Protecting People and the EnvironmentUNITED STATES NUCLEAR REGULATORY COMMISSION
207
1 UNITED STATES OF AMERICA
2 NUCLEAR REGULATORY COMMISSION
3 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
4 ***
5 475th MEETING
6 ***
7 Two White Flint North, Room T2-B3
8 11545 Rockville Pike
9 Rockville, MD
10 Wednesday, August 30, 2000
11 The committee met, pursuant to notice, at 8:30
12 a.m.
13 MEMBERS PRESENT:
14 DANA A. POWERS, Chairman
15 GEORGE APOSTOLAKIS, Vice-Chairman
16 MARIO V. BONACA
17 THOMAS S. KRESS
18 GRAHAM M. LEITCH
19 ROBERT L. SEALE
20 WILLIAM J. SHACK
21 JOHN D. SIEBER
22 ROBERT E. UHRIG
23 GRAHAM B. WALLIS
24
25
. 208
1 P R O C E E D I N G S
2 [8:30 a.m.]
3 DR. POWERS: We will now come to order. This is
4 the second day of the 475th meeting of the Advisory
5 Committee on Reactor Safeguards.
6 During today's meeting, the committee will
7 consider the following; performance-based regulatory
8 initiatives; license renewal guidance documents; operating
9 advance of Indian Point Nuclear Power Plant 2; and, Siemens,
10 our S-RELAP-5 Appendix K small-break LOCA code. We will
11 also examine proposed ACRS reports.
12 The meeting is being conducted in accordance with
13 the provisions of the Federal Advisory Committee Act. Mr.
14 Howard Larson is the Designated Federal Official for the
15 initial portion of the meeting.
16 We have received no written statements or requests
17 for time to make oral statements from members of the public
18 regarding today's session.
19 A transcript of portions of the meeting is being
20 kept and it is requested that the speakers use one of the
21 microphones, identify themselves, and speak with sufficient
22 clarity and volume so they can be readily heard.
23 Do any of the members have opening comments they
24 would like to make for today's session?
25 [No response.]
. 209
1 DR. POWERS: Seeing none, we will move to the
2 first item of business, which is performance-based
3 regulatory initiatives. Sieber, I believe you're going to
4 lead us through this.
5 MR. SIEBER: Yes, Mr. Chairman. By way of
6 introduction to this subject, we did have a discussion of
7 the high level guidelines for performance-based activities
8 in June and we considered a draft letter at that time.
9 On the other hand, we've now had two more months
10 to think about all of this, so we may or may not wish to
11 renew our interest in that draft letter.
12 Also, by way of introduction, this is all based on
13 an initiative in 1998, with the publishment of a white paper
14 on risk-informed and performance-based regulation, which is,
15 to my view, sort of a gospel since it provides all the
16 definitions that anyone could ever want to have related to
17 risk-informed regulation and performance-basing.
18 An outcome of that was an SRM, where the
19 Commission directed the staff through research to develop a
20 high level guideline and the most recent edition of that
21 high level guideline that I have is one that as handed out
22 yesterday, which is a pre-decisional draft from Research to
23 the EDO, which contains the latest revisions and explanation
24 thereto.
25 So with that brief introduction, what I would like
. 210
1 to do is introduce Mr. Prasad Kadambi, from Research, who
2 will lead us through this subject.
3 MR. KADAMBI: Thank you very much, Mr. Sieber.
4 Good morning, Mr. Chairman and members of the committee.
5 I believe our Division Director, acting Division
6 Director, Dr. Farouk Eltawila, has some opening remarks.
7 MR. ELTAWILA: Thanks, Prasad. Yesterday, you had
8 two presentations, one on option two and the other one on
9 option three, and I think we'd want you to look at the
10 performance-based regulation as the third tool for these
11 activities, ongoing activities here.
12 Every activity that the NRC is going to be
13 undertaking in terms of amending the regulation, whether
14 it's coming from internal stakeholders or external
15 stakeholders, we go through the process. The process will
16 be based on the agency performance goals.
17 If it meets these performance goals, we look for
18 them if they are going to be risk-informed and if there is
19 risk-informed, we look for the second option, can any of
20 these be made performance-based. That would be the
21 preferred option that the agency will be taking.
22 So the first thing is we go through a certain
23 screening criteria. If the regulations need to be amended,
24 we'll amend them, and then we will decide whether we want to
25 be performance-based, risk-based and performance-based, or
. 211
1 we go back to traditional ways.
2 So that's the way I would like you to look at them
3 in that regard.
4 The other thing, we apply the guideline, as Prasad
5 is going to present, into the area of hydrogen control,
6 which was presented to you yesterday, and we'll try to show
7 the linkage between how you deal with the issue in a
8 risk-informed and then how you expand on it to be
9 performance-based, and hopefully you look at this from this
10 perspective.
11 Thank you.
12 MR. KADAMBI: Thank you. Again, my name is Prasad
13 Kadambi. I'm in the Office of Research. I'm joined in
14 today's presentation by Bob Youngblood and Chris Smith.
15 They helped us conduct the case studies and they will be
16 presenting information about the case studies, which will be
17 the centerpiece of our presentation today.
18 As Mr. Sieber mentioned, you heard from the staff
19 just a couple of months ago, so I will try not to be
20 repetitive in anything, but I would like to reiterate a
21 couple of points.
22 The word "high level" in the title really has to
23 do with the level of conceptualization in that. It does
24 deal with the sorts of things that are in the white paper,
25 which Mr. Sieber mentioned, and in the strategic plan and
. 212
1 these sorts of things.
2 As a result, we feel that the guidelines can be
3 applied to all three arenas of the agency's activities,
4 reactors, materials and waste arenas.
5 The other point in the title, the word
6 "performance-based activities" comes in, and that really
7 addressed NRC activities at this point. Mostly it deals
8 with developing or changing regulatory requirements.
9 DR. POWERS: What is ISL?
10 MR. KADAMBI: ISL is a company that has provided
11 us technical assistance on this. It used to be part of
12 Scientech. And if you have anymore questions, maybe when
13 Bob Youngblood --
14 MR. YOUNGBLOOD: That's correct. I'm Bob
15 Youngblood from ISL. I work in the Rockville office of ISL,
16 which is essentially equivalent to the former Rockville
17 office of Scientech, which provided NRC support. There was
18 a parting. Scientech transferred us in February.
19 MR. KADAMBI: What I would like to do is begin
20 with an outline of the presentation. As I mentioned, the
21 case studies will be central to this. I will begin with an
22 overview that provides a snapshot of where we are today.
23 We will talk about interrelationships among the
24 regulatory activities that Dr. Eltawila mentioned, and,
25 again, here, the white paper plays a central role. It
. 213
1 provides the definitions and attributes for articulating
2 these interrelationships. And we'll talk about the staff's
3 plans and end with the conclusions.
4 By way of overview, we did speak to ACRS and then
5 subsequently to ACNW on July 25, with similar presentations.
6 We have tried to emphasize that performance-based regulatory
7 approaches compliment risk-informed regulatory approaches
8 and the traditional means of regulation.
9 I believe everybody has a copy of the draft
10 Commission paper, which is now being reviewed at the EDO
11 level. By a strict definition of the interpretation of the
12 SRM, we may have concluded at one time that the guidelines
13 would be enough to meet the SRM, but we have actually gone
14 beyond that and I believe the case studies represent that.
15 Also, in response to the public comments, we have
16 tried to note and respond to concerns about some
17 stakeholders feeling that the focus of the performance-based
18 efforts would emphasize burden reduction at the expense of
19 safety.
20 What I would like to point out is that
21 performance-based regulation does not mean either less
22 regulation or more regulation, but I hope it means smarter
23 regulation.
24 The guidelines themselves are, admittedly,
25 process-oriented and I don't think that necessarily detracts
. 214
1 from their usefulness. And by usefulness, I'm talking about
2 raising the right kinds of questions for an analyst to deal
3 with a regulatory issue.
4 The guidelines themselves are in three groups;
5 viability, assessing change, and the high level regulatory
6 principles that the agency adheres to.
7 The guidelines are not meant to be a cookbook, but
8 what we have found in applying in the case studies is that
9 when used in the context of a simplified expert panel, as we
10 did, they did provide useful information and at this point,
11 because they have demonstrated their usefulness in two
12 distinctly different areas, and as you will hear about it,
13 we feel that it's ready for agency-wide application and
14 that's how we are going forward to the Commission.
15 Meanwhile, the staff is, of course, continuing its
16 efforts in the other initiatives, the risk-informed
17 regulations. We are updating the risk-informed regulation
18 implementation plan. There have been workshops held on the
19 matter of using risk information in the NMSS types of
20 activities.
21 And as we gain experience in these areas, we feel
22 we are getting closer to the objective of having a well
23 integrated approach to regulatory initiatives.
24 At this point, I will request Bob to start with
25 the first of the case studies.
. 215
1 MR. YOUNGBLOOD: Good morning. I'm Bob Youngblood
2 of ISL and I will be talking about case study one. What
3 we've done in case study one is take the body of risk
4 information put together by the option three people, who
5 drew from years of work in this area and summarized it for
6 their own purposes, and taken a piece of that information
7 and basically postulated a need to be addressed by what
8 we're calling here a regulatory framework.
9 We're not using the word requirement here, because
10 we don't mean just requirement, but the whole suite of rule,
11 guidance, technical specifications, because it's useful, I
12 think, to look at the whole safety mission a little more
13 globally than just at the rule level.
14 As preface, risk information, the kind of thing
15 risk information will do generically is identify what's
16 important, tell you potentially how reliable the function
17 you're talking about needs to be, and place the function
18 that you're going to look at in context of what's going on
19 in the scenarios that it is supposed to play a role in.
20 And the following bullet in this case study, what
21 we're pointing out is the risk information from the option
22 three team emphasized the potential importance of station
23 black-out scenarios at some plants, emphasized the
24 importance of addressing core melt phenomenology in the
25 sense that it may create harsh environments that systems may
. 216
1 need to cope with, and the loads that you need to address in
2 preventing containment failure from combustion of gas have
3 to be evaluated in the context of other loads that may exist
4 at the time you get the combustion.
5 At this point, I guess I'm implicitly assuming
6 that you had this discussed a lot yesterday and that I did
7 not need to sort of talk about the threat to containment
8 from combustion of gas.
9 So based on that risk information, the high level
10 --
11 DR. POWERS: Let me ask a question about that.
12 MR. YOUNGBLOOD: Sure.
13 DR. POWERS: The threat to containment. In
14 thinking about the threat to containment due to the
15 combustion of gas, do you distinguish between steel
16 containments and concrete containments?
17 MR. YOUNGBLOOD: Buried implicitly in the next
18 chart, we do, because I believe that their work looks at,
19 for example, the generation of carbon monoxide. We're not
20 going to get into the level of detail that you're going
21 after with that question.
22 DR. POWERS: What I'm thinking of is that we have
23 some evidence that steel containments fail catastrophically,
24 whereas concrete containments fracture, leak, but they don't
25 fail catastrophically.
. 217
1 Wouldn't one want to make a distinction between
2 the two then?
3 MR. YOUNGBLOOD: Yes, but not necessarily in this
4 talk at the level that I'll be working at. I'll be -- well.
5 I think the discussion that you're advocating I think would
6 be naturally raised in the context of the next couple of
7 slides. I'll still resist it, but that's where it will
8 belong.
9 Now, the guidelines start right off looking at
10 parameters that are potentially monitorable parameters, and
11 so the next slide is about beginning to look for those
12 parameters and placing them in some sort of context.
13 And there are different kinds of parameters that
14 one could look at. One kind of parameter is the parameters
15 that talk about the physics and chemistry of getting the job
16 done, which we're here calling capability parameters, which,
17 for fluid systems, would be things like flow rates,
18 possibly, and for this system, for a system to remove
19 combustible gas, will bear on things like the temperature
20 that units achieve and how many of them you've got and where
21 they're located and atmospheric mixing and other figures of
22 physics, chemistry things that bear on whether the function
23 will actually be successful.
24 DR. KRESS: Most of those things I see up there
25 are uncertain and would have distributions. Are you
. 218
1 considering each of these having a probability distribution,
2 like the containment pressure or the severe accident loads
3 or the amount of combustible gas present when you --
4 MR. YOUNGBLOOD: Potentially, yes. I have not --
5 I'm not in the loop on the option three work, but I think
6 that a big part of their recommendation is a very careful
7 evaluation of things like this one here, amount of gas and
8 the rate generated. The combustible gas source term needs a
9 lot of evaluation, for example.
10 I think the answer to your question is yes.
11 DR. KRESS: In that case, your second line up
12 there would be some sort of overlap of probabilities rather
13 than just greater than.
14 MR. YOUNGBLOOD: Yes, yes, yes.
15 DR. KRESS: At some level.
16 MR. YOUNGBLOOD: Yes. Yes.
17 DR. KRESS: That sort of thing.
18 MR. YOUNGBLOOD: Yes. All I meant to imply by
19 that symbol was some sort of comparison.
20 DR. KRESS: Some sort of comparison, yes.
21 DR. APOSTOLAKIS: Coming back to Dr. Powers'
22 question, that would refer to the containment capability, I
23 believe, would it not?
24 MR. YOUNGBLOOD: Yes.
25 DR. APOSTOLAKIS: And is that branch developed
. 219
1 further and does it affect the developed branch or the
2 containment pressure?
3 MR. YOUNGBLOOD: Well, the purpose of even having
4 it there is to place your evaluation of the pressure in the
5 context so that you can make that comparison.
6 In preparing this slide, I put it there because
7 some containments are such that the loads that you get
8 compared to the capability are not much of an issue and
9 really to place in context the fact that we're going to be
10 talking about the kind of plant that needs igniters.
11 So I didn't mean to get into the steel versus
12 concrete discussion here and didn't prepare to do that.
13 DR. APOSTOLAKIS: The point is that everything
14 that you are analyzing on the left is conditional --
15 MR. YOUNGBLOOD: Yes.
16 DR. APOSTOLAKIS: -- over here on the right.
17 MR. YOUNGBLOOD: That's right. And I think that's
18 a strong feature of what the option three people are saying
19 and it's part of what they mean by risk-significant
20 scenarios. You're supposed to look at the scenarios that
21 have appreciable frequency, get the phenomenology right and
22 go from there.
23 DR. APOSTOLAKIS: So somehow I think you should
24 put the right words there or symbols to indicate that.
25 MR. YOUNGBLOOD: Okay.
. 220
1 DR. APOSTOLAKIS: Because right now you have an
2 inequality.
3 MR. YOUNGBLOOD: Okay. Of course.
4 DR. APOSTOLAKIS: People might think that these
5 two are independent.
6 MR. YOUNGBLOOD: No.
7 DR. BONACA: When you indicate the containment
8 capability, are you considering the ultimate capability of
9 containment, as estimated, for example, for the IPEs, or are
10 you considering the committed containment capability that is
11 on the basis of technical specifications?
12 MR. YOUNGBLOOD: It's my sense of the option three
13 people that they mean ultimate, but --
14 MR. ELTAWILA: There is a distribution. I think
15 we will do it exactly like we did it for the IPE direct
16 containment heating, take the containment failure
17 probability function versus the load and the intersection
18 between the probability and the load, and that will give you
19 the ultimate, the failure of that containment.
20 So it's both of them are distributed.
21 MR. YOUNGBLOOD: But I would like to make a point
22 I maybe should have emphasized earlier, and that is that in
23 going through this case study, what we have tried to do is
24 take the risk insights that come in over the transom, they
25 identify a need, they tell you what the risk significant
. 221
1 scenarios are, what the phenomenology is.
2 They point out that in black-out, you may not have
3 power that you might have wanted to run igniters. So you
4 need to address that.
5 So it's the -- and I'm just taking all that for
6 granted, letting them set the safety priorities and I would
7 classify the uncertainty issues that you are raising in that
8 realm.
9 Now, not to say we don't have to address them once
10 they have been set, but it seems to me that many of these
11 questions are related to the risk-informing task and that
12 once that's done, what we're trying to do, once the mission
13 has been stated, we're trying to look for parameters,
14 kicking off the guidelines exercise.
15 DR. APOSTOLAKIS: So I think it would be useful to
16 cull all this risk information as the context within which
17 you are doing this.
18 MR. YOUNGBLOOD: Yes.
19 DR. APOSTOLAKIS: This is a popular word these
20 days.
21 MR. YOUNGBLOOD: Yes.
22 DR. APOSTOLAKIS: PRA defines the context within
23 which performance parameters will be defined.
24 ELTAWILA: Correct.
25 MR. YOUNGBLOOD: Great. Okay. And so the
. 222
1 parameters that come out of the bottom here, they are
2 physical parameters having to do with unit capability.
3 There are configurational parameters and there's just a
4 plethora of capability type parameters that one would --
5 DR. APOSTOLAKIS: Is capability the right word
6 there?
7 MR. YOUNGBLOOD: I'm not sure that it's the
8 perfect word, but it's a word that many people working in a
9 related area under the reactor oversight process would use
10 in this context.
11 DR. APOSTOLAKIS: The only problem is that you are
12 using it also to talk containment capability.
13 MR. YOUNGBLOOD: Well, I don't think the usage is
14 inconsistent, but when we're referring to parameters, we
15 just are talking about the physical mission and I guess
16 containment capability is part of that.
17 It happens that the focus of this discussion is on
18 the igniter capabilities, but all of these things are -- if
19 we were trying to apply the guidelines to containment
20 capability, that would certainly be a consistent usage of
21 the term, I think.
22 DR. APOSTOLAKIS: So these are really the
23 monitoring capabilities that you have to identify in order
24 to assess viability, right? I'm trying to understand.
25 MR. YOUNGBLOOD: Right. So guideline one asks are
. 223
1 there measurable or calculable parameters and what I just
2 tried to do is illustrate what sort of thing you might get
3 in the way of a capability parameter.
4 In the interest of brevity, we will not plan to
5 show a slide about reliability/availability parameters, but
6 those would certainly also be important and will be
7 discussed in subsequent slides.
8 So we have parameters to talk about the mission,
9 parameters to talk about reliability/availability, and all
10 of those aspects are elements of what you try to do with the
11 safety requirement, do a job and do it with some
12 reliability.
13 DR. APOSTOLAKIS: Now, by guidelines, you refer to
14 the four attributes of a performance-based measure.
15 MR. YOUNGBLOOD: The things in this treatment are
16 called the viability guidelines, which really the -- yes.
17 DR. APOSTOLAKIS: Measurable parameters and so on.
18 MR. YOUNGBLOOD: That's right.
19 DR. APOSTOLAKIS: It seems to me that there is a
20 guideline that's missing there. Namely, that we should try
21 to set the performance measures at as high a level as we
22 can, high in the sense that you showed.
23 MR. YOUNGBLOOD: Yes.
24 DR. APOSTOLAKIS: In other words, if you go back
25 to your tree.
. 224
1 MR. YOUNGBLOOD: Yes.
2 DR. APOSTOLAKIS: If I could do something and
3 measure or eliminate or whatever the amount of combustible
4 gas present, then I wouldn't need to worry about the
5 physical parameters and everything else.
6 MR. YOUNGBLOOD: That's right.
7 DR. APOSTOLAKIS: So there has to be some logic
8 progression there that I can't really do that, so I go to
9 the next level.
10 MR. YOUNGBLOOD: That's right.
11 DR. APOSTOLAKIS: Then maybe I can't do that, then
12 I go to the next level. Because if we don't do that, we may
13 end up, again, with an extremely prescriptive system, where
14 we just set performance measures all over the place.
15 MR. YOUNGBLOOD: Now, there is a guideline that
16 does that, although maybe less explicitly than you would
17 like, and that's the flexibility guideline. The third one
18 says you want flexibility. Now, if you let me argue that
19 the higher you are, the more flexibility you have, which I
20 believe would be generally correct, then that guideline at
21 least touches on the point you're raising.
22 DR. APOSTOLAKIS: You are right, Bob, but I would
23 not like it to be implicit.
24 MR. YOUNGBLOOD: Okay.
25 MR. KADAMBI: There is an explicit statement of
. 225
1 this concept in the guidelines to assess change and it
2 occurs in the guideline that says the performance-based
3 approach can be incorporated into a regulatory framework.
4 DR. APOSTOLAKIS: Please put that back up. I'm
5 sorry.
6 MR. KADAMBI: That's okay. That is what we call
7 the Guideline II-F.
8 DR. APOSTOLAKIS: Page?
9 MR. KADAMBI: Page 1-3.
10 DR. APOSTOLAKIS: Roman Numeral F, then what? I
11 mean, not F, F is not Roman. Which one is it?
12 MR. KADAMBI: It is II-F and it is the combination
13 of the four amplifying guidelines, really. But the idea of
14 that --
15 DR. APOSTOLAKIS: Are we looking at the same page?
16 You said 1-3 of --
17 MR. KADAMBI: The Attachment 1.
18 DR. APOSTOLAKIS: Oh, Attachment 1.
19 MR. KADAMBI: I'm sorry. The Attachment 1 is the
20 actual guidelines themselves.
21 DR. WALLIS: So this is what says the
22 performance-based approach can be incorporated into the
23 regulatory framework. Is that the one?
24 MR. KADAMBI: Right. And by regulatory framework,
25 what we are trying to say is there is a --
. 226
1 DR. WALLIS: Why do you use the conditional all
2 the way through? Why don't you say "is" instead of "would
3 be?"
4 MR. KADAMBI: Well, because these are guidelines
5 that hopefully lead the analyst to ask the question and they
6 are not meant to be go/no-go types of conditions.
7 DR. WALLIS: So they are deliberately a little
8 vague in using "would" instead of is.
9 MR. KADAMBI: Yes. There is a level of design in
10 there. Are you there, Dr. Apostolakis? Maybe it isn't as
11 explicit as we would like.
12 DR. APOSTOLAKIS: Point me again to where exactly
13 that is. I'm on the right page, 1-3. F?
14 MR. KADAMBI: 1-3, and it is the guideline that is
15 identified by capital F.
16 DR. APOSTOLAKIS: Right.
17 MR. KADAMBI: And what that is meant to convey is
18 that the regulatory framework which consists of what I think
19 of as a hierarchy in which the Code of Federal Regulations
20 occurs at the top and there are lower level guidance
21 documents, have to be considered in terms of the
22 functionality that is to be provided, and that in applying
23 the feasibility guidelines, you look for the specific
24 elements in the framework and that is where I believe the
25 question that you're asking would be answered.
. 227
1 DR. APOSTOLAKIS: But would it hurt to make it
2 explicit that we would like the measures to be at as high a
3 level as we can?
4 MR. ELTAWILA: I agree with you, Professor
5 Apostolakis. It will not hurt. It's not explicit in the
6 guideline and I think it's a good idea. We'll take it into
7 -- yes, absolutely.
8 DR. APOSTOLAKIS: It is an essential element of
9 removing unnecessary burden, right? Because, you know, you
10 can be as prescriptive in a probabilistic -- in a
11 risk-informed, performance-based system as you are in the
12 old system.
13 MR. KADAMBI: You are right.
14 DR. APOSTOLAKIS: So what I would say then, Bob,
15 is that give prominence to this when you present this, go
16 down the hierarchy. Then the next statement should be can
17 -- now we ask the question, can we have performance-based
18 criteria at the loads. No, we can't. And we go down one,
19 no, we can't. And that really sends a clear message that
20 you are forced to go logically to lower levels, where you
21 finally say, yes, here I can do something about it.
22 MR. KADAMBI: That's exactly the process that is
23 involved.
24 DR. APOSTOLAKIS: But it doesn't come across
25 during the presentation, and I think that would be extremely
. 228
1 valuable.
2 MR. KADAMBI: Thank you for that point, yes.
3 MR. YOUNGBLOOD: With that as preface, we've
4 discussed briefly the idea of capability parameters of
5 reliability/availability parameters.
6 And now the next slides basically just step
7 through guidelines, applying each of the four viability
8 guidelines to parameters of that kind.
9 So here the guideline is recapitulated at the top
10 of this slide, which is number six, measurable or calculable
11 parameters to monitor plant licensee performance exist or
12 can be developed. Capability parameters clearly exist.
13 Once you pick a technology, which we're implicitly
14 here thinking of igniters, many of those things are
15 implicit. Environmental qualification is part of that and
16 those are parameters and here I have tried to list ones that
17 are not necessarily amenable to performance monitoring, but
18 this guideline is only asking part of the question.
19 DR. APOSTOLAKIS: Let's look at the
20 reliability/availability.
21 MR. YOUNGBLOOD: Sure.
22 DR. APOSTOLAKIS: This refers to what, reliability
23 of what?
24 MR. YOUNGBLOOD: The function of getting rid of
25 combustible gas as it's generated in the context of the
. 229
1 risk-significant scenarios that cause it to be generated.
2 DR. APOSTOLAKIS: So it's a function.
3 MR. YOUNGBLOOD: Yes.
4 DR. APOSTOLAKIS: And then division?
5 MR. YOUNGBLOOD: In listing division and unit
6 reliability, I chose -- I assumed that a licensee might
7 choose to have enough redundancy that would be more than one
8 division and that a division would have units in it.
9 DR. APOSTOLAKIS: So essentially, then, what
10 you're doing there, again, is you are following an implicit
11 top-down approach.
12 MR. YOUNGBLOOD: Yes. Exactly.
13 DR. APOSTOLAKIS: Where you say I want this
14 function, this function is performed by these systems.
15 MR. YOUNGBLOOD: Exactly.
16 DR. APOSTOLAKIS: For each system, this. So,
17 again, the previous comment applies.
18 MR. YOUNGBLOOD: Yes, it does.
19 DR. APOSTOLAKIS: That, again, because if I just
20 look at that and I don't talk to it, then, again, I may
21 start putting requirements on all of these.
22 MR. YOUNGBLOOD: That's right.
23 DR. APOSTOLAKIS: I believe that's something that
24 really needs to be emphasized in the report and the
25 presentations, because this is the essence of trying to have
. 230
1 a performance-based system, it seems to me.
2 I know you guys agree. It's just that
3 communication sometimes is very important and you will not
4 be the ones implementing this throughout the agency.
5 MR. YOUNGBLOOD: Right. The first four
6 guidelines, the ones we're talking about here, have been
7 around for a long time and I believe that in agreeing with
8 you --
9 DR. APOSTOLAKIS: It pains you a lot.
10 MR. YOUNGBLOOD: No, no. In this case, it's not
11 painful. But I think that we are also talking about a scheme
12 where there will be several parameters involved and not just
13 one and there may be aspects of the system that are dealt
14 with by means other than performance-based monitoring.
15 In other words, there will be a multi-parameter
16 approach to this system. And I think in the early days of
17 this whole discussion, which is many years old now, I think
18 that the original formulation of some of these guidelines
19 contemplated an all or nothing single parameter approach
20 implicitly, and that one reason it doesn't come across is
21 that we're still using language from five years ago, but
22 didn't recognize that.
23 DR. APOSTOLAKIS: We should.
24 MR. YOUNGBLOOD: We should.
25 DR. KRESS: So if I'm looking for performance
. 231
1 measures on, say, the power supply to these things, I would
2 look under this reliability/availability.
3 MR. YOUNGBLOOD: That's right. And, again, the
4 risk information emphasized the need to have that work in
5 black-out.
6 DR. WALLIS: If you try to answer the question
7 will it work, you said something about mixing, but you can't
8 monitor mixing. Therefore, you ignore it. Mixing is
9 important to radiological work.
10 MR. YOUNGBLOOD: Yes. I foolishly brought up
11 mixing. Actually, in the option three work, it's treated in
12 a separate area and in the SECY, I believe --
13 DR. WALLIS: You went through the things that
14 affect whether it will work or not and mixing is one of
15 them.
16 MR. YOUNGBLOOD: That's right.
17 DR. WALLIS: And then you decided not to monitor
18 it because you can't do it.
19 MR. YOUNGBLOOD: No. Actually, I left it off the
20 slide because it's not part of what I was looking at.
21 DR. WALLIS: Why did you leave it off? There are
22 things that affect performance which you left off, as well
23 as put on. You left them off because they were difficult to
24 measure or for some other reason?
25 MR. YOUNGBLOOD: No. I think that the active
. 232
1 systems involved in mixing would be eligible for treatment
2 as any other active system would be.
3 DR. WALLIS: Like the containment fans.
4 MR. YOUNGBLOOD: Yes, whatever they are, but we
5 just did a piece of the puzzle and all the -- yes. The
6 piece of the puzzle we did is this piece right here and
7 mixing would be part of that somewhere on here, and isn't
8 the piece that we did.
9 DR. WALLIS: To do anything that affects
10 performance, it would seem to me you have to consider and
11 then evaluate in some way and then decide what to do with.
12 MR. YOUNGBLOOD: Yes, you would.
13 DR. KRESS: You're saying if you did this in a
14 complete manner, you would do that.
15 MR. YOUNGBLOOD: Yes, absolutely.
16 DR. KRESS: But you're just giving us an
17 illustration.
18 MR. YOUNGBLOOD: Absolutely. Yes. In fact, this
19 isn't about option three. It's about the guidelines.
20 DR. APOSTOLAKIS: But that raises another
21 question, though. We have agreed that we have to be
22 hierarchical. Now, let's come to the same horizontal level,
23 which involves, say, five parameters, that all affect
24 whatever it is about.
25 MR. YOUNGBLOOD: Okay.
. 233
1 DR. APOSTOLAKIS: Now, if you look at the physics
2 of the problem, these parameters are entering some equation
3 or something. Is there any guideline that says that perhaps
4 controlling two of those is sufficient and I don't have to
5 do something for the others or you would go ahead and
6 control all five and performance measures for all five, even
7 though there will be some redundancy or the end diversity
8 perhaps that way?
9 So what is the philosophy at that level?
10 MR. KADAMBI: I believe, as you work through the
11 availability guidelines, some of the parameters that you
12 choose in Guideline I-A would turn out to be not as valuable
13 as -- there will be some kind of a prioritization, I
14 believe, as you work through the guidelines. This is the
15 sort of thing that we learned as we went through the case
16 study.
17 Now, ultimately, I believe that after you
18 determine a set of parameters that meet the viability
19 guidelines and you say that these are perhaps the parameters
20 I want to work with, then you get into the second set of
21 guidelines that assesses change and you try to do the sorts
22 of things and efficiency and effectiveness and also the
23 guideline that tries to assess net benefit in terms of how
24 much -- what do you have to do to get the performance
25 monitoring for these.
. 234
1 That's what would lead you to say that not all
2 those parameters are equally valuable, that some are more
3 valuable.
4 DR. APOSTOLAKIS: I agree with what you said, but
5 it seems to me that at that level, you will have, again, the
6 perennial conflict between defense-in-depth and a
7 rationalist approach.
8 Because of defense-in-depth or the structuralist
9 interpretation, you would tend to have more, because you
10 would like to have higher confidence that you are really
11 controlling the thing.
12 On the other hand, you may end up imposing
13 unnecessary burden.
14 So there has to be something explicit there as to
15 how one would balance those two. Now, maybe what you just
16 said is good enough, I don't know, but is there an explicit
17 discussion of this somewhere or is it hidden under
18 effectiveness?
19 MR. KADAMBI: Well, for right now, because we are
20 focusing on guidelines that would apply, in a sense, across
21 the board in all three arenas, we have avoided getting into
22 this kind of detail which would work differently if you're
23 dealing with reactors as opposed to materials as opposed to
24 weights.
25 DR. APOSTOLAKIS: But I would be perfectly happy
. 235
1 to have you write down what you just said. I think it's
2 extremely important for the implementers of this to be aware
3 of these inherent conflicts.
4 DR. KRESS: Otherwise, you're going to have an
5 eternal debate between the licensee on how many of these I
6 need.
7 DR. APOSTOLAKIS: But the people who will take
8 your reports and try to implement them, probably they will
9 have to go through this process again and rediscover the
10 wheel, so to speak.
11 MR. KADAMBI: That's the reason why we feel the
12 management directive that we have proposed as one of the
13 immediate follow-on activities will address how these things
14 work in a level of detail that's not covered in this paper.
15 DR. APOSTOLAKIS: But I would be happier if you
16 put down on paper what you told us the last minute. I mean,
17 it's clear to me that you are aware of the problem. It's
18 just that I want to emphasize it.
19 MR. KADAMBI: Understand.
20 DR. APOSTOLAKIS: So others will be aware of it,
21 too.
22 MR. YOUNGBLOOD: On the preceding slide, I listed
23 capability parameters and reliability/availability
24 parameters.
25 The next guideline asks whether objective criteria
. 236
1 for these parameters exist or can be developed. For the
2 capability parameters, it falls out of the phenomenology
3 that needs to be analyzed for those sequences and the nature
4 of the technology itself.
5 And for reliability/availability, you can derive a
6 criterion from a LERF guideline of ten-to-the-minus-five per
7 year, for example, and a challenge frequency of this
8 function, which might be plant-specific.
9 DR. APOSTOLAKIS: Why isn't that part of the
10 maintenance rule? Am I missing something?
11 MR. KADAMBI: It could well be. It could well be
12 part of the maintenance rule.
13 DR. APOSTOLAKIS: Would or is?
14 MR. KADAMBI: We don't know enough about it.
15 MR. YOUNGBLOOD: Could you amplify that question a
16 little bit?
17 DR. APOSTOLAKIS: You're talking about the
18 reliability and availability of trains and components, isn't
19 that what the maintenance rule is supposed to do?
20 MR. YOUNGBLOOD: Yes, it does. I think the
21 licensee gets to derive those and here I think we're talking
22 about a different provenance of those criteria, but, you
23 know, long division is long division.
24 DR. APOSTOLAKIS: I mean, you mentioned earlier
25 the fan coolers. This is an SSE, right? System, structure
. 237
1 or component. Presumably, there will be some maintenance
2 rule.
3 MR. YOUNGBLOOD: There is tremendous overlap
4 between this whole topic and everything in the maintenance
5 rule. Why are you querying it for development of this?
6 DR. APOSTOLAKIS: Because the maintenance rule
7 probably will not care -- in fact, it will not care about
8 surface temperature, so it's not within the scope, is it? I
9 don't think, the way I understand it.
10 So your first bullet is --
11 DR. POWERS: The maintenance rule would care if
12 the surface temperature did not get high enough to do the
13 function of the igniter.
14 DR. APOSTOLAKIS: But it's the igniter that's the
15 focus.
16 DR. POWERS: Yes, whether it can do its job or
17 not.
18 DR. APOSTOLAKIS: Yes.
19 DR. POWERS: That would be something that the
20 maintenance rule would care about, what the specific
21 temperature is.
22 DR. APOSTOLAKIS: Anyway, you are right. There is
23 tremendous overlap.
24 MR. KADAMBI: And I would like to emphasize that
25 this is an illustrative example for the guidelines and not
. 238
1 so much to get into the technical regulatory issues
2 involved.
3 MR. YOUNGBLOOD: One last point to be made.
4 DR. APOSTOLAKIS: By the way, not everything I say
5 is a criticism.
6 MR. YOUNGBLOOD: I understand.
7 DR. APOSTOLAKIS: Don't be so defensive.
8 MR. YOUNGBLOOD: So far it's a good day.
9 DR. WALLIS: When you do this exercise, you say
10 it's illustrative for the guidelines, are you going to end
11 up giving us some indication of why this is better than the
12 present system?
13 MR. KADAMBI: We hope that it will come out as
14 part of the way we --
15 DR. WALLIS: You make some hopeful statements
16 about increasing efficiency and reducing burden, and I can't
17 evaluate at all from the case study whether it increases or
18 decreases burden of effectiveness and efficiency.
19 ELTAWILA: First of all, we have not gone through
20 the whole exercise completely, but I can illustrate, for
21 example, yesterday, in the discussion that we had the
22 situation that the igniter for the ice condenser in Mark 3
23 containment is vulnerable during station black-out.
24 So either you will have to require an additional
25 power supply for this igniter or you have to accept a
. 239
1 risk-based criteria, as Dr. Kress indicated yesterday, which
2 will say I will accept a very low frequency in this case.
3 Here you might have a third alternative, saying
4 can I have certain characteristics or performance that I can
5 measure and I can rely on to substitute for either not to be
6 performance-based, risk-based, or to have additional
7 requirements of the igniter.
8 So that's when you really apply it systematically,
9 you might reach a different conclusion.
10 DR. APOSTOLAKIS: It seems to me that the greatest
11 benefit of this is what I insisted earlier on seeing
12 explicitly, that the principle is that you should set the
13 performance criteria at as high a level as you can and you
14 will give flexibility to the licensee to use methods that
15 the licensee deems appropriate to demonstrate that the
16 criteria have been met.
17 If we are successful at that, then the whole thing
18 will indeed be useful.
19 Now, an additional benefit is that in many
20 instances, you will be able to define the context using risk
21 information, but that is not the only thing you are doing,
22 as I understand.
23 MR. KADAMBI: In fact, the second case study will
24 show that.
25 DR. APOSTOLAKIS: Right. So the most important
. 240
1 thing is to try to define these performance criteria at the
2 high level, so we will not be burdensome.
3 MR. KADAMBI: Correct.
4 DR. WALLIS: Listening to Farouk, it seems to me,
5 also, that you might well increase safety by focusing on
6 this and knowing what you're doing better.
7 MR. KADAMBI: Exactly. That is a very important
8 point.
9 DR. APOSTOLAKIS: Absolutely.
10 DR. KRESS: George, this business of setting
11 parameters at the highest level, it seems to me like it
12 needs additional guidance on what criteria you use to decide
13 which level is the highest level you can, and that's what I
14 see is kind of missing so far.
15 DR. APOSTOLAKIS: In some sense, the four
16 attributes that we have seen many times --
17 DR. KRESS: Measurable doesn't --
18 DR. APOSTOLAKIS: That's right.
19 DR. KRESS: You can live with it if it fails,
20 those things.
21 DR. APOSTOLAKIS: I think those will guide you to
22 say, gee, I can actually have it at this level.
23 MR. YOUNGBLOOD: If we get to slide 1-D, that
24 point will be touched on.
25 DR. APOSTOLAKIS: Do you think you will ever get
. 241
1 there?
2 MR. YOUNGBLOOD: Not if poor Chris is going to get
3 to speak. Okay. I'll move on now to slide eight, guideline
4 I-C, flexibility exists or can be developed.
5 On the previous slide, we concluded for
6 reliability/availability that the top -- that the functional
7 reliability would follow from high level guidance. Now, how
8 well a division or a unit would have to perform is a
9 function of the actual design configuration.
10 In that sense, there's a lot of flexibility
11 potentially and I think in reality for this system in
12 meeting reliability/availability guidance by laying your
13 design out differently.
14 Once you pick a technology, I don't think there is
15 very much flexibility in the capability parameters. The
16 criteria, I think, tend to be fixed there by technology and
17 phenomenology. But I think there's a lot of flexibility
18 below the level of the functional reliability.
19 DR. APOSTOLAKIS: Let me understand this a little
20 better, because I'm not familiar with the phenomena here.
21 You're talking about surface temperature.
22 MR. YOUNGBLOOD: Yes.
23 DR. APOSTOLAKIS: What surface is this?
24 MR. YOUNGBLOOD: It's a surface in the igniter,
25 that the igniter has to get hot to ignite and tech specs
. 242
1 require --
2 DR. APOSTOLAKIS: Is this is a measurable
3 requirement?
4 MR. YOUNGBLOOD: Yes. And current tech specs
5 actually require that it be measured.
6 MR. ELTAWILA: It's not the temperature itself.
7 It's the current and voltage, and that correlated to a
8 surface temperature.
9 MR. YOUNGBLOOD: Yes. In going through, in
10 stepping through the guidelines, I tried to sort of cast the
11 net widely and not bias it toward parameters I knew I
12 wanted. So I tried to throw in other parameters to show
13 them kind of falling off the wagon as we looked at other
14 guidelines. I'm not sure that --
15 DR. APOSTOLAKIS: But I guess the thought that
16 occurred to me was it's one thing to say these parameter
17 should serve as a performance measure and quite another to
18 also -- well, not quite another, but part of this thinking
19 process should be whether there are methods that are widely
20 acceptable for achieving this.
21 MR. YOUNGBLOOD: Yes.
22 MR. KADAMBI: That is definitely part of it.
23 MR. YOUNGBLOOD: Well, Guideline I-D asks the
24 question whether having set the criterion, by the time we
25 realize it's not met, have we tolerated an unacceptable
. 243
1 safety situation.
2 DR. APOSTOLAKIS: Now, Dr. Wallis probably would
3 ask you what is an immediate safety concern. If he doesn't,
4 I do. What is an immediate safety concern? Do you have a
5 definition of an immediate safety concern? I'm learning.
6 MR. KADAMBI: I would answer that question by
7 saying that it is -- you know, context is everything, like
8 in a lot of things.
9 DR. APOSTOLAKIS: You should be working on ATHENA.
10 MR. KADAMBI: Because these are high level
11 guidelines and, you know, what are the safety concerns in a
12 materials area or in a waste area may be different, quite
13 different from a reactor area, where you have the magnitude
14 of what constitutes the level of safety concern, as well as
15 the time element and, you know, does one have enough time to
16 react to something that is considered a safety concern.
17 DR. APOSTOLAKIS: If I were to be extremely
18 critical, which I don't intend to do, I might say there may
19 be someone who really hates PRA and is very happy with the
20 existing way of doing business in his or her branch and says
21 everything we do now is a safety concern for us. So I
22 cannot apply your methodology because everything is a safety
23 concern.
24 So what is it in this approach that would make
25 that person rethink or reassess his or her position?
. 244
1 MR. KADAMBI: My answer to that question has to go
2 to the second set of guidelines, which, if you apply the
3 performance goals that the Commission has given us and you
4 deal with the four pillars, as they are called in some
5 cases, in an honest way, I think it would naturally lead you
6 to ask the kind of question that would make this person that
7 you are referring to question their current thinking.
8 DR. APOSTOLAKIS: How about if I say the
9 following? I have an immediate safety concern if the
10 corresponding node in the event tree, which started the
11 whole thing, is down. Am I defining a safety concern then
12 in a way that would really help people?
13 DR. KRESS: I think an equivalent definition might
14 be if my conditional core damage frequency has arrived at a
15 certain level.
16 DR. APOSTOLAKIS: Yes.
17 DR. KRESS: And I think you could do it the same
18 way.
19 DR. APOSTOLAKIS: In other words, we are both
20 using the event tree.
21 DR. KRESS: Using the event tree to define what
22 your safety concern is.
23 DR. APOSTOLAKIS: Rather than saying, gee, the
24 temperature is lower than I thought it was and that's an
25 immediate safety concern, because that's buried down in the
. 245
1 detail.
2 DR. WALLIS: George, I did highlight the section
3 that you asked the question in my name about. And the
4 answer in the paper that we were given is -- one of the
5 answers is that a sufficient safety margin exists.
6 So you don't have an immediate safety concern if a
7 sufficient margin exists, which simply begs the question,
8 because now you ask, well, how do you know if you have a
9 sufficient safety margin.
10 DR. APOSTOLAKIS: And I think Tom's answer comes
11 close to answering your question; namely, that safety margin
12 in this case would be the conditional core damage frequency,
13 given that that node is --
14 DR. WALLIS: But that has to be, I think, more
15 explicit.
16 DR. APOSTOLAKIS: I believe so. Otherwise, it's
17 open-ended.
18 MR. KADAMBI: I would submit that that's the sort
19 of statement that would occur in the reactor arena when that
20 concept is dealt with for reactors.
21 DR. APOSTOLAKIS: Well, you can find similar
22 things for other applications. But, again, I think you
23 should feel free to elaborate on these things and give
24 examples and say what you just said. In the reactor arena,
25 you have the event trees and you do this.
. 246
1 Now, that doesn't mean that you have to have
2 examples from every single application, but it is you give
3 an idea to people as to what an immediate safety concern
4 might be.
5 MR. KADAMBI: At one time, the Commission paper
6 was twice as long as it is.
7 DR. APOSTOLAKIS: We can triple it.
8 MR. YOUNGBLOOD: I think some of the points that
9 you were just discussing are what's meant by the sub-bullets
10 of the third bullet on slide nine here. In doing this work,
11 again, I was thinking in terms of the LERF guideline,
12 ten-to-the-minus-five per year, and some such guideline as
13 that can be used to answer.
14 DR. APOSTOLAKIS: We stand corrected, in this
15 case, it's the LERF.
16 MR. YOUNGBLOOD: In the spirit of the reactor
17 oversight process, applying CDF and LERF guidelines. I
18 think if you were way beyond those guidelines, then that
19 would be a safety concern.
20 DR. APOSTOLAKIS: As a matter of fact, we also
21 have guidelines in one of the regulatory guides as to what
22 kinds of spikes in the CDF and LERF we can tolerate and for
23 how long.
24 DR. KRESS: Those ought to be --
25 MR. YOUNGBLOOD: Yes.
. 247
1 DR. APOSTOLAKIS: So we can bring those in here.
2 MR. YOUNGBLOOD: Yes.
3 DR. KRESS: And in the defense-in-depth, I
4 wouldn't focus solely on LERF. I would have guidelines
5 associated in both CDF and LERF.
6 DR. APOSTOLAKIS: If it is relevant. But that
7 fits nicely into this.
8 MR. YOUNGBLOOD: Now, for this particular
9 function, just to complete the exercise, we won't go into
10 numbers here, but I think that between the LERF guideline,
11 the CDF challenge frequency, and so on, the reliability that
12 we would be looking for in this function is not extremely
13 high and that it would be relatively feasible as active
14 systems go to do performance monitoring to get the
15 reliability/availability that we needed.
16 DR. KRESS: Now, let me ask you a question on this
17 particular specific one.
18 MR. YOUNGBLOOD: Okay.
19 DR. KRESS: We are really probably only concerned
20 about station black-out sequence. The overall LERF
21 guidelines that we've been tossing around is for the
22 summation over all sequences.
23 MR. YOUNGBLOOD: Right.
24 DR. KRESS: And supposedly you wouldn't want one
25 particular sequence to contribute most of that.
. 248
1 MR. YOUNGBLOOD: Right.
2 DR. KRESS: And the question that I would have
3 then is, is the LERF guidelines that you're talking about
4 here some fraction of the ten-to-the-minus-five or is it
5 ten-to-the-minus-five, or how do you allocate that among the
6 sequences?
7 MR. YOUNGBLOOD: What I was answering was the
8 immediately safety concern question. I would say that if
9 your LERF went from ten-to-the-minus-five total to several
10 times ten-to-the-minus-five as a result of failure of this
11 function, then that would begin to be an immediate safety
12 concern.
13 How you set the parameters --
14 DR. KRESS: If one sequence kicked you over,
15 you're saying.
16 MR. YOUNGBLOOD: If failure of this function
17 kicked you over by an amount equal to several times
18 ten-to-the-minus-five, then that would get into a region
19 that some would want to call a safety concern.
20 Your question, I think, goes really more to how we
21 allocated -- how we got the criteria in the first place,
22 because we want the combustion piece of LERF to be much less
23 than ten-to-the-minus-five, and that figures into our
24 functional reliability that we need.
25 I would argue that that's part of the
. 249
1 risk-informing task. They need to tell me how important
2 that function -- how reliable that function needs to be.
3 And there is allocation involved, I think, and I want them
4 to do it.
5 DR. KRESS: That was my point. There's going to
6 have to be an allocation in here.
7 MR. YOUNGBLOOD: Yes.
8 DR. KRESS: And that's not necessarily the
9 ten-to-the-minus-five LERF.
10 MR. YOUNGBLOOD: That's right.
11 DR. KRESS: Okay.
12 MR. YOUNGBLOOD: So what we conclude from this
13 exercise is that the guidelines fostered a useful discussion
14 and if they didn't before, maybe they have today.
15 I think that big chunks of the regulatory
16 framework that we would talk about here could be
17 performance-based to a significant degree. Not all of it.
18 I think there are pieces of it that can't be addressed that
19 way. I think the guidelines properly say that they can't be
20 addressed that way.
21 And so having carried out this discussion and it's
22 seeming to make sense, I think we violently agree with many
23 of your points, but the guidelines, such as they are, I
24 think, led to a useful discussion of this thing.
25 DR. APOSTOLAKIS: If you emphasize in your
. 250
1 write-up the points that we raised this morning, I think
2 this is a very good piece of work.
3 MR. KADAMBI: I agree.
4 DR. APOSTOLAKIS: How come you don't disagree now?
5 You don't say a word.
6 MR. KADAMBI: You made me an offer I couldn't
7 refuse.
8 DR. KRESS: This whole set of high level
9 guidelines requires PRA input, it seems to me. Now, there
10 are certain regulations in certain areas where PRA just
11 doesn't help any. You're just saying this doesn't apply to
12 those areas. You've got to do something else there.
13 MR. KADAMBI: If we can carry on to the next case
14 study, we'll try to address that.
15 DR. WALLIS: Could I just go back a little bit
16 about the summary slide here?
17 MR. YOUNGBLOOD: Sure.
18 DR. WALLIS: What you've showed us is a change in
19 how the regulations would be implemented in this particular
20 case. Is that what you're showing?
21 MR. YOUNGBLOOD: I carried out the discussion as
22 if there were no rule in place.
23 DR. WALLIS: No change in the regulation.
24 MR. YOUNGBLOOD: As if there were no existing rule
25 and we were trying to make a new one.
. 251
1 DR. WALLIS: Oh, I wasn't aware of that. I
2 thought you were going to show that this could be compatible
3 with an existing rule. You're actually going to propose to
4 change the rule?
5 MR. YOUNGBLOOD: What I did was take risk
6 information that says here is a need to be met, how would we
7 build a regulatory framework to meet that need. Now, there
8 happens to be a lot of ruling.
9 DR. WALLIS: Well, if you're going to change the
10 rule, for me, anyway, you would need more arguments about
11 why this is now a better rule than we had before.
12 MR. YOUNGBLOOD: I agree. I agree.
13 DR. WALLIS: That didn't come across at all.
14 MR. YOUNGBLOOD: Yes.
15 DR. WALLIS: There was no comparison. I wasn't
16 convinced --
17 MR. YOUNGBLOOD: One of many things that didn't
18 come across. I think that we should be expected to quantify
19 the risk that we accept under a given oversight scheme,
20 which is different from the PRA quantification we do today,
21 and compare it, because oversight schemes will have
22 different observation periods, different parameters that
23 they look at, different uncertainties associated with how
24 much you really know about the plant.
25 I think that that's a field of study that we
. 252
1 aren't doing enough of. If that's what you meant, I
2 completely agree.
3 DR. POWERS: If I can interrupt just a second, to
4 understand. Do we have an idea how long this will continue?
5 MR. YOUNGBLOOD: Two minutes.
6 MR. SIEBER: Great. That's how much time is left.
7 MR. KADAMBI: I'd like to just get into the second
8 case study, because it demonstrates the range of
9 applicability of the guidelines.
10 MR. SMITH: Good morning. I'm Chris Smith, from
11 ISL. I will talk briefly about case study two.
12 To start off, this one is very different from what
13 Bob just talked about, in that here we took a more focused
14 look at an existing rule and applied the guidelines to see
15 how performance-based is a rule as it is.
16 Actually, we took a recently revised rule, the
17 requirements to respiratory protection, Subpart H of 10 CFR
18 20. To start off, we looked at the changes that had been
19 made.
20 Now, the rules were revised in October of last
21 year and then became effective in February of this year.
22 They made several changes that resulted in greater
23 licensee flexibility. So decided that this was a good case
24 to throw the guidelines at and see if the guidelines
25 supported this claim that they are, in fact, more
. 253
1 performance-based.
2 We took the viability guidelines and we walked
3 them through in detail a couple of the specific changes to
4 the rule. I guess one thing I should make clear is rather
5 than Bob's looking at the entire regulatory framework, we
6 just looked at the rule part and what was changed there.
7 So in this sense, this is another case where it's
8 different than what you just saw. This is more of taking
9 the guidelines and assessing an existing rule.
10 This table actually shows a fair amount of
11 information all at once, so I'll try to walk through it as
12 briefly as possible.
13 The three specific rule changes we looked at
14 involved -- one of them, the first one was a modification to
15 the requirement that discussed what had to be considered
16 when a licensee performed their ALARA analysis.
17 The new revised rule added text saying that the
18 ALARA analysis performed as part of the Subpart H
19 requirements should consider non-radiological safety
20 factors. Now, this isn't necessarily something that the
21 licensee wasn't already doing. It's something that wasn't
22 in the existing rule prior to the revision.
23 DR. KRESS: Chemical toxins?
24 MR. SMITH: Correct. That would be a good
25 example, chemical toxins or inert air or high heat or any
. 254
1 other kind of OSHA type worker risk that's not radiological.
2 And we took that requirement and we ran through
3 the four viability requirements, do parameters exist, do
4 subjective criteria exist, is there licensee flexibility,
5 and is there sufficient margin to safety.
6 And we looked at the old rule versus the new rule;
7 that is, we looked at the delta; did the change result in
8 increasing the viability of performance-basing this area.
9 And looking through it, the parameters didn't
10 change, the criteria didn't change, the flexibility
11 increased, which might not seem straightforward right when
12 you look at it, because we added the requirement to the
13 rule, but in essence, the requirement was already there. We
14 just allowed them to consider it in one analysis rather than
15 splitting it up into two.
16 So as a result, your ALARA may now result in a
17 newly higher dose based on the way you schedule the work,
18 but the worker risk would be reduced.
19 So in that sense, we looked at it as this
20 increases the licensee flexibility. They've got one
21 analysis to look at worker risk and they can have more
22 options as to looking at something which may result in
23 slightly higher risk -- or slightly higher dose, but result
24 in lower risk.
25 So in that sense, we looked at the change to the
. 255
1 rule resulted in greater flexibility and the safety margin
2 was, if anything, increased. Like I said, looking at both
3 in one analysis, you're going to maximize your margin of
4 safety for total worker risk rather than just dose.
5 So in that sense, we looked at the revised
6 requirement as actually increasing the margin of safety.
7 DR. KRESS: When you do this, in the reactor
8 arena, I don't know whether this was for the materials arena
9 or not, but in the reactor arena, the probability of having
10 to protect yourself against radiological hazard --
11 MR. SMITH: Is small.
12 DR. KRESS: It's small, but it's higher and the
13 probability of having to protect yourself against chemical,
14 I believe, or maybe smoke. I don't know what the
15 probabilities are, but my point is it seems to me like in
16 assessing this whether you increase the margins or decrease
17 has to account for those probabilities some way, and I
18 didn't see that in your assessment.
19 Somehow it's a probability times consequence and
20 all I -- and I didn't see how you got both of those into the
21 equation.
22 MR. KADAMBI: What we are considering here is a
23 combination of the parameters that provide for safety of the
24 worker and what we were told by the staff expert in this
25 area is that when people are actually working in the reactor
. 256
1 environment, it is the non-radiological factors that pose a
2 much higher risk than the radiological factors.
3 And therefore, the end result is that workers are
4 more safe with the new rule.
5 DR. KRESS: So you did include this probability.
6 MR. KADAMBI: Yes.
7 MR. SMITH: Yes.
8 MR. KADAMBI: Let's move on. They've got to get
9 through.
10 MR. SMITH: Okay. Quickly, talking about the --
11 MR. ELTAWILA: Dr. Powers, how long do you want us
12 to continue? Because I would like to keep it on time. So
13 how many minutes?
14 DR. POWERS: Why don't you cut it off at quarter
15 of?
16 MR. ELTAWILA: Okay.
17 MR. SMITH: I just want to talk about this next
18 requirement that we looked at, because I think it brought up
19 a good important point that we didn't necessarily see ahead
20 of time, but learned from the application of these
21 guidelines.
22 The new requirement was revised to include
23 quantitative fit test criteria into the rule itself.
24 DR. APOSTOLAKIS: Can you point to the screen,
25 please?
. 257
1 MR. SMITH: I'm talking about this cell right
2 here.
3 DR. APOSTOLAKIS: Okay.
4 MR. SMITH: The rule was revised to actually add
5 fit test criteria into the rule, whereas before it
6 previously just said the licensees shall have a program,
7 documented program, and all the guidance documents contain
8 the actual fit test requirements that had to be met.
9 Clearly, this is a move into something that is
10 more prescriptive, but when we looked at the -- when we ran
11 the viability requirements across this rule change, we found
12 out that these fit test criteria are being used in
13 generating the dose calculations, which are being used as
14 performance parameters in many areas.
15 So when you're calculating your dose based on
16 somebody wearing a mask, it assumes that the fit test that
17 they had is working, the mask is working as designed.
18 Therefore, having the fit test criteria in the
19 rule is necessary to ensure that those performance
20 parameters are accurately recorded.
21 So even though it is more prescriptive, using the
22 viability guidelines leads you to understand that this
23 portion of the rule is not something you would pursue to be
24 performance-based and it makes sense to have it more
25 prescriptive.
. 258
1 MR. ELTAWILA: Prasad, maybe you want to wrap up
2 the discussion.
3 MR. SMITH: And I think the summary, I think, is
4 already what I've said.
5 DR. WALLIS: Your written summary said they
6 support the validity of the viability guidelines and wasn't
7 telling me why the viability guidelines were being tested by
8 this case. They were being used, but were they being tested
9 in some way?
10 MR. KADAMBI: The point of the case study was, in
11 fact, to test the guidelines to see whether they are ready
12 for prime time, as it were. And in this case study, we also
13 went beyond the viability guidelines. We had enough
14 information, because this was a rule already on the books,
15 we could assess the guidelines as changed, the next sector
16 of guidelines.
17 And what we found is that, yes, if we had gone
18 through these guidelines with the rule change that was
19 proposed back in October of last year, we would have come
20 out with saying that, yes, it makes sense to change the rule
21 in this way and there would be net benefit to be gained from
22 it in terms of safety, as well as in terms of flexibility
23 and, in some cases, actually increasing the safety margin.
24 That's really the bottom line. And if I may, let
25 me just -- sorry to go through these musical chairs like
. 259
1 this, but I think we have already covered this. Farouk
2 already went through the fact that we are going to be using
3 this in other kinds of areas, including risk-informed, as
4 well as the traditional approaches.
5 Our plan, I would like to just focus on this a
6 little bit. We are going to apply the guidelines in ongoing
7 and future changes.
8 Right now, our immediate plans are to apply the
9 guidelines in some of the option three effort. As part of
10 the risk-informed regulation implementation plan, there is
11 going to be a class of regulatory initiatives that are
12 identified as being not appropriate to be risk-informed. We
13 don't necessarily know what they are right now, but whatever
14 that turns out to be, we would examine whether the
15 performance-based guidelines might help us obtain regulatory
16 improvement in those cases.
17 The management directives that we hope to develop
18 will, in fact, focus the questions more towards specific
19 activities and we fully intend to incorporate the kind of
20 specificity that has been brought out in the discussion
21 today.
22 A communications plan, basically, what we would
23 like to do is people know about these guidelines and even
24 people outside the NRC, such as those involved in consensus
25 standards, may find it useful to apply the performance-based
. 260
1 guidelines when they seek to make consensus standards more
2 performance-based.
3 The planning, budgeting and performance
4 measurement process will really take into account a
5 combination of all these assessments when allocating
6 resources for a particular rulemakings or other changes to
7 the regulatory framework that we would undertake.
8 DR. KRESS: Let me ask you a question about that.
9 The examples we gave improved both safety and gave
10 flexibility and reduced burden.
11 Let's suppose you had a case where you reduced
12 burden, but actually increased risk a little. Do you have
13 guidelines on what to do about that particular condition?
14 Is that allowed?
15 MR. KADAMBI: I believe, depending on the
16 circumstances, it would be allowed. If you look at the
17 guidelines in Roman Numeral II, guidelines to assess change,
18 because there the performance goals of the agency are very
19 prominent and if we can substantially maintain safety while
20 decreasing unnecessary burden, that would support the
21 agency's performance.
22 DR. KRESS: It's those words like "substantially
23 maintain safety" that bothers Professor Wallis. Is this a
24 reverse regulatory analysis of some kind?
25 MR. KADAMBI: I hope the words in the guidelines
. 261
1 themselves are better than the ones I used off the cuff.
2 But in concept, that is really --
3 DR. KRESS: But in concept, that's allowed.
4 MR. KADAMBI: Yes. True.
5 DR. KRESS: But we're not sure what the criteria
6 are.
7 DR. POWERS: I guess I'm a little perplexed then.
8 If I come back to the individual respirator protection, how
9 do I use many of the societal goals that I've got to assess
10 that?
11 DR. KRESS: I don't think you can use some society
12 goals there. It's like some performance-based regulations,
13 where you have to have other kinds of goals.
14 DR. POWERS: I wonder what those would be.
15 DR. KRESS: That's another thing. I think you
16 need to establish the goals separately for some of these
17 rules other than just rely always on CDF and LERF.
18 Certainly in the NLSS materials area, you're not
19 going to have CDFs and LERFs. You may have equivalent to a
20 LERF, but you have to have goals right at the start, which I
21 would call acceptance criteria or something, and then you
22 work down from that.
23 But for this case, this is an ALARA goal. I think
24 you have a different -- you apply the different goal just as
25 if you have to meet it like you would a CDF or a LERF.
. 262
1 MR. KADAMBI: To not detain the committee much
2 longer, I believe we did meet the elements of the SRM that
3 we set out to accomplish. We did obtain stakeholder input.
4 We observed that the advisory committees can see these
5 guidelines and practices, they exercise their oversight
6 responsibilities.
7 The guidelines are subject to improvement and
8 certainly the comments that we've received today will be
9 part of what we will use to improve the guidelines.
10 We do expect that there will be better integration
11 as we move forward and gain experience in all these
12 initiatives.
13 Mr. Chairman, thank you very much for your time.
14 DR. POWERS: Thank you very much. I guess we're
15 in a position now to move on to our second topic for the
16 day, which is license renewal guidance documents. Dr.
17 Bonaca.
18 DR. BONACA: Mr. Chairman, as you know, the NRC
19 and the industry have been developing a set of license
20 renewal guidance documents. These documents are intended to
21 help the industry develop applications for license renewal
22 and to help the NRC in its review of individual
23 applications.
24 These documents appear to be close to being
25 released for public comment and the ACRS is planning to
. 263
1 review these documents during the October 19 and 20
2 subcommittee meeting of the License Renewal Subcommittee of
3 the ACRS, and is planning to write a report on the guidance
4 documents at the November meeting.
5 This morning we have the staff coming to tell us
6 the status of these documents and when they will be released
7 for our review.
8 With that, I will leave it to the NRC to take us
9 through the description of the standards.
10 MR. GRIMES: Dr. Bonaca, if I may. My name is
11 Chris Grimes. I'm the Chief of the License Renewal and
12 Standardization Branch, and as part of my introduction of
13 Dr. Sam Lee, who is going to make the presentation this
14 morning, I would like to inform you that just about an hour
15 ago, we notified NEI and the Union of Concerned Scientists
16 that we are shipping to them today copies of Generic Aging
17 Lessons Learned report, our revised Standard Review Plan for
18 License Renewal, a draft regulatory guide that proposes to
19 endorse an NEI guide, Revision 2, dated August of 2000.
20 And the collection of documents is going to be
21 distributed over the next several days. I expect a press
22 release later today announcing the availability of the
23 document and a Federal Register notice should be published
24 tomorrow announcing the availability of the documents for
25 comment and our plans for a workshop to be held on September
. 264
1 25.
2 DR. POWERS: What is the time period you have
3 allowed for the comments?
4 MR. GRIMES: The Federal Register notice requests
5 comments by submitted by October 16, which is a 45-day
6 comment period.
7 But as Dr. Lee will describe, we have already
8 solicited substantial stakeholder input in the development
9 of these drafts that we're issuing for comment.
10 Our purpose today is to walk you through the
11 organization and construction. I would note that Noel
12 Dudley has done an admirable job of carving up the guidance
13 documents to make assignments to the individual ACRS members
14 and we're going to explain to you what you can expect to see
15 in the documents.
16 We have one set of the documents here to show you
17 today and on Friday we would deliver an additional 15 sets
18 for the ACRS.
19 WE also will have all these guidance documents up
20 on the web, if not now, later today. I think that we turned
21 the CIO on just moments ago to go ahead and move the web
22 site over the external server.
23 So we are in the midst of the launch. These
24 represents another major milestone in the license renewal
25 program that fulfills a commitment that we made to the
. 265
1 Commission last fall in response to the issue that was
2 raised by the industry called credit for existing programs
3 for license renewal in SECY-99-148.
4 So we are still on schedule, we are still marching
5 forward to make improvements in the license renewal program,
6 and today we'll describe to you the material that will be
7 provided to the ACRS during the public comment period.
8 DR. UHRIG: Is this material significantly
9 different than the material that we've already been provided
10 with? We have large volumes of GALL.
11 MR. GRIMES: The trick is how significant is
12 significant. I would say that it is significantly different
13 because of the level of effort that has been put into
14 revising the generic aging lessons learned, since we
15 released the initial draft in December, last December, in
16 support of the first workshop.
17 But you aren't going to see the new programs that
18 have occurred or anything. It has been more in terms of the
19 level of detail that we've gone into in identifying which
20 programs do which things. But Dr. Lee will cover some more
21 of that during his presentation.
22 DR. BONACA: Just one comment. When some
23 restructuring of documents has been performed for some
24 specific reason, I've seen, for example, some significant
25 restructuring of the GALL-2 report, and I would appreciate
. 266
1 some information of why and it would help us in using some
2 of the review already performed to date and see how we can
3 expedite the review of the new documents.
4 So with that, I leave it to you.
5 MR. GRIMES: I would also like to note that
6 accompanying me here at the table is Dave Matthews, the
7 Director of the Division of Regulatory Improvement Programs,
8 and Dr. P.T. Kuo, who is the Section Chief that supervised
9 the development of the generic aging lessons learned SRP and
10 reg guide.
11 With that, Dr. Lee?
12 MR. LEE: My name is Sam Lee. I'm from the
13 License Renewal and Standardization Branch. Like Chris has
14 mentioned, this has been a major project by NRC and we have
15 significant work input from the NRR, Division of
16 Engineering, Division of Systems Integration and Inspection,
17 and also our Division of Reactor Improvement.
18 Also, we have significant input from the Office of
19 Research and the work is done mostly by the two national
20 labs, which are Argonne National Lab and Brookhaven National
21 Lab.
22 Also, the reviewers who have been involved in the
23 initial license renewal applications, they are also part of
24 this effort.
25 So we tried to capture a lot of experience, a lot
. 267
1 of effort into this product.
2 At the table, I have people from the License
3 Renewal and Standardization Branch, Jerry Dozier, Tammy
4 Bloomer and Rani Franovich. They are the licensee renewal
5 leads for individual chapters of the SRP, of the GALL
6 report.
7 The four documents we have in here are the GALL
8 report, the generic aging lessons learned report, and we
9 have the standard review plan, which is this, and then we
10 have the reg guide, NEI-9510, which is actually the reg
11 guide, just a couple pages up front endorsing the NEI-9510,
12 which is attached in the back. So these are the four
13 documents that we have.
14 We will leave you this copy today and we will give
15 you more copies, I guess, on Friday. We will also touch a
16 little bit on the schedule and then we have some specific
17 questions, and we'll just go over that review.
18 As background, like Chris said, this started
19 during the review of the initial license renewal
20 applications. Both the staff and industry recognized that
21 many of the license renewal programs to manage aging are
22 existing programs.
23 So that generated an NEI letter and we generated
24 this SECY-99-148, and that resulted in a Commission
25 direction. That directive started to prepare the generic
. 268
1 aging letters report, the GALL report. They evaluated
2 existing programs and document the basis for acceptance and
3 also identify where existing programs should be augmented.
4 DR. WALLIS: Can I ask you about the GALL report?
5 For some reason, we got some reports from Dr. Pollard, of
6 Concerned Scientists.
7 MR. LEE: Yes.
8 DR. WALLIS: Was the intent that he wanted them
9 put in the GALL report?
10 MR. LEE: I guess the GALL reported started a
11 couple years ago when we started looking across --
12 DR. WALLIS: I was wondering, how did you dispose
13 of something like that? Did you put it in the GALL report
14 or did you refute it or what did you do?
15 MR. LEE: In some cases, we actually added it into
16 the GALL report. They submitted -- Union of Concerned
17 Scientists --
18 DR. WALLIS: Is there a mechanism for refuting
19 arguments like that or responding?
20 MR. LEE: We actually had internal documentation
21 to actually document that we reviewed this report, this is
22 what we found.
23 DR. WALLIS: It seems to me that something like
24 that, which looks like a severe criticism, needs to be
25 responded to very carefully and well.
. 269
1 MR. LEE: We will consider that. But the
2 information has been captured into this GALL report now.
3 MR. GRIMES: Dr. Wallis, if I may add that during
4 the December 1999 workshop, David Lochbaum criticized us for
5 developing the GALL report without having considering their
6 perspective in terms of aging issues.
7 So our response to that was as we worked with NEI
8 to review their views about aging management programs, we
9 also undertook a review of the UCS reports to determine
10 whether or not they identified any unique or different
11 perspectives on aging effects that warranted aging
12 management programs.
13 But you are correct, we need to be very careful
14 about the context in which we describe the nature of that
15 review and its results and how they are used in GALL.
16 MR. LEE: And then the second bullet is to develop
17 the standard review plan to focus on areas that we need to
18 augment programs. And the Commission directed us to involve
19 stakeholders early and that's why we have workshops and we
20 actually get input from Dave Lochbaum.
21 Then we have to brief the Commission on -- we have
22 to issue the documents for public comment and then brief the
23 Commission on the public comments received, and these
24 documents are going to the Commission for their final
25 approval next year.
. 270
1 And after we have additional experience, and we
2 will, the Commission directed the staff to go back to the
3 Commission with recommendations, if any, for future
4 rulemaking to improve the license renewal process.
5 These four documents here interrelate, so they
6 kind of work together. The GALL report is the technical
7 basis document for the standard review plan and the standard
8 review plan provides guidance for the staff. In this case,
9 it's a focus on areas where existing programs would be
10 augmented.
11 Then the reg guide endorses 9510. Our goal is to
12 endorse 9510 with no exception. And 9510 lays out a process
13 for the applicant to follow to prepare the license renewal
14 application and it actually lays out what is an acceptable
15 format of the application. And the standard review plan
16 actually uses that format when we formulate the standard
17 review plan.
18 So we have tried to make all these documents
19 consistent. And one of the things that I guess we have to
20 admit is because we the GALL and SRP are evolving, so NEI
21 recognized that they would have to do additional work to
22 9510 to make it consistent with the final GALL and the
23 standard review plan. So you might see some inconsistencies
24 during, but that's not the intent.
25 We had early stakeholder involvement. We had
. 271
1 workshops, we had many meetings, all these meetings with
2 NEI. The Union of Concerned Scientists sent in five reports
3 that we have reviewed and then the next step is a workshop,
4 I guess, on September 25. That's in the auditorium.
5 DR. BONACA: You are telling me one of the things
6 that the ACRS would look at is, in fact, how well these
7 documents are integrated in a way that they provide clear
8 guidance and there isn't conflict and also that both the
9 applicant and the reviewer can have a common understanding.
10 You are telling me that there is still some lack
11 of integration in these documents with the NEI document.
12 MR. LEE: I would not call that integration,
13 because NEI doesn't know --
14 DR. BONACA: Because it's work in progress.
15 MR. LEE: Work in progress. They are trying to
16 catch up. And the SRP they are still changing. We just
17 need to fine tune it and finish that up.
18 DR. BONACA: Okay.
19 MR. LEE: Here is the schedule. We are at this
20 point right now. We are issuing this for public comment.
21 Actually, the Federal Register notice will go out tomorrow.
22 Then we have a workshop and we have to brief the ACRS in
23 October-November in more detail, and we are going to brief
24 the Commission on public comments received in November.
25 The way we understand it is this date is being
. 272
1 changed right now, but that was the last thing we have. So
2 this is being changed.
3 MR. GRIMES: I was notified by the Office of the
4 Secretary late last night that they couldn't get all of the
5 invitees together on the 27th, and so they are now polling
6 the Commission for an alternate date and it now looks like
7 the Commission briefing will be held on December the 5th.
8 MR. LEE: Then we have to brief the committee
9 again next year for the Commission approval in March of
10 2001.
11 The GALL report, this document is an evaluation of
12 existing programs and provide a basis why existing programs
13 are adequate and certainly to be augmented for license
14 renewal. It addresses the aging effects, it identifies the
15 programs, and then it calls for the program evaluation.
16 It has the ten elements, the ten program elements
17 you look at, such as what kind of inspection technique, what
18 kind of frequency, what kind of acceptance criteria, what
19 kind of corrective action, what control do they have.
20 And the GALL report is actually in two volumes.
21 So this might be a little change from what you had seen in
22 December. In December, this was the introduction section.
23 It got pulled out and made what we call the volume one,
24 which is just a summary. So this is basically a summary,
25 which was the introduction in the December version, if you
. 273
1 look at the December version. WE pulled that out.
2 Also, in here, we have provided a cross-walk from
3 the SRP into the GALL. So you find certain things, the SRP
4 said where did that come from, and in GALL, you go through
5 here. There is some table that points you to the particular
6 items in GALL, so you can trace that. So we call that a
7 bridge between the SRP and GALL.
8 It also has some indexes in here. So if you want
9 to look for a particular system, you try to find where it is
10 in GALL, so there is some indexes in here to help you do
11 that. So that's the purpose of this GALL.
12 DR. WALLIS: Independent review of this, is the
13 ACRS the only body that's looking at the whole thing that
14 might be regarded as an independent body?
15 MR. LEE: This document --
16 DR. WALLIS: Do we have the whole responsibility
17 to do the independent review or is someone else also helping
18 you with that?
19 MR. GRIMES: The ACRS has the whole
20 responsibility, until such time as we find someone who
21 volunteers as part of the public comments to take it on.
22 DR. WALLIS: So you haven't got an expert group or
23 anything.
24 MR. GRIMES: No. We do not have a designated
25 expert group or a peer review group for this report.
. 274
1 MR. LEE: I guess some entity, like the ASME, they
2 have assimilated the document, but they will be through the
3 public comment period.
4 Here is the GALL and this is pretty much like what
5 you had seen in December. The majority of the evaluation is
6 in Chapters 2 through 8. So that has not been changed.
7 If you look at the GALL report, the structure is
8 still the same as back in December. I just want to show
9 some pages.
10 So it's still the same as back in December. It goes through
11 the structure and component, what environment it's in, what
12 material it's in, and more aging effects, and more of the
13 programs, and that goes through this ten-element evaluation,
14 and then conclusion, whether it is adequate or you need to
15 augment it. So this hasn't changed since December.
16 But what has changed is sometimes it gets real
17 bulky to keep repeating this ten-element evaluation. So we
18 start putting things in what we call Chapter 11, which is
19 basically extracted this part and put over here, and then in
20 here we just say see Chapter 11 and then we give the number.
21 So that's a format change.
22 And we suggest, for convenience, so if you go to
23 Chapter 11, okay, you will not find all the aging management
24 programs in Chapter 11. Only certain things they kept
25 repeating many things, it's more convenient just to put them
. 275
1 in one place, we thought. Some of these, they are still in
2 the table.
3 And then, I guess, the other new things in here, compared
4 with the December version, is we have this called a TLEA
5 chapter, which is like a few pages. Doing the initial
6 application review, some of the applicants choose to propose
7 an aging management program to address the TLEA, a program
8 to address fatigue, or the nuclear program to address EQ.
9 So based on that experience, we've written up what we have
10 accepted in the initial license applications for aging
11 management of TLEA. So there are examples in there.
12 We also had a new chapter, which is Chapter 1.
13 This is a one page description of the ASME code. The staff,
14 when we go through this, we allow the ASME code a lot. So
15 the staff thought maybe we should provide basis in a
16 description of the ASME code, and that's what they use.
17 And then we go on to the standard review plan.
18 The standard review plan references the GALL report, so it
19 incorporates the GALL report, and then it focused the staff
20 review in the area where some programs would need to be
21 augmented.
22 And some of the features in here, this has not
23 changed from the previous version, if you've seen the
24 previous version.
25 DR. BONACA: This has not changed?
. 276
1 MR. LEE: The structure has not changed. The
2 information has changed, because we're going through
3 comments and the conforming change, so we make all the
4 conforming change.
5 There are tables in each chapter of the standard
6 review plan which captures information from GALL, basically
7 summarizes information in GALL. And that had changed
8 because GALL had changed.
9 And in here, the GALL Volume 1, we provide you a
10 cross-walk from this table back to this paper here. So you
11 say, gee, where did that come from, and go here, go to here,
12 it says this is item such and such, then go here and then
13 you can find it. At least that's the purpose.
14 DR. WALLIS: This is also electronic, so that if
15 you want to make the cross-reference, it's something that
16 helps you do to that?
17 MR. LEE: Right now we don't have that, no. We
18 have this, I guess, on the web and then this is in Microsoft
19 Word.
20 DR. BONACA: The logic on how all these documents
21 come together, it's in the SRP.
22 MR. LEE: Basically, the SRP pulls it together.
23 It takes information from NEI-9510, it takes information
24 from GALL, and from the reg guide. But then the 9510 cut
25 off early in terms of giving the applicant guidance in terms
. 277
1 of how to go through the process.
2 DR. BONACA: I understand that. But if the
3 members have to look for a guiding document that would walk
4 us through, it would be the standard review plan.
5 MR. LEE: It might not walk you through that much,
6 because it's -- this is the place where it ultimately will.
7 MR. GRIMES: I'll be bold enough to say the
8 standard review plan is the guiding document and that that's
9 where you should look for the best articulation we have of
10 the expectation on decisions regarding adequacy of aging
11 management.
12 DR. BONACA: The reason why I'm asking that
13 question is there is our concern is that when the reviewer
14 has to write an SER, there is a complexity here to the
15 amount of volumes and how they are integrated.
16 So that it is important that one document is
17 recognized as a guide through all this set of documents.
18 MR. LEE: Yes. For the staff to write the SER,
19 all it needs is the standard review plan. That's the goal.
20 DR. BONACA: Yes, but so many of the bases are
21 sunk into the reporting documents and certainly we want to
22 make sure the review is a thorough review and looks at the
23 technical issues and doesn't just reference something, but
24 the reviewer is able to understand how to get into the
25 reports and get to the issues.
. 278
1 MR. KUO: I just want to point out, in the
2 standard review plan, there is a section specifically for
3 reviewers that we call the review procedure.
4 DR. BONACA: Great. That's what I was looking
5 for.
6 MR. LEE: Okay. This uses the format from the
7 9510. But like I said, there's some consistency problem
8 with the format in that 9510 does not have the complete
9 benefit of GALL. So in certain places, you might say
10 describe this program, but then GALL says no further
11 evaluation is needed. So the SRP doesn't pick that up. So
12 you see a little inconsistency there, but we will fix that
13 by next March when we issue it for final.
14 And here is the table of contents for the standard
15 review plan and this hasn't really changed from the earlier
16 version, if you're seen it. It goes through the rule
17 requirements, the administrative portion, and then here is
18 the scoping, here is the methodology for identifying the
19 structures and components subject to license renewal
20 requirements.
21 And also here is the guidance in how do you
22 actually check the results, to make sure the results come
23 out okay. And Chapter 3 is the aging management review and
24 here is where the GALL report comes in. So you see an
25 individual section here corresponding to the chapters in the
. 279
1 GALL report.
2 The only exception is that in here we combine the
3 containment and structure together in one section in the
4 standard review plan. GALL has it in two chapters.
5 And here are some of the TLEAs that we have
6 identified based on the initial review of the applications.
7 So their comments, we can put them in there. And then we
8 have technical positions and one of them is on the ten
9 program elements, what does all that mean, and other things.
10 And the details of the table of content is in the
11 last two pages of your hand, so you can look at that.
12 Here is the reg guide and this provides guidance
13 for the applicant to prepare the application. Back in '96,
14 we issued a draft reg guide to endorse NEI-9510 Revision 0.
15 Since then, we have gained a lot more experience based on
16 the review of the initial applications. We have done some
17 scoping reviews, and now NEI has revised that and given us a
18 Revision 2 and the current draft reg guide, we would propose
19 to endorse that.
20 Here is the table of content on 9510. Basically,
21 it goes through the rule requirements. It talks about how
22 to identify your structures that are subject to license
23 renewal requirements, how to do aging management and the
24 TLEA evaluation, and then provides a format of the
25 application.
. 280
1 One of the useful things that this document has is
2 this has an appendix which identifies which structures and
3 components are active or passive. So that has a nuclear
4 aging management review.
5 DR. SHACK: Sam, as you're pulling the string on
6 these things and you go from the SRP to GALL, you're
7 eventually going to end up with references to documents like
8 the owner's group reports which are used to provide
9 justification.
10 Are there going to be publicly available versions of all of
11 those reports?
12 MR. LEE: We are working with EPRI. A lot of
13 these reports are EPRI reports. So now EPRI is giving us
14 public versions, also. The answer is yes, we are providing
15 at least a non-proprietary version.
16 We issue the FRN, the Federal Register notice, we
17 asked for comments on this document and, in particular,
18 we're interested in comments that goes to the extent of
19 which existing programs adequately manage aging. That is
20 the whole focus.
21 Then we have four specific questions; does the
22 GALL report actually provide sufficient -- some programs
23 should be augmented, identify that, and also in here we have
24 a question on being more specific. We have a process that
25 endorses correlations of the ASME code and we explain that
. 281
1 in the FRN and then in here we said they are the codes that
2 are not endorsed by 55(a) or by the process, how should we
3 handle that in GALL.
4 Also, the GALL report identifies the aging effects
5 for certain structures and components.
6 If an applicant decided for their plant that that
7 aging effect is not applicable, should they provide
8 information in the application?
9 And that is the last I have and like Chris said
10 earlier, the comment period ends October 16, a 45-day
11 comment period.
12 MR. GRIMES: Following up on a question that Dr.
13 Wallis asked earlier regarding peer review by other
14 organizations. Sam mentioned that we're continuing a
15 dialogue with the ASME to try to get ASME feedback on GALL
16 and how well it reflects the code programs to manage aging
17 effects.
18 Also, I would say that we're soliciting public
19 comments and when we say provide too much credit, there are
20 two ways that we might have provided unwarranted credit for
21 programs.
22 One is lack of technical basis. That is, there is
23 some technical flaw in our logic. Another aspect is that we
24 didn't articulate it clearly. There's a distinction between
25 whether or not we had good cause and whether we explained
. 282
1 good cause.
2 So we will be trying to seek out expert peer
3 review from other organizations; in particular, the other
4 code bodies. And to the extent that we can get other
5 recognized organizations to contribute voluntarily, we will
6 do that.
7 The ACRS, fortunately, we can ask of you, but some
8 of these other folks, we have to get them to volunteer their
9 attention.
10 DR. SEALE: Some time ago, I gave a reference back
11 to Mr. Dudley of some people in the DOE establishment who
12 are doing work in the aging management area. Some of the
13 materials they're talking about are not exactly the
14 materials you're interested in, but some of them are.
15 These are the people at Los Alamos. Have you
16 gotten any feedback from them?
17 MR. GRIMES: We haven't even contacted them yet.
18 I do have the name of the contact, but we consciously
19 decided that we would let Los Alamos recover from a fire and
20 security questions and we were pressing to get GALL issues.
21 But we do intend in contacting Los Alamos and sharing their
22 experience for the weapons programs.
23 DR. SEALE: In the elastomers area, in particular,
24 I think you have a lot of common interests.
25 DR. BONACA: I have a couple of questions regarding -- is
. 283
1 there a way in which we can understand how the experience
2 from the Oconee and the Calvert Cliffs applications have
3 been folded in the report? Is there any supporting
4 documentation that you have, a summary? You may have
5 developed something for just your use that shows a checklist
6 of how you addressed some of the issues in the report.
7 MR. LEE: We don't have such a summary.
8 DR. BONACA: Don't have such. Okay.
9 MR. GRIMES: The primary means of incorporating
10 the lessons from Calvert Cliffs and Oconee was drawing on
11 the same talent pool. I think the easiest way for us to be
12 able to respond would be that if you have a question about
13 particular -- whether or not Calvert Cliffs or Oconee
14 contributed to a particular element of GALL, just ask us and
15 we'll --
16 DR. BONACA: I'll give you an example of why I ask
17 the question. I reviewed the latest SRP version I had, not
18 this one, and I was looking at the scoping and screening
19 portion and it's a very important one.
20 I was left with the impression that, and maybe I'm
21 wrong, I haven't gone through in detail, but that although
22 the Oconee experience is reflected somewhat, the next plant
23 that comes up will have the same difficulties in sorting out
24 the accidents that are a part of the licensing basis and so
25 on and if that is the impression and that impression is
. 284
1 correct, then one could say that this is not a standard
2 review plan and that will allow for different applicants to
3 end up with different sets of assumptions, of accidents,
4 actually, that are going to be in their basis.
5 In that sense, I would say then the Oconee
6 experience is not properly utilized in support of the SRP.
7 Now, again, I'm not making a judgment. That was
8 an impression I had.
9 MR. GRIMES: I can see how you might get that
10 impression and I would tell you that when we reflected on
11 the Oconee experience, we considered Oconee to be relatively
12 unique with respect to the scoping issue and we attempted to
13 translate that experience into something that we considered
14 to be more typical or generic.
15 And we have specifically asked Arkansas to give us
16 feedback because they pattern themselves after Oconee. They
17 do not have the same kind of quality assurance
18 categorization process.
19 So we will be attempting to test whether or not
20 our translation is clearly articulated when we do our --
21 DR. BONACA: Let me give you an example. Clearly,
22 our intent is not one of helping write the SRP, but
23 certainly it would have helped me if, for a plant licensed,
24 fully licensed to current requirements, at least to have
25 been provided the set of accidents that you would expect to
. 285
1 be there, just as a reference, for example, and I didn't see
2 that.
3 And because it would help me to understand how an
4 older plant or understand even the applicant for an older
5 plant in what accidents he has to explain that are not part
6 of his own licensing basis.
7 It's just a suggestion, and I didn't see that
8 help.
9 MR. GRIMES: Actually, we are hoping that the ACRS
10 will provide us comments on how to write the report so it's
11 clear and we will consider that as one of the comments as we
12 go through the public comment process. That is, the
13 possibility of adding a list of accidents to consider as
14 part of the scoping process.
15 DR. BONACA: And the other question I had was
16 there were a number of license renewal generic issues that
17 were significant, I believe about 100 or so, and I imagine
18 that they have been resolved and have been addressed in the
19 body of the guidance documents.
20 MR. GRIMES: We went through the inventory of
21 generic renewal issues and we disposed of over two-thirds of
22 them. We have document -- we've got a database that we can
23 go back to identify where we expected to -- where we think
24 we resolved them in GALL or the SRP. We've got about a half
25 a dozen generic renewal issues that we're going to continue
. 286
1 to pursue; for example, improving guidance for fatigue
2 management, considering environmental effects. The issues
3 that continue to warrant dialogue.
4 DR. BONACA: Okay.
5 MR. LEITCH: How are they proceeding? Are they
6 more or less on hold until this new program is approved?
7 What is the status of the plants that are in the pipeline
8 now?
9 MR. GRIMES: The plants that are under review
10 right now are being reviewed using the practices that we
11 established on Calvert Cliffs and Oconee and the staff will
12 be able to take benefit of this improved guidance to develop
13 its safety evaluation, but the standard for the safety
14 evaluation is going to -- it will have to stand on its own.
15 It can't use this for reference. It can only use
16 this as a knowledge base.
17 MR. LEITCH: Thank you.
18 DR. BONACA: Any other comments from the members or
19 questions for the presenters?
20 [No response.]
21 DR. BONACA: If not, thank you very much for the
22 presentation and for the effort. We've come a long way.
23 Mr. Chairman, back to you.
24 DR. POWERS: Thank you. I echo your sentiment
25 that this has been an extraordinary undertaking. I think I
. 287
1 need to go through all the challenges that it poses.
2 I will emphasize that it's an extraordinary
3 undertaking. You're asking the ACRS to read a phenomenal
4 number of pages. It's at least ten percent of what we're
5 usually asked to read each month. So it's significant.
6 I would remind the members that this is part of
7 the statutory charter, so we really have to do it this time.
8 With that, I will recess us until quarter of the
9 hour, in which we will get to hear very interesting
10 discussion of the operating events at the Indian Point
11 Nuclear Power Plant.
12 [Recess.]
13 DR. APOSTOLAKIS: Back into session. Indian
14 Nuclear Power Plant Unit 2, the cognizant member is
15 Professor Seale.
16 DR. SEALE: Well, I guess you could almost
17 subtitle this "Where is Washington Irving when you need
18 him," because all hell seems to be breaking loose there for
19 a little while in the Hudson Highlands and Ichebod Crane's
20 old country. I think that's where it was, anyway.
21 A year ago tomorrow, there was a reactor trip,
22 which has been indicated as having complications that took
23 place at Indian Point 2. It was a trip. There were
24 problems with the diesels and one thing and another, and I
25 think we're going to hear this from the folks there.
. 288
1 Then that will be a relatively short presentation,
2 I understand, and then we're also going to hear about a
3 steam generator tube failure that took place on the 15th of
4 February.
5 So you had an exciting time, I guess, here
6 recently anyway. Mr. Marsh, are you going to be the --
7 MR. MARSH: I'm right here.
8 DR. SEALE: You're hiding.
9 MR. MARSH: I'm hiding from you. Good morning.
10 My name is Tad Marsh and I'm the Chief of the Events
11 Assessments and Generic Communications Branch in NRR.
12 DR. SEALE: Okay.
13 MR. MARSH: I have a couple of introductory
14 comments and then we're going to turn it over to our team up
15 at the front here to do the presentation for you.
16 This is a joint presentation by NRR and by
17 regional management and staff on the two events that you
18 described. Now, with the regions, NRR reviews, as you know,
19 and screens operational events as they occur. We also work
20 with the regions to decide what's the best way for the
21 agency to respond based on the severity of the events.
22 Our branch looks at the events and whether there's
23 any generic implications or any generic communications that
24 are needed. So that sets kind of the backdrop for who the
25 players are in this team for you today.
. 289
1 As a further backdrop for you, I'm sure you know
2 that operational experience is showing us that plants are
3 behaving better lately. There haven't been very many
4 significant events.
5 In the last perhaps 20-21 months, we've had three
6 AITs in this country. What you are going to hear about
7 today are the events surrounding two of those AITs.
8 So we look at these events with some significance,
9 as they do represent some fairly significant occurrences.
10 I'm going to turn it over to Eric now, who is
11 going to introduce the team and begin the presentation.
12 MR. BENNER: Thank you, Tad. As Mr. Marsh
13 indicated, we are here to discuss two events which occurred
14 at the Indian Point Unit 2 Nuclear Power Plant;
15 specifically, an August 31st trip, loss of off-site power to
16 four 480 volt vital buses, and then a subsequent loss of
17 emergency power to one of those vital buses; also, a
18 subsequent draining of a safety-related battery.
19 The second event will be a February 15, 2000 steam
20 generator tube failure. Again, as Mr. Marsh indicated, both
21 of these events resulted in the NRC sending an augmented
22 inspection team and we have the leaders of the two steams,
23 specifically Jimmy Yerokon and Ray Lorson from Region I to
24 discuss the particulars of each event.
25 Their presentations will provide sequence of
. 290
1 events, licensee response, root cause investigations and
2 safety significance. Also, as part of each event, we have
3 Jim Trapp, the Senior Reactor Analyst from Region I, to
4 discuss risk implications. We also have representatives
5 from the licensee and members of the technical staff here to
6 help answer any questions.
7 Mr. Goodwin and myself, who work for Tad, can
8 assist in any discussions of what we're doing generically in
9 response to these events.
10 Brian Holian, Deputy DRS Director from Region I
11 here, will conclude.
12 So at this point, I will turn it over to Jimmy to discuss
13 the first vent.
14 MR. YEROKON: Good morning. My name is Jimmy
15 Yerokon and I'm from the NRC's Region I office, and I led
16 the 1999 August event AIT at Indian Point 2.
17 I have been with the NRC for about 11 years.
18 Prior to that, I was in the industry, with about ten years
19 of experience, mostly as a startup engineer. Before that, I
20 finished degrees in mechanical engineering and physics.
21 Before I get to the event, I just wanted to point
22 out, I have an electrical simple drawing on my right side
23 here to help make my presentation easy. As we get to the
24 various points where I need to refer to the diagram, I will
25 do so and that should be very helpful.
. 291
1 Basically, for the event of August of '99, I will
2 group the discussion into three phases. The first phase,
3 I'll talk about the initiation process, how the event
4 initiated, and then I will get into the complications, which
5 is the bulk of the discussion, things that went wrong, and I
6 will finish up with the end result of what the event
7 eventually resulted into.
8 On August 31, there was an automatic reactor trip
9 at IP-2 and the automatic reactor trip was caused by a
10 spurious actuation of one of the channels of the reactor
11 protection system, OT Delta T. That is the over
12 temperature, delta temperature channels.
13 What happened is at that time, maintenance people
14 arrived, who had some pre-planned maintenance activity going
15 on in one of the channels of that OT Delta T circuitry, and
16 that channel or that circuit had a spurious actuation or a
17 spurious spike, and that that gave the 204 logic and there
18 was an automatic reactor trip.
19 So the trip was as designed, when you have 204,
20 the logic. And as designed for a plant trip, there are four
21 6.9 KV buses that provide the on-site power, supply that
22 switch to the off-site power supply.
23 The on-site power supply comes through this unit
24 auxiliary transformer and that feeds the buses one, two,
25 three and four that have the station power, station service
. 292
1 or like the RCPs, the main feedwater pumps and stuff like
2 that.
3 And you have the off-site power line coming here
4 through this station auxiliary transformer that provides
5 power to bus five and bus six, and those two buses, through
6 step-down transformers, provide power to the vital 480 volt
7 buses, buses 5A and 6A, although bus 3A and 2A also get
8 their power supply stepped down from bus two and three.
9 But anyway, when you have a plant trip as designed
10 for the site, these buses switch because you've lost the aux
11 transformer from the unit. They all transfer to the
12 off-site power supply.
13 And when that happened, there's obviously a
14 voltage dip on the bus and the buses are filling the station
15 power supply. And all this was as designed for plant trip.
16 The problems or the complications started shortly
17 afterwards, within about three minutes of the plant trip.
18 Suddenly the four vital buses all had their supply breakers
19 trip open. So there was a disconnect from the off-site
20 power line.
21 And as designed for a loss of off-site power to
22 those four buses, the diesels, they have three diesels,
23 those three diesels started as designed and they started to
24 load up on the buses.
25 Now, the next issue that occurred that was
. 293
1 unexpected was within about 10-12 seconds of the diesels
2 trying to load up on the buses, diesel 23, which was the
3 power supply to bus 6A, suddenly the breaker for that diesel
4 opened up.
5 Now, I will discuss later the reasons for all the
6 unexpected events that occurred. But the breaker opened up,
7 leaving bus 6A without any power supply. It didn't have
8 off-site power to it nor emergency power supply to it, also.
9 With the loss of power, complete loss of power to
10 bus 6A, the loads obviously were without power and one of
11 the loads on that bus, the one that gets into the aspect of
12 my discussion, is for battery 24. So with the power supply
13 lost, the battery 24 on bus 6A was physically handling the
14 load on a 125 volt A/C instrument bus.
15 After about seven hours, with several activities going on,
16 and I will touch in the course of my discussion, the battery
17 was depleted and with the battery depleted, there was loss
18 of power supply from that 125 volt A/C instrument bus and
19 several control room enunciators apart from that instrument
20 bus, and there was a loss of those control room enunciators.
21 Now, although those control enunciators are not
22 safety-related, but they do serve -- there are some safety
23 systems that have enunciators that reflect their conditions
24 as part of those enunciators.
25 As part of the management procedure for the site,
. 294
1 with the loss of 75 percent of enunciators for safety
2 systems, it calls for declaration of unusual event, which
3 was declared at 9:55 p.m. on that day of the event.
4 Several hours later, they were able to restore
5 power to bus 6A. The diesel was -- the breaker was closed
6 back up. They had some troubleshooting activities, trying
7 to determine the reason for the breaker of the diesel
8 opening to the bus.
9 That breaker went open and over-current. So they
10 took all the loads of the bus, they did some measuring, some
11 troubleshooting, and several hours after this, they were
12 able to satisfy themselves that it was safe to close the
13 breaker back to the bus. They did so. The diesel was
14 restored to the bus.
15 And after a short while, they were able to restore
16 the off-site power supply to buses 5A, 2A and 3A. And a
17 couple of hours later, they restored off-site power to bus
18 6A and secured the diesel.
19 With that, the event was basically over.
20 Now, the safety significance for this event; obviously, you
21 have the loss of bus 6A, loss of complete power to bus 6A,
22 and off that bus, you have one of the motor-driven auxiliary
23 feedwater pumps. You also have a train of ECCS pumps,
24 injection, RHR, service water, cooling water. So those
25 pumps were lost for that duration.
. 295
1 Although ECCS was not required, that train was
2 lost. Also, there was the loss of battery 24, which
3 resulted in loss of instrument bus 24 and loss of the
4 control room enunciators analysis for safety systems. That
5 factored into the safety significance of this event.
6 Also, there was increased burden obviously to the
7 operators. They had to deal with trying to restore off-site
8 power to the buses and also emergency power to bus 6A.
9 Some equipment that was lost during this event, one of them
10 was there was a control valve from the turbine-driven
11 auxiliary feedwater pumps, one of the generators that failed
12 open, and with that, they lost one motor-driven auxiliary
13 feedwater pump and a valve for the turbine-driven failed
14 open. So they had to cycle the turbine-driven auxiliary
15 feedwater pump on and off for level control in one of their
16 areas for some period of time.
17 After spending some few days in there, the AIT
18 found several root causes and contributing factors for this
19 event and we were able to group these into four main areas.
20 We grouped them on configuration control, which
21 mostly included the hardware issues. We also grouped some
22 of them in management oversight. There were some obvious
23 technical support problems both during the event and some
24 that were there before the event occurred, and we found some
25 inadequacies in some corrective actions, mostly some
. 296
1 problems that existed before the event that had not been
2 adequately corrected.
3 In the configuration control area, which includes
4 most of the hardware issues, the first one they had was a
5 problem with a load changer that works all the station
6 auxiliary transformer.
7 The changers are supposed to work if left in the
8 automatic control mode, is when you have a loss of power
9 from the on-site power source and the on-site loads transfer
10 to the off-site power, you obviously expect a dip in voltage
11 and the changer, what it does, it controls voltage upward
12 from this transformer, such that it maintains it constant.
13 So you don't have the voltage set point at the 480 volt, you
14 don't sense that to cause the loss of off-site power to the
15 buses.
16 In short, when you lose off-site power and you
17 transfer to the off-site, the design is such that you don't
18 lose the off-site to the 480 volt buses.
19 The changer had been maintained in the manual mode
20 as opposed to the automatic mode. So when that transfer
21 blew to the off-site power source, there was no automatic
22 control to step up the load on the secondary side and the
23 voltage dip was such that the degraded voltage relays for
24 the safety buses, the 480 volt buses, since -- the voltage
25 dip went below the set point and the relays picked up. So
. 297
1 that was the reason for the breakers opening up from the
2 off-site power source to the vital buses.
3 DR. SHACK: Now, is this a procedure problem, an
4 error problem?
5 MR. YEROKON: The top changer was the design, it
6 was supposed to be in the automatic mode. There were some
7 previous problems that needed to be repaired and for some
8 hardware issues and also scheduling issues, the repair had
9 not been made for about a year or so.
10 So it was in manual mode.
11 DR. SHACK: All this time.
12 DR. SEALE: This was a known and approved action.
13 That is, they knew that they had it in the manual and that
14 that was not according to the design specification.
15 MR. YEROKON: That's correct. So that was the
16 first real complication. The second one, with the buses,
17 the vital buses being powered by the diesels as expected for
18 the loss of off-site power, you also had the instance where
19 the breaker from EDG-23 tripped open on over-current, and
20 the reason for that is one of the over-current -- the
21 protection on the breaker, the short time over-current relay
22 had been set at 3,200 amps instead of the required 6,000
23 amps.
24 And the setting was -- they thought they had it at
25 6,000, but actually it was 3,200 and there were some
. 298
1 problems in the instrument used to measure the set point for
2 the trip current.
3 So when you had a loss of off-site power to the
4 buses, the diesel startup, the loads started coming onto the
5 bus. There were three heavy loads. I think it was
6 component cooling water system, aux feedwater, and there was
7 a third one that were trying to secure on the bus, and the
8 combination of currents was enough to trip the amps on that
9 breaker that was set at 3,200 as opposed to 6,000.
10 DR. UHRIG: Are there separate loads on the three
11 generators?
12 MR. YEROKON: Yes.
13 DR. UHRIG: But it's not a common bus.
14 MR. YEROKON: That's correct.
15 DR. UHRIG: In other words, you could not have
16 transferred those loads to the other two generators.
17 MR. YEROKON: That's correct. It's separate
18 trends, a trend A and a trend B.
19 MR. HOLIAN: Jimmy, while you're heading to the
20 next slide, one clarification on that tab changer. In the
21 past, Jim Trapp, when he introduces himself, he's a previous
22 licensed operator. Also, in the past they had operated in
23 manual on the tap changer years before. So that had been
24 done.
25 But when they had a license amendment and it
. 299
1 tightened up their degraded voltage set points for that
2 year, it should have been probably in automatic.
3 MR. LEITCH: Did the load sequence onto bus 6A, as
4 designed, or was there also a problem with the timing of the
5 sequencing?
6 MR. YEROKON: There was also some minor problems
7 with the sequencing. The sequencing relays, they had for
8 the three heavy loads I talked about, which are the aux
9 feedwater, the component cooling water pump and the service
10 water pump, there was a combination of electro pneumatic and
11 electronic type sequencing relays for those.
12 So with the stop in the electro pneumatic relay, plus or
13 minus for the sequencing, there was some overlap such that
14 it was possible for all three loads to be coming on at the
15 same time, are supposed to be sequencing at different times.
16 Now, if the breaker had been set correctly at
17 6,000 amps, that will have not been enough to cause the
18 problem, to cause the breaker trip. But the combination of
19 the overlap and the erroneous setting of the breaker --
20 DR. WALLIS: It's just a transient surge from
21 starting up three things at the same time.
22 MR. YEROKON: Right.
23 DR. UHRIG: Isn't that the purpose of the sequencing,
24 though, is to avoid that, to separate the transients?
25 MR. YEROKON: Yes. But like I said --
. 300
1 DR. UHRIG: But the combination of the three
2 steady-state or two steady-states and one transient were too
3 much.
4 MR. YEROKON: Yes. The overlap from the three
5 loads was possibly too much for the wrong setting. If it
6 was the 6,000 and you go up to 4,000, it would still be
7 okay. So you had all that combination.
8 DR. BONACA: Did they realize that the battery charger was
9 re-energized and, therefore, the -- it took a long time
10 before the actual enunciators were lost, right? Seven
11 hours?
12 MR. YEROKON: Right.
13 DR. BONACA: But did the plant realize that the
14 battery was discharging?
15 MR. YEROKON: Yes.
16 DR. BONACA: Because they knew this was taking
17 place.
18 MR. YEROKON: Yes. They knew it was going on, but
19 several things. They thought they had a real over-currents
20 on the loads on bus 6A. So they were trying to find the
21 reason for the over-current. It wasn't clear. It was then
22 before they found out it was some error in the setting.
23 So they didn't want to look back on the loads on the bus for
24 what the problem was.
25 They also had the loss of off-site power to the
. 301
1 other buses, so they were checking the loads on those other
2 buses also to see if there was some real -- there were two
3 issues. They thought they had real problems. They were
4 doing all this for several hours. That's why.
5 They were aware the battery was draining. They
6 knew -- as a matter of fact, the FSAR described I think it's
7 like a two, maybe 2.5 hour duration expected on the battery
8 and it went on for much longer than that.
9 So that was not much they really could do about
10 it, I guess.
11 I grouped the other three areas, they kind of go
12 together, the management oversight, the technical support,
13 and the corrective actions.
14 During the event, there were some real problems, I
15 guess, we find in the management oversight aspect. There
16 was a planned outage scheduled to occur shortly, later on, I
17 think a few weeks or shortly after this event occurred.
18 So the focus while this even was going on we found
19 was on shutdown work plans. The plant was tripped. I guess
20 there was some preparation going on. You know, we have a
21 trip, we go into the outage we had planned. So we thought
22 that was a misplaced focus.
23 There was also problems with the coordination and
24 use of resources. There were several things that were
25 happening. The battery was draining. There was a concern
. 302
1 with bus 6A, off-site power to the four buses really.
2 So what equipment do you go after first and what
3 do you fix, all that was going on.
4 Also, there were some issues before the event,
5 obviously, that caused all these things to line up. The
6 setting for the degraded voltage relays for the vital buses,
7 when the voltage dips enough, the relays pick up, but also
8 there is a time delay before you actually open the -- before
9 the breakers are opened and that time delay is to allow for
10 the recovery of the voltage dip.
11 Well, the licensee tested the pick-up for the
12 relays, but the drop-out of the -- the time delay comes in
13 about three minutes. If it's recovered, there's another
14 setting that does not allow the relay to trip the breakers.
15 They checked the pick-up for the relays, but they
16 never checked the set values. So there was some question
17 about even if the voltage had recovered enough, the relay
18 would probably still have opened the breakers.
19 DR. WALLIS: So I think what you're saying is that
20 they set up all these things waiting to happen. They were
21 set up for some time before. They were just waiting for the
22 initiation.
23 MR. YEROKON: Absolutely.
24 MR. SIEBER: Is it station personnel who sets the
25 relays or is it some off-site group?
. 303
1 MR. YEROKON: It's the station personnel do the
2 setting on the relays and also on the over-current amp.
3 It's done by the site.
4 As part of the technical support, also, before the
5 event, the procedures they had for emergency declaration,
6 there were some weaknesses we found in those. Actually,
7 when they lost off-site power to the four vital buses, for
8 some time duration, they should have declared an initial
9 event, instead of waiting for the control enunciators. So
10 that was some weaknesses in their procedures.
11 They didn't have a procedure for recovery of a
12 vital bus. With bus 6A losing all power supply, there was no
13 procedure in place to help in the recovery of that bus. So
14 they didn't have a procedure for recovery of a single vital
15 bus loss of complete power.
16 Also, we had some instances of inadequate
17 corrective actions. The initiator for the event, the
18 spurious actuation of the OT Delta T, that spike on channel
19 four wasn't the -- they had some instances, I think four or
20 five days before the event, where there were some spikes on
21 that channel and knowing that there was still some planned
22 maintenance going on on that same circuitry, where the spike
23 occurred in 204, that happened.
24 So there was some deficiency with the RPS, I think
25 we thought they didn't have timely corrective action to it.
. 304
1 Also, the load top changer, as Brian mentioned
2 earlier, that was known. There was a licensing basis that
3 relied on that being in the automatic mode and it was in
4 manual for about a year.
5 DR. SEALE: Could I ask? After the fact, did they
6 go back and make sure that there weren't some other breakers
7 that were mis-set because of the instrument that had given
8 them the false reading on 6A?
9 MR. YEROKON: Yes, they did. Yes. The check --
10 the breaker, I think, was a DB75 type of breaker. So they
11 checked similar breakers and they also expanded to some that
12 are now even DB75s, but they had some grouping methods and
13 they must have seven or eight breakers they looked at.
14 DR. SEALE: And the meter, I guess, that gave them
15 the bad reading, was that bad technique or was it a bad
16 meter problem?
17 MR. YEROKON: It was a bad technique.
18 DR. SEALE: So you had to re-train some of the
19 electricians.
20 MR. YEROKON: That's correct. What they were
21 doing, the secondary side of the amp current was where they
22 had been checking as opposed to the priority side. The
23 station next door was doing the right testing, so they were
24 able to go there and borrow some equipment to do the right
25 check.
. 305
1 DR. SEALE: And maybe an instructor.
2 MR. YEROKON: That was corrected.
3 MR. SIEBER: A question on your OTDT. Spikes are
4 pretty common on those in all plants that have them and
5 typically utilities will reduce power when they're going to
6 work on -- put one channel in trip, so that spike isn't as
7 high and doesn't come to the set point.
8 Did they do that before they did their maintenance
9 on the second channel?
10 MR. YEROKON: No, that was not done. And the
11 spike was not because of the maintenance.
12 MR. SIEBER: That's true, it was spurious. It
13 happens all the time.
14 MR. YEROKON: There was no power reduction to
15 implement that maintenance.
16 MR. SIEBER: Ordinarily, that would be what a
17 utility operator would do to avoid the possibility of the
18 trip and maybe five, six percent power is how far down they
19 would go. Typically, that will avoid a trip from that.
20 DR. WALLIS: You said spurious spikes occur all
21 the time?
22 MR. SIEBER: Yes.
23 DR. WALLIS: Then they're hardly spurious.
24 MR. SIEBER: A single spike is not predictable.
25 DR. SEALE: Random spikes, not spurious.
. 306
1 DR. UHRIG: It started at 2:30 in the afternoon.
2 When did the situation come under complete control here?
3 Did it exceed the seven hours that the battery --
4 MR. SIEBER: It was three in the morning.
5 DR. UHRIG: Three in the morning.
6 MR. YEROKON: The battery was -- it took about
7 seven and a half hours for the battery to deplete. The
8 unusual event occurred, it was about -- at 3:30 the
9 following morning.
10 DR. UHRIG: So it's 13 hours. How about the availability of
11 personnel? Did the people who were on the day shift stay
12 over?
13 MR. YEROKON: When the event occurred in the
14 afternoon and the battery depletion started going on, they
15 kept personnel behind to deal, because they had, like I
16 said, various issues unknown, and they kept -- they brought
17 extra people in.
18 But there was still the issue in the management
19 oversight with the focus on the right issues to deal with,
20 the resource loading, getting to do the right task.
21 So they had the personnel, but it was just the
22 management of the personnel. And also the focus on outage
23 planning that was also taking place.
24 DR. UHRIG: Because the people involved in outage
25 planning probably would not have been those capable of
. 307
1 handling this type of problem, would it? Or the management
2 personnel.
3 MR. YEROKON: There were some management personnel
4 that should have been involved with trying to deal with the
5 problem that were also involved with the outage planning.
6 DR. APOSTOLAKIS: I have a question. I gather
7 from your description of the event that there were several
8 latent errors, like the switch was not on automatic, it was
9 on manual and so on, that preexisted, of course, and the
10 trigger of the trip.
11 MR. YEROKON: Yes.
12 DR. APOSTOLAKIS: And you have identified a number
13 of issues here on this slide. Are you familiar with the
14 revised reactor oversight process?
15 MR. YEROKON: Yes.
16 DR. APOSTOLAKIS: If that process were in place,
17 would it have identified these issues?
18 MR. YEROKON: The load top changer, I'd say no to
19 that. There's no reason why the oversight process would go
20 look into the top changer. So that probably wouldn't have
21 been picked up.
22 DR. APOSTOLAKIS: You would not have picked up the
23 weak response during the event, would you?
24 MR. YEROKON: I'm not sure I --
25 DR. APOSTOLAKIS: Because there is no way to pick
. 308
1 it up.
2 MR. HOLIAN: In the response to the revised
3 oversight program, you would still have an event response,
4 similar. At the end slide, I'll compare these two events a
5 little bit and the fact that the steam generator tube
6 failure event did occur prior to initiation of the revised
7 oversight program, but the region's response to both events
8 was similar.
9 Now, if your question is back on prior to the
10 event --
11 DR. APOSTOLAKIS: Right.
12 MR. HOLIAN: -- would it have been picked up under
13 the revised oversight program, I guess I would comment that
14 prior to this event, under the old program, portions of it
15 were missed. I think that's what Jimmy is mentioning. That
16 thing was in manual. It was in a backlog. That was an
17 extensive backlog for the plant.
18 So even under the old program, there were aspects
19 maybe that the NRC inspectors also didn't realize the issue
20 of the top changer in manual.
21 So I don't know if we've done a formal review. I
22 know we've done a review of looking at the significance of
23 the events under the revised oversight program, that was
24 part of a Commission paper, to look at would they be
25 classified similarly under the revised oversight program,
. 309
1 and that has happened.
2 DR. APOSTOLAKIS: But the revised program does not
3 wait until something happens. It would have the performance
4 indicators and the baseline inspection program.
5 DR. KRESS: But one of those performance
6 indicators is the number of trips.
7 DR. SEALE: None of these would have shown up on
8 the screen.
9 DR. APOSTOLAKIS: The number of trips is not a
10 problem here. Probably is the integrity of the safety
11 systems.
12 DR. KRESS: I know, but this would have been
13 entered into the number of trips performance in that case.
14 DR. APOSTOLAKIS: But before the trip, I'm looking
15 for the latent errors before the trip.
16 MR. HOLIAN: Unless it was in the inspection
17 program where an inspector would have picked it up, a PI
18 wouldn't have picked it up, I don't think.
19 MR. TRAPP: When we talk risk significance,
20 though, if we had put these conditions through the SDP
21 process, which we would have done post-event, then these
22 would have come up to be risk significant issues.
23 DR. APOSTOLAKIS: I'm talking about pre-event.
24 MR. TRAPP: Pre, we wouldn't have been able to
25 identify them.
. 310
1 DR. APOSTOLAKIS: So I think the conclusion is
2 that parts of this would not have been picked up.
3 MR. HOLIAN: Right.
4 DR. APOSTOLAKIS: Now, there is a disagreement
5 between this committee and the staff regarding the so-called
6 safety conscious work environment, what we call safety
7 culture. And it seems to me that your first two bullets at
8 least there, the weak response during the event and focusing
9 on shutdown work plans and so on and coordination, and the
10 weak technical support before the event are part of the
11 safety culture, because that's a very broad area.
12 The disagreement is that the staff claims that
13 there is no need to pay special attention to the work for
14 safety conscious work environment, because if it's not good
15 enough, we will see the results either in the performance
16 indicators or something will happen and the baseline
17 inspection program will pick it up.
18 This committee claims that there is some validity
19 to this argument, but it's an untested assumption and there
20 are aspects of safety culture that probably will be too late
21 when you see them.
22 I am under the impression, and please correct me
23 if I'm wrong, that we are right. When you say weak response
24 during the event, I just don't know how a baseline
25 inspection program can pick that up or the performance
. 311
1 indicators are completely out of the question.
2 When you say weak technical support before the
3 event, again, I have difficulty seeing how the revised
4 reactor oversight process can pick that up.
5 So you are providing, I think, ammunition to this
6 committee with this description that these untested
7 assumptions are actually not very good assumptions.
8 Now, how do you respond to this?
9 MR. HOLIAN: The first comment I would like to
10 make is under the revised oversight program, one of the
11 keystone inspections is the problem identification and
12 resolution inspection. I know that that inspection has been
13 beefed up for looking at safety conscious work environment.
14 There is one aspect of that that must be addressed in each
15 of the inspection reports.
16 So there has been, from my view, a little bit of a
17 shift towards documenting a review on an annual basis of
18 that aspect and that's -- so that's an appropriate
19 inspection, from the staff's view here, to put it in.
20 A lot of the problems here at IP-2 that you will
21 see at the summary slide are corrective action problems that
22 have existed for years and if you classify that under the
23 broad term of safety consciousness at the station, I know
24 it's used in other terms of is an employee willing to raise
25 safety concerns, that's more often the case.
. 312
1 But if you were to broadly state it for an
2 effectiveness in the corrective action program, that shows a
3 long history at IP-2 and is something that I think both
4 programs were tracking and the revised oversight program, I
5 would say, would still be able to track that adequately.
6 DR. APOSTOLAKIS: That's why I was careful not to
7 include the third bullet in my argument, because I was sure
8 you would give me that answer. But for the first two
9 bullets, I don't think that helps.
10 And what makes matters worse is that we don't
11 think or some of us on this committee do not think that
12 there is a common understanding even of what a safety
13 conscious work environment is, and yet we see various events
14 or things that one would use the word weak, as you have done
15 there, to describe parts of that safety conscious work
16 environment.
17 So it's a mystery to me why the agency is not
18 paying more attention to this thing when you gentlemen, when
19 you actually do your job and you're doing a very good job
20 analyzing incidents, time and time again, get into
21 management issues, you identify these weaknesses, but then
22 when it comes to the oversight process or worse, when the
23 Office of Research tries to do something in this area,
24 somehow we don't do anything.
25 You probably have already suspected that there is
. 313
1 a lot of background to what I'm saying.
2 MR. YEROKON: The last slide just sheds some light
3 on the risk significance associated with this issue.
4 MR. TRAPP: I'm Jim Trapp and I'm one of two
5 reactor analysts in Region I, and I'm going to discuss a
6 little bit about the risk significance. And all the risk
7 significance that I discuss are a collaboration between
8 Region I SRAs and our PRA Branch in the Office of Research,
9 and we've also got input from Con Edison.
10 The conditional core damage probability for this
11 event was estimated to be about 2E-to-the-minus-four and
12 this is the NRC estimate. The licensee's estimate was about
13 1.88E-to-the-minus-four. And to put it in perspective --
14 very close. WE thought we were very close, until NRC said
15 we were not very close at all. That was a surprise to us.
16 DR. APOSTOLAKIS: What is the core damage
17 frequency that Indian Point 2 has estimated?
18 MR. TRAPP: Baseline is, with internal flooding,
19 3.3E-to-the-minus-five. That's internal events.
20 MR. YEROKON: I think he meant estimates.
21 MR. TRAPP: I'm sorry. That's their baseline CDF.
22 DR. APOSTOLAKIS: And what is the uncertainty they
23 have around it or is that the mean value?
24 MR. TRAPP: That's the mean value.
25 DR. BONACA: For internal events only?
. 314
1 MR. TRAPP: That's correct.
2 DR. APOSTOLAKIS: Does it have a --
3 MR. TRAPP: With flooding.
4 DR. APOSTOLAKIS: Well, you would classify this as
5 an internal event, though.
6 MR. TRAPP: Right.
7 DR. APOSTOLAKIS: Do you have any idea what their
8 upper bound is?
9 MR. TRAPP: No. No, I don't. I'm not even
10 certain that they have -- that they've done that. I
11 shouldn't say, but I'm not certain that they have that.
12 DR. KRESS: Typically, it's a factor of ten on
13 CDF. Just under ten.
14 DR. APOSTOLAKIS: Three or four usually.
15 DR. POWERS: If you want an estimate on it, you
16 can always look at the design Indian Point study, because
17 they did an uncertainty analysis there.
18 DR. APOSTOLAKIS: So you are talking about almost
19 an order of magnitude change here.
20 MR. TRAPP: It would be like operating at full
21 power for approximately four years.
22 DR. APOSTOLAKIS: And, again, coming back to your
23 findings about weak response and inadequate management
24 attention and so on, I know for a fact, even though I
25 haven't looked at this particular PRA, that these kinds of
. 315
1 issues, organizational factors, are not in the PRA.
2 Should I conclude then that if I put them in the
3 PRA, my core damage frequency would be significantly
4 affected, not necessarily go up to ten-to-the-minus-four,
5 but here is something that happens because of all these
6 things, latent errors and so on, and it changes my core
7 damage probability by this much?
8 I mean, is my PRA in serious trouble because I
9 have not included these kinds of events?
10 MR. TRAPP: I'm probably not the best person to
11 answer you.
12 DR. POWERS: He honestly didn't expect an answer.
13 DR. APOSTOLAKIS: I might note that there was a
14 report issued by the Union of Concerned Scientists recently
15 and you certainly support some of their arguments with your
16 presentation.
17 MR. TRAPP: Undoubtedly, right. Because it was
18 those latent errors in the design, in the tap changer and
19 the latent error in the breaker setting that was really the
20 cause of this event, and if you factor that into the IPE,
21 what you really have is every time you have a reactor trip,
22 you're going to lose power to the --
23 DR. APOSTOLAKIS: So this is a real omission then
24 in the PRA.
25 DR. BONACA: For the review of the corrective
. 316
1 action program or the conditions of it, the amount provides
2 some insight?
3 MR. TRAPP: But as you know, I guess, when you do
4 the PRA, you have to start with something. You have to
5 assume that the design is the design, you have to assume the
6 tech specs are going to be followed. If you don't follow
7 the tech specs and you don't have the design, then I don't
8 know where you would start.
9 DR. APOSTOLAKIS: Do you really have to?
10 DR. BONACA: I don't know.
11 DR. APOSTOLAKIS: Somebody ought to check.
12 DR. BONACA: I asked the question before because
13 you mentioned before that there were some issues with the
14 corrective action program, a big backlog, and so there would
15 be a correlation there that one could make regarding this
16 issue of human performance or some aspects of safety
17 conscious work environment.
18 MR. HOLIAN: Yes. I don't know if, in real time,
19 to put a percentage on that, drop their IPE number, to try
20 to quantify that.
21 DR. BONACA: I didn't mean that. But there seems
22 to be a reliance in the new oversight process on the
23 corrective action program.
24 MR. HOLIAN: Yes.
25 DR. BONACA: And that seems to be the most direct
. 317
1 indication of human performance and safety conscious work
2 environment.
3 MR. HOLIAN: Yes. And one other aspect of that,
4 probably just a reminder, is that the revised oversight
5 program, on the problem identification and resolution, also
6 each inspection report throughout the year tries to address
7 an aspect of that and that's supposed to be pulled up in the
8 annual.
9 DR. BONACA: But you believe that one could
10 develop a performance indicator based on some metrics of the
11 corrective action program of an individual unit.
12 DR. APOSTOLAKIS: Is that an unfair question?
13 MR. HOLIAN: It would be worthwhile. I think
14 based on the revised oversight program's reliance on that, I
15 think it is one of the key inspections that we do and so
16 from a regional perspective, we put a lot of emphasis on
17 looking at that aspect.
18 So if it could be done, it would be worthwhile.
19 DR. BONACA: Because the units have developed some
20 metrics that they look at for their own corrective action
21 program and they seem to be pretty reliable. The
22 categorization of the significance.
23 MR. HOLIAN: Significance categorizations,
24 trending.
25 DR. BONACA: How long it takes to reduce it down,
. 318
1 the backlog. I mean, so many -- there are a number of
2 indicators that are pretty significant.
3 DR. APOSTOLAKIS: But, again, I would like to
4 note, in this context, that on slide ten, you're only
5 addressing the third bullet. These indicators you are
6 referring to would not handle the first two bullets.
7 So just relying on the corrective action program
8 leaves out a significant part of what we think we understand
9 by safety conscious work environment.
10 MR. LEITCH: I think it may get at the second
11 bullet. If you take a look at the corrective action
12 program, the age of the corrective actions and the backlog,
13 it may, although not have identified these two specific
14 issues, it may give some indication of the technical support
15 prior to the event.
16 For example, if you look at the number of operator
17 work rounds, which were, I think, significant in this
18 particular situation, that should give you advance warning
19 of a problem in the technical support area.
20 DR. APOSTOLAKIS: And I agree that this may very
21 well be true in many instances. But still, all these things
22 are done in an ad hoc manner. Nobody -- I have yet to see a
23 report or a paper or something that says here is a way to
24 correlate these things to understand, to look for these
25 things when you see this and that.
. 319
1 For some reason, we have decided as an agency not
2 to pursue that. I mean, there is no question that there is
3 a lot of stuff one can do with these things and perhaps the
4 answer will be a corrective action program that would be
5 expanded to include other things that are now unfortunately
6 called organizational issues there, where the agency should
7 not touch them.
8 DR. BONACA: There are indicators that the
9 utilities still use that look more at the two top ones.
10 There are indicators of impediments to the operators.
11 DR. APOSTOLAKIS: But this agency has not looked
12 into this.
13 DR. BONACA: I know, I understand.
14 DR. APOSTOLAKIS: And this is a utility. So
15 obviously these indicators are not 100 percent successful.
16 DR. KRESS: When they test diesel generators to
17 see what their reliability is to start up and carry load, it
18 wouldn't have picked up this undercurrent problem with the
19 relay.
20 MR. SIEBER: No.
21 DR. KRESS: I would have thought it would, but I'm
22 asking the question, would it?
23 MR. SIEBER: For start, run and load tests, it
24 would, provided it was done in automatic.
25 MR. TRAPP: My impression is that when they --
. 320
1 Jimmy, help me here. When they test the diesels in that
2 manner, they don't load them to that set point.
3 DR. KRESS: I see.
4 MR. TRAPP: That's a pretty heavy load to test the
5 diesels at.
6 DR. KRESS: It may be a deficiency in the test
7 procedure.
8 MR. SIEBER: Yes.
9 MR. YEROKON: That's correct. The one test that
10 could have come close to picking up on this issue is the one
11 that's done every refueling outage, and I'm not sure how the
12 sequencing is applied during that testing.
13 If the sequencing works fine, then you won't pick
14 it up. It has to overlap for that current. So it's
15 possible once every refueling to test that, but, also, it's
16 possible the testing would not reveal the inadequacy with
17 the set point. The other tests monthly don't go to that
18 area.
19 DR. POWERS: If I could touch on a question on
20 this CCDP of two-times-ten-to-the-minus-four. If memory
21 serves, the recovery process itself on the event involves
22 some under-cooling.
23 MR. TRAPP: I'm sorry?
24 DR. POWERS: Involved under-cooling.
25 MR. TRAPP: I don't recall that, but it certainly
. 321
1 would have been factored into our analysis.
2 MR. YEROKON: I think if you discuss the
3 assumptions for the --
4 MR. TRAPP: We are going to zip through the
5 assumptions that were made for that. It might help. But
6 I'm not sure we did that aspect of it.
7 DR. APOSTOLAKIS: Was this presentation, Professor
8 Seale, supposed to be shown?
9 DR. SEALE: Yes.
10 MR. TRAPP: To put the CCDP in perspective, two
11 years ago, in '98, David Besse had a tornado loss of
12 off-site power, some problems with a diesel. Their CCDP,
13 issued in the ASP and NUREG, was 5.6E-to-the-minus-four. So
14 they were roughly a little higher than we were. They
15 operated about a day on the diesel. So that's just to give
16 some perspective. So this was about three times less than
17 what was seen at Davis Besse.
18 The dominant sequence, when you lost 6A, you lost
19 -- the key here was you were moving decay heat through the
20 steam generators. So the important aspect from an operator
21 point of view was to keep aux feed or feed going into the
22 steam generators.
23 The other key here is you lost primary
24 once-through cooling, because the Indian Point operates with
25 the block valves closed and they're electric. So you would
. 322
1 need 6A to open one of the PORV block valves to get
2 once-through core cooling.
3 So the way it usually works is if you lose aux
4 feed, you go to once-through core cooling. They didn't have
5 that option.
6 They lost the 23 aux feed pump, they lost the PORV
7 block valve, and they lost 6A. The diesel, 22 diesel was
8 supplying 21 motor-driven pump. They have two motor-driven
9 pumps, aux feed pumps, one turbine-driven pump.
10 So that piece of equipment was key to them to keep
11 the 21 aux feed pump running.
12 Jimmy expressed some of the problems they had with
13 the turbine-driven pump, but certainly that was functional.
14 So really the dominant core damage sequence for
15 them is that they lost the 21 motor-driven aux feed pump,
16 they lost the turbine-driven pump, and they failed to
17 recover feedwater, then they would have gone to core damage
18 because they didn't have once-through core cooling.
19 One of the interesting aspects of this loss of
20 off-site power that would be different is the 6.9 power
21 stayed on. The key is here you lost the power going to the
22 480 volt buses. The 6.9 power was available, so you still
23 had a condensate pump and you still had the feedwater train.
24 So that was a little different than your typical loss of
25 off-site power, and we had to factor that into the risk
. 323
1 analysis.
2 Of course, if you lost the 22 diesel, you'd also
3 lose the 23 motor-driven aux feed pump. So that cut set
4 would also be factored in.
5 Conservatisms in the analysis is no credit was
6 given for 480 volt bus recovery and the key here is that the
7 operators would -- if need be, if they lost the 23 or the 21
8 aux feed pump and the turbine-driven aux feed pump, they
9 could have repowered 6A. That breaker is set -- the amps
10 were set about half, but if you just put on the 23 aux feed
11 pump, put everything else and pull the lock, close the
12 breaker, that pump would have been re-energized. So that
13 wasn't part of our analysis. We didn't give them credit for
14 that.
15 The other part of the analysis is the 21 aux feed
16 pump has a separate supply of power through Unit 1. This is
17 the Appendix R alternate A/C power source and there's some
18 switches down locally that you can switch in alternate power
19 from the old retired Unit 1.
20 That was also not factored in, that they could
21 have repowered the motor-driven pump had the 22 diesel
22 failed. That wasn't factored in.
23 And, of course, since the 6.9 buses were still
24 energized, there was a chance that they could have restored
25 power from the normal off-site power supply. That evolution
. 324
1 wasn't proceduralized and required some manipulations in the
2 logic, the black-out logic.
3 So that recovery action would probably not have
4 been able to be done very quickly.
5 The other key assumption here from a risk point of
6 view was, of course, that the -- as I discussed, the
7 feed-and-bleed requires two out of two for success. That's
8 another conservatism in our assumption, in that if the
9 turbine-driven aux feed pump or the motor-driven had worked
10 for a period of time, the one out of two PORV might have
11 been adequate to do once-through core cooling, and that
12 analysis hadn't been performed.
13 DR. BONACA: Why is the PORV blocked? Is it
14 because of spurious safety injection concerns?
15 MR. TRAPP: It would be the PORV block valve, so
16 it would either be that you're afraid your PORVs are leaking
17 or you're afraid that -- you know, it does eliminate one
18 path for a small break LOCA in that if your PORV
19 inadvertently opens, now you won't have the LOCA, so there's
20 some balance there.
21 DR. BONACA: So a temporary block is not -- for a
22 period of time, these plants were running with a blocked
23 PORV because of concern with spurious safety injection.
24 That would be a significant impact on the IPE values they
25 have presented, because I think they have relied on
. 325
1 bleed-and-feed in those assessments.
2 MR. YEROKON: The blocked valves are closed, but
3 you also realize they have safety paths supplied to them.
4 So it wouldn't be a significant impact, because you have a
5 safety source.
6 MR. TRAPP: There is a risk balancing going on
7 there. You're taking away your small break LOCA probability
8 before they fail open, but you're losing now, that if you
9 lose the bus like they did, you lose feed-and-bleed.
10 DR. BONACA: Most of these PWRs have credited
11 bleed-and-feed and that's a significant capability that if
12 you deny that, that would be a significant contribution
13 increase to core damage frequency.
14 MR. YEROKON: The credit is okay as long as you
15 have the safety power supplied to it, which they did.
16 MR. TRAPP: From the model, as long as you model
17 it with the PORV block valves closed, so you don't lose the
18 bus. As long as you do that correctly, then you're okay.
19 DR. SEALE: Could we move along now?
20 MR. YEROKON: Thank you.
21 MR. LORSON: Good morning. My name is Ray Lorson.
22 I'm going to follow the short presentation and discuss the
23 steam generator tube failure event which occurred at Indian
24 Point 2 on February 15th.
25 I was the augmented team inspection leader that
. 326
1 followed up inspection.
2 I've been with Region I for approximately nine
3 years.
4 I've been a resident inspector at several sites,
5 and currently, I'm the senior resident inspector at Seabrook
6 Station.
7 I'm going to discuss for the sequence of events
8 that occurred, talk a little bit about the safety
9 significance of the event.
10 We'll look at some of the root cause areas which
11 complicated the event during the recovery, and Jim Trapp
12 will talk about the risk significance of the event.
13 We're going to start with a short video that the
14 licensee made of the conditions inside the failed tube.
15 The tube that had the failure was the row 2,
16 column 5 tube in the number 24 steam generator, and we'll
17 start with the tape.
18 Okay.
19 Here we're inside the row 2, column 5 tube. You
20 can see the beginnings of the crack. The crack is
21 approximately 3 inches long and approximately 1/8th of an
22 inch in width in some locations.
23 The crack will come into --
24 DR. WALLIS: We are inside a tube?
25 MR. LORSON: We are inside a steam generator
. 327
1 U-tube, that is correct.
2 DR. SIEBER: This is up in the U-bend?
3 MR. LORSON: Yes. You can see the crack here.
4 Some of the widest points in the crack was approximately an
5 eighth of an inch.
6 DR. POWERS: Looks like a hole.
7 MR. LORSON: And we'll get to the end of the crack
8 here in a second.
9 As I mentioned, it was approximately 3 inches
10 long.
11 DR. UHRIG: That was obviously taken at no
12 pressure, right?
13 MR. LORSON: That is correct. This was taken by
14 the licensee.
15 DR. UHRIG: If there was any pressure at all, it
16 would probably spread it more.
17 DR. WALLIS: Why does it suddenly get bigger? It
18 just moved?
19 MR. LORSON: You're asking about the shape of the
20 crack?
21 DR. WALLIS: Yeah.
22 MR. LORSON: The probe is actually entering into
23 the tube.
24 So, it's just strictly the characteristics of the
25 crack are changing.
. 328
1 DR. WALLIS: Changed its focus or something?
2 MR. LORSON: No, it's the probe itself actually
3 transversing through the tube.
4 DR. WALLIS: We're just looking at one piece of
5 the crack here, then.
6 MR. LORSON: We actually started at the beginning
7 of the crack and we went over to the end of the crack.
8 DR. WALLIS: So, this is some sort of medical
9 technology they're using?
10 MR. LORSON: Actually, what you didn't see -- and
11 we tried to cue the tape right up, but just shortly before
12 that part, you could have saw the instrument that was
13 actually inserted into the actual tube sheath itself, and it
14 was basically a small-diameter-type camera.
15 The event was caused by primary water stress
16 corrosion cracking of the row 2, column 5 tube in the number
17 24 steam generator.
18 We just saw an indication that was present in the
19 tube.
20 The initial primary to secondary leak rate was
21 approximately 150 gallons per minute. That was at the
22 initiation of the event.
23 DR. WALLIS: Did it get bigger later on?
24 MR. LORSON: No, actually, that was the maximum
25 primary to secondary leak rate, and the actual leak rate
. 329
1 went down over time.
2 DR. KRESS: You find this leak rate by measuring
3 the make-up water required?
4 MR. LORSON: I'm sorry?
5 DR. KRESS: How do you determine this leak rate?
6 By measuring the make-up water required to go into the
7 primary system?
8 MR. LORSON: The licensee did a mass balance on
9 the primary plant.
10 Based on that, they could determine the make-up
11 volume that was required to operate the plant.
12 DR. UHRIG: They did this after the plant was shut
13 down.
14 MR. LORSON: That is correct.
15 DR. UHRIG: Is this a four-loop or a three-loop
16 plant?
17 MR. LORSON: Indian Point 2 is a four-loop
18 Westinghouse pressurized water reactor.
19 DR. UHRIG: It's about 1,100 megawatts?
20 DR. SIEBER: Typically, for Westinghouse steam
21 generator, they totally plug rows 1 and 2. Did Indian Point
22 plug either 1 or 2?
23 MR. LORSON: Not prior to the event.
24 DR. UHRIG: This is the original steam generators.
25 MR. LORSON: Yes.
. 330
1 DR. SIEBER: Not annealed.
2 Did they have indication prior to the event that
3 there was some kind of leakage? Usually you can pick that
4 up by analysis.
5 MR. LORSON: Yes, and I'll talk about that
6 briefly.
7 DR. SIEBER: All right.
8 MR. LORSON: During the event, there were several
9 complications associated with the attempt to place the plant
10 into cold shutdown.
11 That was one of the focus areas for the team, and
12 we'll talk about that during this presentation.
13 The results of the tube rupture are that the plant
14 remained in an alert status for approximately 24 hours, and
15 there was a minor radiological release.
16 The next slide, we're going to discuss the
17 sequence of events.
18 On February 15th, at approximately 7:17 p.m., the
19 operators detected secondary plant radiation monitors, which
20 was indicative of a primary to secondary tube leak, and they
21 also noted a decreasing pressurizer level.
22 The operators carried out the immediate actions
23 that were identified by the emergency operating procedures
24 and by the abnormal operating procedures.
25 As you can see through the sequence of events, the
. 331
1 affected steam generator, the number 24, was isolated by
2 8:31 p.m.
3 At 9:02 p.m., the operators began a manual primary
4 plant cool-down to reduce reactor coolant system
5 temperature. The cool-down that they initiated was very
6 large, and it resulted in a rapid drop in the pressurizer
7 level such that the operators were required to manually
8 initiate safety injection at 9:04 p.m.
9 At approximately 11:38 p.m., the tube leakage was
10 initially stopped as the steam generator pressure and the
11 reactor coolant system pressure were approximately zero, the
12 differential pressure between those two pressures.
13 As I mentioned earlier, there were some
14 complications in placing the plant shut-down cooling system
15 in service.
16 However, on February 16th, at 12:39, the licensee
17 did place the shut-down cooling system in service, achieved
18 cold shut-down at 4:57 p.m., and terminated the alert
19 declaration at 6:50.
20 The next slide will discuss briefly the safety
21 significance of the event.
22 The team determined that the operator -- the
23 initial operator response was prompt and appropriate and
24 noted that the licensee was successful in achieving a cold
25 shutdown condition.
. 332
1 However, as I mentioned previously, there were
2 several operator performance and procedural issues,
3 equipment issues that were identified which delayed
4 achieving the cold shut-down conditions.
5 There were also some problems noted in the area of
6 emergency response, and as I mentioned earlier, the event
7 did have a minor radiological release. However, the team
8 determined that there was no measurable off-site
9 radiological release from this event and concluded there was
10 no impact on public health and safety.
11 DR. KRESS: At 150 gpm's is a fairly sizeable
12 leak. Did anybody go in and examine the adjacent tubes to
13 this crack to see if there was any measurable effect on the
14 adjacent tubes?
15 MR. LORSON: I think the issues related to the
16 condition of the steam generator were looked at in a
17 subsequent or a follow-up NRC inspection. I think Brian may
18 talk about that a little bit at the conclusion of my
19 presentation.
20 The purpose of the augmented inspection team that
21 we formed following the event was primarily to focus on the
22 event.
23 DR. BONACA: I have a question.
24 There is a set time, I believe, to balance primary
25 pressure and secondary pressure. That's one of the
. 333
1 objectives of the procedure, of the emergency operating
2 procedure. What's the time?
3 MR. LORSON: The time when they initially --
4 DR. BONACA: No, no, no, I'm talking about the
5 target time for which the EPGs, the emergency operating
6 procedures are set.
7 MR. LORSON: I'm not aware of any specific time
8 that's listed in the emergency operating procedures.
9 However, I think the --
10 DR. BONACA: It's part of the licensing of these
11 procedures, I believe. You have to demonstrate that you can
12 achieve balancing of primary and secondary pressure within a
13 given time.
14 If I remember, it's less than an hour, I thought.
15 MR. LORSON: I'm not aware of any requirement.
16 DR. BONACA: But I didn't see any discussion of
17 that.
18 MR. TRAPP: We'll discuss that a little when we go
19 into risk significance.
20 DR. WALLIS: The leak is flashing steam water when
21 it comes out of the hole?
22 MR. LORSON: Yes.
23 DR. WALLIS: It's a mixture of steam and water.
24 MR. LORSON: Yes.
25 DR. WALLIS: And is it going into a sub-cooled or
. 334
1 a saturated liquid?
2 DR. KRESS: It depends on where the break it.
3 DR. WALLIS: The worst situation would be steam
4 going into sub-cooled water.
5 MR. LORSON: Right. The break would have occurred
6 up at the upper portion of the U-bend. There I would expect
7 primarily going into a steam environment.
8 I did mention there was a minor radiological
9 release.
10 Here, we see the major pathways.
11 A majority of the release that occurred through
12 this event basically occurred as the activity from the
13 primary plant entered the faulted steam generator and
14 transversed into the -- basically the air injector off-gas
15 system.
16 The licensee -- or I should say Indian Point 2 has
17 a design feature such that there's the radiation detector
18 downstream of air-injector release point, and if it detects
19 a high-radiation condition, it automatically redirects the
20 air jet to release -- instead of going into the atmosphere,
21 back into the primary plant containment, and this action
22 occurred very, very promptly during the event, and
23 approximately 90 percent of all the activity that was
24 released during the event was basically initially bottled in
25 the primary containment. It was subsequently released
. 335
1 through an elevated monitor release point, subsequently.
2 DR. WALLIS: Root cause doesn't include anything
3 about why the crack occurred?
4 MR. HOLIAN: This inspection, the AIT, did not
5 have that as its charter.
6 There was a subsequent inspection and then, of
7 course, the NRR review looking at that was also something
8 I'll address at the end.
9 DR. WALLIS: Is this a fairly common event?
10 MR. HOLIAN: No. I'll address it right now. The
11 issue in the licensee's root cause event was the noise
12 levels in the steam generator.
13 DR. WALLIS: Was there tube vibration?
14 MR. HOLIAN: No, no tube vibration. We'll touch
15 on that at the end. They had indications of hour-glassing.
16 I'll touch on some of that at the end.
17 MR. MARSH: There have been about 6 steam
18 generator tube ruptures in this country since the '76
19 time-frame.
20 One of the things we did in determining how the
21 agency should respond was look at the manner in which we
22 responded to those events, and we were consistent in terms
23 of the agency's response team evaluations in terms of how we
24 looked at it.
25 MR. LORSON: Okay.
. 336
1 Getting back to the radiological releases, I did
2 mention that a majority of the release pathway was gaseous
3 release from the containment that was released approximately
4 a day following the event through an elevated monitored
5 release pathway.
6 There were other gaseous releases that were
7 attributed to the event, primarily due to leakage through
8 secondary plant components.
9 Also, there was a liquid release through the steam
10 generator blow-down lines.
11 Basically the initial -- during the initial
12 portion of the event, some of that water became contaminated
13 with primary system activity.
14 However, as I mentioned, there was no measured
15 radioactivity at any of the off-site monitoring locations,
16 and consequently, we determined that the event had no impact
17 on public health and safety.
18 DR. KRESS: The activity measured -- the air
19 injector that injects the air outside --
20 MR. LORSON: That is correct.
21 DR. KRESS: Is that a gamma monitor?
22 MR. LORSON: There is a radiation detector that's
23 downstream of the air injector, and I believe that is a
24 gamma detection instrument.
25 DR. KRESS: So, it would have picked up, what,
. 337
1 krypton?
2 MR. LORSON: It would have picked up one of the
3 noble gases, and it repositioned during the very initial
4 stages of the event.
5 It repositioned very, very quickly from the
6 atmosphere to the containment.
7 DR. UHRIG: How long was the waste held in
8 containment before it was released?
9 MR. LORSON: It was held in containment, as I
10 recall, approximately one day prior to release.
11 DR. UHRIG: There would be some decay.
12 MR. LORSON: There would have been some decay,
13 correct.
14 We have a pretty good idea, though, of the
15 activity that is released through the containment pathway,
16 because there is, as I mentioned, a radiation monitor on
17 that particular pathway, and so, for sampling and flow rates
18 and monitor indications, you can get a pretty good estimate
19 of the activity released.
20 DR. KRESS: There is no way to determine how much
21 of I-131 was released, I guess.
22 MR. LORSON: Not that I'm aware of. I don't know
23 whether or not there was more sampling done inside the
24 containment prior to that.
25 DR. KRESS: They would have to sample to determine
. 338
1 it.
2 MR. LORSON: Not to my knowledge.
3 DR. UHRIG: Did they have any fuel leakage in the
4 core at the time of the incident?
5 MR. LORSON: Not that I recall.
6 As I mentioned, one of the areas the team looked
7 at was we noted that there was some delays associated with
8 achieving cold shut-down conditions, and we identified
9 several root cause areas which were attributed to the
10 delays.
11 Specifically, there were issues associated with
12 operator performance, procedural adequacy, equipment
13 performance, and also, emergency response of the licensee's
14 organization.
15 The first area we're going to focus on is operator
16 performance.
17 As I mentioned earlier, the team determined that
18 the licensee's initial response was prompt and appropriate
19 and felt that the operators had good procedural adherence
20 overall.
21 There were some deficiencies, however, in
22 establishing the plant into the cold shut-down condition,
23 and that was basically attributed to a couple of factors:
24 One, as I mentioned, that the initial plant
25 cool-down was very rapid.
. 339
1 That led to the need to manually initiate safety
2 injection, which then complicated further operator response
3 during the event.
4 DR. UHRIG: Was that a design deficiency, or was
5 that an operator deficiency?
6 MR. LORSON: I believe it was a combination of the
7 two factors, and we primarily determined it was an operator
8 deficiency in that, when the operator manually initiated the
9 cool-down, he achieved a great cool-down than what he had
10 intended.
11 DR. UHRIG: I know that several sister plants have
12 this problem of too much cold water coming in in a situation
13 like this.
14 MR. LORSON: This is not a question of too much
15 cold water coming into the plant. It's a question of
16 releasing too great a volume of steam through the manual
17 steam bypass valves.
18 DR. UHRIG: All right.
19 MR. LORSON: So, the design or equipment problem
20 might be, is that these valves can be difficult to operate,
21 and there might not be good feedback of the specific valve
22 position to the operator.
23 So, it's really a combination of the two issue,
24 operator performance and also the equipment probably wasn't
25 the easiest equipment to operate.
. 340
1 MR. LEITCH: This plant has N-16 monitors. I
2 wonder if you have any comment about the value of those N-16
3 monitors in early appraisal of the situation.
4 In other words, my question basically is, had they
5 not had the N-16 monitors, would it have seriously inhibited
6 their analysis of the situation early on?
7 MR. LORSON: I believe the N-16 monitors were a
8 value to the plant, because it gave them a good indication
9 prior to the event that there, in fact, was a leak from the
10 number 24 steam generator.
11 However, I think that, even without the N-16
12 monitors, the operators had adequate information prior to
13 the event and at the initiation event to make the proper
14 plant operating decisions.
15 Also, there were some issues associated with
16 operator recognition of plant configuration.
17 A couple examples include the component cooling
18 water valve line-up to the heat exchangers. There was some
19 confusion with the operators with respect to the valve
20 configuration necessary to place the heat exchangers in
21 service.
22 That led, again, to a delay in placing the
23 shut-down cooling system in service, and also, at one point
24 during the event, when the operators were attempting to
25 lower reactor plant pressure to basically stop the continued
. 341
1 primary to secondary leakage, there were some issues
2 associated with the line-up of the auxiliary spray system
3 such that it took them a longer period of time to lower
4 plant pressure than what would have been expected.
5 In the area of procedural quality, the team
6 determined that the procedures, the emergency operating
7 procedures and the abnormal operating procedures to guide
8 the initial response were good. There were, however,
9 several procedural deficiencies that challenged the operator
10 during the plant cool-down phase.
11 The first issue, there was a delay in placing the
12 shut-down cooling system in service because of an apparent
13 procedural discrepancy over the initiation set-point at
14 which to place the shut-down cooling system in service.
15 Also, the system configuration problems that we
16 had talked about on the previous slide, where operators had
17 some difficulty in recognizing the proper configuration of
18 some components necessary to continue on and place the
19 systems into -- shut-down cooling system in service -- we
20 felt that the procedures could have done a better job of
21 stepping the operators through the actions necessary to
22 place some of these systems in service following the safety
23 injection even.
24 And then, finally, the team determined that the
25 licensee's procedure for monitoring reactor coolant system
. 342
1 temperature while the plant was on the shut-down cooling
2 system -- we identified a potential deficiency associated
3 with the method that the procedure uses to calculate reactor
4 coolant system temperature.
5 The team reviewed equipment performance during the
6 event.
7 We determined that the necessary event mitigation
8 systems worked properly, and that included the reactor
9 protection system, the auxiliary feedwater system, and the
10 safety injection system.
11 The team noted, however, that there were several
12 preexisting equipment problems that did challenge the
13 operators in placing the shutdown cooling system in service.
14 I have four of the problems listed here.
15 The first two problems involved the ability to
16 maintain condenser vacuum. There was two times during the
17 event where condenser vacuum was lost because of equipment
18 problems.
19 These were pre-existing equipment problems, and
20 the operators were challenged, then, to basically recover
21 condenser vacuum to continue using the condenser as a heat
22 sink to continue with the primary plant cool-down.
23 There was an existing design deficiency associated
24 with the containment valve seal water system. This system
25 is designed to minimize leakage through containment
. 343
1 isolation valves.
2 There's only a few plants in the country that have
3 this particular system, and there was a pre-existing design
4 deficiency such that, after the safety injection event
5 occurred, the operators were challenged to go out and take
6 some actions necessary to maintain the system in a operable
7 or stand-by condition.
8 Also, the final bullet there talks about the
9 pressurizer power-operated relief valve design problem.
10 The licensee has a modification in place to
11 correct this particular problem.
12 However, during the event, the operators were
13 required to enter the containment prior to establishing the
14 plant in cold shut-down conditions to basically install a
15 back-up nitrogen supply to the pressure-operated relief
16 valves.
17 DR. POWERS: Is this something that would be
18 recognized in their IPE?
19 MR. TRAPP: It was an LTOP issue to get nitrogen.
20 MR. LORSON: The next area that the team looked at
21 was emergency response.
22 The team determined that the licensee's emergency
23 response protected the health and safety of the public. In
24 particular, the event was classified promptly, and we
25 determined that the licensee performed a good critique of
. 344
1 the event.
2 There were, however, several problems associated
3 with the implementation of the emergency plan, including the
4 augmented emergency response facility staffing was not
5 completed within the time requirements that were specified.
6 There were problems associated with the
7 accountability of personnel on-site to respond to the event.
8 The emergency response data system was not
9 operable for several hours, and that was a preexisting
10 equipment problem.
11 There were some problems noted involving
12 implementation of the media response plan such that there
13 was -- in one case, there was a local town official who was
14 not notified of the event per a existing licensee
15 commitment.
16 There were problems associated with some of the
17 infrastructure in the emergency response facilities, and
18 then the team also identified issues associated with the
19 timeliness and quality of technical support that was
20 provided to the operators.
21 Following the event, the NRC, Region I, conducted
22 a supplemental emergency preparedness inspection.
23 There were several white findings issued
24 associated with this event, and Brian will talk about that
25 in his remarks.
. 345
1 DR. POWERS: You have indicated a variety of
2 problems here.
3 Have they been all through the significance
4 determination process?
5 MR. LORSON: Following our inspection, there was a
6 follow-up inspection in both the emergency preparedness area
7 and also for the issues that were identified by the
8 augmented inspection team.
9 There was a follow-up inspection, and each of the
10 issues was characterized through the significance
11 determination process.
12 As I mentioned, in the emergency preparedness
13 area, there were some white findings associated with this
14 inspection.
15 In the non-emergency preparedness follow-up
16 inspection, my understanding is that there were primary
17 green findings.
18 MR. HOLIAN: Six or seven green findings. And
19 while we're on the topic, I can mention the three white EP
20 issues, if you'd like, now.
21 They were the three issues that had been tracking
22 on a history of EP problems.
23 They were untimely augmentation, accountability
24 issues, and finally, the dissemination of information to the
25 joint news center.
. 346
1 DR. POWERS: I guess I'm more interested in the
2 fact that we can go to an excessive cool-down and that still
3 stays in the green category.
4 How do I understand how we came up with that -- I
5 mean the documentation I have is it's a green.
6 I don't see how the logic went to arrive at that
7 conclusion.
8 MR. TRAPP: I guess one way to look at it,
9 characterize it, is the bar is core damage.
10 Opening up the steam dumps to the condenser and
11 cooling down too quickly, while it was an operational
12 perturbation, really didn't increase their -- or decrease
13 their margin to core damage that significantly that it would
14 reach our threshold of a white.
15 As you know, the human performance STP aspects are
16 still somewhat under development, so you know, maybe there
17 will be more in the future on that.
18 MR. LORSON: I think, also, to try to put that in
19 perspective, when the team determined that the initial
20 cool-down was excessive, there are technical specifications,
21 limits associated with the amount of temperature change you
22 can have on a cool-down event per unit in time, and the
23 licensee -- the number in the tech specs for Indian Point 2
24 is 100 degrees in a one-hour period.
25 I believe the final number was that they achieved
. 347
1 approximately a 104-degree cool-down in, I think, the
2 worst-case rolling one-hour period.
3 So, they slightly exceeded the limit that was in
4 the technical specification.
5 However, that does mean to say that there weren't
6 transient cool-down rates during that cool-down event which
7 exceeded the 100-degree-per-hour average rate, and that
8 particular cool-down analysis was looked at by the licensee
9 in a separate engineering analysis, and they concluded there
10 was not an adverse impact on the structural integrity of the
11 reactor vessel as a result of that cool-down.
12 MR. TRAPP: It was an hour after the initiating
13 event, too.
14 They had already stabilized, equalized, and pretty
15 much -- I wasn't in the emergency center at that point, but
16 I think the event was pretty much contained by that point,
17 and then they got on the steam dumps a little hard.
18 DR. POWERS: Well, there's a reason for those
19 technical specifications on the cool-down rate.
20 MR. MARSH: If I can add a licensing perspective,
21 it's not to say that cooling down excessively is an
22 acceptable way of achieving cold shut-down, but from the
23 standpoint of the licensing analysis, the steam generator
24 tube rupture licensing analysis, that event assumes that
25 there is a double-ended guillotine break of the tubes, you
. 348
1 have a full off-set, and you end up with a lot of flow out
2 the primary end of the secondary, and you actually get an
3 automatic scram, you get full safety injection, and you end
4 up with a pretty rapid cool-down in the primary. That's the
5 stylized -- and loss of off-site power, and you achieve cold
6 shut-down in a pretty short time-frame.
7 That's not to say that events like this, where you
8 end up with a longer time to achieve cold shut-down, which
9 get aggravated by excessive cool-downs and by safety
10 injection, are the way that this unit should have been
11 operating post that event, but just from a licensing
12 perspective, it was within that analysis.
13 DR. POWERS: What I'm dancing around and avoiding
14 saying bluntly -- and I'll say it bluntly -- is I'm worried
15 about PTS, and I'm not sure our significance determination
16 process takes us down the PTS pathway here, and so, we end
17 up with a green finding on something that shouldn't be
18 green. That's what I'm worried about it.
19 MR. MARSH: I think the STP process, as we pointed
20 out, has as its benchmark core damage --
21 DR. POWERS: Yeah.
22 MR. MARSH: -- as opposed to PTS limits.
23 MR. LORSON: As I mentioned earlier, Jim Trapp is
24 going to discuss the risk significance of the event.
25 MR. TRAPP: This is an interesting event to do a
. 349
1 risk analysis on, because we looked at the CCDP of the event
2 and then we're also applying the revised oversight program
3 to these issues, the steam leak issues, and the 1997 steam
4 generator tube inspection.
5 So, I'm going to get kind of Dr. Jekyll and Mr.
6 Hyde on you.
7 I'm going to tell you that the risk of the event
8 was pretty low and we applied the ROP, initially we'd come
9 up with fairly significant risk, and I also want to caveat
10 in this the inspection report on the steam generator tube
11 inspection hasn't been issued yet, and it's still a process
12 being developed.
13 We still need a regulatory conference, and there
14 will be input by the licensee, and we haven't had all that
15 put together yet.
16 So, things may change.
17 DR. POWERS: Well, when you do your calculation of
18 the CCDP, do you have a chain of events that result in
19 vessel fracture that are built into your modeling, so you
20 can look at -- suppose that the cool-down rate was a little
21 higher in transitory periods and judge whether -- what the
22 probability of getting into an event R is.
23 MR. TRAPP: The way our model would work is we
24 would have the steam generator tube rupture as the
25 initiating event. The human actions that you need is
. 350
1 equalization and, you know, isolation of the feedwater.
2 DR. POWERS: You're going to come to a success
3 path in this model when, in fact, it's a compound event.
4 It's something that PRAs just don't handle very well right
5 now, and if -- suppose that we had a transitory period of
6 1,000 degrees an hour in there and got into a PTS event --
7 it just doesn't show up in the way the PRA is designed.
8 So, you don't get a good CCDP on this event. It's
9 too complicated.
10 MR. TRAPP: I guess on the positive review point,
11 I'll just throw this out.
12 They don't have high-pressure injection here. The
13 safety injection pumps are about 1,200 pounds, and they
14 already had the break in the RCS, and the pressurization
15 part of it there would be challenging.
16 MR. LORSON: The other thing I think that might
17 affect the PTS concern is the initial cool-down was rapid,
18 but from normal operating temperature to about 100 degrees
19 below normal operating temperature.
20 So, at the time when the safety injection
21 occurred, the reactor vessel --
22 DR. POWERS: Yeah, you're not in the vulnerable
23 period, but when you're calculating the CCDP, you're going
24 into a what-if scenario, you know, what if things were just
25 a little different, and I don't think you've got the tool
. 351
1 available to you to explore this what-if the way we'd really
2 like to explore it.
3 I presume that a good calculation of CCDP takes
4 things both ways.
5 DR. APOSTOLAKIS: No, it's biased toward bad
6 things.
7 DR. UHRIG: Was there any impingement on other
8 tubes of the water coming out?
9 MR. HOLIAN: I was going to address that just
10 briefly.
11 They did do a visual examination of the adjacent
12 tubes, and I believe there was no indication.
13 MR. TRAPP: The initial CCDP estimate we did the
14 night of the event based on the models.
15 Using the licensee's IPE data, just dividing the
16 contribution from tube ruptures divided by the initiating
17 event frequency of tube ruptures, you came up with 7.7 to
18 the minus 5, and this was largely the basis for us
19 dispatching the AIT.
20 We knew these numbers were conservative, because
21 we knew the leak rate was nowhere near the double-ended
22 guillotine rupture, you know, four to six hundred gallons
23 per minute that would be assumed in the PRA, and since the
24 tube rupture event is very human action-dominated, you know,
25 the smaller leak rate gives you a lot longer times to do
. 352
1 equalization and other evolutions for the operators, and we
2 knew the risk would be much lower.
3 The licensee revised their CCDP number based on
4 the longer times available for doing the human actions, and
5 now the charging pumps were available to back up the SI
6 pumps, if needed.
7 So, making these two assumptions, they came up
8 with a CCDP of 2.2 to the minus 6.
9 So, this would be a low-to-moderate
10 risk-significance event.
11 The other aspect of this was to look at the 1997
12 tube inspection and do sort of a what-if. Our premise here
13 would be, if the inspection was deficient, if the tube could
14 have ruptured, what would be the risk if you had a
15 double-ended guillotine rupture of a tube?
16 The way this risk analysis worked -- and this is
17 what we call in the new STP the delta CDF, not the CCDP, and
18 this new process, the way it works is the delta CDF that you
19 calculate is assumed to be equivalent to the delta LERF.
20 So we calculated a delta CDF and a delta LERF, and
21 some of the assumptions we made were that the steam
22 generator tube failure frequency was -- due to the
23 degradation of '97 -- was once per reactor year, and this
24 was down from like a 1.3E to the minus 2 per reactor year,
25 was the baseline.
. 353
1 We assume that half the tube failures would result
2 in a rupture, rather than a leak that was seen at Indian
3 Point.
4 I mentioned before we assumed that the delta CDF
5 is equivalent to the delta LERF, and the analysis that was
6 done largely by NRR also included induced steam generator
7 tube ruptures and steam line breaks and primary
8 over-pressure events such as ATWS events, and the results of
9 this analysis, like I said, are still pending, they should
10 be out shortly, but the risk significance of doing a delta
11 CDF calculation came up with high risk.
12 DR. BONACA: I understand the risk significance,
13 and I want to point out again that I believe condenser
14 vacuum was lost twice.
15 So, there was a significant reliance on
16 atmospheric dump, and so, it was fortunate that we had a
17 clean core, so there was no activity in the primary side.
18 I mean if you had some contamination or activity
19 actually in the primary side, you will have releases out
20 there, and you know, the public probably is as concerned
21 about these releases as we are about risk significance when
22 we do these evaluations.
23 Again, I think that, if you go back to look at the
24 goals that you have for the emergency operating procedure,
25 you'll find there is a time that you have to shoot for for
. 354
1 balancing the pressure and stopping releases, and here there
2 was a significant issue with the potential for releases
3 lasting over hours, right? It lasted several hours.
4 MR. LORSON: That's correct.
5 The plant remained at pressure and temperature
6 longer than what you would have liked to because of the
7 delays that we discussed in terms of achieving the cold
8 shut-down condition.
9 MR. TRAPP: But of course, off-site doses wouldn't
10 generally -- that analysis wouldn't rely on the condenser.
11 MR. HOLIAN: The final summary slide, quickly.
12 There are supplemental inspections and activities
13 going on for both of these events.
14 Quickly, on the August '99 event, we did have a
15 inspection team that followed up to the AIT and found that
16 the short-term corrective actions were adequate for
17 re-start. There was also enforcement action taken, which
18 was a separate inspection that followed those issues.
19 It did indicate some mixed performance in the
20 recovery efforts.
21 There are some longer-term corrective actions that
22 are still being picked up, and I'll pick that up under the
23 agency focus aspect.
24 Emergency preparedness issues have been evident
25 for a couple years at Indian Point 2. They were evident
. 355
1 during the steam generator tube failure event.
2 I summarized the three white findings, which is a
3 degraded cornerstone under the EP issue, and they did do, in
4 the June time-frame, a remedial drill to demonstrate
5 adequate performance.
6 Steam generator tube failure root cause -- we
7 talked earlier about that.
8 I'd just remind the ACRS that at the point after
9 this tube failure, NRR took on the aspect of working with
10 the safety evaluation for, at that time, would plant
11 re-start be allowed or not.
12 The region took on the aspect of following up with
13 the AIT and looking at the root cause of the event.
14 The new program, in particular, asks you to look
15 at was there performance issues that related to this root
16 cause? The licensee's root cause evaluation submitted was
17 the noise in these older steam generators, in effect, masked
18 the signals in the '97 time-frame.
19 As Jim Trapp has mentioned, the inspection report
20 is due to go out shortly from Region I. It's in the final
21 stages of concurrence.
22 There was a quick-look report put out in July of
23 this year, just a month or so ago, and it did highlight
24 three performance issues that the special inspection found,
25 and that was -- noise was one of them, high levels of noise
. 356
1 in these steam generators, the old steam generators.
2 There were issues with this -- in the '97
3 time-frame now -- this being the first PWSCC, a PWSCC
4 indication was found in the apex, in the U-bend area, and
5 which was a new degradation mechanism for this utility, and
6 secondly, and related to that PWSCC indication, was they did
7 have evidence of denting at the top tube support plate,
8 which could lead to hour-glassing and further lead to PWSCC.
9 So, those three issues are summed up in that
10 quick-look inspection, and there will be more detail -- much
11 more detail coming out in the inspection report on previous
12 root cause, and that was one reason why that additional risk
13 estimate was done on the condition of the generators being
14 put back in service from the '97 time-frame to when the tube
15 failure occurred.
16 DR. UHRIG: Was this plant old enough to have had
17 phosphate treatment?
18 MR. HOLIAN: Yes. They did switch over. I don't
19 know the time-frame, exactly, early on.
20 I did want to add one other item on root cause.
21 You know, the Surry plant had a tube rupture early on, and
22 this is akin to that.
23 This plant did have some tech specs put in prior
24 to the Surry time-frame to look at hour-glassing and this
25 phenomenon.
. 357
1 So, it was something that they were tracking in
2 some ways.
3 I do also want to mention that, in the quick-look
4 report, which is just a two-page summary, the licensee has
5 disagreed with the inspection findings, the performance
6 issues in particular, and as we go down here -- it's not
7 mentioned here, but on the steam generator tube failure root
8 cause, there's tentatively a meeting with the licensee under
9 the new program, a regulatory conference in the region
10 that's tentatively been set up for September 26th.
11 That's usually about a month after we get an
12 inspection report out, where the licensee will come in and,
13 once again, debate or discuss those performance issues and
14 the risk assessment that was done.
15 Finally, there was an information notice that was
16 put out on the steam generator tube failure event, in
17 general.
18 Agency focus plant -- I'd just like to mention
19 that, as of the last senior management meeting, Indian Point
20 was one of two plants, an agency focus plant, D.C. Cook
21 being the other aspect in a special category.
22 The items that you see there are key categories
23 that have been issues that the region has been tracking for
24 several years at Indian Point 2 and were categorized in that
25 agency focus letter that went to the utility in May.
. 358
1 The agency focus meeting has been scheduled for
2 September 11th on-site up at Indian Point 2 to check the
3 status of Indian Point's business improvement plan in these
4 areas.
5 MR. LEITCH: Do I understand correctly that the
6 licensee has now decided to replace the steam generator?
7 MR. HOLIAN: Oh, I'm sorry. That's a good summary
8 statement.
9 Yes, they are in the midst of replacing them right
10 now.
11 The inspection has started.
12 That decision was made three weeks or a month ago.
13 They have a team on-site that's done a couple down in
14 Florida and is scheduled to do Calvert Cliffs.
15 So, they're on a schedule, they're well into it,
16 and hope to have them replaced by the end of
17 November-December time-frame.
18 MR. MARSH: I'd like to also add that, in terms of
19 a generic communication, the agency is still considering the
20 need for anything coming out of the Indian Point 2 event,
21 whether any of the particulars of the performance issues
22 need any further actions.
23 DR. SHACK: For the other PWRs, have they plugged
24 their row 2 U-bands?
25 MR. HOLIAN: Most have done row 1 and 2, I
. 359
1 believe, and this utility had done row 1 and had not done
2 row 2.
3 DR. SHACK: But most have done both rows.
4 MR. HOLIAN: Yes.
5 DR. UHRIG: Did they have significant cracking in
6 the support plates due to the phosphate treatment
7 originally?
8 MR. HOLIAN: Originally there was a wide variety
9 of indications, and some of it was support plate cracking.
10 They actually had -- the hour-glassing was
11 crunching of the tubes, and actually, the flow slot deforms
12 in an hourglass.
13 They did have a tech spec for monitoring that.
14 That was one of the issues that will be brought out in the
15 inspection report.
16 DR. SEALE: I understand we have a couple of
17 people from the utility here.
18 Would you like to add anything to the comments
19 that have been made?
20 MR. GROTH: Good morning. Thank you.
21 I'm John Groth. I'm the chief nuclear officer at
22 Indian Point 2. I appreciate the opportunity to come and
23 talk with you today.
24 The August event we have used to our advantage to
25 improve our ability to do self-assessments and to learn some
. 360
1 lessons that will stand us in good stead, and we are making
2 progress in correcting those problems.
3 For the February 15th event, I'd like to point out
4 a couple of things.
5 In the process of cooling down the plant, you use
6 the non-faulted steam generators, not the faulted steam
7 generators, so you're not releasing anything to the
8 environment.
9 So, please remember we've got a four-loop plant
10 and we have three other ways to reduce the temperature in
11 that plant, not relying upon the one that is faulted.
12 I'd like to also share with you some of the
13 lessons that we've learned in this case.
14 As has been indicated by the inspectors, the
15 difficulty came not in the initial response but, rather, in
16 the stabilization period.
17 As we look into our training programs, not only
18 our own but across the industry, we've found that all of us
19 have concentrated on the initial actions: How do you
20 respond to the casualty? How do you get the plant into
21 condition where you're stable?
22 We really didn't practice that next phase, which
23 is, in fact, the more difficult, because we've practiced the
24 others so well.
25 We're now practicing the other. We've shared that
. 361
1 with the industry, and around the industry, we're doing
2 that, because we all found that, as we tried to do that
3 follow-on activity, it is very complex, and it becomes very
4 challenging, and it takes a long time, and that's something
5 that is very important, particularly in how we, as an
6 industry, train the emergency plan area.
7 I'm sure all of you have, at one time or another,
8 watched an emergency plan drill.
9 We impose a series of casualties that are
10 unrealistic to drive the plant very quickly to a point where
11 you have a release.
12 The scenario is not a realistic one.
13 It is time-related, and it's designed to have it
14 done in a very short order, and what we've done over these
15 years is pre-condition our audience to believe that this is
16 what's going to happen.
17 This becomes a very difficult problem for all of
18 us, because the understanding that one of these events
19 requires time to develop, time to analyze, and time to
20 respond to is not there, and consequently, you have an
21 expectation within our audiences for things to move more
22 quickly, to be more dramatic, to be more, I'm going to say,
23 difficult and challenging, and it's very, very hard for them
24 to watch grass grow.
25 We need to think again in our industry on how
. 362
1 we're doing these things and how, in fact, we teach our
2 audience just how badly we deform the plant in the first
3 case to get to a condition where we can practice the release
4 piece, and then the next piece is the understanding and the
5 training that's involved in how one of these events really
6 transpires.
7 We thank you for the opportunity to come and make
8 comment today, and we would be pleased to take any questions
9 you might have that might help, but thank you for the
10 opportunity.
11 DR. SEALE: Are there any questions?
12 [No response.]
13 DR. SEALE: You reminded us of something just now
14 that I think is very important, and that is that, as we have
15 these experiences, we always want to be sure that we listen,
16 because there are some messages that are being passed along
17 to us, and you certainly have pointed out this protracted
18 response problem, and I can certainly appreciate what you're
19 saying there, that that could be a distortion that could
20 bother you in the long run.
21 So, that's a very good point.
22 Does any of the members of the committee have
23 anything to ask any of the NRC people or the people from the
24 utility?
25 [No response.]
. 363
1 DR. SEALE: Well, here again, I think we find
2 that, anytime we have the opportunity to hear from the
3 operators and from the NRC's inspection people, we learn
4 something, and I know that we have today.
5 George, your comments, I thought, in connection
6 with the new plant evaluation process, were very much to the
7 point. We may want to talk about this a little bit more in
8 some frame or mode with the Commission.
9 All of you, I'd like to thank you for your
10 preparation -- you obviously were well-prepared, and we
11 appreciate that.
12 I know you're going to have fun talking to the
13 people up there in that public meeting that's coming up, and
14 so, maybe we helped polish up some of that, too.
15 So, with that, I'll pass it back to the Chairman.
16 DR. POWERS: Thank you, Bob.
17 I'd echo my sentiment that this is quite a nice
18 presentation that you put on here, information-packed, and I
19 think it's particularly significant for this committee,
20 because we're getting to see some of the theoretical aspects
21 of the new reactor oversight process in the harsh light of
22 reality, and so, we're going to -- these are view-graphs
23 that aren't going to get thrown away.
24 They're going to be used over and over again, not
25 so much on this incident but in the overall evaluation of
. 364
1 how we're doing on this new reactor oversight process, and
2 we really thank you for taking the time to put this material
3 together for us.
4 Mr. Wallis, may I ask, will you need transcription
5 for your presentation after lunch?
6 DR. WALLIS: I think that's up to you, sir.
7 DR. POWERS: Okay.
8 In that case, I think we can bring the
9 transcription to a close for the day, and I will recess us
10 for lunch until 1:30.
11 [Whereupon, at 12:27 p.m., the meeting was
12 concluded.]
13
14
15
16
17
18
19
20
21
22
23
24
25
