? MR. BRADLEY: Yeah, that's probably a date we can -- I can't speak for you guys. MR. FLOYD: You might have to cover it, Adrian. DR. APOSTOLAKIS: Okay, so maybe then we can give you some time. DR. KRESS: That would be helpful. That would be in time before we wrote our letter. DR. APOSTOLAKIS: Yeah. MR. FLOYD: As you know, it's hard to schedule all of these meetings that are going on and all the parties that are interested in what's going on in risk informed regulation. That is a challenge to get the schedules to match. DR. APOSTOLAKIS: That's right, but especially for issues like Option 3 -- I mean, on Option 2. It seems to me we really have to go through and scrutiny every detail because these are major pieces of regulation. MR. FLOYD: Right, significant changes. DR. APOSTOLAKIS: Now, the second bullet -- MR. FLOYD: Yeah, let's get to that one. DR. APOSTOLAKIS: What is it that bothers you there, the fact that they are using the safety goal numbers -- MR. FLOYD: What's bothering us -- DR. APOSTOLAKIS: -- as given by the Commission, or the idea of using safety goals? MR. FLOYD: No, no, not the idea of using safety goals. It's the way in which the safety goal is being applied. Our understanding is that the current policy statement is that safety goals are to be applied on the general fleet of plans and are not to be used on an individual plant basis. Our read of the framework document seems to be driving it more towards the intent to use it on an individual plant basis, and that hasn't been a policy decision that has been made yet. DR. APOSTOLAKIS: If you use 1174, you're using it on an individual plant basis. DR. KRESS: That's much more so there than you are in this. DR. APOSTOLAKIS: But those numbers there come from the QHO. MR. FLOYD: Well, 1.174 is really for evaluating individual changes that you're making -- DR. APOSTOLAKIS: But individual plants -- MR. FLOYD: -- but it doesn't specifically say anywhere in Reg Guide 1.174 that you cannot exceed either the minus four or ten to the minus four core damage frequency. DR. KRESS: My reading of the framework document would say that they're using these safety goals just like they were intended, and that is to gauge how well they're writing the regulations to get the fleet of plants on an average down there. They won't say each individual plant, you have to meet this safety goal. They're going to say we're going to write the regulations so that on the average, there is a tendency to, if they follow these regulations, there is a tendency that they might meet the goals, but not each individual plant. That was my reading -- DR. APOSTOLAKIS: Yeah. DR. KRESS: -- that they were just using them just like they were intended to be. DR. APOSTOLAKIS: I had a different reading. I thought that before you know it, if you use them like that, they will become measures of adequate protection, and that ws not the original intent. MR. FLOYD: No, it was not the original intent. MR. BRADLEY: Well, I guess the fact that we have different readings says something, that maybe -- DR. APOSTOLAKIS: That we should have a subcommittee meeting. DR. KRESS: To clarify this. MR. FLOYD: It's an issue that needs to be in here. DR. APOSTOLAKIS: No, this is an important issue in my view. MR. FLOYD: Right. DR. APOSTOLAKIS: This is one of the most important issues -- MR. FLOYD: Right. DR. APOSTOLAKIS: -- because just because the numbers are there and you don't have any other numbers, it doesn't mean you have to start using them to do certain things for which they were not intended. I thought the issue of the applicability of the goals in the individual plants had been settled. We were discussing this. MR. KING: It was settled in 1174. It was raised as a policy issue, and the Commission said yes. They said yes in the context of 1174. They haven't said yes in the context of revising the regulations. DR. APOSTOLAKIS: So the Commission is allowed to have selective implementation. MR. KING: But you're only going to raise this as a policy issue. I mean, I think clearly, this is something the Commission needs to weigh in on. MR. FLOYD: Especially if you're going to be risk informed and you realize that not every plant needs a ten to the minus fourth core damage frequency to meet the quantitative health objectives and the safety goal. DR. APOSTOLAKIS: That's my concern, too. MR. KING: That's our concern. DR. APOSTOLAKIS: It may turn out to be a measure of inadequate protection, not adequate. That's not the original intent. See, if it's adequate protection, it's risk based. If it's a measure of inadequate, then it's still risk informed. MR. BRADLEY: And 1.174 was written to go beyond adequate protection. DR. APOSTOLAKIS: I can play the word game as well as you can. I'm sorry, Biff. MR. BRADLEY: 1.174 Contains a discussion of the fact that it was intended to establish measures beyond adequate protection. It's not intended to define what is adequate protection. DR. APOSTOLAKIS: That's true. MR. BRADLEY: Yeah. MR. FLOYD: Our third bullet really goes -- is a corollary to the second bullet, and it's the policy issue. We think that is a departure to apply again the subsidiary objectives of the safety goal on an individual plant basis, and that was our read as to what it looked like it was going towards. The last bullet, I think, is perhaps one that's a little bit interesting, and it goes to what Tom was talking about in his presentation about the application of the backfit rule. I don't think anybody in the industry disagrees that if there is an outstanding issue out there that does show a cost benefit for being a provision in one of the Option 3 risk informed changes that it's appropriate to put it in, that's fine. We don't see any clear guidance yet in the framework document, which would suggest that additions that want to be made to requirements that aren't included today would, indeed, have to pass some sort of threshold criteria, and what is that criteria before they would be imposed under the Option 3 framework. I mean, our understanding is that under the Option 3 framework, we are looking for, you know, improvements in the effectiveness and the efficiency of the regulation by removing unnecessary burden and by focusing the regulation on that which is important. While our understanding is that with this double edged sword, we're only going to be allowed to make minor changes, we think if there are significant additions to the requirements, they ought to have a criteria of what level of significance do they have to add value for. Otherwise, they just become potentially another candidate for a reduction at a later date because they're not effective, and we don't know what that criteria is at this point. Overall, and these words probably more come from our working group than anybody else. What we're really looking for in Option 3 is a pragmatic versus a more theoretical approach. We really think that the best approach in Option 3 is to use generic risk insights. In fact, when we think about what the role of the regulations is to be applicable to a fleet of plants, we're not sure that you need a great deal of plant specific detailed PRA information to adopt Option 3. What you ought to be looking at are broad, generic insights from PRA results across a fleet or a subset of the fleet of plants rather than being too focused on the actual plant specific PRA results. Now, there may be certain cases where that's appropriate, depending on how the regulation is written, but in general, we think it ought to be a generic application for a generic rulemaking. As we've said, we want to make sure that we do preserve the risk informed philosophy and make sure we don't get too risk based in this by focusing too much again on numbers and plant specific PRA results. DR. SHACK: Would that mean, for example, that you don't particular like the alternative two approach to 5044, where it would be very plant specific? MR. FLOYD: Yes. We think it would be better to have a generic rulemaking which stands back and looks in the big picture, what are the important things for -- maybe it doesn't apply to all the fleet of plants, but maybe you have to break it up into subsections of a fleet of plants, that these are the items that our insides tell us are important to pay attention to from a public health and safety perspective, and then write a generic rulemaking around that framework. DR. SHACK: Just coming back to your bullet in the previous page where you've objected to the quantitative basis. Now, is that the particular number or just the notion of a quantitative basis? MR. FLOYD: No, I think it's the particular number focused on a plant specific result as opposed to -- yeah, certainly we think it may not be inappropriate to have a quantitative basis for it. That's applied generically based upon insights from a population of plants. Biff and Adrian, jump in here is you have any problems with that. No, okay. Our recommendations under Option 3, as we had in our letter, were to complete the ongoing efforts that are going on in hydrogen control and the fire protection, 50.48 Appendix R effort. The greatest -- we did a survey of the industry last fall, and all our indications are is that the results of that survey are not any different today. The feedback we're getting is that people think the areas that are most right for improvement are codes and standards under 50.55(a) and the large break loca analysis, 50.46. DR. KRESS: Can I ask you a question about that? Let's take the large break loca. We've got the 50.46 rule and Appendix K, which imposes certain requirements on the plants. Now, if they're going to risk inform that, then there will be another kind of rule with different requirements imposed. In order to get a benefit estimate, you must have had some sort of notion of what this new rule is going to look like in order to determine what you're going to be allowed to change and do. I'm not sure, without your having the actual new risk informed rule, I'm not sure what assumption was made in, say, the NEI letter where these showed up. How were these cost estimates arrived at? That's my question. MR. HEYMER: On the 50.46, most of the responses were from -- not all the responses, but most were from the Westinghouse plants, who used the data that they developed under the owners group approach to define what the benefit might be from the Westinghouse owners group activity associated with the large break loca. In that, they'd made some preliminary assumptions and from their preliminary work, they thought that they could get down to a specific break size in redefining the large break, and based on that. DR. KRESS: They made some assumptions. MR. HEYMER: Yes, they made some assumptions. Now, obviously, the larger the break size, perhaps the lower the benefit, but I mean, that's what they based it on, and some of the other owners groups who provided information to us said they didn't think they could go that way, but they felt there may be some benefit if they applied the follow-on activities from a large break. Whether or not they came down to the same size at the Westinghouse owners group was a matter of debate. So, there was some provisional -- it was preliminary. We acknowledge that, but at least it gave us some indication of where we might go. DR. KRESS: It gave you an indication. MR. HEYMER: Yes. DR. KRESS: Probably good enough numbers for doing prioritization, anyway. MR. HEYMER: Yeah, and then once you get into it, you know, you re-assess that as you go along. DR. KRESS: As I said, after the rule just some conformity. MR. HEYMER: And obviously before the executives say move on. They want some harder numbers still. DR. KRESS: Thank you. MR. FLOYD: Some of the non-Westinghouse plants also were looking at elimination of the requirement to consider a loss of off site power coincident with the double ended guillotine break as one approach, as well as removing some of the conservatisms, the known conservatisms in the Appendix K analysis. DR. KRESS: Eliminating the loss of off site power. Is that because you've got it down to such a low probability that it's not -- MR. FLOYD: Yes. If you look at the probability of the large break size coincident with the loop, you're pretty far down. DR. KRESS: Pretty far down. Okay, thank you. MR. HEYMER: Just a quick point on the codes and standards. It wasn't really clear from the survey when we went back to people and asked questions whether the benefits that they were quoting were really just from what they might achieve under Option 3 or what they might achieve through Option 2 as well. Some people who put numbers in there hadn't implemented ISI or IST-type of activities. MR. FLOYD: And then obviously, the desired approach that the industry really has is they think if we could focus on the codes and standards and large break rule, that that will probably flush out the preponderance of the policy issues and the degree of rigor that will have to be demonstrated to make these changes, and then they will be in a better position at that point to look at prioritization for remaining regulations. We actually think while that sounds like that's a slowing down of the effort, we really think that would actually result in a speed-up of the overall effort because we'll either know the benefit is there and as the desire to move forward or not. When we get the results from these first two efforts, then we'll know the magnitude of the policy issues that will have to be likely dealt with. This next one, again, this is just a perception. Perhaps it's because we've had a lot more dialogue with the staff on Option 2 than Option 3. The timing is a little bit different on the two efforts, at least with respect to our involvement and interactions. We do see, at the front end, maybe a little bit different and fundamental approach. We see the research effort perhaps, again our view being a little bit more focused on the plant specific PRA results and the importance of the number. That's an impression that we have, and we see the NRR side. While that's certainly an element, we see them very much more concerned, perhaps, about some of the other elements that are not covered by the PRA being brought in. Perhaps, as I said, it's a function of the level of dialogue that we've had to date on Option 2 versus Option 3 that's driving that, but that's an early perception that we have. If that fundamental difference does indeed exist, then we think it would be of great benefit to sort that out and make sure that we have a consistent approach from both offices in the agency. I will comment, and I think this is just the nature of the beast that we're dealing with on the second bullet, that most of the meetings we've had to date on Option 2 really wind up focusing on what do you do with the stuff that your risk categorization process tells you is not very important, and sometime I know we're going to have to get over that, but at some point in time, we really need to start paying attention to what's the stuff that we all think is important and making sure that we're dealing with that correctly and figure out how to not spend so much time worrying about the unimportant or less important stuff. DR. SHACK: How comfortable are you with you agreement over what it's going to take to decide what's important and what's not important? MR. FLOYD: Do you want to address that, Biff? MR. BRADLEY: How comfortable we are with the agreement? DR. SHACK: The categorization. MR. BRADLEY: Well, I think really that's the driver of this issue. To some degree it's a degree of confidence in the categorization process. DR. SHACK: Oh, well I thought the argument was on the actual treatment that you would give to the low risk significant. MR. BRADLEY: Right. DR. SHACK: But not whether on what you decide. MR. BRADLEY: We're more comfortable with the categorization element than we are with the treatment element. It seems that -- I think, you know, we've done categorizations of more of a limited scope thing, a maintenance rule, ISI, IST. Those are, you know, categorization, graded QA. There's some sense of you can do that and success. There seems to be more controversy and more difficulty trying to decide what do you do with it once you've binned it, you know, and there's not a -- you know, you can always be really conservative in the categorization process to take care of any disagreement, but then once you've done that, you know, that in itself doesn't give you anything. You've got to do something with the two bins, and that's where the difficulty is, more controversy. So, it's also where the whole benefit would evolve. So, just having an agreement and a mutual understanding on how to categorize doesn't tell us everything we need to know to have success, you know, and widespread buy-in into these efforts. I think we'll get there. It's always difficult, and there's a long history to some of these things, and change is difficult, but you know, I believe we can get there. Steve's right, you know. There seems to be an inordinate focus on things that everyone seems to agree on low safety, you know, significant in the categorization process. MR. HEYMER: When the NRC staff are planning some visits to plants to look at some of the programs that will be applicable to the low safety significance, and I think once they've seen those, I think it will ease the path of resolution. There will be a better understanding from the NRC side of what we really mean by commercial or commercial nuclear practices. I think one we've got through that phase, there will be a much better understanding, and we'll have a better idea of where we're going and how far we can go. MR. FLOYD: And just a final observation, and it really goes to what Dr. Apostolakis referred to just a little bit ago, and that is there is, we think, a limited window of opportunity for getting general support within the industry for these initiatives. It's expensive to be a pilot plant. It's expensive to kind of go out on your own, as South Texas has. The longer that review takes, the more resources that are applied to it. If we could show some timely resolution on 5044 and action on the South Texas exemption request, I think that will spur the rest of the industry on to be much more interested and engaged in applying risk technology, and will probably promote better PRA models and a desire to upgrade PRA models and get better information, which personally, we think, is to the benefit of the overall industry and public health and safety. We think there's just so much information that's available from the insights that you get from risk analysis, that we'd like to see a strengthening of those abilities within the industry. That concludes our formal remarks. DR. KRESS: We certainly thank you and ask if any of the committee members wish to ask any more questions. Any comments from research or NRR? Anybody in the audience want to make a comment? Seeing none, I wish to thank the NEI representatives. We hope you can make our full committee meeting on the review of the framework document. With that, I'm going to declare this meeting adjourned. [Whereupon, at 3:05 p.m., the meeting was concluded.]
Page Last Reviewed/Updated Tuesday, July 12, 2016
Page Last Reviewed/Updated Tuesday, July 12, 2016