466th Meeting - September 30, 1999
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
***
MEETING: 466TH ADVISORY COMMITTEE ON REACTOR
SAFEGUARDS (ACRS)
U.S. Nuclear Regulatory Commission
11545 Rockville Pike, Room T-2B3
White Flint Building 2
Rockville, Maryland
Thursday, September 30, 1999
The Committee met, pursuant to notice, at
8:30 a.m.
MEMBERS PRESENT:
DANA A. POWERS, Chairman, ACRS
GEORGE APOSTOLAKIS, Vice-Chairman, ACRS
THOMAS S. KRESS, ACRS Member
MARIO V. BONACA, ACRS Member
JOHN J. BARTON, ACRS Member
ROBERT E. UHRIG, ACRS Member
WILLIAM J. SHACK, ACRS Member
JOHN D. SIEBER, ACRS Member
ROBERT L. SEALE, ACRS Member
GRAHAM B. WALLIS, ACRS Member
P R O C E E D I N G S
[8:30 a.m.]
DR. POWERS: The meeting will now come to order. This is
the first day of the 466th meeting of the Advisory Committee on Reactor
Safeguards.
During today's meeting the committee will consider NRC
Safety Research Program, Combustion Engineering Owners Group proposal to
eliminate the post-accident sampling system requirements, proposed
rulemaking and review plans associated with the development of risk
informed revisions to 10 CFR Part 50, proposed resolution of Generic
Safety Issue 148 -- "Smoke Control and Manual Fire-Fighting
Effectiveness," a joint ACRS/ACNW report on the proposed framework for
risk informed regulation in NMSS, discussion of proposed ACRS reports.
This meeting is being conducted in accordance with the
provisions of the Federal Advisory Committee Act. Dr. John T. Larkins
is the Designated Federal Official for the initial portion of the
meeting.
We have received no written comments or requests for time to
make oral statements from members of the public regarding today's
session. A transcript of portions of the meeting is being kept. It is
requested that the speakers use one of the microphones, identify
themselves and speak with sufficient clarity and volume so that they can
be readily heard.
I want to begin this morning with a few items of interest.
All the members should have before them a memorandum from Mr. Singh
concerning our visit to Davis- Besse and Region III. He has a tentative
schedule for those dates. He really does need to have a confirmation on
your availability to attend those visits.
Also, members should have before them a package entitled,
"Items of Interest." I will call your attention to the Water Reactor
Safety Information Meeting that will be held later this month or
actually in October, and I will also call to your attention a news
release on Jack Sieber being officially one of the fold here.
Today we are going to do a somewhat different plan of attack
on the morning session. We are going to hear some words from Ashok
Thadani on the vision statement from the Office of Nuclear Regulatory
Research. We may mark time a little bit until Ashok gets down here,
because we are a little ahead on the agenda. I think we are going to go
off the record after Ashok's presentation and discuss more about the
Research report and our obligations and what Professor Wallis has
invented for us in that regard.
Do any other members have opening comments that they want to
make?
[No response.]
DR. POWERS: Seeing none, I will turn the floor over to
Professor Wallis and you can begin your introduction.
DR. WALLIS: Well, I am very eager to hear what Ashok
Thadani has to say, and so I would like to invite him to come up here
and tell us.
MR. THADANI: Thank you. Good morning. As I had promised
you a couple of weeks ago, I would like to come back and share with you
further developments in terms of the vision for the Office of Nuclear
Regulatory Research. Let me go over some of the background that you
probably know but I think it is appropriate to repeat.
We briefed the Commission in May and it was on planning,
budgeting and performance management issues, and as a result of that
briefing, the Commission issued a Staff Requirements Memorandum on June
9th of 1999, and as you'll note here exactly what the SRM asked the
Office of Research to develop in terms of the vision and role. They
focused on three things -- how do we complement the front-line
regulatory activities, how we independently examine evolving technology
and anticipated issues and to the extent that we have in place a center
of excellence, and of course it was noted in the SRM to make sure that
these activities and efforts flow from the strategic plan.
We owe the Commission a response. The paper is due middle
of October and our intention is to send this paper and seek Commission
endorsement what we believe the appropriate vision ought to be.
It is very important to us to make sure that we have a sense
of what you think of this vision statement, and so we would very much
like your views and suggestions on this.
I think you probably know best, other than the Office of
Research, what the Office of Research is up to and I very much value --
DR. WALLIS: But that can't be true -- excuse me.
[Laughter.]
MR. THADANI: I very much value your views and thoughts on
this matter. I think it is important not just for today but for the
future for the Office.
Going on to -- I think you have seen this before, but I just
wanted to make sure and put it up again, what the NRC's vision statement
is and this appears in the strategic plan as stated here, and we wanted
to make sure of course that there was a link in terms of what the Office
of Research does to the vision, NRC vision, as stated in the strategic
plan.
We had an interesting discussion last time on this vision
statement. You saw the earlier drafts. You provided some thoughts and
in fact there was a revised statement prepared here and I solicited
input from others and in particular from all the Staff of the Office of
Nuclear Regulatory Research. I sent them that draft with some changes
we made after our meeting with you, and I did get feedback from a fairly
good number of Research staff with suggestions and changes and we made a
real effort to try and accommodate the comments that we received. Some
of them are very good. Some were probably not what we could accept, but
we made a real attempt to try to revise the statement to be reflective
of what the Staff views were.
DR. KRESS: Do you want comments on this one now?
MR. THADANI: I'm sorry?
DR. KRESS: Do you want comments on this one, the new one,
now?
MR. THADANI: Yes. Yes, I would appreciate your feedback,
thoughts and -- we tried to capture the key elements that the Commission
had indicated.
DR. KRESS: It probably is -- developing technical basis
probably means the same thing but I think you ought to have the words
"conduct research" in there somewhere. It doesn't say you conduct
research anywhere.
MR. THADANI: It doesn't --
DR. KRESS: No.
MR. THADANI: -- say conduct research --
DR. KRESS: Developing technical basis involves that.
MR. THADANI: Yes.
DR. KRESS: But it involves a lot more.
DR. APOSTOLAKIS: Replace "analysis" by "research."
MR. THADANI: I could use --
DR. APOSTOLAKIS: Conducts independent research and develops
technical basis.
DR. WALLIS: I guess that the line organizations also
develop technical basis, don't they?
MR. THADANI: Yes.
DR. WALLIS: And they also do analyses.
MR. THADANI: Yes.
DR. KRESS: So that doesn't distinguish you too much from
the line organizations.
DR. WALLIS: Independent is the word.
DR. KRESS: Yes.
MR. THADANI: Independent is the key here. There are some
key words in my mind -- independence in terms of doing the research and
developing technical basis for realistic -- realistic is an important
word.
DR. KRESS: I think that is a key word, too.
MR. THADANI: There is a tendency to say that I can make a
conservative decision --
DR. KRESS: Well, you can.
MR. THADANI: -- and I don't need to develop a better
technical information base. I personally think that is the wrong
mindset and we have got to change that and when we talk about it, you
know, you'll see a little bit more, we want to remove excessive
conservatism or reduce unnecessary burden. Well, how do you go about
doing that -- the technical base.
I don't know how many of you were there when John Ahearne
briefed the Commission on the Center for Strategic and International
Studies report and made a very strong statement about -- and, by the
way, it is in Inside NRC, if you haven't see it -- about what has
happened to research and that as an agency when you are going through
significant change that is when research becomes even more critical, and
there was some interesting dialogue, so if you weren't there I certainly
recommend reading the transcripts.
There were two very important issues that were discussed.
One was research. The other was the whole issue of safety philosophy
and the role of adequate protection and so on.
They are both very significant issue to the agency, I think.
Anyway, go ahead.
DR. APOSTOLAKIS: I think the vision should be as short as
possible and it seems to me the words that "protect the public health
and safety" are redundant. All regulatory decisions do this so I would
say, you know, independent research and develops technical base for
realistic regulatory decisions and prepares the agency for the future.
DR. KRESS: I might be tempted to break that into two
sentences.
DR. APOSTOLAKIS: Well, if you delete the "protect the
public health and safety" then one sentence is all right. It is a
little long but it is all right. Now it is too long.
DR. KRESS: Yes.
MR. THADANI: It is long. It is interesting. We put this
in as a result of some comments we got. The concern was reflected that
the vision statement ought to at least say why do we do something, why
is it that we are doing it, and it is really in the end somehow, as you
see --
DR. APOSTOLAKIS: Well, but this is really the mission of
the agency itself.
MR. THADANI: Exactly.
DR. APOSTOLAKIS: So to repeat it really doesn't add
anything to your own vision.
MR. THADANI: It doesn't -- it doesn't in a --
DR. APOSTOLAKIS: It really should be a short, crisp
statement and I think if you delete those words --
DR. SEALE: I disagree.
DR. APOSTOLAKIS: Oh?
DR. SEALE: I think that --
DR. APOSTOLAKIS: I would be surprised if all of you agreed
immediately.
DR. SEALE: I understand but I think that with the extent to
which research is and has been under fire in recent years, it strikes me
that it is important that we not lose sight of that as a goal not only
of the Commission but specifically of the research.
You could delete "protect the health and safety" and put in
"keep the industry happy" but I don't think we want to do that --
necessarily.
DR. APOSTOLAKIS: No, but the point is that protecting
public health and safety is the mission of the agency --
DR. SEALE: That's right.
DR. APOSTOLAKIS: -- and as such is a part of that agency,
so it is understood that that is why you are doing these things.
I mean your mission is a subset of the mission of the
agency.
MR. THADANI: It is.
DR. APOSTOLAKIS: Yes -- so -- because I think there is
something to be said, there is a lot to be said about short and crisp
statements, that people glance at them and say, yes, this is what they
are doing.
DR. KRESS: Well, I would -- in contrast to that, I would
add another word instead and the word I would add is "for the future
by" -- I would put in the word "anticipating and evaluating" -- that is
word I think you need in there.
DR. APOSTOLAKIS: Anticipating and evaluating -- yes, that
is good.
MR. THADANI: Yes, that is a good word.
DR. APOSTOLAKIS: So we are now making it longer.
DR. KRESS: Yes, that's what I was saying.
DR. APOSTOLAKIS: Soon we are going to need an executive
summary.
[Laughter.]
DR. APOSTOLAKIS: Come on, guys, that is not the mission
anymore. If you have an executive summary, promise to drop that
"protect the public health and safety."
DR. WALLIS: I guess that there should be another thing
somewhere, and you have a vision statement or maybe NRR has a vision
statement and NMSS has a vision statement. They are all part of the same
agency, so I would like a vision of how they fit together from
somewhere.
MR. THADANI: I think you can certainly ask the Commission
for that.
DR. WALLIS: That seems to me what is often missing, and if
you read some of the assignments and description of the various
divisions that's fine, but how do these guys interact? That is never
made very clear.
MR. THADANI: I think this is frankly -- I am very pleased
that we were asked to do this by the Commission and I think we know the
way we have been working in the past. It is implicit, but it does need
to be I think explicit.
After I get done talking about all these really good ideas
and so on, I do have some views on so what does it really mean. We talk
about this vision and this and that, but in the end what does it really
mean in where we are, so again I would certainly be interested.
We will try and take a look at the statement again given
what you just suggested and see if we can't revise it.
This is sort of a pictorial way of really -- I mean this
only -- I am only talking about the Office of Research. As the
Commission has said, you have to make sure that this all flows from the
strategic plan, and indeed from vision statement flow the goals and the
goals, I will talk about that a little bit, and then from those goals
flows the concept of what would be our accomplishments, how they relate
to the goals, and if we were to accomplish what we say what kind of
issues and activities we are going to be involved in.
When you get into areas of issues and activities, you are
certainly getting into details. You are not on the level of the
strategic plan. Certainly some of the information will get in the
performance plan and then clearly all of it gets in the operating plan
for the Office.
So there is the relationship --
DR. APOSTOLAKIS: What is the difference between a goal and
a planned accomplishment?
MR. THADANI: For example, the goal is maybe maintain
safety. It is a performance goal and there are strategic goals in the
strategic plan and there are performance goals, and the planned
accomplishment could be improvement in certain safety factors, so to
speak. Issues under that may be generic safety issue and the activity
under that may be, okay, we have got to run experiments for sump
blockage and so that is sort of the logic.
If you recall, we had eight, if I remember correctly I think
there are eight planned accomplishments, and under that we have issues.
If you remember, we went through the discussion doing a budget and we
talked about how we prioritize the activities. It was clear to us that
we couldn't prioritize at the issue level. We had to really get down to
the substance of real changes to be able to prioritize.
DR. APOSTOLAKIS: Would it be more real to put the issues
under the goals and then the planned accomplishments? In other words,
in attempting to meet your goals, you are identifying a number of issues
and then you identify specifics that you need to do to resolve these
issues, and then at the bottom you have the activities that will help
you do it.
DR. WALLIS: That was my comment. I don't think they should
really be called issues. I think they should be called needs. In order
to meet the goals, someone has to figure out what needs to be done.
That I think is one of the weakest parts of the whole process.
MR. THADANI: Issue may be integrity of steam generator
tubes, as an example, of issue, but now under that I mean there are
various degradation mechanisms, for example. Well, what is it that we
are talking about? What will we do? So that was --
DR. APOSTOLAKIS: So you just argued for placing the
issue --
MR. THADANI: With the planned accomplishments.
DR. APOSTOLAKIS: Underneath. You have this big problem
that you are saying, well, it has to go A, B, C, D to do it.
DR. WALLIS: That's right. I agree. Someone has to look at
maintain safety --
MR. THADANI: Right.
DR. WALLIS: -- and say, well, okay, we're maintaining
safety. What are the challenges to safety, what are the threats to
safety, or what are the future threats to safety? What do we need to do
in order to meet those. That has to be articulated in enough depth so
that when you do the work you know that you are doing the right work.
MR. THADANI: Yes, and this is really, I believe that when
you get to enough depth I believe that you have to get to the activity
level, because I don't think -- I mean to say, for example, reactor
pressure vessel integrity -- that's wonderful, it's important, we all
agree it is important, so what is it that we are going to do? In the
end I think you really -- well, we will take a look at this -- in the
end I think you really have to get down to the specifics.
DR. APOSTOLAKIS: As a matter of fact, the issues and the
planned accomplishments is really what distinguishes you from other
Offices. You are taking more time. You are going into more detail.
You are creating something that perhaps did not exist before, either an
approach or new understanding, so this is really what makes you
different because the goals in Division -- I mean everybody wants to
have realistic regulatory decisions, right?
MR. THADANI: Well, I think so.
DR. APOSTOLAKIS: Depending on what reality is -- but I
think that is where you differ from other people.
MR. THADANI: Yes, yes, yes, yes. In fact, yes, you are
right. At the issue level, if you really were to look at that
information, it does separate us in a way, yes.
DR. WALLIS: I don't think it is just realistic. I think it
is regulatory decisions which are robust and are not likely to come up
against some failure mechanism which was not anticipated. It is not
just being realistic. There is something else that regulatory decisions
must be besides just --
DR. APOSTOLAKIS: But if you are realistic, aren't you
covering yourself?
DR. WALLIS: Well, I am not sure realistic really covers --
realistic to me means based on real facts and not fantasy, and that is
part of it, but also you have got to sort of anticipate where the
threats might come from and so on.
DR. APOSTOLAKIS: I would make that part of realistic.
MR. THADANI: I think you will see this, that that is
clearly when you go through and break it down, the parts of the
statement, break it down that if it doesn't cover, Graham, what you
said, then we will come back and see what we need to do.
DR. APOSTOLAKIS: Would it be better instead of realistic to
say "rational"?
DR. WALLIS: No, everything is rational.
DR. APOSTOLAKIS: We don't use that word here.
MR. THADANI: If I may --
DR. APOSTOLAKIS: Because then it implies that everything is
PRA.
DR. WALLIS: I would rather not have the word "realistic"
because it implies that there are unrealistic decisions.
MR. THADANI: Realistic -- and I am sorry that this chart
has an error in it and it troubles me that this has an error in it,
because my favorite piece is missing on this, and I am looking at my
Staff now.
This is the three key elements, I think, in the vision --
independence, realism, and being forward-looking. Those are important
elements.
Realism to me doesn't mean that we are not going to make
sound decisions but that developing information base to understand what
the realistic response would be and then making a very conscious
decision if one wants to add whatever margin, it is a conscious decision
at the end.
I will share with you that I feel very strongly about it
because I know how it is done today. If you analyze transients you
start with limiting conditions upfront. Those are the initial
conditions. You assume maximum worth rod is stuck. You assume you will
not get credit for the first scram signal and you go on and on and
different people review these analyses for different purposes. At each
end you have a conservatism ended on. At the end you just don't know.
If you were to train operators using the analyses in the old FSARs,
you're in trouble, I think, because that is not how plants behave.
It is important to get realistic understanding of what the
responses would be and make the decision at the end what margin do we
want, are those -- you know, whatever the safety limits are, maybe I
will go back to this word George used -- some kind of a rational thought
process that this is what we want given the consequences from these
kinds of transients and so on. It is important.
I think we have got now 40 years, 30-40 years of experience.
It is important to move and say we are going to move and say this is
what we are going to do.
DR. APOSTOLAKIS: But let's think a little bit about this
word "realism." You can have a decisionmaker who makes a realistic
decision based on what is generally known and what is available to him
or her, so he might say, well, we don't understand this phenomenon very
well, but the prudent thing to do would be to add this -- like this
committee did with the containment spray system for the AP600.
MR. THADANI: I was there.
DR. APOSTOLAKIS: So realism perhaps is not the right word
for you because the Office of Research would look at that situation and
would say, well, we can do better than this because we can develop this
methodology or we can do some experiments, so I am not sure that realism
really conveys what the Office of Research does.
In NRR people are realistic based on what they have and they
say, well, gee, you know, this is the way we will have to do it. Now
they know it is a conservative decision, but it is a realistic decision
because they have to make sure nothing happens based on what they know.
DR. BONACA: You used the word "prudent" and I believe it is
a good word for qualifying the word "realistic." I mean you want to
have a realistic but prudent decision, so I don't know if there is
another one we can coin for soundness, a sound decision.
DR. KRESS: The trouble with the word is we all know exactly
what he means -- you know, like we don't know what the words "best
estimate" means -- that means best estimate versus conservative input
and you are putting in your conservatisms on the end.
The problem is though we know what that means, but does
everybody else know? And the word "realism" has so many meanings to
it --
DR. APOSTOLAKIS: That's right.
DR. KRESS: -- to the outside world that it is a good
communication between us but I am not sure it is a good communication
for the outside world.
DR. APOSTOLAKIS: One thing that bothers me with both the
vision statement and this is that I don't think we have really captured
what research is all about. We are talking about realism. We are
talking about technical basis, and so on, but as somebody said earlier,
I mean other offices can claim that as well.
What makes this Office different?
DR. WALLIS: Let me just say something --
DR. KRESS: It states future needs in the prioritization --
that is in the vision statement.
DR. APOSTOLAKIS: And that's it?
DR. KRESS: I think that is one of the main things.
DR. WALLIS: Let me suggest something, George. I don't
think you make decisions, so you can't say that you want realism in
decision-making. Someone else needs to be doing that and worrying about
that.
What you do is you evolve a technical basis for
decision-making and this technical basis continually evolves for a lot
of reasons. Your job is to stay on top of that continual evolution of
the technical basis for the decision-making which other people are going
to make, not RES primarily -- the Commission, NRR, a lot of people will
make decisions in the field. You have to give them a sound technical
basis, which evolves all the time. That is what your job is.
DR. APOSTOLAKIS: Just from my own benefit, when the agency
was reviewing the AP600 design and the issue of passive systems came up,
the agency developed some approach.
MR. THADANI: Right.
DR. APOSTOLAKIS: Who did that?
MR. THADANI: Much of the work was done, as you know, by the
Office of Research. What NRR did was to lay out -- I was at NRR, as you
know -- and we had, Research came to NRR with their ideas on what should
be done. There were a whole bunch of back-and-forths, and in the end
there was a very extensive agreement on the work that the Office of
Research was going to do.
There was very strong support for Research to develop
technical basis and that is really, as you probably know, led to the
APEX facility, ROSA, and the use of SPES that Westinghouse came in with.
The Office of Research also looked at the reliability of passive systems
and so on, things like that.
DR. APOSTOLAKIS: So it is consistent with this.
MR. THADANI: Yes.
DR. APOSTOLAKIS: But coming back to Graham's question about
decisionmaking, it seems to me what Research does, and not necessarily
the Office, but Research, it contributes to understanding and it
develops tools for doing analysis that will help you make decision.
MR. THADANI: Yes.
DR. APOSTOLAKIS: And we are not really saying those
things --
MR. KING: Tools and data.
DR. APOSTOLAKIS: Well, tools I would take in the broadest
sense, but sure, and data.
MR. KING: I mean take the extreme case -- license renewal.
Without a Research program, what would the agency's position be on
license renewal?
DR. APOSTOLAKIS: But then don't we use those words, Tom,
somewhere? I haven't seen them. I have seen realistic decisionmaking
and I think Graham's point is very good. Others are really making the
decisions. You are not "Understanding" -- the word "understanding" and
the word "tools" and "data," analytical tools to support rational
decisionmaking, I think that is really your mission.
DR. WALLIS: So let's go back to Number 1. What do you mean
by independence? I think you are dependent.
MR. THADANI: No.
DR. WALLIS: Unless you supply the tools that other people
need, you are not doing your job. You have a dependence. They have a
dependence on you.
MR. THADANI: But when I say --
DR. APOSTOLAKIS: There's a different --
MR. THADANI: -- when I say independence, the agency makes,
where appropriate, develops a fair amount of technical basis for its
decisionmaking process and we become an integral part in making sure
that we have developed the right technical basis.
For example, if it is high burnup fuel, is what the industry
proposing to do acceptable? Should it be acceptable? We will
independently develop the technical base.
DR. WALLIS: Independent of whom? What do you mean by --
MR. KING: The industry.
MR. THADANI: Industry.
DR. WALLIS: Oh, independence from industry. That's what
you mean.
MR. KING: Yes, so we make our decisions with --
DR. WALLIS: Well, I hope the agency is reasonably
independent of industry anyway.
MR. THADANI: Now wait a minute. The agency can accept and
review and say that's okay, and what we are saying is we have a
responsibility as the Office to Research to make sure that we would even
identify and as we did for high burnup fuel, as we did MOX fuel, we
would identify areas where we need to develop independent technical
basis to be able to support what the industry may be doing.
DR. APOSTOLAKIS: Why don't you say independence from the
regulated, to make it clear? Because "independence" alone can be
misinterpreted.
MR. THADANI: Yes -- let me suggest, and I think part of the
difficulty is maybe I have more information, if I can go forward I think
it would really help to see what some of these words mean and maybe then
we can come back and say do we now need to revise, what is meant, what
do we really mean by some of this.
DR. APOSTOLAKIS: But the point, Ashok, is not really
whether you understand these things. The point is does the written
statement or the viewgraphs or the writeup convey what you want to do,
and what you are getting now is responses that are indicating that
perhaps it doesn't convey an accurate picture.
I mean we are not questioning --
MR. THADANI: No, I understand what you are saying. To me
the critical pieces are when you look at the vision statement itself and
then some of the specific statements that you will see later on in this
presentation, what is it that we mean by various pieces.
I have to go back, still stay fairly close to what the
Commission's SRM said, unless we have a huge concern about it.
DR. WALLIS: Ashok, let me say why we are worried about
this, why I am worried about this. If these are really critical
aspects, then when you look at what you are doing, you should say are we
really doing what we say we are doing.
MR. THADANI: Right.
DR. WALLIS: So if these are really three critical aspects,
every time you do anything you are going to be worrying am I being
sufficiently independent, am I being sufficiently realistic in my
decisionmaking, and am I being forward-looking, and I don't see any
other than forward-looking is really helping me in evaluating what I am
doing as a member of RES.
I think that there are other things, like, as I say,
evolving a technical basis -- is what I am doing helping the agency
evolve a technical basis. Yes, I can do that. I can ask that question.
MR. THADANI: But to be realistic means you have to develop
technical basis, to understand what realism --
DR. SEALE: The problem is that realism not only means
different things to different people, it means different things to
different problems.
We have talked in here recently about the difficulties
associated with handling defense-in-depth and risk informed regulatory
process, and when we talk about what is realistic in the
defense-in-depth assessment, that is one thing. To ask what is
realistic when you compare between options based on a risk assessment in
terms of a 50.59 assessment or something like that, that is entirely
different. That is a different kind of realism that is involved.
MR. THADANI: Yes.
DR. SEALE: And you don't make that distinction.
MR. THADANI: Realism -- excuse me -- realism, really the
point we are trying to make is get the best available information.
Get the best available information and then you'll make the
decision. Don't just say a priori go forward, and I think
defense-in-depth is certainly part of the thought process that you have
to go through. If I may go through this --
DR. WALLIS: I think you have to -- I am going to interject
some realism and forward-looking aspects here, and I want to have some
time to present what I am going to present this morning, after you --
DR. SHACK: Can I just add one comment? Developing
technical bases is an endless open-ended task. You know, it seems to me
these three aspects are the critical ones. What decisions do you need
independent information on, what do you need for realistic
decisionmaking and what do you need for forward-looking? Develop
technical bases is just -- it's endless. I could do research forever.
DR. APOSTOLAKIS: But what the Office does, really, is
sometimes it advances the state-of-the-art, contributes to our
understanding, and sometimes develops tools that implement the
state-of-the-art -- but nowhere here do we say that. We are always
hiding it under realism, technical basis.
Why don't we come out and say that the mission of Research
is to advance the state-of-the-art and to develop analytical tools that
reflect the current state-of-the-art. That is really what you do.
MR. THADANI: The Commission asked us how we independently
examine an evolving technology and anticipated issues.
DR. APOSTOLAKIS: But this not the mechanics of how you do
it.
MR. THADANI: But these are the critical elements.
DR. APOSTOLAKIS: No, but the first one says the Commission
asked the Staff to describe the vision -- comma -- and what we are
addressing now is the vision.
MR. THADANI: Yes, but I am saying -- no, no, I am trying to
capture everything in there now.
The vision has to reflect some of these thoughts. That's
all I am saying.
DR. APOSTOLAKIS: Yes. The point of view I am taking is I
should be able to look at two or three viewgraphs without Mr. Thadani
present --
MR. THADANI: Yes.
DR. APOSTOLAKIS: -- and understand the mission.
MR. THADANI: Yes.
DR. APOSTOLAKIS: And unless you tell me that you are really
advancing the state-of-the-art and you are developing tools and data,
collecting data, you are not really telling me what you are doing.
I mean these statements are very general, in my opinion.
Anybody can claim that they are contributing to realistic
decisionmaking.
MR. THADANI: You really have to go to the next level.
DR. POWERS: No, no, no. I mean anybody can claim to be
contributing to realistic decisionmaking. So?
DR. APOSTOLAKIS: So you are not really telling the world
what this Office is all about.
DR. WALLIS: What distinguishes you, in other words.
DR. APOSTOLAKIS: Unless you tell them that you are
advancing the current level of understanding in some instances or you
are implementing -- you know, you are developing tools, like I would
view for example the reactor safety study not as a research project,
although it is on the boundary there, it is really pulling everything
together, right?
MR. THADANI: Yes.
DR. APOSTOLAKIS: Most of the stuff there was already out
there. In other instances though you are actually doing research. You
are advancing the state-of-the-art. I mean you are developing an
understanding that did not exist before.
Now these kinds of thoughts, it seems to me, we should not
rely on live persons to --
MR. THADANI: No, we shouldn't. We shouldn't. I agree with
that, but I don't know -- I guess I am having difficulty when it says
prepares the agency for the future by evaluating safety issues involving
current and new designs and new technologies.
DR. APOSTOLAKIS: I am not arguing to put those in the
mission statement, but where you start elaborating on the mission --
MR. THADANI: Yes?
DR. APOSTOLAKIS: -- I would like to see the words like
"understanding," "state-of-the-art" -- then I know you are a research
organization.
MR. THADANI: And if you look at this, this is what it
says -- improving the agency's knowledge. Isn't that what you are
saying?
DR. APOSTOLAKIS: Yes. That is a good part of it, yes.
MR. THADANI: And I am saying underneath this vision there
are -- I know I am not going to have time to go through it, but I would
urge to really, I would really appreciate it if you would take a look at
the viewgraph. Is that for the substances in my view, from the vision
statement which, as you said, has to be shortened, sort of punchy, you
have got to go down to the next level.
DR. WALLIS: Well, I think this is very interesting and very
useful and it is very helpful as a preliminary to what I have to say,
but you can't take all my time, and we're forcing you to do that, so I
think we could have another half-hour though, because I think we may
actually go into questioning these, and I think we should, so why don't
we -- I will just shorten my part of the presentation to suit it.
MR. THADANI: Okay.
DR. WALLIS: Let's go to a quarter of 9:00 then -- this is
just no longer than a quarter of 10:00.
MR. THADANI: I think you are in control, not me, on the
time.
DR. WALLIS: Who is in control here, is it me or the
Chairman?
DR. SEALE: It's you.
DR. POWERS: Well, I am going to control the total
DR. WALLIS: The total.
DR. APOSTOLAKIS: He provides a vision and we provide the
mission.
DR. POWERS: And I hasten to add that you have got a
leadership function here not solely to present your material but to
marshal your troops so that they can provide additional things.
DR. WALLIS: It seems to me --
DR. POWERS: I would very much question the need to parse
the vision statement down to microscopic detail because I think far more
informative than the mission statement is going to be the issues and the
goals that you set forward, because I think you run into a trap. Your
vision statement has to be so short that it necessarily has some
generalities to it. Specifics come the next level down in your little
pyramid.
DR. WALLIS: That is exactly my point too. I don't know
what you are going to do with that, Ashok, and we've got to shut up and
listen to you, but in your handout in your mission statement it said
"role" and you have some bullets where you elaborate on the mission
statement.
That is where you should become more specific and then we
can look at each of those bullets and say are these things which can
then be evaluated in terms of what you do? Can we test what you do
against what you say you are going to do in those areas?
MR. THADANI: Really that is where the substance is in the
end, and so let me just go through a few points and then hopefully we
will have a few minutes -- at the end I would really like to have a few
minutes to sort of get broader feedback from you.
In terms of what do we do, George, you just touched on it,
trying to improve the agency's knowledge where we think that knowledge
needs to be improved -- there are significant issues there.
I mean there are some examples on this chart -- plant aging,
fire --
DR. WALLIS: It's not just knowledge, it's got to be useful
knowledge --
MR. THADANI: I mean what does it mean to say may be
significant to risk then? I can have a lot of abstract discussion, I
mean we can, but it clearly articulates why.
DR. WALLIS: No, it doesn't. It says may be significant to
risk.
MR. THADANI: Upfront sometimes you don't know and you have
to look. Anyway, we clearly have responsibility to take a systematic
look at our requirements and if it is appropriate to change those
requirements, we have examples here -- risk informing Part 50 of our
regulations. We have anticipate transients without scram rule and
station blackout rule. As you know there's been a question whether the
regulations have been effective in achieving the level of safety we
thought they will achieve, and was the cost consistent with what we
thought the cost was going to be, so it's aspects like that.
Again, as I say, these are just examples to try to
illustrate the thoughts that go behind this, and then of course focusing
on operating experience and any results that may come out of research
activities. For example, we look at reliability, availability of
systems and components and from that sometimes you learn things and you
go back and change.
You recall MOVs, motor-operated valve issues with thrust
issues, service water system issues and so on, so it's an area that
provides a great deal of very useful information.
DR. APOSTOLAKIS: I suggest that you add a fourth bullet --
developing an analytical for computerized tools. For example, these
three bullets would not accommodate something like Sapphire.
MR. THADANI: Yes. If you let me go all the way through and
then if I haven't done what you want me to do, I promise you that we
will come back.
DR. APOSTOLAKIS: All right. This is a long way of saying
shut up.
MR. THADANI: No, this is one way of saying I think I have
got that point covered.
DR. APOSTOLAKIS: I am not questioning whether you know,
Ashok.
MR. THADANI: No, I am saying --
DR. APOSTOLAKIS: I am just saying the viewgraph doesn't say
that.
MR. THADANI: No, but because there is another viewgraph
coming and if it doesn't -- look, I completely agree that is a very
fundamental role of the Office of Research.
When you get this so-called "center of technical excellence"
and so on, you will see what do we mean by that. This could be captured
differently. You could put it different places. My fundamental issue
is have we captured all the important elements, and if not, then I want
to be sure we do.
DR. WALLIS: I have a question. I'm sorry, Ashok. Who is
the customer for all the things you do on page 6? They are good things
to do but who is the customer?
MR. THADANI: That is a very interesting issue. Let me pull
up the chart.
Let's take the first one, plant aging. There was a time --
this is research started by the Office of Research -- and it was
considered anticipatory research. There wasn't strong support for it.
What I said is very important. There wasn't strong support for it, but
the Office of Research said to the EDO we do want to go forward and went
forward.
I tell you that if that Office had not done that, we would
have significant issues on license renewal. Fire -- think with fire it
is recognized that the Office of Research has responsibility to make
sure that the methodologies develop if we are really going to go and
apply risk analyses in our decisionmaking process. We have a
responsibility to first say where the weaknesses and why they are
important and then develop the methods.
Methods will be used by others. We use them, of course, in
the analyses that we do ourselves. We certainly work with other
stakeholders. As you know, we have been working on the standards and
this plays into that, and so the use of this is not just necessarily
Research, NRR and NMSS, Regions and whatever, but it is broader use.
I can go on, if you like, on this chart. Would you like for
me to elaborate on that?
DR. WALLIS: No, I just asked who is the customer and you
are saying there is some broad use out there.
MR. THADANI: In the end we believe the customer is going to
be NRR, Commission, and in some cases the industry.
MR. KING: And NMSS.
MR. THADANI: And NMSS, yes. Yes.
MR. KING: They may not know they are the customer yet.
That is the anticipatory part. That is our role, to look on the horizon
and make sure that they are aware of these things coming, and get them
ready to deal with those things. In a lot of cases they do know they
are the customer because they send us user needs and say we need help,
so it is a whole range of things.
MR. THADANI: Again, since I know how much time I have, I
won't dwell on it, but these are some of the examples, ways that we do
support the regulatory activities.
Sometimes -- a lot of the things we do are not really
research and I think that needs to be recognized, and that is why we
have to be careful about the words in the vision statement also. When
we consolidated with that part of AEOD, we picked up the operating
experience evaluation portion and so on.
These are many of the areas that we are providing support
and in many cases in terms of regulatory activities, but some of the
ideas that come out of here, for example IPE, IPEEE, they support you
might recall the generic safety issues that were sort of -- were not
resolved but it was noted that IPEs and IPEEEs would help resolve those
issues, so when I say "reviews and associated perspectives," that is
part of what it means, and that those insights then are used in other
ways including by the Office of Research.
Clearly certification of the advanced light water reactors,
you know about, and we need to be looking ahead. Are there going to be
some new designs coming down the road beyond the light water reactor
technology? As an agency are we going to be ready when the time comes?
Decommissioning is of course in terms of trying to develop a
database for some of the codes, for analysis -- here I would say more
realistic analysis than we do today with DandD code, for example.
DR. POWERS: Let me ask you a question about the third
bullet there.
You don't have a whole lot of new designs in the pipeline
coming down to us. We have kind of gone through and done that, but I
see signs that the industry is looking at some fairly radical changes in
the existing reactors. We certainly have quite a few power uprates, of
a 5 to 7 percent range, but I also see things of fuel being modified to
get more heat transfer out of it and things like that.
Would that fall under this bullet or is there some other
place that you put that kind of advanced use of the existing reactors?
MR. THADANI: In terms of using new technology, let's say,
that would be more efficient, better fuel designs and so on, clearly
would be in support of regulatory activities, but it will also,
depending on what it is, we have characterized it as basically preparing
the agency for what is about to come down the road, and that may be
exactly what you say for current reactor designs.
I just go one step further than you did, and that is I am
beginning to see a change in climate in this country, I think, and
beginning to see more optimism in terms of the role of nuclear power in
20, 30, 40 years from now. A variety of factors play into that, but it
may be that at least the Office of Research -- we need to be somewhat in
tune to what may be happening out there, and to be sure that our
recommendations to the Commission come in a timely fashion, and this may
be beyond the light water reactor technology.
DR. SEALE: But Ashok, wouldn't it be wise to include in
that bullet, perhaps making that bullet a little bit more general,
responding to the kind of upgrade of existing technology that the
industry may be looking at directly, because otherwise that is a very
tempting target to the people who keep wanting to cut your budget and so
forth. In that format, it is very tempting.
On the other hand, if you can enfold into that the advances
in existing technology associated with modified fuel designs and things
of that sort, then I think that is a much less vulnerable area.
MR. THADANI: Yes. I captured this -- a little bit of
repetition, I think -- but I think that is a fair comment. You will see
it another place in this, but we need to probably integrate these a
little bit better than we have done.
You see, the bottom bullet here sort of -- we need to do a
little --
DR. SEALE: MOX fuel is not new geometry.
MR. THADANI: No --
DR. SEALE: Or new thermal hydraulics.
MR. THADANI: No, but there are some unique aspects to MOX
fuel.
These are some other areas, and I am going to rush through
because I do want to make a few comments at the end.
These are some of the -- again, trying to go back to the SRM
and trying to say in a bullet fashion what is it that we do that
supports what I think is the high level vision statement.
DR. KRESS: Ashok, just one comment on this. I love these
examples. They really show we're hard-working and being useful, but in
your writeup that you provided us, you have similar examples scattered
throughout different parts. If it were me, I would get those out
because they are the example of the day and a year from now or two years
from now they are no longer appropriate and you would have to change
his, and, you know, you want to make this document sort of a timeless
thing.
MR. THADANI: Yes. One of the things -- I know exactly what
you are saying and if I were in an environment where I could make
statements and have acceptance immediately it would be one thing. I am
not in that environment.
I have to be able to say, well, what does it mean, what have
I done in the past, how did lead to something, and Graham's question I
think was very valid -- what have you really done and why did you start
aging research and what was the value.
Sometimes, and I will give you some examples. Source
term -- how many people would have started on NUREG-1465. You tell me
today, would I get support to start source term research?
Absolutely not. I don't believe I would. What is the
value? Today we know, but we didn't know then, for sure. You know, PRA
technology, I mean -- 1975, 24 years ago. Actually the draft came out
25 years ago. It takes some time --
DR. WALLIS: It takes a generation.
MR. THADANI: It takes time in some case. It takes vision
and you have to fight for it, in my view, because it is not going to
come easy, and that is why I am extremely happy that ACRS by far has
shown a lot of vision, in my view, and that is why I am very anxious to
make sure that I talk to you frequently enough, make sure I get your
advice and recommendations.
I am not going to go through the last charts, but these
things we have talked about are just wonderful. They are nice things to
do. They are good things and with some changes, as you have suggested,
maybe they reflect what the Office of Research should be doing.
But now I am concerned that we are really not where we would
like to be. We are not there. This is good, it's wonderful to say all
these things. As a country we are losing leadership in many areas. I
mean it may be debateable whether we are leaders in the thermal
hydraulic area. We are certainly leading in the severe accident arena,
no. There are many areas -- we may not agree with the specific details
in terms of the core capabilities and so on. In many areas we are below
what one might call core capability and I think we can get into a debate
on that.
The point I am trying to make is we are not leaders in many
areas in the world anymore, like we used to be. We used to be able to
do things on source term. We used to be able to do things on material
aging in a bigger way.
A lot of the facilities in this country, as you know, have
closed down.
France and Japan are two countries we tend to go to
oftentimes when we need information. We did a functional comparison of
research between U.S. NRC, IPSN and others, and we spend about 25
percent of what they do, and obviously we have twice as many facilities,
almost twice as much as France and Japan and so on.
You have noted in your recent report there are certain
things that we are not doing and you are right, we are not doing some
things for a variety of reasons. I am concerned about the direction and
I have so indicated in my interactions with the Commission and in terms
of the budget activities.
John Ahearne briefed the Commission and I certainly urge you
to take a look at the transcripts because there is a lot of information
there that is helpful.
Every time we close a facility, we walk away from things --
there are times you have to do that, obviously. I mean when you have a
$40 million budget versus $200 million in the '80s and $104 million five
years ago, a lot has to give, but I think we are slowly but surely
eating more and more of our seed corn -- I believe -- as we move on, and
I think this talking about vision and what we do is very helpful, but I
also think the broader implication of this is are we where we should be
or are we not, and I am concerned about where we are in terms of what we
do and how we do that.
So I did want an opportunity to tell you my own views after
having talked about what I think is a lot of nice things to do and so
on, that I am concerned where we are.
DR. WALLIS: Ashok, I thank you very much. I think the
ACRS, whatever we say is really very supportive of what you are trying
to do, and what I would like to ask you now is why did you come here
today?
Two reasons, it seems to me. One is so that you can
influence the report that we are going to write on Research, and the
other one is for us to influence what you say and do, and particularly
this vision statement. Do you want us to influence this vision
statement, and in that case, what we started to do for awhile would be
appropriate maybe by some other mechanism or some other time, where we
actually go through and help you produce a really good, saleable,
impressive vision statement which helps both external relations you have
and also your internal operations as a guidance for your people, not
just telling the rest of the agency and the world what you do.
So what did you hope to do about influencing us and how can
we influence -- help you, in terms of specifics?
MR. THADANI: I guess -- two parts I was looking for.
First, I wanted to make sure and get through this dialogue a
sense of what some of your concerns might be, and to try and fold in, as
long as we are in general agreement.
I did want to make sure that you had a fairly good sense of
where we are, what we are going to be sending to the Commission. I
would like to be able to tell the Commission that we have discussed this
with the Advisory Committee. I am not sure that I can say much more
than that, that you certainly provided your comments and so on during
these discussions.
We would like to solicit endorsement from the Commission, as
I said. Now I don't know if the Commission -- but I wouldn't be
surprised if the Commission were to come to the committee and solicit
its views, but I wanted to be sure that you were really pretty much
up-to-date on where we stand and to take advantage of your thinking, and
we have a few days, a couple of days I think really, to try and see what
revisions we need to make and get this to the Commission.
DR. WALLIS: So you are going to send this vision statement
with amplification on the role --
MR. THADANI: Yes.
DR. WALLIS: -- in the first bullets here in a few days?
MR. THADANI: Probably, Jack, early next week.
DR. WALLIS: So then it becomes something --
MR. THADANI: Let me say this -- early next week for our
concurrence process. It goes to Commission -- it goes to the EDO or the
Commission October 15th? I am not sure which.
MR. KING: The Commission October 15th; EDO October 8th.
MR. THADANI: So we have a week to get concurrences. That
means whatever changes we make we will probably make today and tomorrow.
DR. WALLIS: Then it becomes a document which we can refer
to in our research report and if there are some things that we feel are
very good we will cite them. If there are some things that we feel
should be expressed other ways, then that would be the place we would do
it, rather than doing it to you before you finalize the document.
MR. THADANI: I think that is really -- since we have such
little time to get this out -- we have probably a week -- when it goes
into concurrence process, it doesn't mean that we can't make changes
during that time.
DR. WALLIS: I think if you actually talked to at least half
a dozen of these members here you would get significant comments on most
of these bullets.
MR. THADANI: Yes.
DR. WALLIS: I am not sure you have a mechanism for doing
that.
MR. THADANI: I would prefer a more unified view on this.
DR. KRESS: Rather than individual.
MR. THADANI: It's difficult. I mean I tell you I am very
happy we went through this process of getting -- I wanted feedback from
Office Staff -- every one. I encouraged them to give us their feedback,
and I think it added value. We didn't necessarily agree with everything
but I think it added value and so it would be important, I mean it is
important to us as an office to make sure we, if we are too far off, I
certainly would like to make sure we have a sense from you.
DR. WALLIS: How would you get that? Have you got that now?
Do you feel satisfied you got it this morning?
MR. THADANI: I got a sense that there are some areas there
may be -- I don't have, I am not sure I have a sense of the committee.
DR. KRESS: One possibility might be -- I know you don't
want individual members, you would rather have a consensus, but one
possibility is that each of us mark up this letter, this product, and we
discuss it shortly and maybe send the markup to it, on it.
MR. THADANI: If you can do that, that would be --
DR. KRESS: Just margin comments and things of that nature.
MR. THADANI: If you can do that, that would be of great
value to us, great value, because there are some big issues of
disagreement I would like to make sure we deliberate and think through.
DR. WALLIS: I would ask the Chair whether it is appropriate
for the committee to act that way, and for each one of us to send marked
up versions of the document? Does the committee have to act as a
committee in this case?
DR. POWERS: Well, what you get is a bunch of individual
comments. You don't get a committee position.
DR. WALLIS: That's right. This is appropriate?
DR. POWERS: I would certainly be more comfortable sending a
committee position.
DR. KRESS: Well, we could probably do that too.
DR. POWERS: I think Tom actually said he wanted to do
that --
DR. KRESS: Yes.
MR. THADANI: That's what I thought --
DR. KRESS: That is what I had in mind.
DR. POWERS: A bunch of marginalia and then just boil it
down into a single --
DR. KRESS: Yes, that is what I had in mind.
DR. POWERS: I am loathe to say that we have time on the
agenda to do that --
DR. KRESS: That is the real problem.
DR. POWERS: But we can look as the hours pass by to see
what time we have.
On the other hand, it seems to me that honing and refining a
vision down to the minimum number of words and the maximum coverage
possible without having gone through at least the next three in your
pyramid goals, so that you can go back and iterate them, I think it is
not useful to look at the vision by itself. I think you need to look at
your goals, your issues and your planned accomplishments and develop
that part of the pyramid, and then go back and make sure your vision
statement is okay.
I am not sure -- I am quite certain you can work the vision
statement for months and it becomes a very effete exercise. I think the
meat lies down lower.
MR. THADANI: As always, yes. My greatest interest really I
have is does the vision statement reasonably capture what the Office is
all about, and underneath elaboration, because you can never -- I mean
there are always going to be, someone is going to say I understand that
is your vision but I really want to know what it is that you do and how
do you satisfy this expectation in the vision, and so you do have to go,
at least in my view, and that is why in this paper we tried to go to the
next level of bulletizing what it is that we would do in the different
pieces of it.
DR. KRESS: I particularly thought that the set of bullets
on the roles of research captured pretty well what you do. I really
thought those were good.
MR. THADANI: And we had to do that. When I asked the Staff
for their feedback, it was clear to us that we couldn't just send them a
vision statement. We had to give them some additional information, so
we provided some bullets in bullet form, some of the ideas.
DR. WALLIS: My advice on the bullets is look at each bullet
and say when I plan the work, when I do the work, when I have done the
work and I look back at the bullet, does it help me to assess what I am
doing, and I think you should separate the bullets which are like that,
where you actually an objective like confirming regulatory decisions or
something, and the bullets which describe how you go about your
business, like ensuring processes. When they are separate things, then
the objectives actually have an output.
It is now a quarter of 10:00. Are you ready to leave?
Thank you very much.
MR. THADANI: Yes, thank you.
DR. WALLIS: I now have to make my own presentation. I will
hand over to the esteemed Chair.
DR. POWERS: Okay. I will hold you to your time. You have
half an hour, sir.
DR. WALLIS: Can we go off the record, please?
DR. POWERS: Oh, yes. We can now go off the record.
[Recess.]
DR. POWERS: Let's come back into session.
Our next session is on the PASS system and the Combustion
Engineering Owners Group's ideas on how to make modifications to the
point of elimination of the PASS system.
Dr. Kress, I think you will lead us through this?
DR. KRESS: Yes, thank you, Mr. Chairman.
That is exactly what we are talking about. Our Subcommittee
on Severe Accident Management heard this basically last week, I guess it
was, and you recall that we had a previous similar proposal from the
Westinghouse Owners Group, and wrote a letter on that where we had
questions about the possible need for pH measurements in sumps and the
possible need for fission product, certain fission product measurements
in containment, so when you hear the presentations from CEOG and the
industry you might keep in mind those reservations we had before, but
also ask yourself the questions are these -- is the information that is
supposed to get out of the PASS system really needed for anything or are
the needs met better by other sources of information?
The other question you might think about with respect to our
letter, possible letter we may write, is is the PASS an appropriate
configuration or appropriate way to get such information in the first
place.
With that as a little preamble, I guess I will turn it over
to the Staff to introduce it -- Mr. Schneider? Okay. The CEOG people
first.
MR. SCHNEIDER: My name is Ray Schneider. I am the Project
Manager for the CE Owners Group activity on PASS elimination, PASS
reduction, and what I would like to talk about today is an overview of
the process as to what the PASS Elimination Project is, why we are
interested in removing PASS from the systems, and why we believe our
proposed relaxations and alternatives meet the needs of all the
prospective stakeholders involved with the need for the information for
accident management.
The purpose of the effort is to eliminate the Post Accident
Sampling System in all the CEOG plants. Essentially that means removing
the PASS requirements from the licensing basis, discontinue operation
and maintenance activities on PASS components and identify where
appropriate alternate non-PASS equipment, basically in-plant equipment,
and establish other procedures as appropriate to meet the general
objectives of emergency planning, accident management, core damage
assessment and all the other needs that go along with beyond design
basis events.
I would like to talk a little bit about background as to the
environment in which PASS was initially instituted and other activities
that have both come before and after PASS. Many post-TMI changes were
instituted to help the plant staff better respond to post accident, to
beyond design basis accidents, and to increase operator preparedness for
those events.
Among those changes were improved emergency operating
procedures which went from very design-basis oriented to now functional
operating procedures as well as optimal operating procedures, so there
became diagnosis phases, more of an ability for non-typical events to be
dealt with rather effectively, and a much higher level of training.
In addition, instrumentation was placed in the system to
help the operator with previously unanticipated events. The reactor
vessel level monitoring system is one, the head vents for avoiding
noncondensables is another, and then there was a lot more emphasis and
knowledge placed on the use of the in-plant instrumentation and better
interpretation of equipment like core exit thermocouples and neutron
flux monitors, so a lot has changed in terms of how the operators use
the information, interpret it, and what is available to them.
Operator training has increased substantially in beyond
design basis events including severe accident management and emergency
preparedness, and then in the same earlier timeframe concurrent with
some of these there was a Post Accident Sampling System that was
implemented and we will talk a little bit about that in a minute.
Following the years of TMI there's been very detailed
reviews of the plants from a PSA perspective and the introspective look
at severe accident vulnerabilities to identify situations where the
plants become more vulnerable and to eliminate those and to better learn
how to deal with potential higher risk events or high frequency events.
Most recently in the end of 1998 all plants in the United
States including all the CEOG owners have implemented severe accident
management guidance which includes calculation age for severe accident
management, detailed guidance as to how to monitor severe accidents and
how to select procedures or actions given a beyond design basis event.
The Post Accident Sampling System was intended to acquire
information, initially primarily for accident management and accident
recovery actions. The sampling of the reactor coolant chemistry and
containment atmosphere generally included items such as hydrogen for the
containment atmosphere and actually reactor coolant, boron for the
reactor coolant, pH -- well, pH in the reactor coolant in a different
way -- we can talk a little bit about that -- and radionuclide sampling.
The PASS information was primarily tied to, in the original
regulation, core damage assessment. As we will discuss throughout the
presentation, the need of the system to meet these goals we believe is
unnecessary and clearly the one stated goal of one of the major
components of the PASS system, which is radionuclide assessment, has
just recently been determined by the NRC to be a non-useful item, and
they have granted the ability of Westinghouse to remove radionuclide
assessment for purposes of core damage assessment.
So, as we will talk about, a lot of the features, actually
all of the features of the PASS we believe are no longer needed and
actually never were.
The necessity of PASS requirements has been an open question
for quite some time. NRC did their own introspection and they funded
contractors to look at an internal review of the PASS system, and the
contractors basically determined that PASS was at best marginal to
safety. It recommended relaxation of most of the PASS requirements but
because of what we believe was a faulty cost benefit analysis of the
times, they tabled the ideas.
In 1993 the CEOG independently went in to the NRC with a
request to change a number of the PASS requirements due to the fact that
we believed they were no longer -- they were not needed or there were
better ways of doing the function without forcing a sample. NRC
concurred in our recommendation to use safety-grade containment hydrogen
monitors, which were required by Reg Guide 1.93 to replace the hydrogen
grab sample.
They also concurred with our position to delete the pH --
there weren't even requirements. The pH turned out to be a Reg Guide
1.97 recommendation which a number of plants had opted for but they
agreed with the recommendation to delete the pH commitments.
In addition, with respect to relief on boron, they suggested
that neutron flux measurement could deal with the -- replace the boron
sampling and they put an arbitrary time at that point of eight hours on
it. Later additional relief was granted in the System 80 Plus
certification process in 1995 for the operation of PASS so they can
design a much more benign -- a much simpler system.
The Generic Letter on PASS relief has been under discussion
of the NRC we believe for several years, since about 1996, so there is a
long history behind concerns over the use of PASS.
Given the background, why are we revisiting the PASS relief
issue? Well, we have learned a lot in accident management. We have
implemented the severe accident management guidance. We have gone
through the detailed PSA reviews. We have looked at all our operating
procedures. What we find is that the information is not needed for
accident management, emergency planning or any aspect of the way the
utility does its business which deals with the state.
The use of PASS was not -- began, as we learn more about it,
began to be viewed as not risk beneficial --
DR. WALLIS: Could I comment? What do you mean "viewed
as" -- either it is risk beneficial or it isn't.
MR. SCHNEIDER: We believe it is not risk beneficial.
DR. WALLIS: Is that based on some numerical calculation or
something?
MR. SCHNEIDER: It is based on the fact that the information
you are going to get out of the PASS system will be biased and
nonconservative --
DR. WALLIS: Based on some sort of weird logic where you say
what's the need, does it meet the need, no, it doesn't -- that sort of
thing?
MR. SCHNEIDER: Yes, it is based on how accidents are
managed and what information we are expecting --
DR. WALLIS: So it is not risk that is being evaluated, it
is really something else, is it?
MR. SCHNEIDER: Well, there's multiple ways of looking at
risk. One is that you have to send someone in to take a sample. If you
are not using the sample you are exposing the person to radiation
unnecessarily.
Another aspect of viewing risk is if you get information
that is not going to be helpful and could confuse or obfuscate the issue
or possibly slow down evacuation --
DR. WALLIS: It isn't beneficial for any purpose whatsoever
then? It's not just risk. Is that what you are saying?
MR. SCHNEIDER: Right. And the system maintenance costs are
high and are beginning to rapidly grow through obsolescence, and what we
will talk about -- we have a few slides where we basically go over some
of these issues.
Again, operator training guidance and resources have
substantially improved over the years. A detailed evaluation of PASS
use within at least the accident management and emergency planning
structure shows that alternative or alternate in-plant instruments and
guidance are more appropriate for taking the actions and making
decisions. PASS measurements are too slow for decisionmaking. The PASS
measurements are not relied upon in our emergency operating procedures,
so the operators don't use them.
The site emergency plans establish the emergency action
levels on area radiation levels, core exit thermocouples without using
PASS so they are not used by the emergency planners.
Emergency planners do not require the PASS data for
protective action recommendations and guidance and we have confirmed for
a number of the state agencies, whoever contacted the state agencies in
our group -- I think there's been three or four contacts with the state
agencies -- have all agreed that they don't see a need for that
information.
SAMGs -- Severe Accident Management Guidance -- indicate the
candidate high level actions are independent of PASS so they are not
used in the beyond design basis severe accident management structure.
The core damage assessment methodologies we use are geared
primarily with thermocouples, use of thermocouples, hydrogen, and
containment area radiation, and similar to the WOG methodology which
just recently has been agreed to by the NRC as being independent of
PASS.
The issues of risk-benefit or the risk issues come to the
fact that PASS requires personnel to obtain the sample, analyze the
sample, and to interpret the sample. So you're diverting the resources
during an accident. We believe that's a negative action.
Plus, in order to obtain the sample, you have to send
someone down to physically take a relatively highly radioactive sample,
that we don't feel is going to be used or that isn't going to be used,
and, therefore, there is an ALARA impact that's unnecessary, plus the
sample line itself runs the potential of possibly clogging or having
leaks and as a result, you can end up having aux building radiation
leakages directly from the containment.
DR. WALLIS: As I said, I think, before when you presented
this, the fact that it's not a very good design doesn't really mean you
should go do away with it. If the function is needed, then you might
argue for a better design that didn't have these disadvantages. But
we've been through that, I think, with the subcommittee.
MR. SCHEIDER: Right, but the function is not needed.
DR. WALLIS: These are weaker arguments than your argument
that you don't need the function anyway.
MR. SCHEIDER: Right.
DR. WALLIS: You can establish that, then that overcomes all
the arguments we'd have about why don't you just do it in a better way.
MR. SCHEIDER: Okay. Understand. We felt that the system
was not providing a benefit and, in addition, it had these detriments,
if you wanted to use it.
So to take --
DR. WALLIS: But the fact that you've got a bad design is
not really an excuse. If the function is needed, then the obvious
response should be, well, make a better design of the system.
MR. SCHEIDER: Right. The function is not needed.
DR. KRESS: I think if you looked at the functions for PASS, if you did
not currently have the PASS system at all and NRC was considering
passing the rule requiring a system that gave you those functions, they
would have to make a back-fit analysis in that case.
What I think you're saying is it would not pass the back-fit analysis,
the regulatory analysis would fail.
MR. SCHEIDER: Right.
DR. KRESS: So it's sort of an inverted use of the
regulatory analysis. If you couldn't have passed the rule in the first
place, maybe it's a good reason to not have it in the second place, I
guess.
MR. SCHEIDER: There is no benefit for the system, period,
and I guess maybe that's the strongest argument. But in addition, we
believe the use of the system as it currently exists could be
detrimental and causes its own set of problems.
Last of all is the fact that you're getting non-conservative
estimates of the sample.
The last item is the item of cost. Surveillance and
maintenance costs are relatively high. It's the most highly maintained
system in the plant. There are going to be capital cost issues
resulting as the equipment ages and becomes obsolete. Some of the --
DR. WALLIS: When you say it was the most highly maintained
system in the plant, this sort of --
MR. SCHEIDER: Requires the highest level.
DR. WALLIS: -- rather unnecessary, superficial thing which
is only called upon -- never called upon -- is the most highly
maintained system in the plant?
MR. SCHEIDER: The standby equipments have a lot of very
sensitive components that fail, I guess, rupture disks. Dave, do you
want to discuss about why the system has as many problems as it does?
DR. WALLIS: It just seems very surprising to me.
MR. BICE: I'm Dave Bice, representing Arkansas Nuclear 1,
Unit 2, our Energy Operations. Any required system, which this one
presently is, by NUREG-0737, requires us to continuously surveil it, to
make sure it will be available upon a severe accident.
Well, during those surveillances, we're going to find the
problems in the required maintenance, due to valve leak-by, since we're
going from a 2000 pound system down to a zero pressure sample system,
you're going to blow rupture disks over time as you get very small
amounts of leakage, slowly pressurizing systems.
You have pressure control valve failures. Instruments are
so -- I don't know how you call it -- technical and expensive that they
constantly need calibration to be maintained available for your
accidents. So we do spend quite a bit of time, manpower and money on
this system.
DR. SEALE: Could I ask the nature of the Y2K issues that
are pertinent to this particular system?
MR. SCHEIDER: There's a handful of plants -- not every
plant has a Y2K issue with it, but there's a handful of plants that have
software for -- calibration software for analysis and they would need to
basically modify their software to address the -- I guess it's the
assessment. Not the sampling itself, but in the actual radiation
assessment of the -- but it's -- and that varies.
The more automatic the systems, the more the Y2K problems.
The less automatic the systems, it diminishes. So that's a variable
among the plants.
DR. SEALE: Does this have to do with recalibration
sequences and that sort of thing, where a calendar is involved?
MR. SCHEIDER: I don't know the precise details of why there
is. I know that they have claimed that they're going to have to modify
the software that does the analysis. Whether it's just a formal -- to
formally go through to make sure that there is no Y2K chip that would
cause a problem, it depends on how they -- when they purchased the
software and what software they're using and if there is a clock issue.
DR. SEALE: It's intriguing; most of the things I've read
indicate there aren't any Y2K issues. It's interesting that this one
suddenly turns out to be.
MR. BICE: I don't the issue, Bob, is there are. I think
the issue is that the plants have done the analysis and have identified
those issues that need some corrective action, and this is probably one
of them.
MR. SCHEIDER: Right.
MR. BICE: The computer that does the analysis probably
needs a Y2K chip.
DR. WALLIS: It's going to present data that you says you
had iodine in your plant in the 1900s.
MR. SIEBER: Right. Shouldn't all the Y2K issues be solved
by now? It's getting pretty close to New Year's Eve. If it doesn't
pass, then I guess there's a problem.
DR. WALLIS; You're saying that they should have done it
already?
MR. SIEBER: They should have fixed this already.
MR. SCHEIDER: My expectation is that if this doesn't go
through, there's backup. They're prepared to make whatever changes have
to be, or they'll make fix-arounds to get around the Y2K. But that
wasn't the only -- again, that's not the only issue. It's really the
long-term maintenance, the perennial oversight, the additional capital
costs of just the obsolescence.
And in addition, there were some plants that the sooner the
better, because they face an incremental capital cost additionally.
DR. SEALE: One other thing on that. You indicate that
there are obscure procedures that will no longer be necessary, I take
it, if you implement this change. Is that a universal problem among the
plants or is it just certain plants that have this and how many of them
are there?
MR. SCHEIDER: Well, core damage assessment is like a ten to
20 page procedure, going down isotope by isotope, assessing exactly what
the status of the cost is supposed to be, and that's not used often. So
training on that still probably goes on and that we are strongly
recommending elimination of that full procedure. So that would be --
right now, all our CEOG plants have it, because we always had a
four-tiered core damage assessment methodology, radiation, hydrogen,
core exit thermocouples, and radionuclides.
We're recommending that that one basically be cut out and
eliminated, and that's the most complicated and most cumbersome of the
lot.
DR. WALLIS: That's a bad argument, though. If the
procedures are obscure, then make them clearer.
MR. SCHEIDER: It's tedious. It's a very tedious and
detailed assessment to get information that can't be used. But that's
-- and so how much training do you want to give to someone?
DR. WALLIS: You mean complex and time-consuming and so on.
MR. SCHEIDER: Right.
DR. WALLIS: You don't mean obscure, I hope.
MR. SCHEIDER: Sorry. Okay. No. Okay. Yes. It's clear.
They know what they have to do, but it's a very complicated and complex
procedure.
What the CEOG is recommending is basically -- we're
recommending currently three items; eliminating RCS dissolved gas
sample, eliminating the reactor coolant boron sample, and replacing the
radioisotopic sample with a commitment for field team monitoring using
iodine-131 monitors in the field.
But there's two other items that are on here for
completeness. When the PASS is said and done, we have already, through
a granted SER, eliminated the sump pH sample and we've already replaced
the containment hydrogen samples with Reg Guide 1.97 in containment
hydrogen monitors.
So in this case, it's a replacement, and what I would like
to do is walk through what we really mean by eliminating the sample in
terms of -- in many cases, we're really replacing it with something that
is doing the better job.
DR. KRESS: At this point, I'd like to point out to the full
committee that in our previous concern about the Westinghouse, one of
the items was the sump pH. As far as I'm concerned, it's not an issue
with the CEOG because the elimination of that function has already been
granted.
So it's not really an issue with respect to their proposal
to eliminate their PASS system, because they've already eliminated that
part of it, basically.
MR. SCHEIDER: Okay. I guess I'll use this slide pretty
much for the remainder of the presentation and work through the items
one by one, and I'm willing to talk about what we believe the bases to
be and what we've done with respect to the work.
The RCS dissolved gas sample, what we have is a table which
has basically the originally measurement that we're talking about, its
initial purpose, our recommendation, the basis for the recommendation,
and its current status. So we'll cover all five on the table.
The RCS dissolved total gas measurement was initially
intended to identify non-condensable gas potential as a result of the
TMI bubble in the head. But at the same time, post-TMI results --
post-TMI regulation also had the utilities install reactor vessel level
monitoring systems to pick up the presence of initial voiding and vents
to basically deal with the voiding.
All of our procedures rely totally on the RVLMS assessment
and the reactor vessel head vents to do that job. It's quite an
effective method. It's a redundant Class 1E system, and we see no
benefit in requiring a dissolved sample.
So that's proposed.
DR. BONACA: Does the reactor vessel level measurement
system span the whole -- it only measures up to a certain point in the
vessel, doesn't it?
MR. SCHEIDER: Right. It measures the void down to the --
in our system, down to the top of the fuel plate.
DR. BONACA: So right now, there is no procedure in place to
define this input from PASS.
MR. SCHEIDER: The PASS assessment -- right now, we're
required to have a method to basically take the sample. The sample is
taken, but there is no process by which they would -- there is no way to
use it.
DR. BONACA: So there is no procedure that says get
assessment of non-condensable from PASS.
MR. SCHEIDER: Right.
DR. BONACA: There isn't.
MR. SCHEIDER: The operator has the option of saying if he
needs it, but the thing is, if you have a visual on the void, you know
that you have the void there. And if you fill up without any voids,
it's not a concern. So the -- right.
In terms of the sump pH, all but one of the CEOG plants have
TSP control buffered sumps. The pH is controlled to greater than seven.
We've looked at a number of issues recently, but the concern -- the
issue, we believe, is that out of pH of seven, based on radiolytic
iodine evolution curves and based on looking at the capability of
degrading the TSP, you have a lot of margin before you get to any kind
of significant pH, any kind of significant iodine evolution.
So we think there is a lot of margin there already. The
existing methodologies are done according to the current standard review
plans, and we've previously gotten granted approval for the removal of
the system.
DR. POWERS: When you choose the pH-7 as the dividing line for when you
don't get significant iodine in re-evolution, have you concluded that
based on just looking at I-2 partitioning or have you looked also at
organic iodide partitioning?
MR. SCHEIDER: The pH-7 was based on the regulation. It
defines pH-7 as the point to stay above. What we've looked at is the
iodine gas evolution for pH. Basically, we've looked at the iodine gas
evolution. In the old MHA, you were allowed or you were required to
consider four percent organic iodine in the containment atmosphere,
regardless, and we believe that based on current -- the new source term,
the organic iodine is predicted to go down substantially, but at four
percent elemental iodine, the pH would be around five, that would allow
that to evolve.
DR. WALLIS: When you say pH control, to me, controlling
something means that you measure it and you control it. If you're not
measuring it, you're really predicting it. You're not controlling it.
MR. SCHEIDER: Yes. We've put in enough TSP. TSP is
relatively --
DR. WALLIS: You're predicting that it will be maintained
about seven, but if you don't measure it, I don't think you're really
controlling it.
MR. SCHEIDER: Right. We're predicting it will be
maintained above the level sufficient to limit iodine evolution to
significant levels. You may go below seven, but the iodine evolution is
not going to --
DR. WALLIS: One could argue, in some sort of
defense-in-depth method, that, well, you've predicted it, but to be sure
you've really got it there, you really ought to measure it.
MR. SCHEIDER: Unless you're going to do something with that
measurement, I don't know what you can do. If you -- we don't -- we
believe that the design basis is adequate and provides a lot of margin.
We also believe that the TSP, as you start to neutralize it, it only --
the example that we've run just recently is to estimate the impact of
TSP degradation on the pH and the conclusion we came to is that a
typical plant that will have like 6,000 pounds of TSP in the sump,
designed to a pH of seven, would only require 1,500 of those pounds to
maintain a pH of six and a half.
So we believe there's a lot of margin. We also believe --
DR. WALLIS: What form is it in? What form is this TSP in?
MR. SCHEIDER: TSP is in solid baskets. It's a solid in
baskets.
DR. WALLIS: It's in granular form or something?
MR. SCHEIDER: Right.
DR. WALLIS: So if it's somehow sintered together or forms
some layer on the outside so the middle is not accessible?
MR. SCHEIDER: There were a lot of tests for solubility.
There is a lot of evidence to indicate that it's going to be soluble in
the water we're putting it in. I don't think that's going to be a
problem. And solubility occurs in a matter of hours.
DR. WALLIS: Have you got it covered up with something which
came loose from some other part of the system? And there are things
that could be postulated.
DR. KRESS: In a previous remark, just a few seconds ago,
you related the pH-7 somehow to the new source term, and I fail to see
any connection at all between those two things.
Could you go back and clarify that just a bit?
MR. SCHEIDER: Yes. The issue is -- okay. Let's go back.
This is basically a radiological -- a radiolytic iodine evolution from a
water pool pH versus the fraction of iodine that will radiolytically be
produced.
When we did the -- when you do the --
DR. KRESS: Where did you get that curve, by the way?
MR. SCHEIDER: This is an ORNL report. I have the report
here, but it's ORNL, on chemical forms of iodine, issued maybe five or
so years ago.
DR. POWERS: What was the dose rate?
MR. SCHEIDER: I'd have to go and double check. I don't
know that off hand.
DR. POWERS: Is there, in fact, no dose rate dependence at
all in that curve?
MR. SCHEIDER: No. There's -- the curve may not show the --
well, there is scatter. There is a little data scatter on here, that
maybe the dose rate dependence, we took the mean, but there's a couple
of -- it's a radiologic -- it is a radiolytic curve. The dose rates are
accounted for, whether -- but I don't have the exact information as to
how sensitive the actual curve is to the dose rate.
But the point that we might want to focus on is that when we
do the -- when you look at the original source term, all maximum
hypothetical accidents are based on the old source term, which required
-- which had like a 91 percent elemental iodine, four percent organic,
five percent particulate.
And the significance of the organic is that the organics
never got into the sump and basically remained in the containment
atmosphere for the full duration of the event.
And what that basically says is that regardless -- in the
old source term methodology, regardless of where you started, you're
always going to have at least four percent iodine in the atmosphere.
With the new source term, the iodine contribution has
dropped to like .15 percent, some negligible -- the organic iodine is a
negligible level. And what we're basically demonstrating here is that
even if you were to move back to a pH of five, the equivalent effect
would be no worse than what you would have with the old -- with the
existing source term with the organic iodine contribution that you've
already analyzed.
So in essence, the general feeling is you're relatively
bounded. You can't -- you're not going to have an uncontrolled pH
condition. You have a lot of margin. The other curve we have, that our
chemist has, who is sitting in the back here, basically demonstrates
that as the TSP degrades from 6,000 pounds to about 1,500 pounds, even
if I only have one-quarter of the TSP left, I'm only going to move my pH
from seven to six and a half.
So there's a lot of margin in that because it's a relatively
effective buffer. We've also kind of looked at the possibility of
precipitating this stuff out and if the phosphates do leave, you still
have the sodiums in the solution.
So for what we -- what we seem to see --
DR. KRESS: That's a mystery to me. The phosphate is an
anion and the sodium is presumably a cation. So how is it that the
sodium compensates for the loss of the phosphorous?
MR. SCHEIDER: I'll have my chemist get up for a minute.
MR. LURIE: I'm Steve Lurie. I represent the CE Owner's
Group Project Office. I work for ABB.
I've been working with Ray on digesting this information and we've
looked at the possibility of a reaction such as precipitation of some of
the tri sodium phosphate from other sources of chemicals, such as
calcium hydroxide, I think has been mentioned, and there may be others.
If something like calcium were to evolve into the post-accident sump and
reacts with the tri sodium phosphate, it will precipitate out the
phosphate ion. So calcium phosphate will precipitate.
What is left behind is sodium and in the case of calcium hydroxide
leaching out, of course, the counter ion from the calcium hydroxide is
hydroxide. So you really don't lose any alkalinity by that type of
precipitation reaction.
Only if the sodium were to somehow be removed from solution
would you then begin to lose alkalinity. Now, the sodium will react
with other acids. It reacts certainly with boric acid.
DR. POWERS: You've lost me completely.
MR. LURIE: It would react with other acids.
DR. POWERS: You've lost me completely. Sodium has no
effect on pH at all.
MR. LURIE: Sodium hydrolyzes in water to form sodium
hydroxide.
DR. POWERS: I see.
MR. LURIE: Only if there is another counter ion, such as
chloride or fluoride or some other acid will it then start to neutralize
and the pH becomes lower.
DR. POWERS: So it's really the hydroxide that you're
worried about, and you've lost all buffering capacity here.
MR. LURIE: Well, you may lose some buffering capacity if
some of the TSP precipitates out, but you won't lose alkalinity, if
that's the concern.
DR. POWERS: What you're concerned about is going acid -- it
goes acid naturally, because you're producing nitric acid all the time
here.
MR. LURIE: Yes. There will be some sources of acidity
which will tend to lower the pH, but I think as Ray has been trying to
point out, all sources of such acidity are not expected to remove the
majority of the TSP that has been placed into the sump by design to
neutralize the boric acid.
DR. POWERS: Can I understand how you came to that
conclusion?
MR. LURIE: I think some of the plants that have looked at
the advanced sources of acids have included that in their recent
calculations. I know plants like Millstone and perhaps others have done
some very recent calculations for sizing their tri sodium phosphate
content.
I don't have any particular numbers at hand, but I don't
believe those other sources of acids that have been taken into
consideration are anywhere near the magnitude compared to the amount of
boric acid that you have to design for neutralizing post-accident sump.
DR. POWERS: It seems to me I recall the ORNL report, if not
the specific one, one of its companion documents, speaking of literally
tons of cable insulation that could be radiolytically destroyed to
release HCL.
MR. LURIE: It could be that there is that quantity
available, but then the question becomes, over time, how much of that
really decomposes and how much acidity does that really create over the
long term and what does that mean in terms of the final pH in the sump.
DR. POWERS: And what did you conclude? You said, gee, it
can't possibly all come out at once, so I don't need to worry about it,
or what did you do?
MR. SCHEIDER: The instances we've looked at it, we know --
we looked at it for System-80 Plus in the design and it had an impact on
the pH, but only to move it slightly into the high R-6 category. So
it's not that the effect is not there, it's just that it's not
overwhelming based on the cases we've looked at.
But either way, we don't view it as an instrumentation
issue. It's part of the design process.
DR. KRESS: Do you have a model for the rate of
decomposition of cable and production of hydrochloric acid?
MR. SCHEIDER: I had the results of that analysis. I think
that I didn't do the analysis, per se. So I know what the conclusion
was. I don't -- I don't know the details of the specific modeling.
But they did consider timing, where you'd have substantial
decay of the -- and at that point, you have substantial decay of the
iodine, so the impact would be negligible beyond that, if there is
anything beyond that.
MR. LURIE: And needless to say, we don't consider this an
issue for PASS relaxation, since, as Ray has pointed out, this has
already been granted previously.
MR. SCHEIDER: But we felt that they were points that we
felt it was worthwhile for us to explore a little more.
DR. KRESS: Please continue.
DR. POWERS: I would just comment that I remain perplexed
about how we come to conclusions that the curve is adequate and the
acidity is controlled here and what happens as we move on in the days
following an accident.
DR. KRESS: I share your perplexity.
MR. SCHEIDER: In terms of the hydrogen, containment
hydrogen monitors, the hydrogen monitors were used for basically two
purposes; atmosphere monitoring for combustibility and the potential for
use in a CDA. What we've recommended here is basically replacing the
reg guide, the sample itself, which is delayed and slow, to analyze with
a continuous monitoring system, based on the Reg Guide 1.97 monitors.
It's real-time, adequate for the assessment of
combustibility and core damage assessment, and we rely on the monitors
in the severe accident management guidance, and, again, this was also
previously granted in 1993.
MR. BARTON: How reliable are those monitors and what is the
maintenance history on maintaining them operable?
MR. SCHEIDER: They're Class 1E. They have to be operable.
There doesn't seem to be any major issue with the maintenance and they
are covered in the tech specs. So the plant's willing to maintain
operability of those.
DR. KRESS: How many of those monitors typically are in
containment?
MR. SCHEIDER: Redundant, the two. They're probably not in
containment. They may take a sample from containment.
DR. KRESS: Sampling.
MR. BARTON: They draw out of containment, but they're
outside.
DR. KRESS: Could that line also be used to monitor gamma
radiation if it's a sampling line that's continuous?
MR. SCHEIDER: I wouldn't know what's involved.
DR. KRESS: It's just a speculation.
MR. SIEBER: I'm not sure what that would tell you, because
you would have all the plate-out and everything going on in the
sampling.
DR. KRESS: You may have a lot of plate-out, which, by the
way, can be accounted for by differential readings.
MR. SIEBER: Right.
MR. SCHEIDER: The reactor coolant system boron is
essentially intended to assess return to criticality. Right now, the
wording in the EOP is as the last resort action to establish this
capability. The reliance is primarily on the negative startup rate
monitors, the Class 1E neutron flux monitors.
And for the events, the beyond design basis events, that are
the more common, like boron dilution due to steam generator tube
ruptures and stuff like that, they have specific procedures in the EOPs
that are conservative and provide guidance as to controlling boron
dilution.
So we're recommending the removal of the need for the
sample. It doesn't respond in the time of accident mitigation and is
backed up by other equipment.
The last item is the radioisotopes. The original purpose of
the radioisotope assessment, the only one we were able to actually
locate based on NUREG-737 was in core damage assessment. We're
recommending that that -- the radionuclide capability be removed, delete
the sample, and replace it with increasing or providing field teams with
I-131 monitoring capability.
DR. KRESS: That's already outside of containment then?
MR. SCHEIDER: Right. It's to assess where the radiation
has gone and what the levels are outside of containment.
MR. BARTON: They're out in the countryside someplace doing
this.
MR. SCHEIDER: Right. There is no way, even if you sample
it inside containment, that you're changing where the radiation is
going. The thing is that you do want to know regardless where it has
gone and you want to be able to deal with that.
We already have field teams going out. What we're going to
commit to is make sure that they have the 131 sampling capability with
the teams.
DR. KRESS: Do the CEOG plants have any capability to vent
the containment?
MR. SCHEIDER: There may be a few that have some gerry-rig
capability.
DR. KRESS: It would be gerry-rigged.
MR. SCHEIDER: There is one that has a purge that could do
it. It is possible for some of the plants. Otherwise, the lines are
kind of convoluted for others.
MR. BARTON: Do the EOPs ever direct the operators to vent
containment?
MR. SCHEIDER: EOPs have -- okay. When you deal with EOPs,
you're dealing with events that have a certain level of radiation
releases, and the issue there is, no, they don't. You wouldn't consider
venting a containment unless the thing is basically on the verge of
failing or you expect a catastrophic failure of the containment. That's
well outside the space of the EOPs.
Severe accident management has a challa to vent containment,
with instructions to spray down the containment, to remove fission
products from the atmosphere, and guidance as to what the pros and cons
would be. But the assumptions would be to assume that high levels of
radiation previously exist in the containment and not to basically
underestimate the consequences of it, but to take definite actions,
spray down to make sure you remove all the particulates you can remove,
and do very short duration control vents.
But you would monitor the consequences of that with the
field team, as well, but that's a last resort action basically to
prevent containment total rupture.
MR. SIEBER: There's a fair amount of uncertainty associated
with using field monitoring as a way to determine whether or not there's
been a release or what the size and extent of that release is. Is that
not correct?
MR. BARTON: You've got to be right in the plume at the
right time.
MR. SIEBER: Right.
MR. SCHEIDER: You do know, regardless, that -- well, one,
you're trying to identify the location of the plume and, again,
monitoring -- you know you have radiation in containment. You know the
levels are high. And you know that you're leaking from the containment.
The goal here is to try to find out where the radiation is going.
So it doesn't really help you to even pinpoint and say, oh,
radiation level is high, my monitors in containment are reading
off-scale, then, yes, you have a potential for high levels of iodine
releases.
What you really -- sorry.
MR. SIEBER: The potential for release is determined by the
radiation monitors in containment, right? And that's good enough from
the standpoint of coming up with a PRA protective action recommendation,
PAR, without considering iodine.
MR. SCHEIDER: Right, right. The other issues, plus, if you
have -- there are a number of radio -- of very serious events that
basically bypass the containment where the radiation goes directly out
without having to be filtered by the containment structure itself.
And those you totally rely on iodine-131 monitoring in the
field. So like for the team generator tube ruptures and the IS LOCAs,
which admittedly are not the most high frequency events, but are clearly
the most significant events, you're totally relying on this, because
PASS does nothing for you.
PASS only -- if it does anything at all, all it's going to
be able to do is tell you that there's some iodine residing in the
atmosphere and the assumption would be, with the high levels of
radiation, you'd expect some iodine residing in the atmosphere.
And the actions you would take, basically spraying down the
containment, the containment atmosphere and taking the -- and using your
filters as much as you can are basically the same actions that you would
take regardless of what the levels were.
DR. KRESS: On what basis do you claim that high radiation
automatically equates to some iodine?
MR. SCHEIDER: There are two pieces to the puzzle. You
can't look at one instrument by itself. You have the structure of the
transient. You know that you've uncovered the core for a finite amount
of time and assume that you're dealing with something like a half-hour.
DR. KRESS: So you have the thermocouple readings and you
have the high radiation and you have the hydrogen, put that together,
then you can say you can pretty well infer there's iodine.
MR. SCHEIDER: Right. And there's a rough correlation
between radiation release and the level of core damage, and the level of
core damage could be approximately correlated to the amount of iodine
that's been released.
And all the procedures are to basically determine where the
iodine is going, understand what the risk is to the public, and take the
appropriate actions, and all of those procedures and processes are in
place and are being met.
DR. KRESS: I'm sure you can infer that iodine got into the
containment by looking at hydrogen plus high radiation plus
thermocouples. How about an hour later, two hours later? You still
have high radiation. Can you say anything about the content of that
radiation due to iodine then?
MR. SCHEIDER: The action should still be to periodically
spray down the containment to remove airborne particulates and any of
the elemental iodine. Whether or not you actually precisely know what's
in the atmosphere is not as important as knowing is the radiation -- is
iodine leaking out of the containment in some way, because you might
have other sources of radiation releases that you're not anticipating.
You still have to rely on the fact that you could be leaking through
steam generator tubes, you could be leaking through leaky main steam
safeties, you could be having some leaks through the aux building,
through some other piping. You've got to count on your iodine-131.
No matter what you're doing, the field team monitors have to
basically protect the public. Other information is irrelevant in this
process. You don't want to focus on a piece of data which basically
pre-prescribes what you think might happen. That's your concern, right?
But the issue is that you have to basically make sure that
the stuff that's getting out to the public is known, you've got to know
where it is, you've got to know when it is, and you've got to know how
much it is.
So I think the only other item we probably covered in the
discussion is that we believe that the current sampling systems, at
least, when you basically have the -- when you basically do the sample,
because of deposition and plate-out, you're not going to be picking up a
lot of the iodine leading to possibly a false sense of security.
Also, if you count -- prior -- when you talk about venting,
prior to venting, if there's a lot of iodine in the pool and you do a
measurement, you may not pick a lot of it up in the atmosphere, but the
moment you vent, you're going to depressurize it and re-volatilize it
again.
So you've got to be real careful what decisions you're
actually making and you've got to make sure that it's not just based on
an instrument that's relatively unreliable and gives snapshots in time.
Also, there is a concern, the general one, that you're
sending people to get doses over and above what they normally would be
getting by just -- you still have to send your field teams out. They're
going to get dosed regardless.
But the team that has to basically suit up and collect
samples, if you're not going to use the information, that we believe
have no attendant benefit, it makes no sense to do. You have a limited
number of resources. You're diverting them to that area.
DR. KRESS: Could you give the committee a feel for if you
did do the PASS sampling, say, for radionuclides, how frequent of a
sample -- how often would you have a data point?
MR. SCHEIDER: It would vary on the utility. The way the
regulation reads is you have to take a sample three hours from the time
you determine it's needed.
DR. KRESS: That's the starting point.
MR. SCHEIDER: But it's three hours from the time you
determine you want to take it, which may be some time down, and then --
DR. KRESS: Once you've started taking it, how often could
you get a sample?
MR. SCHEIDER: It would depend on the system. The grab
sample systems would take, my guess is, more than an hour per sample.
DR. KRESS: Per sample.
MR. SCHEIDER: That's probably -- do you have a better feel
for it?
MS. PARTRIDGE: I'm ReNae Partridge, from ANO Chemistry,
Superintendent. We have an automatic system, and so it takes about ten
minutes for a flush and then we just do it automatically. So we can do
it over and over again. So we don't have a grab sample system, and I'm
sure it would take much longer for plants that have that.
MR. SCHEIDER: Because the grab sample, you have to send
someone into the room and pull another sample out and do the initial
flushing.
DR. KRESS: The requirements in the regulations don't go to
the extent that you would have to have this automatic system, you could
just have a manual grab.
MR. SCHEIDER: Right.
DR. KRESS: And some of the plants have just a manual.
MR. SCHEIDER: I would say 90 percent of the plants have
manual.
MR. BARTON: Yes, most of them have manual, I believe.
DR. SHACK: There's two here that have the in-line,
apparently, from the table.
MR. SCHEIDER: So the -- okay. And in essence, the
information that you would get isn't being used in any part of the
procedures. So I'd like to basically just get to the conclusions. We
believe that the -- well, the industry has matured over the years.
They're equipped with more appropriate accident management tools,
accident management guidance, accident management instrumentation,
better prepared to interpret the information the in-plant
instrumentation is telling us.
We have more of a heightened awareness and understanding of
the severe accident challenges we're going to be confronted with or at
least we could be confronted with.
A review of the PASS system indicates that PASS has not
proven useful in any of these environments. PASS is not required in the
emergency operating procedures. So the operator staff doesn't use it.
PASS is not relied upon in the severe accident management guidance, so
the tech support center isn't going to be using the information.
PASS is not needed to establish the early emergency alert
levels, emergency action levels, nor is it used for the long and short
term protective action recommendations. So the emergency planners
aren't using it.
In the re-entry procedures, PASS is not used for re-entry.
They have other mechanisms for doing that, long-term sampling, with the
normal sampling system. So PASS is not used for the re-entry.
We believe PASS can mislead and misinform, on top of
everything else, and we've done two things in the process to expand the
normal sampling system limits at the lower end to make sure that we have
more capability under the more common potential events and ensuring the
iodine-131 site survey capability, we believe, enhances plant safety.
And the final conclusion is the elimination of PASS, as
discussed here, is acceptable and will not degrade the plant's emergency
preparedness.
There's a number of slides that follow this slide in your
packet and they're basically more detailed discussions of the tabled
information.
DR. KRESS: Are there any questions, before we hear the
staff's side of it?
DR. SHACK: I just have one, coming back to the thing that's
really not on the table, which is the sump pH.
DR. KRESS: Yes.
DR. SHACK: If the buffer fails and the pH goes above, what
do you do? What is your action? Does it make a difference?
DR. KRESS: Well, you're aware that the potential is there
for iodine re-volatilization and you may -- it gives you information
that puts you on alert, anyway. I don't know that you can do anything
about it. You can't go in and --
DR. SHACK: The guys can go right back in there with some
more TSP.
DR. KRESS: You can't go in and throw in some more TSP.
MR. BARTON: Not hardly, Bill.
DR. WALLIS: But you might be more reluctant to vent your
containment or something.
DR. KRESS: You might be reluctant to re-enter, you might be
reluctant to vent, or you might have your emergency response people put
on alert again. It's information and you use it for your
decision-making process. There's not much you can do about it.
MR. BARTON: You use it for long-term recovery.
DR. KRESS: Yes.
MR. BARTON: Because you've got high acid, it's going to
affect a lot of components and stress corrosion cracking.
DR. KRESS: You may be able to figure out ways to -- you may
want to turn the sprays back on it. You may want to turn the sprays
back on it.
MR. BARTON: It goes into the decision of rebuilding the
plant or TMI-2'ing it.
MR. SCHEIDER: The procedures themselves, the severe
accident guidance would naturally send you to turning sprays on,
regardless, because you're always going to want to continually wash the
atmosphere.
DR. SHACK: I would have thought, yes.
DR. POWERS: But the problem is if you start recirculating
from the sump, your sprays, all you're doing is continuously
recontaminating the atmosphere.
DR. KRESS: So it may not be very effective.
DR. POWERS: As soon as that -- you switch to recycle mode,
spray is fine for aerosol particles, but it's no good for iodine.
DR. KRESS: It just helps put the iodine back into
containment, that's about all there is to it.
DR. POWERS: It makes a real good mass transfer.
DR. KRESS: Yes, it's a real good mass transfer.
DR. POWERS: Now, in fairness, you have to remember there is
a school of thought, which I think I'm actually a member of, that says,
no, we've got it wrong on this partitioning business, that what happens
is you're not putting up iodine vapors into the atmosphere, because
what's happening is you're partitioning the vapors out, they're reacting
with the ozone and the atmosphere and creating very tiny little aerosol
particles.
And the partitioning acts as a pump, if we're right about
that, then the sprays could be effective in that case, except,
unfortunately, the particles are so small, they're down in the minimum
decontamination zone from sprays.
DR. KRESS: They're down in that valley.
DR. POWERS: They're in the valley.
DR. KRESS: Okay. With that, are there other comments?
Otherwise, we'll just turn it over to the staff for their presentation.
We're running about 15 to 20 minutes behind.
DR. POWERS: So all they'll do is eat into our lunch hour,
and we wouldn't that against them, would we?
DR. KRESS: Not at all.
DR. POWERS: I didn't think so.
DR. KRESS: Not at all.
MR. O'BRIEN: The last two times I was up here, the battery
went out. So I'm expecting that.
DR. POWERS: Well, I expect that to happen, too. We told
you to bring your own batteries from now on.
MR. O'BRIEN: Good morning. This is the staff's
presentation of its review of the CE topical report on justification for
elimination of PASS system.
Chris and Mike, we've got an extra chair or two, you guys
come on up.
The PASS system has many different samples being taken which
really fall into different areas of the NRR organization, so we had
several reviewers on this task, including, from the Materials and
Chemical Engineering Branch, Chris Parczewski. Come on up, Chris, we've
got a chair over here, too.
From the PSA Branch, we have Mike Snodderly; Reactor Systems Branch,
Lambros Lois; and me from the Emergency Preparedness and Health Physics
Section.
And added to our team and really a member from way back
when, as Ray mentioned, there --
[Loudspeaker interruption.]
DR. KRESS: That always happens when you're here.
MR. O'BRIEN: And as I say, really not a new member, because
Rick Hasselberg, from the Incident Response Center, was involved in
looking at this issue when NRR was putting together its generic letter
to grant some of the relaxations that are actually being requested by
these owners' groups.
His presence at this meeting is in particular because of
questions that were asked related to the RRC, the last presentation.
This follows along the familiar path of my presentations
previously. I won't spend too much time on items that have been hit or
discussed by the CE owner's group already, but just a little bit on the
background. NUREG-0737 is really where the PASS system criteria or
requirements, as you may want to call them, were initiated.
And the only thing I want to highlight there in this point
is that licensees were required to have a capability to take these
samples. There were certain samples that they had to be able to
analyze, but it was the capability that they were required to have.
This capability included that they be able to obtain and
analyze these samples within three hours of the decision to do so.
DR. WALLIS: Actually, it didn't say they've got to take
samples. It said they had to provide measurements or quantification of,
and there could have been maybe more methods to do it, conceivably.
Just as a clarifier.
MR. O'BRIEN: Okay. The issue there is, the capability I
wanted to point, the second thing, is three hours after the decision to
do so. So the concerns about exposure of people would be taken into
account in any decision to take these samples, based upon the need of
it.
DR. WALLIS: It seems to me there might have been
intelligent ways to devise a system which would make some of these
measurements without the necessity of exposing people. So some of these
arguments are not very impressive, the reasons for doing away with the
system.
MR. O'BRIEN: Yes.
DR. WALLIS: If you can argue that the function is not
necessary, fine.
MR. O'BRIEN: Yes. And I think we'll be hitting that on a
number of the samples, and there's a couple of the samples which we are
still working on.
Reg Guide 1.97 also discusses taking samples, and, actually,
it does discuss samples, and it includes -- the NUREG-0737 did not
include samples from the containment sump. The Reg Guide 1.97 did and
it also discussed measurement of pH.
This slide I won't spend much time on, because it has been
discussed by the CE owner's group. It just says some more background,
NUREG-CR-4330, a contractor report, didn't look at the PASS system,
along with other requirements that were considered to perhaps have
marginal importance to safety and made its determinations, as was
discussed by the CE owner's group.
This issue was also looked at when we were doing advanced
plants and some relaxations to the timing criteria and the sample
criteria were provided. In particular, extension of boron samples,
date, hours, dissolved oxygen samples for 24 hours, radionuclides to 24
hours, and the elimination of the chloride samples.
As you're aware, Westinghouse came in with a similar topical
report. The difference between the Westinghouse topical report and the
CE topical report is Westinghouse stated that or justified keeping the
system, but taking it out of the licensing basis, which we then reviewed
as if it was out of the plant, based upon how we look at -- we have no
control and they can make changes as they please.
CE was direct in saying that they were planning on not using
the system at all and provided that justification.
Now I'm going to go through each of the samples that are
being justified to be eliminated by CE and provide our staff positions
on those. The first ones are the ones that are less controversial, I'd
say, the ones that we are pretty confident were acceptable.
DR. KRESS: if this proposal is approved, is it likely that
Westinghouse will come back and say, wait, we didn't mean it, we decided
we don't want to keep our system either?
MR. O'BRIEN: I meant to bring up this point as far as the
approval of these topical reports does not change --
DR. KRESS: Doesn't change --
MR. O'BRIEN: -- anybody's plan at all. Then they'll have
to come in for a license amendment, which references the topical report
and expedites our review and approval. So even though CE has a safety
evaluation on their topical report considering the containment sump pH,
that's not necessarily an approval of any given plant.
DR. KRESS: It's a good first step, though.
MR. O'BRIEN: It's a very good first step, but I think when
you say the issue is not on the table, it really is on the table,
because we're going to be addressing it as far as this topical report
and any future licensing actions. So determinations made on this issue
would affect actions taken on individual plant applications.
DR. WALLIS: This is an odd one, because the RIVLS tells you
whether or not you've got voids in the vessel. These voids could be
full of steam, they could be full of something else. If they have
hydrogen in them, there's no way that measuring dissolved gases tells
you how much void you've got.
And if the purpose was to know whether you've got a void a
vessel, then measuring dissolved gas content doesn't tell you that. It
could be it tells you if it's saturated or not, but beyond that, it
doesn't tell you how much of it came out.
So I guess you're arguing that the real purpose was to
determine if there might be voids in the vessel. Now you have a much
more direct way of measuring voids in the vessel, you don't need the
sample.
MR. O'BRIEN: That's correct.
MR. LOIS: Lambros Lois, Reactor Systems Branch. Yes. From
the systems point of view, the thing is to try to avoid uncovering the
core by suppressing the water level and RIVLS test that one, and, at the
same time, give you a way to mitigate the situation by letting whatever
is in there, condensables, steam, or whatever else.
DR. WALLIS: The purpose of this requirement was to know if
you might have voids in the vessel. It was not for some other purpose.
If you've read the statement of considerations for measuring RCS
hydrogen, it was not for some other purpose.
MR. O'BRIEN: I believe that's so, and, also, for
interruption of cooling.
MR. LOIS: Yes. The other consideration, again, has to do
with the amount of void. If you eventually have to resort to natural
circulation, you don't want voids in the upper part of the core, which
also can propagate in the upper part of the U-tubes in the steam
generators.
Therefore, on both counts, you try to avoid having voids in
the upper part of the core or the vessel.
MR. O'BRIEN: All right. The next issue is a chemistry
issue reviewed by Chris Parczewski.
DR. WALLIS: Excuse me. This is certainly not important for
accident management. I mean, if you wanted to know, we've gone into a
situation where we may want to depressurize the system and
depressurizing it would bring gases out of solution. You might want to
know how much you have in solution. Do you want to assess what would
happen if you did statement? I'm trying to think of why you might want
to know how much dissolved gas you have, other than just knowing that
you might now have voids, you might know that you might create them by
doing something.
MR. LOIS: Indeed.
DR. WALLIS: It's changing the temperature or the pressure
or something.
MR. LOIS: Depressurization is one of the major concerns,
because you have it pressurized going --
DR. WALLIS: The RIVLS won't tell you that, will it?
MR. LOIS: It will tell you a level.
DR. WALLIS: But it won't tell you that if you depressurize,
you will make voids. It won't tell you that.
MR. LOIS: No. But it will tell you when you have them --
DR. WALLIS: It will tell you after you've done something.
MR. LOIS: Yes, indeed. Absolutely. Yes.
DR. WALLIS: So you wouldn't want to know the dissolved gas
content for that sort of accident management.
MR. LOIS: The EOPs don't count on that. In other words,
they don't call on that ahead of time.
MR. O'BRIEN: RCS oxygen, conductivity, chlorides in the RCS
and containment sump were justified to be eliminated by CE and we've
found that to be acceptable, and the rationale for that is information
used to evaluate -- the information is used to evaluate the potential
for stress corrosion cracking of stainless steel components. The
chloride concentration can be fairly accurately predicted from
concentrations of chlorides in incoming water and the only function of
--
DR. WALLIS: What's the incoming water? Excuse me.
MR. O'BRIEN: The way that the requirement was put into
place had to do with cooling water systems.
DR. WALLIS: This is in the RCS. It's not in the
containment. Yes, all right. I agree.
MR. O'BRIEN: And the only function of conductivity
measurement is to confirm the other analysis, and it was never required
by 0737 or Reg Guide 1.97, but some plants have that, so it was
addressed in their report.
DR. WALLIS: So you're not addressing the containment sump
chlorides. There are two things. There's RCS chlorides and containment
sump chlorides.
MR. O'BRIEN: Yes. We're addressing the containment sump
chlorides, also.
DR. WALLIS: Where is that addressed?
MR. O'BRIEN: It is in the same place. It will also be --
Chris?
DR. WALLIS: That's not incoming water. I thought there was
insulation you could dissolve and things like that.
MR. PARCZEWSKI: What we are thinking is the water coming
from the borated water storage tank, which would -- if you have some
contaminant over there, the contaminant would be, of course, brought
through the RCS, the injection.
DR. WALLIS: There's no chlorine in insulation and things
which is going to be eaten up?
MR. PARCZEWSKI: There might be some.
DR. POWERS: There is a huge amount of chlorine in these
insulations. They're polyvinyl chloride, for heaven's sakes. They must
run about 12 percent by weight chlorine.
DR. WALLIS: So it can't be estimated from just knowing
what's in incoming water.
MR. PARCZEWSKI: You have to have additional source
activities, yes, but really the only action of chloride would be stress
corrosion cracking.
DR. WALLIS: So your argument is you don't need the
function, not that -- right?
MR. PARCZEWSKI: Yes.
DR. WALLIS: You don't need to know the containment sump
chlorides.
MR. PARCZEWSKI: Precisely, exactly.
DR. POWERS: I think it's important to understand that there
are things that chlorine does besides stress corrosion cracking. It
will affect the radiation chemistry of iodine, which plays -- it's an
overwhelming amount compared to the iodine, and it will cut down on the
amount of peroxide that you have available in that solution, because
it's reacting with the peroxide, and peroxide is what keeps your iodine
from -- your I-2 in solution gets pushed back to I-minus, because of
reactions with the peroxide.
DR. KRESS: So it impacts this pH issue.
DR. POWERS: Yes, yes. And which direction it goes depends
a lot on where you are.
MR. O'BRIEN: Is this a separate effect from pH?
DR. POWERS: Yes.
MR. O'BRIEN: If you measure -- this would measure both
adequate to --
DR. POWERS: Don't get me wrong. I don't think there is any
reason to be measuring chlorides to understand the radiation chemistry.
I think that's one of the things the radiation chemists have to
understand is they're going to have chlorides they have to deal with and
measuring them is not a high yield activity.
I just point out that it has other things besides the stress
corrosion cracking, but, again, I don't think there is any need to
measure it.
If you said, gee, let's measure the chloride, then I'd say,
oh, well, if you're going to measure that, great, let's measure the
ozone, that's far more important to me, and if we're going to measure
the ozone, gee, let's measure the nitrous oxides.
DR. KRESS: You're going to have a whole list of things.
DR. POWERS: There's a long list of things before I hit
chloride.
DR. WALLIS: Well, let's discuss that. You're saying that
there was a requirement to measure containment sump chlorides, which
really wasn't necessary. In light of what we know today, was there
something else which is more important to measure? Doing away with
measurements which now you think are not important, but maybe in light
of extra wisdom, there might be something which is important to measure.
MR. PARCZEWSKI: You mean in the sump.
DR. WALLIS: Trying to cover the scope of your inquiry.
When you said we don't need to make these samples, take these samples,
you're doing away with some things, but probably the original intent was
to measure anything you need to know which might help you in making
decisions, and, at that time, they thought that chlorides might be
important.
MR. PARCZEWSKI: Yes. But our main is the fact that this
only really -- the main purpose, the main action of the chlorine or
fluoride will be the --
DR. WALLIS: I'm saying that nothing else -- did your review
consider what other things which -- other than the things we now know we
don't need to measure, might be more important to measure, and how do
you decide that issue?
MR. O'BRIEN: That's a good point.
DR. WALLIS: Did you do that?
MR. O'BRIEN: Well, I think that as I brought up in the very
first, it's a capability. There's capabilities in being able to take
the sample and then there's capabilities in being able to measure
certain parameters.
And what is being addressed at least in this specific item
is eliminating the capability to measure these items.
DR. WALLIS: I think it would be reassuring if you could
also tell us that there is nothing else which really needs to be
measured for which we need to sample, in the light of what we know
today.
MR. O'BRIEN: Yes. I think that -- well, at least the
question then becomes, I think, in the minimum, if we keep the
capability of sample, that's one thing. If we eliminate the capability
of sample, then we have to address everything, I think.
If we keep the capability of sample, is there extra work
right now to say, well, what else do we need to look at that we might
not have thought of.
DR. WALLIS: I thought that they proposed, because it's so
expensive to maintain these pipes and valves and rupture disks and
stuff, to do away with the whole thing.
MR. O'BRIEN: That is correct, but we have not necessarily
approved that at this point. They're still looking into it. And one of
the things as far as the cost that I thought was interesting and we
probably need to do more work is everything that was mentioned had to do
with a high pressure kind of system. There's containment sump and
containment atmosphere won't be high pressure systems. The RCS
obviously, if it's designed to take pressurized samples, will be, but
the other two -- maybe you could address that, if there is the cost
aspect of that.
Is it that the lower pressure systems are also as costly?
MR. BICE: This is Dave Bice, again. The lower pressure
systems are still going to keep us maintaining the same structural
design we've got now, unless we did an entire new redesign, which, of
course, would be costly.
But let's keep in mind, what we came up was, first, the year
2000 approaching, obsolescence issues, where we couldn't replace
equipment anymore with older models of equipment, we would have to
address the obsolescence issues, and the constant maintenance costs.
At that point, when we saw the cost it was going to cost the
utilities, we said, well, let's see what the design basis is for this
and then looking at the design, we said, well, aren't we meeting that
design through other methods.
So really the issue that came out here in these meeting
isn't the cost. It's the fact that it's really not needed anymore at
all.
So that's the issue we're pursuing.
MR. O'BRIEN: Okay. The next sample is RCS boron and
containment sump boron and Lambros, again, looked at that item. We
determined it was acceptable to eliminate from PASS. Boron sampling --
these samples are used as a backup to other indicators of criticality
concerns.
Do you want to add anymore on that, Lambros?
MR. LOIS: Not really from what you said. We do have other
means. We know -- boron is really not necessary.
DR. WALLIS: Does containment sump boron ensure that there is no
re-criticality after the arrangement of this core into a molten form
that spills on the floor?
MR. LOIS: Yes. Well, in that case, once you go past about
five percent damage to the core, when the core is no longer -- the
integrity has been jeopardized, there are two major problems. One of
them is the core has fallen and if you don't have -- do not maintain
geometry, the criticality issue essentially disappears, because --
DR. WALLIS: All other geometries are worse for maintaining
criticality.
MR. LOIS: Exactly. Precisely.
DR. KRESS: Pretty much so, unless you can maintain it
spherical.
MR. LOIS: Right. In addition to which you will melt the
control elements or the control rods, which will fall -- this has been
proven by -- it then becomes decay heat, but that's a different issue
than control.
DR. KRESS: That's probably the best thing that could happen to the
core, because it would go re-critical. It's much more cool. It's not
going to have enough energetics to release it.
MR. O'BRIEN: Next is containment atmosphere hydrogen and
containment atmosphere oxygen, which Mike Snodderly reviewed for us, the
DSSA, and we determined it was acceptable to eliminate from PASS. The
rationale for that is that there is real-time indication available
through the monitor and oxygen samples are not required for PWRs.
I'm not sure why we put that in there. I think it was
addressed in the CE report and, just for completeness, we addressed it
here.
MR. SNODDERLY: Also because oxygen sampling is more
important for a BWR. So this logic that we use for a P would not
necessarily apply to the B.
DR. WALLIS: Sample to know if you had a fire or something
that's eaten it? Why would you sample oxygen? There's so much oxygen
there already.
MR. SNODDERLY: No. Remember, BWRs are inert.
DR. WALLIS: Okay.
MR. SNODDERLY: So we acknowledged -- that's how we do
hydrogen control for the B's. Basically, Jim, sorry, the basis for the
conclusion was that if you have a gas chromatograph, it's nice to have
as a backup or a confirmatory of the continuous monitors, but the
continuous monitors are the preferred method. So, therefore, if there
is no other need for it, if they're there, that's great, but if they're
not, it's sufficient to support the EOPs and the severe accident
management guidelines.
MR. O'BRIEN: Those are the samples and the conclusions that
we have reached on those ones that we feel are justified for
elimination.
Now are some that are perhaps more difficult and we put this
in this order to get to these at the end, not because it's nearer to
lunch and you're hungrier, but to focus perhaps more on it.
The samples that really we were focusing on and we are
focusing on as a staff is the pH samples in the containment sump and the
radionuclides, and first we're going to address the pH sample.
Chris Parczewski and myself, we met with --
DR. WALLIS: Why do you have to sample to measure pH?
MR. O'BRIEN: Excuse me?
DR. WALLIS: Why do you have to sample to measure pH?
MR. PARCZEWSKI: The pH is measured by grab sample. There
is no instrumentation for continuous pH measurement.
DR. WALLIS: There is no instrumentation for continuous pH
measurement.
MR. PARCZEWSKI: No.
DR. KRESS: Actually, there is in some of them.
DR. WALLIS: I believe there is, isn't there? It's not
installed.
DR. KRESS: It's not installed.
DR. SHACK: You can buy such instrumentation.
MR. O'BRIEN: Okay. We met with Drs. Powers and Kress
yesterday. The ACRS committee, after the review of the Westinghouse
topical, issued a letter that we reviewed and responded to, and now we
have a new letter from the committee and we are currently working to
address those concerns and that's one of the purposes of our discussion
yesterday, was to get amplification on the concerns.
At this point, what we have as status is that we are
continuing to evaluate those concerns.
I'm not running the show, but --
MR. BICE: Dave Bice, again. Just for an added input from
the utility perspective here on the pH issue, there's been some great
discussion brought up about it. They were all good ideas and points
brought up.
How they relate to PASS we're still not sure, from the
utility aspects. It almost seems like a separate issue from this
format.
In addition, I think we've all agreed that it's a long-term
issue, which, again, gives us different alternatives out there. If we
can wait until radioactive decay levels get down farther, a week, a year
even, if we're getting into cleanup phases, we don't have to necessarily
maintain this PASS system in order to assess that pH, if you decide that
there needs to be a pH.
But, again, it's almost a design basis issue, in our mind.
We're almost questioning, well, it's a TSP we've got in there now, we
screw up on the design basis. Back when the NRC approved this many
pounds of TSP for Arkansas Nuclear 1, Unit 2, PASS is really not part of
those questions that we're bringing up.
They may be good questions, but the utility would like to
see, just as you continue to think about it, Jim, and the ACRS staff,
that this issue be separated away from PASS and just look at the PASS
purpose for ourselves, especially in the accident mitigation phase,
which really you won't really see pH as a factor there.
Thanks.
MR. O'BRIEN: The last sample type, I guess I'll say, is
radionuclides from the reactor coolant system, containment atmosphere
and containment sump. As we described in the subcommittee meeting, that
we are considering options for granting, I guess, in part, or perhaps
all of CE's request, we put together four different options, and I'll
discuss them at the next slide.
But the issues that we're looking at related to this sample
is how the information is used and how it may be used. We see use for
it in input to dose assessment codes to determine the potential off-site
consequences of an accident, and then these off-site consequences lead
into protective action recommendations.
We recognize that the initial protective action
recommendation will always be, or we hope it is, on plant conditions and
not based upon an analysis of the dose consequences. However,
subsequent to the initial protective action, what the plant does is
continue to assess the potential consequences to the public and the
environment and that this is one of the inputs which could be provided
to a dose assessment to determine its impact; in particular, the amount
of iodine that could be released.
We have identified, of course, as has been discussed here,
there's a lot of limitations with PASS samples, including the timeliness
of the samples. It will take three hours to take it, up to three hours.
It may be shorter for some sites.
Accuracy of the samples, the issues related to plate-out of
iodine. I guess this really is a limitation, but there are other
indicators which we could use, including containment rad monitor, core
damage assessments based upon thermocouples and so forth.
The accident sequences discussed by the CE owner's group,
there's different accident sequences in which the PASS information will
be more or less useful.
We consider perhaps an intact containment with a potential
for breaching of the containment, that that may be one -- during a
severe accident, that may be one of the situations in which the PASS
containment air sample would be very useful to obtain.
The options that we are considering are, one, to simply
reject the proposal and keep the existing criteria. To relax the sample
time, it previously had appeared at least at sample -- the time to take
the sample was what was driving some of the concerns with this. So we
relaxed it to eight hours for sites that can determine radionuclide
concentrations in the field, as was proposed by CE.
The third option is CE's option to eliminate the PASS
criteria and have this radionuclide -- ability to determine
radioisotopes out in the field, using field data, field team data.
The last one is to just simply eliminate the PASS criteria
and put no additional restrictions, don't require these additional
monitoring capabilities.
DR. WALLIS: You are still considering four options and you
want a letter from this committee. If you had made a decision, we could
comment on what we thought about the decision. If you're just looking
at options, we could simply say the staff seems to be doing good work
and hasn't yet made a decision.
MR. O'BRIEN: What we're doing is we want to provide these
options. These are what we consider a spectrum of things that --
DR. WALLIS: It would be hard for us to -- unless you can
tell us which decision you want to make, it's rather hard for us to
evaluate the validity of it.
DR. KRESS: We could tell them which decision we wanted to
make.
DR. WALLIS: Yes. Or we could tell them -- yes. We don't
have all the evidence they have.
MR. O'BRIEN: That's one thing I wanted to bring up. We're
looking for a new piece of evidence, and that's the input from the
off-site emergency response organizations, the states, and we're working
with FEMA and our states programs office to issue a letter that
describes these four options, describes the limitations, describes a --
with PASS, and to see if they have any concerns with any of these
options that it may affect their emergency response.
We want to consider that in our decision-making here.
DR. WALLIS: That came up at the subcommittee meeting.
MR. O'BRIEN: Correct.
MR. SCHEIDER: Ray Scheider. May I interject a couple of
things? Two. As you're aware, we've already been talking to a number
of the state regulators for a number of our utilities and they've
generally agreed that elimination is fine to do.
The concern I have about the list is that when you say relax
sample time, for example, eight hours, my concern is that there is no
criteria as to what you are looking for for the system to do. You're
telling us how it's to behave, but you don't tell us what you want out
of it.
MR. O'BRIEN: I can address that. The reason for the three
hours is that it was -- as quickly as it was considered as capable to
do, given the technology back when it was first created or this criteria
was created. The reason for eight hours that we're considering is
actually more involved than that. We'd be looking for analysis from
plants to say if the situation where you have your severe accidents
bottled up, what is the timeframe in which containment integrity may be
challenged, and that's the time we'd be looking to relax it to.
MR. SCHEIDER: Well, containment integrity for -- so for
each transient, they'll know what containment integrity is threatened,
but the issue of containment integrity is a totally separate issue from
monitoring radiation releases through an intact containment.
If you have a breach of containment integrity, the concern
is you're going to then revolatilize the stuff you have in the sump
anyway and you have to go with what you're -- what the transient has
told you, especially in -- these are relatively still short time-frames
for a relatively ongoing event.
You have the information --
MR. O'BRIEN: So I figure you don't like option two.
MR. SCHEIDER: Well, I think you have ample information to
make the decision already. I guess that's our position.
MR. O'BRIEN: Okay. Any questions? Well, I guess I should
get to the final. I was waiting for some more questions.
MR. BARTON: Get to the bottom line, then we'll give you
some questions.
DR. SHACK: Three and four for the CE owner's group are the
same thing, right? Because if they do, they're all going to put in the
field measurements, is what I understood.
MR. O'BRIEN: They're going to -- well, the idea is what
does that actually involve and does that play in our decision-making
process on our reviews. So, yes, there's a little bit extra there.
DR. BONACA: Option four, that being the abandonment of the
system.
MR. O'BRIEN: Correct.
DR. BONACA: That's a problem.
MR. O'BRIEN: Just in summary, then, as I discussed
previously, the sample that we find acceptable to eliminate from PASS,
the measurement capabilities, is -- I have them listed there, dissolved
gases, oxygen, chlorides, conductivity, boron, and containment
atmosphere hydrogen and oxygen, and chlorides in containment sump.
Under evaluation, we're still reviewing --
DR. WALLIS: That's two words.
MR. O'BRIEN: I made it to the last slide. Good point.
DR. WALLIS: I didn't realize that at first.
MR. O'BRIEN: It's not a different slide. I was here for
other ones, too. Okay. Under evaluation is two words and our
containment sump pH, and the options for relaxing the radionuclide
sampling criteria and as we discussed before, for that sample, we are
going to the state emergency response organizations to get their input,
if they have any concerns with us proceeding in that regard.
DR. KRESS: Are these the state people where the CEOG plants
are located?
MR. O'BRIEN: No, it would be all. See, these decisions
don't really -- we've got to look at it, are the decisions specific to
CE or are the decisions generic that we're making here. So we're doing
it for all.
That's the last slide.
MR. BARTON: Well, it's a week after the accident, things
are settling down a little bit, and you, the NRC person on-site, tells
the utility he wants to draw an RCS sample. It's a CE plant and they've
eliminated their PASS system. And they go to draw an RCS sample from
their normal means and the lines are all plugged up with all the crapped
that happened from the severe accident. How do you get an RCS sample if
you eliminate the PASS system?
DR. BONACA: Effectively, the decision, in fact, would be
that you would not be able to draw a sample.
MR. BARTON: At least Westinghouse kept the system. So it
gave you another PASS to get an RCS sample.
DR. BONACA: I don't know what I need it for, but I would
like to have it.
MR. BARTON: Your long-term management.
MR. O'BRIEN: I think that was one of the things I was
trying to raise. It's a capability --
MR. BARTON: The Senators will ask you what's the RCS, and
it will be embarrassing if you can't tell them.
DR. KRESS: Do you want to have it bad enough to maintain it
and test it and keep it operable all the time?
DR. BONACA: Well, I've got to think about that, Tom.
That's a good question. But the fact is it makes me uneasy to think
that I will not be able to get a sample anymore from RCS or sump.
MR. BARTON: The normal sample path is probably going to get
all plugged up, because it's small piping.
DR. BONACA: I saw the list of the things we don't need it
for, but I'm trying to figure out is there anything else that I would
likely need it for, and I know somebody, if I were in charge of it,
would ask me a question, go and get a sample.
MR. BICE: I'd like to comment on that real quick, from a
utility perspective, knowing my system that we have at ANO-2. It's the
same sample line. If one is clogged up, the other one is clogged up.
Basically, when you penetrate containment, you are going to one single
sample line from the reactor coolant system or wherever and then it
splits up from there and you either go to the normal sample system,
which doesn't contain as much shielding as the PASS, or to the PASS
system, which, or course, has the old 737 shielding requirements on it.
DR. KRESS: That's because you don't want too many lines
going through your containment.
MR. BICE: That's correct. It would be another leak path.
MR. HASSELBERG: May I comment, from the incident response
point of view, please?
DR. KRESS: Yes, please.
MR. HASSELBERG: My name is Rick Hasselberg. I'm with the
incident response organization, that's what they call us at this
particular point in time.
I coordinate response teams. I manage the incident response
training program and I've worked for about 20 years with the Office of
Public Affairs in trying to keep our public affairs officers up to date
on the technology.
I'd like to give you a very short, probably less than five
minutes, hypothetical situation that you can mull over lunch, if you
like.
Here is the situation. You walk out of here at lunchtime.
Somebody grabs you and says upstairs on the fourth floor, we're
monitoring an event. Actually, we're responding probably in the
expanded -- in the initial activation mode. The region is sending a
team and we're in our full glory trying to follow an event that's
occurred over the last four to six hours.
Some background for you. Actually, the chairman has
requested you to come up and give her some assistance in preparing
responses to some questions.
Pressurized water reactor, several ECCS components have
failed as a result of loss of coolant accident. It doesn't really
matter why. Some equipment operated intermittently, but based on core
exit thermocouple readings, reactor pressure vessel water level
indication, source range indication and other installed instruments, we
have a pretty good idea that the reactor core, the top of the core was
uncovered for 90 to 120 minutes, and the level probably dropped down to
one or two feet above the core support plate for maybe 45 or 50 minutes.
So there core has been uncovered for some time. About three
hours ago, the reactor coolant --
DR. WALLIS: The collapse level.
MR. HASSELBERG: Yes, sir. Right. Three hours or so ago, a
reactor coolant pump was restarted. The core was re-flooded.
Subsequent to that, however, there were several loud bangs heard inside
the containment.
The current situation is that a general emergency was
declared early on. A five-mile radius was evacuated, but ten miles
downwind. About 45 percent of the core exit thermocouples are out of
service because they've basically been melted, being exposed to
temperatures between 1800 and 2500 degrees Fahrenheit.
Other thermocouples are reading now because the core has been re-flooded
between 150 and 325 degrees Fahrenheit. As far as we can determine, no
major off-site release has occurred. The containment is isolated, but
not fully depressurized. Containment pressure is about 16 pounds gauge.
Containment radiation is steady at about 6000 rem per hour.
Now, the chairman has asked for your learned experience in
helping her prepare answers to questions such as this, and these
questions were generated in response to the subcommittee's questions
last week.
And I don't know the answers to these either, but I do know
that you're going to want the best possible information from every
possible source. The more data points you have, the better you can
answer questions.
DR. WALLIS: What does the ACRS want to know, too?
MR. HASSELBERG: I couldn't say. I couldn't speak for the
ACRS, but I know that if the chairman knows you're here, she might ask
for your assistance in the preparation of something like this.
Can we demonstrate that there hasn't been an uncoolable
geometry generated? I'm not sure that PASS could or couldn't help with
that. I think probably it could, with the right analysis.
Has the evacuation gone far enough? Based on the material
that's in there that could conceivably further get out, might a further
evacuation be required.
On the other hand, the people who have already been out of
their homes would like to go back in. Which way do you want to go with
this?
The President wants to know if the national laboratories can
be of some assistance in recovering a sample from the reactor core.
Wouldn't it be a good idea if that was possible? And the Board of
Education wants to know how long it's going to be that they reassign
children to other schools outside the emergency planning zone.
Now, we obviously can't generate answers to all these
questions in a few minutes, but the more information that we have, the
better we can fill in the gaps of information. You remember, from Three
Mile Island and Chernobyl both, that when we can't provide the
information or some knowledgeable source provide the information, the
news media consultants will fill in the gaps, or, even worse, the public
interest groups. The various self-interest groups will fill in the gaps
and there will be a huge range in the predictions.
The more we can narrow it down, the more certainty we can
apply and speak -- this is the government, this is what we say, the
better off I think we will all be.
DR. WALLIS: You hope that that reassures people.
MR. HASSELBERG: We have to hope so. We do know that once
you lose your credibility, you don't get it back for a decade.
DR. WALLIS: Even governments call fall, as in the past
Soviet Union, when --
DR. POWERS: The problem that I see with using this as a
vehicle for deciding what kind of instrumentation you want is that every
answer is better with the more information I have and there is no end to
thing and what you would request if you knew absolutely somebody was
going to ask you these questions would be PASS multiplied by factors of
a thousand.
This just doesn't provide a cap on approaching the issue.
The answer is, yes, you would like to do that, but the fact is most of
those questions we could answer just the way the licensee is arguing
now; based on the analyses that he's done and using conventional
modeling tools.
Now, the issues that we come up with in connection with pH
are different, because we know that those things are not in his modeling
tools, and they do affect the answers for the questions at the bottom of
the -- particularly at the bottom of the page, how long are people out
of their homes and do we have to do something peculiar about children
because of their unusual susceptibility to iodine releases.
MR. SCHEIDER: Ray Scheider. If I may, one comment and
there is one concern. When you have the feeling that more data is
always better is not necessarily true if it's not valid data.
The concern that I would have is that if you really are
monitoring the existing systems and you really have the level of aerosol
plate-outs we're expecting and the level of deposition and things we're
expecting, you'll be confronted with a situation where your PASS will be
telling you that you have a few percent iodine being sampled.
And everyone else will be saying I'm not quite sure, based
on the other measurements, that maybe that's not really right, and then
you'll have the people saying I want to go back to my home; if it's only
one percent iodine in containment, why not let me back. And you say,
well, we don't really believe the information; well, you put the system
in, why don't you believe it. Then you're going to have two dissenting
opinions say, well, but we put the system in knowing we really weren't
sure we were going to believe it, but we just wanted to have it in case
we wanted to not know what to do.
I guess I'm a little concerned that you, with the existing
system, the more information you're going to have could actually
undermine the proper actions to take and you've got to be careful how
you plan on using it, given the way it's designed. It's not an idea
world, very difficult stuff to monitor.
DR. WALLIS: Do you mean more ignorance, you'd be better
off?
MR. SCHEIDER: no.
DR. POWERS: Sure, you'd be absolutely positive in your
opinions, if there is no data to contradict you.
DR. WALLIS: At no time they're in any danger, you can make
statements like that.
MR. SCHEIDER: But there is a lot of information that's
available to you. It's not that you're going blind, but the risk of
having information that you have a high degree of confidence is not
going to be valid and the risk of using that to make decisions of having
people return as opposed to having people evacuate I think is kind of
the -- it goes to the end where you're using information now in a
non-conservative way, and you've got to be real careful to make sure
that you want -- that that's the information you want to basically --
DR. WALLIS: Well, I think the question is whether any
questions of this nature have any impact at all on the way we look at
the PASS system. We did raise it. The subcommittee asked this. We
said, well, not the sub-questions that people like this might ask, for
which PASS might be useful. And if the answer is no, then okay.
DR. KRESS: The answer, I think, does fall down to a
question of measurement of radionuclides and the measurement of pH. The
pH is already sort of a moot point for the CEOG, so it boils down to
this bullet on the staff's thing where they're still evaluating the
requirement needs for measurement of fission products.
I think part and parcel of that deliberation is going to be
are the regulations on the books and the resultant installed systems to
meet those regulations very useful to us at all in this respect, and I
think that's going to be part of the decision.
And I think I hear from the industry that -- what I hear
about the PASS design to meet the regulations is not very useful in this
respect. That's what I'm hearing.
MR. O'BRIEN: I just want to point out that we are being
asked to make a decision on containment sump pH as part of our action
here. So we do have to make a decision in that area.
DR. KRESS: Even though it's already --
MR. O'BRIEN: Because there's other plants that --
DR. KRESS: There's other plants.
MR. O'BRIEN: -- could come in and we think it sets the
stage for it. Are there any more questions?
DR. BONACA: No question. Just a comment on that. I agree
with the perspective, in general. And I wasn't at the committee
meeting, but looking at this presentation, I am convinced that PASS is
not needed for all the things that I was told.
But I am left with another question in mind, which is now
that I am supposed to abandon it -- there is another decision being
proposed here, which is abandon it. So in my mind, I keep asking this
question, well, what could I need it for, and possibly the answer would
be nothing.
But it would be nice to have a couple of slides addressing
just that, helping me in the process, and I would say that I haven't
seen it.
DR. KRESS: I guess with that, I'll turn the floor back over
to you, Mr. Chairman.
DR. POWERS: Thank you, and I will promptly recess us until
1:30.
[Whereupon, at 12:32 p.m., the meeting was recessed, to
reconvene at 1:30 p.m., this same day.]. A F T E R N O O N S E S S I O N
[1:30 p.m.]
DR. POWERS: We will come back into session.
Our next subject we're going to deal with is rulemaking and
review plans associated with the development of risk-informed revisions
to 10 CFR Part 50, and Professor Apostolakis is the cognizant member.
I will ask if you could try to carry this meeting forward so
that we can break no later than 2:45.
DR. APOSTOLAKIS: We had a subcommittee meeting last Friday,
in which we discussed the plans that the staff has to implement options
two and three. Several of you were not there and since we are supposed
to finish this early, I have a proposal.
How about if I summarize quickly what the staff told us and
then ask the staff some questions that seem to be of concern to the
members, as stated at the end of the meeting, instead of going through
the formal presentation? Do the members agree?
MR. BARTON: Fine.
DR. APOSTOLAKIS: Unless the staff objects.
MR. NEWBERRY: That's fine. We've got the principals here
to do just that, Mr. Chairman.
DR. APOSTOLAKIS: Who is going to be up here? You might as
well come now.
MR. KING: Mr. Chairman, do you want the handouts anyway,
just for information?
DR. POWERS: Please.
DR. APOSTOLAKIS: Yes. In fact, we may refer to those for
example. The usual suspects are here. Maybe you could identify
yourself quickly, for the record.
MR. REED: I'm Tim Reed, from the Division of Regulatory
Improvement.
MR. BERGMAN: Tom Bergman, from the same division.
MR. KING: Tom King, from the Office of Research.
MR. CHEOK: Mike Cheok, from the PRA Branch in NRR.
DR. APOSTOLAKIS: So if you can put up there the overview
slide three. We see here what the staff is proposing regarding option
two. Basically, they will propose to have a new rule, 10 CFR 50.69,
that will allow the licensees to -- this is for option two, not the new
one -- that will allow the licensees to use the risk-informed
methodology to categorize systems, structures and components, and
Appendix T will list the criteria and the methods to be used that are
acceptable to effect such a categorization.
The staff is proposing to develop a two-by-two matrix, shown
on the next viewgraph -- next, please -- I'm sorry. That will have, on
the horizontal axis, the two categories that already are in the books,
safety-related and non-safety-related components, structures and
systems, and on the vertical axis, the new element, the new
consideration, which will be risk-informed. It is a measure of the
significance, of risk significance of the components.
The staff is currently proposing two categories, high safety
significant, low safety significant. Then, of course, using the methods
of Appendix T, the licensee will place the SSCs into one of these four
boxes.
The staff has not decided yet what the relief from the
special treatment requirements will be for the four categories. That's
correct?
MR. BERGMAN: Yes. We haven't discussed it.
DR. APOSTOLAKIS: You are thinking about it.
MR. BERGMAN: Yes.
DR. APOSTOLAKIS: You are still thinking about it.
DR. POWERS: Why do we pick high safety significant, low
safety significant as two categories? It seems to me that everybody
that goes through these categorizations ends up with four; high safety
significant, middle of the road safety significant, low safety
significant, and has nothing to do with anything and I'm surprised it's
even in the plant significant.
It seems like that happens every time and you run into
problems because low safety significant, when you only have two of them,
low safety significant runs from almost high safety significant down to
I'll be damned if I know why it's even in the plant.
DR. SEALE: It confers status on things that shouldn't have
it.
DR. POWERS: Yes. It does it both ways. It always seems to
be a problem. So I'm just wondering why we stick with it.
MR. CHEOK: Dr. Powers, actually, if you had used three
categories or four categories, you're right, we would have -- it would
have made our writing of Appendix T a lot easier. But right now, we are
sticking with two categories because that's the way the pilot plants are
going and that's the way the industry is suggesting that we go.
DR. POWERS: It seems to me that that -- my past experience
is that that's not the way the industry is going. Certainly, the South
Texas project, if memory serves at all correctly, is one that's come up
with four.
MR. CHEOK: That's correct, but the way they apply it is
that they do only have the high and low. There are no risk significance
in the medium category. It's applied in a similar fashion in the low
category.
DR. APOSTOLAKIS: I was under the impression that they were
proposing so-called targeted special treatments for the medium category
and it was not the full treatment that they thought one would get.
Right? But some sort of relaxation.
MR. BERGMAN: We're starting to intermingle the South Texas
graded QA approach with this.
DR. APOSTOLAKIS: That's special treatment.
MR. BERGMAN: The category that industry is calling the
Commercial T does correspond to our third box there. The T, I believe,
stands for temporary in their approach, meaning between options two and
three.
Under option two, because we must maintain the current
design basis functionality, there will need to be some assurance that,
in fact, the components can perform their function.
Once option three was implemented and the design basis
itself was made risk-informed, the equipment in the third bin there
would come out, basically out of the scope of our treatment regulations.
So it's reflecting the fact that we are looking forward to
an option three scenario. This would not preclude a licensee for
creating additional categories and modifying varying the treatment based
upon safety significance. As much as South Texas has done within the
current QA program, this wouldn't preclude that approach.
DR. APOSTOLAKIS: But why do you have to use only two?
MR. REED: I think it's almost simplicity, more than
anything else, at this point. I think as Tom said, we can accommodate
more categories and grading. I do agree with the comment that it
certainly does give a little bit more too much attention perhaps to box
two, by saying high there, but we are currently wrestling with
terminology. We're still wrestling with terminology.
We understood your comments, heard them loud and clear at
the subcommittee.
MR. BERGMAN: We have, in fact, discussed more than two
categories.
DR. WALLIS: I have a comment about this picture. I thought
you were sort of insightful in saying there were two dimensions here,
deterministic and risk-informed, like X and Y. And I expect to see X
going up, when it goes to the right, and Y going up when it goes up,
but, in fact, it's the other way around. The deterministic is inverted.
I want to see things that are important on two scales of
high on the right; in fact, they're skewed around. So why don't you
have deterministic with two -- well, as many boxes as you like, but
increasing in importance to the right.
MR. BERGMAN: We've already rotated the axis once. We can
flip them, as well.
DR. WALLIS: Yes, but you don't have arrows going both ways.
What you're really saying is, is it important in a deterministic sense
or is it important in a risk-informed sense. This is a measure of
importance on two axes.
MR. BERGMAN: One of the advantages to this construct is
that when you look at it from a treatment standpoint, the equipment in
box one gets more treatment than the equipment in box two, which gets
more treatment than the equipment in box three.
DR. WALLIS: There is a highest level of what you're doing.
You're representing, in a two-dimensional surface, two different
influences or ways of characterizing things or measuring or
representing, the deterministic ways and risk-informed ways. Then you
could have two or four or 25 boxes, but it doesn't matter, you've got
the same pattern.
MR. BERGMAN: We will consider that.
DR. POWERS: Let me ask a question that forever perplexes
me. They've got some risk information about a particular component and
it's not very important during the normal operation of the plant. They
can have it or not have it. The plant runs fine without it. Even if
another equipment fails without it, the plant still runs; not too well,
but it still runs.
Now, the same component is very important during shutdown
operation, because it maintains my heat sink during shutdown operations.
Shutdown operations only operate for a little short period of time.
How do I decide whether to make that a high safety
significant or a low safety significant?
MR. CHEOK: For right now, the plan is that you should model
all initiators and all operating modes in the PRA. For those initiators
and operating modes not modeled in the PRA, it is suggested that you
deliberate in the fashion that for each initiator and each operating
mode, you determine the accident sequences that may happen and determine
that you have at least one success path for each initiator to be kept as
high safety significant.
DR. APOSTOLAKIS: What? That doesn't --
DR. POWERS: Well, it certainly didn't help me.
DR. APOSTOLAKIS: When you use the importance measures,
first of all, is it thought or assumed that the primary risk insight
will come from the importance measures? This is not the determining
factor for the decision, but the risk insight is in RAW and
Fussel-Vesley.
MR. CHEOK: Correct.
DR. APOSTOLAKIS: So the question is now how do you
calculate those, and I think Dr. Powers' comment had to do with when you
say CDF and LERF and then you go and calculate RAW and F-V, are you
considering the average core damage frequency, for example, from all
modes of operation, all configurations, and you create an aggregate that
will include the average amount of time that the plant spends in each
configuration or are you doing several calculations; one for normal
power operations, one for outage periods or maybe several are for outage
periods, and then you find the corresponding importance measures?
What is it that you are doing? Because you gave us a much
more detailed answer. You say you're going to look at each initiating
event, which I doubt.
MR. CHEOK: But the answer I gave you basically is if you do
not have a PRA.
DR. APOSTOLAKIS: Pardon?
MR. CHEOK: If you do not have the PRA.
DR. APOSTOLAKIS: Well, let's say you have the PRA.
MR. CHEOK: If you do have the PRA, at shutdowns, in the
past, what the licensees have done is determine the importance measures
at power and then the importance measures at shutdown, and then they had
looked at the components that were important in both cases and generally
they have stuck with the fact that if it was important either at power
or at shutdown, it would keep it as HSSCs.
DR. APOSTOLAKIS: And you are proposing to do the same or
you haven't decided yet?
MR. CHEOK: We haven't really decided how we're going to
aggregate the shutdown and at power.
DR. APOSTOLAKIS: Or whether you have to aggregate. You may
want to do it separately.
MR. CHEOK: That's correct. That's right.
DR. APOSTOLAKIS: But now, let's pursue this point a little
more. Let's say you're doing power operations and you know that for
fires, for example, the methodology is not as well developed as the
standard event tree with internal events. So fire risk assessment may
have some conservative assumptions to be able to be done. Otherwise,
you can't do it.
Similarly, if you look at the recovery actions, again, the
methodology is not all there, so people may be making some conservative
assumptions regarding the ability of the plant operators to recover from
some unfortunate incident.
So now we put everything together. So there's a number of
PRA terms in the CDF and some terms contain this particular system,
other terms have recovery actions under certain conditions, other terms
are fire risk assessment produced and so on.
If I find now the importance measure of this system, this
will be affected by all the conservatisms that I have put in these other
terms that probably have nothing to do with the system.
Now, people hear the word conservative and say that's great
that this is conservative. Well, no, because these other terms are
higher than they should be, the importance of the system now is smaller
than it should be. So it's a non-conservative result.
So the question is now what do you do? What do you do? In
other words, can we conclude from this that because the PRA tools are
not developed to a satisfactory level in some of these other areas, the
whole thing collapses? Because the importance measure has to take into
account everything else.
Or you might say, give you a way out, I'm going to break up
the CDF into pieces and say I'm going to do one for internal events, one
for seismic, one for fires. So at least I will have some level of
internal consistency and then the expert panel will have to resolve the
differences.
But that will multiply the effort here tremendously. And
then on top of everything else, you have the oversight process that says
this agency worries about initiating events. It goes even down to more
detail. So what do I do now? Do I consider initiating events alone and
decide that this system is important because it contributes -- may
contribute to some initiator, whereas if I put it as part of the bigger
PRA, it may not be as important. I don't know.
But these are the issues that bother at this point, or some
of us, and we're wondering whether you have thought about them or you
plan to think about them.
There are two fundamental issues, it seems to me. Maybe
only one. That importance measures based on CDF and LERF depend very
much on how the CDF and LERF are calculated. So if you say I'm going to
do only internal events, then you get a certain ranking. If you say I'm
going to include all external events, you get another ranking, which may
be very much distorted by these conservatisms or omissions.
For example, smoke is not part of the fire PRA assessment.
Right?
MR. CHEOK: Right.
DR. APOSTOLAKIS: So that distorts the results. So what to
do?
MR. BERGMAN: I think that's something we'll need to address
during the development of the rulemaking. We don't want to come up with
a scheme that is so complex that licensees can't implement it. I mean,
ultimately, the rules we're talking about do affect thousands of people
on a daily basis. So simplicity is a good thing in some ways and we
need to be able to translate the categorization into what do we expect
from a treatment perspective.
But we do need to consider all modes. We haven't thought in
that detail as to exactly what would be enough in one particular
condition to force it to be declared high safety significant across the
board.
DR. APOSTOLAKIS: It's not just all modes. It's also all
contributions.
MR. BERGMAN: Right.
DR. APOSTOLAKIS: At power, you have fire, earthquakes,
tornadoes, and these are not done at the same level of rigor.
MR. CHEOK: Dr. Apostolakis, in the past few applications,
what the licensees have done is determined what was important at power
for internal initiating events. Then they come up with lists for
importance in terms of shutdown, importance in terms of fire risk, and
importance in terms of seismic risk, if they are PRAs. If not, they'll
do it with the expert panel.
All these lists are now given to the expert panel. The
expert panel has to deliberate each and every SSC that has been
determined to be low safety significant by all these different PRA
models, and determine why they are low safety significant, and if indeed
they are low safety significant.
In other words, there is a PRA member on the expert panel,
and there should be one, he can then explain the parameters which made
this SSC low safety significant and whether it should be higher based on
other considerations.
DR. APOSTOLAKIS: I have two problems with that. First of
all, I agree that something like that will have to take place. But
first, if you don't point out in your documents that one -- or raise the
possibility that one has to look into different pieces of CDF and LERF,
they will not do it, most of them will not do it.
Second, speaking of this observation that the PRA expert
will have to explain to the panel, there was a thought, this side of the
table, that perhaps we should formalize this. Because there are so many
subtleties in these importance measures, there ought to be perhaps a
training session where the expert panel would be trained, say, for half
a day on what these importance measures are, what are the pitfalls,
what's going on, and this is very similar to what we do when we elicit
expert opinion, where, as you know, and the NRC has sponsored several of
those, before you elicit anything, you spend half a day training experts
on elicitation, what does it mean to do this, what does it mean to do
that, and maybe that should be something that the agency should require,
because there are so many subtleties here that I just don't believe that
having a PRA person as a member of the expert panel will actually
educate the rest of the members.
MR. RUBIN: Dr. Apostolakis.
DR. APOSTOLAKIS: Yes, sir.
MR. RUBIN: Mark Rubin, from the staff. When we went up on
the maintenance rule inspections, we generally found that there was
generalized training in PRA methodology for the expert panel. It
probably varied quite a bit and certainly formalizing it, specifying
exactly what was put -- what should be put into it will be an important
part of this activity.
One of the things that the staff's baseline inspection team
looked at was the transfer of risk knowledge to the expert panel as a
whole.
DR. APOSTOLAKIS: So are you agreeing with me or not?
MR. RUBIN: I am agreeing with you, but also just trying to
point out that it seems to have been done.
DR. APOSTOLAKIS: Not exactly what I have in mind. I'm
thinking more in terms of doing what you just said, but then going
beyond that and going to the subtleties and the possible pitfalls, yes.
But, you know, I'm sure that people who have never heard of PRA, when
they go to such a panel, they say, well, what is it.
The issue of importance measures is something, for some of
the reasons that we just mentioned, is something that I think you should
really spend a lot of time thinking, and, in particular, there is some
concern about the use of these measures and consistency with Regulatory
Guide 1.174.
Now, that guide uses as a basic criterion of screening, for
screening requests, delta CDF and delta LERF. Now, Appendix T says that
we will look at Fussel-Vesley greater than .005 and RAW greater than
two, and then there is this wonderful paper, which Mr. Cheok is a
co-author, and I quote now what you say there, "It is clear that it does
not make much sense to define a universal criterion based on RAW. If we
are interested in controlling the change in risk in an absolute sense,
it does not make sense to have a universally fixed value of
Fussel-Vesley as a criterion for risk significance."
Now, you should be careful when you write those papers,
Mike.
MR. CHEOK: And I still believe in those words.
DR. APOSTOLAKIS: Good. So let's hope that Appendix T will
reflect that belief.
MR. CHEOK: And I think if you look at Appendix T, we still
say that the ultimate decision criteria is that you shall determine that
risk in terms of CDF and LERF shall be small.
DR. APOSTOLAKIS: But you are proposing something that you
say -- you call here or you say that it doesn't make much sense. So if
I didn't know that paper, I couldn't tell from Appendix T that you think
that way.
All I'm saying is please revise it appropriately, because it
doesn't make much sense.
MR. CHEOK: Okay.
DR. APOSTOLAKIS: I mean, you may want to end up with some
sort of a importance measure based approach, but you need a lot of
introduction there.
MR. CHEOK: Sure.
DR. APOSTOLAKIS: Any other questions from the members? Do
you want to say anything?
DR. POWERS: Let me ask this question. Suppose a licensee
comes to you and says I've got the major twinger PRA, it does all modes,
all operations, it does fire, shutdown, earthquakes, it's really a good
PRA, and I've gone through and there are only these three fire
components that fall in the number one category.
What do you have to assure that this wonderful PRA that he
has is indeed wonderful and makes reliable -- gives you reliable results
for this process?
MR. CHEOK: Right now we are depending on the ASME consensus
standards.
DR. POWERS: But that doesn't address things --
MR. CHEOK: Bias right now, that's correct. We do hope
eventually that all these initiators will be addressed, but in the
meantime, what we do, I guess we will have to discuss among the staff
what the proper decision would be. We haven't come up with interim
criteria.
DR. POWERS: It seems to me that if a guy did a shutdown
PRA, you're not really in a position to know whether he did a good one
or a bad one. Is that right?
MR. CHEOK: Right now, that's correct.
MR. KING: But I would expect by the time this rulemaking is
completed, which is several years down the road, the standards effort --
DR. POWERS: At the rate that you're investigating shutdown
phenomena, I'm not so confident, several years down the road, that you
will have the information you need to review a PRA.
MR. KING: Well, that's a subject for a separate meeting
we've got scheduled for December 1.
DR. POWERS: I don't think it's a separate meeting. It
seems like this is the strategy to make things risk-informed in Part 50,
and you guys don't have the risk information. How do you make it
risk-informed if you don't have the risk information?
MR. KING: But we're out trying to gather the risk
information and identify what's not there, how do we go about and get
it, whether we do the studies ourselves or rely on somebody else to do
the studies. We recognize we're short on low power shutdown risk
information, maybe on methods as well. We don't have a standard. All
of that has to be done, but I expect all of that will be done before
this becomes a final rulemaking and we can discuss how much is enough
and so forth.
Those discussions have to take place, but I think, at least
the way things are currently laid out, the sequence, to me, looks like
they'll come together and support each other.
DR. KRESS: Assuming this demarcation line between safety
significant, high and low, is based on importance measures, do the
determinations of these importance measures include some measure of the
uncertainty in the actual importance measure?
I've never seen an uncertainty distribution of RAW, although
it certainly ought to be straightforward to get.
DR. APOSTOLAKIS: It's figure four of that wonderful paper.
Rick Sherry did that.
DR. KRESS: I haven't gotten --
DR. APOSTOLAKIS: Rick Sherry has done that.
DR. KRESS: Okay.
DR. APOSTOLAKIS: And his conclusion was that -- well, he, I
mean the three authors, is the point values of the measures, I think
it's Fussel-Vesley, but I don't remember, one of those, and the ranking
that you get based on that. Then for each point, they plot the 90
percent range on the mean value.
DR. KRESS: Okay.
DR. APOSTOLAKIS: So if you did it rigorously and you used
the mean values, the ranking would be the same and there is concern only
when the uncertainties become very, very large.
DR. KRESS: You're talking about uncertainty in CDF and
LERF.
DR. APOSTOLAKIS: In the parameters that go into the
calculation of CDF and LERF, which are also propagated to the measures.
DR. KRESS: I'm concerned about now individual uncertainties
in the RAW value for a specific component.
DR. APOSTOLAKIS: Yes, that's what I'm saying. So you
calculate RAW for one component. That expression has now all the other
parameters in the PRA, right? Because it's based on CDF. All these
have uncertainties. So you can propagate them rigorously and find the
uncertainty in RAW and find the mean value of RAW. That's what Rick
did.
DR. KRESS: That's what I was --
DR. APOSTOLAKIS: So unless the uncertainties of some
parameters are very, very large, the ranking is fairly robust.
DR. KRESS: That was my -- that pretty much answers my
question.
DR. APOSTOLAKIS: It's figure four.
DR. KRESS: I'll have to read that.
DR. APOSTOLAKIS: It's really a nice figure.
DR. WALLIS: Is this one that we don't know about?
DR. APOSTOLAKIS: But it will be handed out to the members.
Could we have it today? You have it now.
DR. SHACK: It's in a packet, a supplementary packet.
DR. APOSTOLAKIS: In here?
DR. WALLIS: Trying to find a way of classifying components
in terms of how much attention they receive or something in terms of
safety evaluation, is that what you're trying to do?
MR. BERGMAN: Yes.
DR. WALLIS: In order to cut down the amount of work, to
only focus on the ones that matter.
MR. BERGMAN: The real purpose of the effort is to modify
the treatment to the structures, systems and components, such that those
components that are more important get more treatment than those that
are less important.
DR. WALLIS: So basically this is a continuum. It's not
boxes. The boxes are simply for convenience.
MR. BERGMAN: Yes. There is gradations. It's not nice,
clean boxes, although ultimately they will declare it low or high. But
there is subjectivity at that boundary.
DR. WALLIS: That's all you're doing?
MR. BERGMAN: Sounds simple, doesn't it?
DR. WALLIS: Are they sheep or goats, is that what you're
doing?
MR. REED: That's all.
DR. WALLIS: One has long horns and the other doesn't.
That's all you're doing.
MR. BERGMAN: It needs to be pointed out, though, that the
number of components in three, in the third bin there, is a much larger
number than the number expected to be in two. In box two is where
you're adding burden. You're taking non-safety-related components and
you're going to add treatment that they may not have previously had.
In box three, for the most part, you're relaxing
requirements on them. So there is burden reduction there, and the net
is expected to be in the reduction end.
DR. WALLIS: I haven't been on this subcommittee long, but
I'd like to see -- it's probably all there -- separation between the
sort of intellectual framework of your approach and then the regulatory
convenience, what is the reason and some way for classifying why you're
choosing this particular criteria to classify. Then for regulatory
purposes, you want to have some boxes, because you want to tell people
to do something to things in the boxes in certain ways, the procedures.
I don't quite see the more intellectual level, why this is
necessary at all.
DR. APOSTOLAKIS: To do what?
DR. WALLIS: To have boxes to separate, to classify and all
that.
DR. APOSTOLAKIS: Because the deterministic dimension
preserves what exists now. This is what they have now. So you can't
ignore that. The new element is the risk-informed.
DR. WALLIS: So it's a transition.
DR. APOSTOLAKIS: It's a transition, exactly.
MR. REED: The current regulations use discriminators, like
safety-related and non-safety-related, in the rules themselves to
determine whether something gets any treatment, EQ or Appendix B or
whatever. So you're stuck with that system today.
Now we're trying to translate that system into something
that's risk-informed, and that's when you get into four boxes with two
sets of terms in each box, as we try to transition.
DR. APOSTOLAKIS: Let's say I own a plant and I do all this.
Several years down the line, we have a beautiful shutdown PRA and all
that. We do everything. And you -- and I put my boxes, my SSCs in the
boxes, you approve my plant.
Can I then drop the terminology safety-related,
non-safety-related, and work with a new system?
MR. REED: No, no. This option, one of the reasons we kept
safety-related is that in option two, we're not giving up functional
requirements and those functional requirements stem from design basis
events, and that classically means safety-related. So we're holding
onto that piece still.
DR. APOSTOLAKIS: So even if I am in box three --
MR. REED: If you go -- you're thinking of option three.
DR. APOSTOLAKIS: Then you will still tell me -- yes,
because they are safety-related, you still have to meet those
regulations dealing with functional requirements.
MR. REED: They're expected to function.
DR. APOSTOLAKIS: So it's not an interim matrix. It's
permanent. That's it.
MR. REED: It's less assurance --
MR. BERGMAN: Until option three goes through.
MR. REED: You were down to option three. You were at the
new set of design basis events and the whole thing there. That's where
you are.
DR. APOSTOLAKIS: I understand that better now. In the
interest of time, any other questions from the members? Does the staff
want to say anything?
DR. SEALE: I think you're being too timid. I still think
you need to identify a zero.
DR. APOSTOLAKIS: A zero. Zero meaning?
MR. BERGMAN: Not significant all.
DR. SEALE: Non-significant.
DR. KRESS: Non-risk significant.
DR. APOSTOLAKIS: Yes, yes.
DR. KRESS: Negligible.
DR. APOSTOLAKIS: South Texas is proposing what?
MR. BERGMAN: Non-risk significant.
DR. SEALE: Yes. I think you're too timid.
DR. KRESS: Isn't that basically what box four would be?
MR. BERGMAN: No. Box three and four are the same from a
risk ranking perspective. They only differ in terms of whether it's
safety-related or non-safety-related. But the low safety significant
category for us goes from zero to wherever the dividing line is
ultimately defined.
DR. APOSTOLAKIS: That's why you made the comment earlier
that you may want to rethink that strategy of using only two categories,
because there may be -- I think the reason why you're not too hot about
it right now is you haven't really thought about what to do with the
SSCs in each category. When you start thinking about it seriously, you
may realize that you need more categories. So that's fine.
DR. WALLIS: How can you design a system without knowing
what you're going to do with it? What's the purpose?
DR. APOSTOLAKIS: There will be another box that says
non-risk significant and there will be different requirements for those.
DR. KRESS: The concept would be the same.
DR. APOSTOLAKIS: Yes.
DR. KRESS: You could add another box down there and it
wouldn't matter to the concept.
DR. SHACK: It also depends on how precisely you think you
can slice the cheese.
MR. REED: And that's a function of the appendix and those
technical issues.
DR. APOSTOLAKIS: So do we have any other comments or
questions? Would you gentlemen like to say anything before you --
MR. BERGMAN: No, we're fine.
MR. NEWBERRY: Mr. Chairman, at the risk of my staff being
unhappy with me for having them -- put up the issue slide. There are
significant questions that still remain. We could see if the committee
wants to --
DR. APOSTOLAKIS: As opposed to the ones we raised?
MR. NEWBERRY: In addition.
DR. WALLIS: I have a problem of knowing when you've
succeeded. It's really when people seem happy? Is there some way you
can tell when you've done what you set out to do?
MR. BERGMAN: I think we've recognized early on that, in
fact, that people may not be happy and that does -- isn't affecting our
thinking. We certainly have objectives, but those objectives may not be
met. We may not, in fact, reduce burden at all. If that's the way it
works out, that's the way it works out.
I think under that case, you don't see a whole lot of enthusiasm for it
in the industry, but we're focused on first trying to do the job right,
come up with what's the appropriate treatment for the equipment in each
box. We're assuming that there will be burden reduction in the end, but
that is not our overriding goal.
DR. WALLIS: It's a kind of experiment, saying what's the
simplest way we can bring in risk information; two categories is the
simplest; if we have one, there isn't anything at all; and see if it
makes any difference to what industry does, see if there is some payoff
eventually.
DR. APOSTOLAKIS: But we know that, Graham, because the
industry has done some of it already and South Texas is claiming savings
of --
DR. WALLIS: But you're trying to create the minimum
framework to let them exercise the option of using risk-informed
information to help --
DR. APOSTOLAKIS: As a pilot.
DR. WALLIS: Really. Unless you get more knowledge, you may
go back and say, well, there ought to be three categories.
DR. SEALE: In the area of 50.59 type.
MR. REED: We're in a rulemaking plan stage here. This is
going to evolve.
MR. BERGMAN: We have not excluded more categories.
MR. REED: It's evolved in the last week, since the last
time we talked to you.
DR. APOSTOLAKIS: Since Friday?
MR. REED: Yes.
DR. WALLIS: That's because you had to make a presentation
today.
MR. REED: Well, that's because we keep cranking away every
day.
DR. APOSTOLAKIS: How often do you make presentations on
this?
MR. BERGMAN: To all parties?
DR. APOSTOLAKIS: Yes.
MR. BERGMAN: Oh, my God. Several a week.
DR. APOSTOLAKIS: A week?
MR. BERGMAN: Yes. A few include --
DR. WALLIS: They're spending too much time on presentations
and not enough time doing the work.
DR. APOSTOLAKIS: So what do you want us to do with those?
MR. NEWBERRY: Let me -- I know the least, so I'll just --
the first issue, selective end point, implementation. This is a
voluntary activity, so plants that would choose to pursue a
risk-informed approach, option two, would do that, and we're looking at
to the extent by which they could do that system by system rather than
for the whole plant.
We think that there is likely merit and no reason to prevent
a plant from tackling this in increments, and that's something we're
going to look at further, but that's a rather significant issue.
So you could have some systems that use safety-related,
non-safety-related terminology, which would still be in the regulations,
and then the 50.69 terminology and Appendix T on a voluntary basis for a
fraction of the plant.
That's pretty scary. I see members shaking their head
already. But that's issue number one and I thought that was worth
mentioning, because it's a real implementation issue.
The -- let's see here -- third bullet, a major
implementation issue, a cost issue for the licensee or applicant and the
staff to go this way, how much staff review is going to be necessary; to
what extent will we have to look at the PRA, will this be a licensing
submittal that we look at in a rigorous way.
Our hope would be it would be minimal. They could implement
it. That's why Mike has put together -- Mike Cheok has put together a
pretty robust Appendix T. The idea would be that the methodology would
be rather clear, so the staff review would be minimal.
DR. APOSTOLAKIS: You mean he will put together.
MR. NEWBERRY: Thank you. That he will put together.
Thought I'd mention that one. Then the last one, the identification and
control of attributes, that gets to the -- we used to have busier boxes.
We used to have more words in the boxes that talked about not just the
importance, but also the nature of the functionality attribute for those
boxes, which now gets into not deterministic aspects, but the aspects of
beyond design basis in severe accidents.
So that's something that we're going to have to work on.
DR. BONACA: I just need to ask a question, which is given
the discussion we had on importance measures and the need for doing work
there, because we all agree that there isn't a clear understanding of
how this should be applied, how do you get comfortable about current
applications, like, for example, the South Texas project, that
characterization is properly done and what are we learning from those
kinds of applications of importance measures?
MR. BERGMAN: There is a fair amount if discomfort, I think,
among the staff. The staff is a lot of people and not everybody is
comfortable with the South Texas approach. I think the people who have
the most familiarity with the PRA are probably -- I'm generalizing here
-- are the most comfortable with it and those least familiar are the
least comfortable with it.
But we have to address many of the same issues in the South
Texas exemption request that we'll have to address in this rulemaking.
The two efforts are very similar.
DR. BONACA: But, you know, we've got to understand that,
because, also, clearly the deterministic approach had all these
shortcomings, but at least it was conservative, let's say, in the sense
that if there was any suspicion that a system was safety-significant,
they would put it as safety-related, and in doing so, you include a lot
of components that shouldn't be there, but, hey, you just did it.
Defense-in-depth.
MR. REED: Yes, exactly. You're starting to get to the --
DR. BONACA: Well, the lack of information --
MR. REED: -- risk-informed --
DR. BONACA: -- just simply has a tag on it.
MR. REED: We have wrestled with that quite a bit and this
is going to be a risk-informed process and we've been wrestling and are
having some good discussions on what defense-in-depth is going to mean
and what safety margins are going to mean and how we're going to build
that into the process, because sometimes it's not too obvious in the
PRA.
DR. POWERS: What does defense-in-depth mean?
MR. REED: You guys have talked quite a bit about it and
I'll ask you.
DR. APOSTOLAKIS: I'll tell you what it means. There is a
pragmatic approach, defense-in-depth at high level, take CDF and LERF
contributions from the cornerstones and apply your importance measures
to each one, then defense-in-depth is built into what you're doing.
This is one member's view.
Then it's built in, because you are doing it for different
categories. Anyway, Scott, what do you want us to do with the issues,
just pointing out that they exist?
MR. NEWBERRY: Yes, we're done. I wanted to point them out
to the committee, that's all.
DR. APOSTOLAKIS: And we will have other opportunities to
discuss them. I mean, not today, but in the future.
MR. NEWBERRY: Yes, certainly. Thank you. I think we
missed a major point. We're proposing to the Commission that we go out
with an advanced notice for rulemaking at this point. We're a long ways
away from resolution. That's all.
DR. POWERS: Suppose you guys do your job, you're really
successful, you've got a new set of rules, an appendix all of your very
own, and we say, okay, what's adequate protection now.
MR. REED: I was expecting that from Graham Wallis.
MR. KING: The approach we're taking is to work around the
current plant risk profile. Similar to Reg Guide 1.174, the criteria
that are in there allow small changes around the current risk profile of
plants. The current plants have met the adequate protection standard by
the current means by which we declare adequate protection.
I think as long as we're just talking about small changes
around the current risk profile, the feeling is that we're maintaining
adequate protection. We're not allowing CDF to rise significantly and
we're not trying to drive everybody down to some lower level. It's,
from a safety standpoint, about where we are and get the things out of
there that don't make sense from a safety standpoint, and maybe plug a
few holes that do make sense to plug. That's the approach we're taking.
MR. REED: I think of this in terms of like a Reg Guide
1.174. I know that's not going to make you happy, but it's not a
significant change.
DR. POWERS: Well, if it were a reg guide, I'd feel very
comfortable and I wouldn't worry about it very much, but when you start
putting in rules, then I've got to worry about it, because there's a
nice definition, a definition I personally just love, because I think
it's a nice finesse of the problem, which is if you're in substantial
compliance with the rules and regulations, then we presume that you're
providing adequate protection.
I think that's -- whoever came up with that had a stroke of
genius on that. But then you're going to put in another set of rules
and I'm not sure --
MR. KING: And if you comply with those, the same thing.
DR. POWERS: -- now all of a sudden what compliance with
those particular rules means.
MR. KING: It means the same thing, from the adequate
protection standpoint.
MR. BERGMAN: We have to make the same finding as part of
the regulatory analysis.
DR. WALLIS: It seems to me you have an opportunity down the
road to do something much better than that and all these boxes and these
sort of adequate protection things which are circular arguments are
because of the deterministic roots of the rule and regulation.
Risk measures have the potential to not have boxes, but have
a continuum. You can integrate it, you can take averages, you can do
all kinds of things with it which are meaningful. You're not just
making judgment calls about how you add up the importance of
regulations, what's adequate, what's not. You can actually do something
rational with it in a much better way than was done in the past.
So whatever you do I think should be a transition from one
of these ways of doing things to the other.
MR. SIEBER: I guess another issue is when you look at the
four boxes, the horizontal dividing line sooner or later has to be
quantitative, it seems to me. Otherwise, you can't fit this continuum
into these categories and the question of if you don't quantitatively
define adequate protection, how do you define the line.
If the net result is to reduce requirements, quality
requirements, for example, then I guess you struggle a little bit to say
that you haven't adversely affected risk profile.
DR. APOSTOLAKIS: No. I think the way the staff has handled
that, starting with 1.174, is they are taking as a point of reference
the existing risk profile of the plant, not what is adequate protection.
So all these analyses go back to that.
And what they did in 1.174, for example, delta CDF, the
allowed delta CDF is the same for plants that have a
ten-to-the-minus-six core damage frequency and plants who have
ten-to-the-minus-four, because they always look at the change in an
absolute sense from the current profile.
Now, if you bring in the adequate protection business here,
then the ten-to-the-minus-six plants should be allowed to do more
things, right?
MR. SIEBER: That's right.
DR. APOSTOLAKIS: So that's how they have handled it so far.
MR. SIEBER: Yes. I guess I understand that, but when I get
down to the next level, the second order effect, which is the horizontal
line that separates the boxes, somehow or other there has to be a
definition there as to how it relates to adequate protection in order to
determine --
DR. WALLIS: And it could be an average. It doesn't have to
be horizontal.
MR. KING: I think as long as the horizontal line is drawn
to represent small changes in risk, then you don't have to
quantitatively define adequate protection and come in and make a finding
again that you've got adequate protection. That's really the concept
behind this, because you haven't changed the risk very much.
DR. APOSTOLAKIS: I have a question that relates to what we
discussed this morning, and Mr. King was present, so you can serve as a
bridge. This morning we heard that the Office of Research conducts
independent research and develops technical bases for realistic
regulatory decisions.
Then we also heard that one of the Research Office's roles
is to improve the agency's knowledge in area where uncertainties exist
and may be significant to risk and so on.
Is this an area of regulatory decision-making where there is
a need for a technical basis where uncertainties exist that may be
significant to risk? And yet we don't hear that Research is going to
develop a research program to develop the technical basis. We hear
that, well, if PRA does not exist or is very poor, then we'll figure out
another way of handling it and maybe we'll ask the expert panel to take
care of it.
Isn't there a contradiction there? I mean, why don't we
just say, well, gee, this is a very important activity for the agency,
the role of the Office of Research is to provide a technical basis for
these decisions, so we have to do shutdown risk, we have to have a
research project to understand the importance measures, the
uncertainties, if you are conservative here and non-conservative there,
what does that do to your measures and so on.
So we are not self-consistent within the span of about three
hours.
MR. KING: Well, I don't agree with the way you've
characterized it. I think there are a number of things going on in the
office to reduce uncertainties. Low power and shutdown risk we've
talked about, fire risk methods to be improved, human reliability
methods to be improved. We're trying to take a look at the effect of QA
on risk.
But by the same token, there are things that are not being
looked at, organizational factors influence on --
DR. APOSTOLAKIS: I thought the shutdown and low power risk
were eliminated from the budget for next year.
MR. KING: No.
DR. APOSTOLAKIS: That's not correct?
MR. KING: No. They are in the FY-2000 and 2001 budget.
They've been eliminated from the 2002 budget.
DR. APOSTOLAKIS: So the year 2000, Dr. Powers will not be
perplexed anymore. Can you live for three years being perplexed?
DR. POWERS: Well, we'll have a chance to hear what their
plans on that are.
DR. APOSTOLAKIS: You may go down to puzzled.
DR. KRESS: Never confused.
DR. WALLIS: I think that to make another bridge with this
morning --
DR. APOSTOLAKIS: We're running out of time.
DR. WALLIS: When you take on a job like this, you should --
do you have to make decisions based on what you know, have to be
functional, have to work and all that, but I think, at the same time,
you ought to lay out for the future or something what it is you would
like in the way of tools or something, so that when it is done next
time, it will be done better. Identify if there is any technical base
which the agency would like to have in order to address this problem in
some better way.
It doesn't have to be a lengthy thing, but just specify it.
MR. KING: And I think that's going on. In parallel, we're
trying to figure out how do you make these decisions in light of what we
know today. We're also trying to improve things in a number of areas.
DR. APOSTOLAKIS: Okay. Good.
DR. WALLIS: Do you think they're good enough for what you'd
like to do in the future?
MR. KING: In terms of research and improving things?
DR. WALLIS: Where would you go from here in the future?
MR. KING: We have a road map on identifying those areas we
think are worthy of doing research to improve our understanding, whether
it's improved method --
DR. WALLIS: No, I don't mean that. I mean, how would the
research and the regulations evolve together, one building on the other.
MR. KING: I think we could put together such a road map
fairly easily. I think it's been thought about, but I don't have a
viewgraph to put up on the machine. But it's been thought about, what
we're doing and the timing of what we're doing and how it fits into --
DR. APOSTOLAKIS: That's the PRA implementation plan, isn't
it?
MR. KING: Well, the implementation plan is sort of a
catalog of things that are going on, but I wouldn't call it a road map.
DR. APOSTOLAKIS: Okay. Unless there is a burning question
or comment from around the table, I propose we go on. Thank you,
gentlemen.
The Nuclear Energy Institute, come to the table, please.
Mr. Riccio, how many minutes do you need, ten?
MR. RICCIO: I can go through it in that.
DR. APOSTOLAKIS: Okay. So can you keep your comments to
ten minutes or less?
MR. FLOYD: Certainly can.
DR. APOSTOLAKIS: Please identify yourselves.
MR. FLOYD: Steve Floyd, from Nuclear Energy Institute.
MR. HEYMER: Adrian Heymer, from NEI, and I work for Steve.
MR. FLOYD: We'll abbreviate our presentation and just hit
the highlights that are in the presentation materials.
I want to probably start out with what we see as the
benefits of this. What we're really looking for, and I think Tom King
summed it up very nicely, is we're really not looking for significant
departures from what the current risk profile is across the industry.
We're looking for what can be done to eliminate unnecessary requirements
that add burden, but don't add safety value, eliminate those from the
requirements, while, at the same time, we look for where we may have
missed things in the past in terms of requirements that are important
from a safety perspective to add.
So the real objective here is to improve the safety focus of
the plants, while, at the same time, we maintain the current levels of
safety in the plants, and so that in a manner that we think can result
in the elimination of unnecessary burden.
And I think people often talk about eliminating regulatory
burden, but I think the key word in that is really unnecessary, and our
definition of unnecessary is if you can show that the elimination or the
change of that requirement does not have a negative impact on safety,
then that requirement really doesn't belong in the regulatory program
and can be eliminated.
DR. WALLIS: Let me ask you. The staff wants to maintain
safety and adequate safety is defined as meeting their regulations.
MR. FLOYD: Correct.
DR. WALLIS: What is your definition of maintaining adequate
safety or safety performance?
MR. FLOYD: Well, I think it's pretty similar to that. The
way we look at it is we agree that the adequate level of safety, however
you want to define that, is as a result of compliance with the current
set of regulations.
So we've been able to measure, we think, what that
performance level is and it is a combination of risk insights and what
is the current state of performance of the plants in key areas.
What we're saying is when we make any changes to the
regulatory requirements, as long as those changes result in a
significantly small, if any changes in the risk profile of the plant or
the performance profile of the plant, then you have not changed what the
current level of safety is.
DR. WALLIS: I find this logically problematic. If the
meaning of the regulations is maintaining adequate safety, then you want
to change the regulations while maintaining adequate safety. But if
adequate safety is defined by the current regulations, what do you
appeal to, to show that nothing is changed?
MR. FLOYD: What you have to do, I think, is look at what is
the current level of safety being provided by the regulations.
DR. WALLIS: Measured by what?
MR. FLOYD: Measured by a combination of risk insights and
current plant safety performance.
DR. WALLIS: Measured by what quantitatively, that I can put
my hands around and say this is my measure of safety?
MR. FLOYD: I think there are several measures. You've got
-- as you have been talking about, we've been talking about delta CDF
and delta LERF as some quantitative measures on the risk side and I
think you have to look at safety performance, initiating event
frequencies, things of that nature for what may be current level of
performance of the plants as measures.
DR. WALLIS: Would it be useful to agree on certain measures
of safety like that which are to be maintained at the level established?
MR. FLOYD: Yes, and I think a lot of that work has been
done. If you look at the structure of the new oversight process and the
cornerstone approach, it does provide a clear statement of the objective
of each of the seven cornerstones and each of the seven cornerstones are
intended to provide what is --
DR. WALLIS: Does it say what level is adequate at each one
of them?
MR. FLOYD: Yes, it does. Yes, it does. What we have on
each of the performance measures in the new oversight process is a
delineation of what are acceptable performance thresholds. Similarly,
when you find a problem in the plant that's not directly measured by one
of the performance indicators, there is a significance determination
process that has been established that also has quantified thresholds.
DR. APOSTOLAKIS: Now, since time is short, maybe we can
also go to questions. Of course, you will be free to raise points that
you really feel are important.
You mentioned measures, Steve, and you were here when we
were discussing importance measures and so on. What is the industry's
view on that? Is the industry satisfied that the current importance
measures are good enough to do these calculations or do they have the
same concerns that were raised a few minutes ago and if they have those
concerns, what are they doing about it?
MR. FLOYD: First of all, I think where a lot of people in
the industry are is that they're reaching the awareness that when we get
too wrapped around the metrics and too wrapped around the fidelity and
quality of the PRA, then you're really trying to be a little bit more
risk-based and less risk-informed.
Some very smart people, much smarter than I am, go out to
plants that are knowledgeable of how to do PRAs and knowledgeable of
deterministic requirements in the plant and without doing a PRA, they
can pretty much come up with a ranking of the SSCs that are important
without running any numbers.
DR. APOSTOLAKIS: But that may have a problem being approved
by the NRC.
MR. FLOYD: I understand that.
DR. APOSTOLAKIS: You have to have certain rules how to do
it.
MR. FLOYD: You have to have certain rules. So what we are
really proposing in our approach is that the PRA insights are just one
of the tools that you consider in ranking the SSCs. You don't make any
go-no go decisions on the basis of the PRA alone.
DR. APOSTOLAKIS: The problem that I see with this, the
potential problem is that those insights may be distorted in ways that
unless you have spent a lot of time working on these measures and
reading about the papers that people write about them, you may not
appreciate, if you are a casual user, or more than a casual, but you're
an engineer, you don't really care about the mathematical foundations of
these things, so they tell you these are the measures and you use them.
Then they warn you that there are many judgments, so be
careful. Okay, I'll be careful. But I doubt that there will be a
warning that says this measure can be completely off, because you
haven't done a good job there and there and here are the reasons.
So it's this distortion of the risk information that worries
me and it seems to me that it would behoove all of us if somebody did
something about it and said, look, this is the situation now, we can do
this and improve things, here it would take too long, but here's another
way of handling that, rather than saying that the expert panel will.
I mean, the expert panel very quickly will have to consist
of Superman to be able to understand all the subtleties of these things
and come up with decisions and that worries me a little bit, and I would
like to sensitize a little bit the industry about it.
I mean, now we are using these things in a serious way. We
have to take them seriously and ask ourselves is the information that
this measure gives me of reasonable quality so that an informed expert
panel can make the decisions.
I'm not asking for high precision, but if you have things
that may distort the information very much, then I don't know.
MR. FLOYD: Our experience with the maintenance rule was
that the way people really did it was they looked at the importance
measures from their risk analysis and they added items into the
significance category on the basis of that, but you never really removed
an item from significance on the basis of just a PRA insight.
Every item that was not a candidate per the risk ranking
also went through the expert panel where other factors came into play;
did the operator really rely upon this, could it increase the initiating
event frequency for an event at the plant, that the plant is likely to
respond to. Those types of factors that came into play.
What we found was that you started out with a set of SSCs
based upon the PRA analysis and you only added to that when you went to
the expert panel. You did not take things away.
And the expert panel did not overturn any of the risk
ranking items that were called high significant and in most cases, they
agreed with the ranking of it.
DR. APOSTOLAKIS: The question is how robust the ranking is.
MR. FLOYD: But nonetheless, it stayed in the
risk-significant category, so it was treated as if it were.
DR. APOSTOLAKIS: Essentially, your answer is that, yes, you
recognize that there are limitations, but you are confident that the
expert panel will be able to come up with a reasonable ranking for
boxes.
MR. FLOYD: Yes, and we think they did that under the
maintenance rule, which is using a somewhat similar approach to what
we're talking about here.
DR. APOSTOLAKIS: I think this is much more ambitious,
though, than the maintenance rule that we're talking about here. It
affects many other rules and special treatment --
MR. FLOYD: Yes, it does. Yes, it does.
DR. APOSTOLAKIS: So the panel may not be as willing to be
conservative here or as conservative as in the maintenance rule.
MR. SIEBER: Perhaps you could clear up an element of
confusion for me. If I have a component, for example, to move from box
two to box three and I reduce the quality requirements for the QA
system, that should reflect itself in the PRA as a change in the
unavailability, right?
MR. FLOYD: If it has an impact, yes.
MR. SIEBER: Well, I guess the question is, how would you --
if you're going to look at the change in risk from before to after, how
do you know what that change in unavailability would be? There's no
experience on that.
MR. FLOYD: What we would propose is that there be a
monitoring program put in place for items that change the classification
scheme so that you can track what the changes are. You'd also keep
track under the new oversight process any adverse finding that you would
have. You have to judge whether or not that condition might, down the
road, if it occurred, affect the classification that you previously put
on it through the process.
So you have to -- this whole process, we see it as being a
living process. It's a feedback loop that has to be built into it to
make sure that you do monitor to see if there are any negative impacts
and if there are, that that has to come into your judgment as to whether
or not you misclassified an item.
MR. SIEBER: Will that be part of the rule or is that just
something we ought to do because we're approached?
DR. APOSTOLAKIS: I believe monitoring strategies are always
required in these things. Isn't that right? And there will be some
monitoring program somewhere there along the line.
MR. NEWBERRY: We're in box three right now. I'm sorry, I
was reading something. That's a question. We're talking the basic
desire is to maintain functionality and monitoring could be one
attribute, along with some others. Monitoring or even that presumes
testing of some sort of surveillance.
DR. APOSTOLAKIS: But essentially you are right, Jack. A
lot of these things, it's a judgment call, and, in fact, that's why we
asked the staff to investigate whether quality assurance measures make a
difference in the failure rates, and Tom referred to that research
project. There are some people who believe that the failure rate
distribution doesn't change much, if at all.
I don't know now whether one can make that statement
universally for all components.
MR. SIEBER: Is there data that shows that or is that a
belief?
DR. APOSTOLAKIS: We don't know, we just don't know.
MR. SIEBER: It would be good if we did.
DR. APOSTOLAKIS: What is your last and best slide you want
to use, because we're running out of time. Is there a point you want to
make, that you want to leave with the committee? We appreciate that
this is an important thing for the industry, so you don't have to worry
about that.
MR. FLOYD: I'm going to ask Adrian to run through this
slide. I think this one kind of sums up where we are on it.
DR. APOSTOLAKIS: And then we are done.
MR. HEYMER: As regards option two, I think this slide just
points out some of the differences that we see compared with where the
NRC staff put up for boxes, we've got three. I think the overall
approach is the same. We start off, we look at all SSCs, we look at all
events.
You do a deterministic and a probabilistic risk evaluation.
You go through the expert panel and you come out with what we see as
three categories. They're safety significant, and that's a mixture of
what was formerly safety-related and also formerly non-safety-related,
but which, through the risk evaluation methodology, you have decided to
be what we call a safety significant.
Now, for those SSCs that were safety-related, we see that
there would probably be no change in the way you treated them. No
change at all. For those that come out, that were non-safety-related
and now are safety-significant, we think that you would identify the
attribute in that component or set of components that makes it
safety-significant and the regulations would apply to that.
On the Commercial T, which I think equates to the staff's
box three, which are SSCs that are not categorized as
safety-significant, but still have a link into the regulation, not
necessarily because they're called safety-related, but because they are
governed by technical requirements in the regulation, core spray for a
boiler might be such a system that comes to mind, or some elements of
that.
And on those, as Steve said, what we believe would happen is
that you would maintain functionality and have -- implement some
commercial controls, make a commitment to implement commercial controls
on those SSCs to provide some reasonable assurance that that
functionality would continue, and there would be a monitoring program.
The gist of this slide is that we think it is important to
change the term safety-related to safety-significant. One, it gets
through some of the cultural issues and, two, I think it helps you out
down the road as you move into option three.
DR. APOSTOLAKIS: But the problem with that, Adrian, is that
the regulations now use safety-related and non-safety-related. To
effect these, you would have to change the whole Part 50 or a good part
of it.
So the staff says the reason why they want to preserve that
categorization is that they can tell the licensee, when it comes to
functionality requirements, go with safety-related or
non-safety-related; when it comes to special treatment, use the
significance, the safety significance.
MR. HEYMER: We would have an option. So the current Part
50 would stay in place and there would be option B to Part 50, which,
for option two, for phase two, would be very similar in language to what
is existing, but you would term safety significant in place of
safety-related, important to safety, important to safe shutdown,
significant to safety, et cetera.
DR. APOSTOLAKIS: Wouldn't it be better to implement this
after option three has been implemented?
MR. HEYMER: I think if you do this prior to option three,
it sets you up for identifying option three, and as regards what are the
--
DR. APOSTOLAKIS: The candidates.
MR. HEYMER: It tells you the candidates, helps you to
identify the candidates to really look at. It also emphasizes it's not
what it's called in the terms of safety-related, but it's the technical
element in that regulation that links back to the components which are
in Commercial T or box three from the NRC staff.
MR. FLOYD: The category that really gets picked up in is
the Commercial T box. As we see items being rolled into what is the
staff's third category or our Commercial T box, those become the likely
candidates for the phase three effort to change the technical
requirements to improve the safety focus.
DR. APOSTOLAKIS: Do you have a report on these things? Do
you have a written document that contains this?
MR. FLOYD: No. We're in the process of developing a
guideline, implementation document that's going to be how you go about
doing the risk evaluation methodology, blending the risk insights with
deterministic requirements, operating experience, et cetera, and we have
a commitment to the staff to get a draft of that guideline document to
them by the end of the year.
DR. APOSTOLAKIS: Any other questions? Well, thank you very
much.
MR. FLOYD: Thank you.
DR. APOSTOLAKIS: We haven't touched option three yet. Mr.
Riccio.
MR. RICCIO: I'll try to blow through this quickly and get
us out of here on schedule. Good afternoon. My name is James Riccio.
I'm the staff attorney for Public Citizens Critical Mass Energy Project.
It's a pleasure to once again present our views to the
Advisory Committee on Reactor Safeguards. As I mentioned to the ACRS
subcommittee last week, I've been attending ACRS meetings for over a
decade, because it's the only venue that the public can hear legitimate
debate about the issues that affect the lives, families, homes and
communities.
I would like to thank Dr. Apostolakis for inviting me to
present Public Citizens' views on PRAs in the regulation of nuclear
reactors.
We're not here today discussing the use of PRA because of
any new revealing insights into the risks posed by nuclear reactors, nor
are we here because of any revealing insights into or any great leap in
the ability of PRA to model reality.
We're here because the nuclear industry has come to the
realization that it's forced to compete -- pardon me -- they are forced
to comply with the regulations that are currently on the books. Their
nuclear reactors will be unable to compete in any deregulated
electricity marketplace.
Public Citizens views NRC's efforts to risk-inform nuclear
safety regulations contained in Part 50 as yet another in a series of
attempts by the agency and the industry to deregulate safety standards
based not upon safety, but upon costs. We've already witnessed NRC's
reduction of requirements marginal to safety, cost-beneficial licensing
actions, the use of notices of an enforcement discretion to avoid
shutdowns and allow restarts, and the new and improved technical
specifications which reduce limiting conditions of operation by 40
percent.
Unfortunately, it appears that the NRC and the nuclear
industry have reverted to a pre-Three Mile Island mind set, where they
do not believe that another meltdown will occur. The NEI testified
before the Senate Oversight Committee that the deregulation of Part 50
safety standards was possible because of the improved safety record of
the nuclear industry.
I do not believe that the nuclear industry has actually
improved safety of the reactors it operates. I believe that they and
the NRC have become more adept at manipulating the performance
indicators.
DR. WALLIS: Do you have a basis for this belief?
MR. RICCIO: Yes, I do, as a matter of fact.
DR. WALLIS: Is there some factual basis that's citable?
MR. RICCIO: Yes. Actually, I'll do it right now. You can
take a look at how safety system actuations are treated by the agency.
When they were unable to get that performance indicator to down-trend,
they redefined it as only those actuations that were actually needed,
and they dropped off very nicely after that.
In my old report, I didn't bring a copy this time of the
Lemons report, at least half the old indicators in one way or another
have been manipulated over time to make them down-trend. But like I
said, we won't go into that, because we'll be here for the rest of the
evening.
The fact that the nuclear industry has not melted down a
reactor in the last 20 years is not a sufficient reason for deregulating
the requirements that help achieved that record. NEI's assumption is
based on the specious argument that operating without a meltdown for a
finite period of time means that your safety is adequate.
Hal Lewis, a former member of the ACRS, recognized this
fallacy when the ACRS took up the original license renewal rule, and
I'll try to get this out coherently, since I stumbled over it last time.
Mr. Lewis stated that the general argument that the fact
that one has operated safely for a finite period of time proves that the
safety level is adequate is just not statistically right, because there
isn't that much history in the industry and it's a trap, because other
agencies, for example, people that have used the argument that they had
24 successful shuttle flights to show that the level of safety was
adequate, and, in retrospect, after one disaster, it turned out not to
be.
The Soviets, after Chernobyl, suddenly discovered that their
level of safety that they had before Chernobyl was not adequate, but the
day before Chernobyl, they would have said that it was adequate on the
basis of operating history.
So it's a trap, a psychological trap.
DR. WALLIS: You do exactly the same thing. I mean, you
extrapolate limited experience in the way you wish to extrapolate, on
the next page. So I think we're both in the trap.
MR. RICCIO: I can only work off of what you people give me.
DR. WALLIS: I'm just saying if you apply criticism to the
NRC or the agency, you should apply the same criticism to yourself, to
be consistent about it, because if this -- if you've got some principle
that you're using, it should be applied fairly to both sides. That's
all.
MR. RICCIO: I want to say I just think your use of PRAs are
basically unsupportable.
DR. WALLIS: Not my use. I'm just --
MR. RICCIO: The industry's. So it's a trap, a
psychological trap to believe that because something has happened, that
you're doing just fine.
We have other reasons for concern with the use of
probabilistic risk assessments in the regulation of nuclear reactors.
PRAs are premised upon the supposition that a nuclear reactor has been
designed, constructed, operated and maintained in compliance with its
design basis.
This supposition is not based in fact. The shutdown of the
Millstone reactors and the subsequent closure of Haddam Neck and Maine
Yankee are evidence of this fallacy.
Maine Yankee had cable separation problems that dated back
to original licensing. Haddam Neck was finally closed when it was
revealed that the ECCS would not have performed its function for the 28
years that it operated.
These examples are documented in my report, Amnesty
Irrational, a copy of which has been provided to the committee.
However, if you prefer not to accept my conclusions of my
report, because of my anti-nuclear credentials, I suggest that you read
the Office of Analysis and Evaluation of Operational Data, AEOD's
report, on undetected failures in safety systems.
In a review of 33 events from the accident sequence
precursor database from '91 to '93, the AEOD found that the failures
existed in systems important to safety and, in some cases, remained
undiscovered for long periods of time.
Four failures for a period of one to ten years, another four
for more than ten years up to 18 years, and though the information is
not definitive, an additional four events may have gone undiscovered
since initial startup, while two others existed since the plant
modification.
And this is an aside, but it's surprising to me at this
point that the NRC, while it's moving toward risk-informed
performance-based regulation, has distorted its ability to actually
track performance. After the AEOD's data was used to prove that senior
managers at NRC were failing to do their jobs, the NRC wiped out the
AEOD, scattering its personnel throughout the agency.
Unfortunately, NRC reminds of Shakespeare's Othello; their
eyes have offended them, and so they've torn them from their body. I
believe that wiping out AEOD and relying upon INPO or other industry
data is a grave mistake that will eventually come back to haunt this
agency and this industry.
Another problem we see with the use of PRAs is that their
assessments do not adequately reflect reality. NRC and NEI have been
tossing around core damage frequencies of one in 100,000 to one in a
million. The slide used last week had a core damage frequency on it of
one-times-ten-to-the-negative-seven or one in ten million.
I wish nuclear reactors were that safe. Unfortunately,
nuclear reactors are a lot more dangerous than the NRC and the nuclear
industry would have the public believe.
After the meltdown of Three Mile Island and the explosion at
Chernobyl, irreparably altered the image of the nuclear industry in this
world.
The industry was quick to point out that Chernobyl couldn't
happen here. Their rationale for this statement is that no reactors of
the Chernobyl design were operating in the United States, but that's not
the point. The reality is a nuclear accident can occur in the US that
would have off-site consequences comparable to that of Chernobyl.
In testimony before Congress, Commissioner, former
Commissioner Asselstine said that while we hope that these occurrences
are unlikely, there are accident sequences for US plants that can lead
to the rupture or bypass of containment in the US reactors, which would
result in off-site releases of fission products comparable or worse than
the releases estimated by the NRC staff to have taken place at
Chernobyl.
That is why the Commission told Congress recently that it
could not rule out a commercial nuclear power plant accident resulting
in tens of billions of dollars of property loss and injury to the
public.
In 1990, NRC was asked the probability of a severe core melt
accident in a US reactor. However, they refused to provide the National
Academy of Sciences' National Research Council with the number they were
seeking.
In NRC's response to the National Research Council, the
agency stated that we'd strongly encourage your committee not to use any
number based on assuming an average severe core damage frequency;
rather, the NRC suggested that the National Research Council state that
there is a reasonable evidence that the ensemble of US reactors meet
NRC's safety goals and that there is reasonable assurance that the
health and safety of the public will be adequately protected.
They are now relying on those numbers that they failed to
give the National Research Council in the beginning of the '90s.
Unfortunately, the NRC seems to be in denial of the fact that meltdowns
have occurred in US reactors and that this probability makes the nuclear
industry anything but safe.
The NRC's latest probabilistic risk assessments don't even
account for the meltdown at Three Mile Island or the earlier meltdowns
at Ferme 1 and other test reactors. I've listed the number of core
damage accidents that have already occurred. The EBR-1 at Idaho
Falls in November of '55, the Westinghouse test reactor on April 3 of
'60 at Walts Mill, Pennsylvania, the SL-1 low power reactor on January
3, '61 at Idaho Falls, again; Ferme 1 on 10/5/66, and Three Mile Island
on 3/28/79.
Even if you exclude the core melt accidents at the test
reactors, the fact is that the US commercial nuclear power industry has
a core damage frequency of two in less than 2,500 reactor years.
My final concern is that I've already witnessed how NRC
intends to use the risk assessment in the regulation of reactors.
They're not being used to improve safety. They're being used to improve
the economics of this failed technology.
I know that individuals on this committee and in this agency
have grave misgivings about how risk assessments were used to avoid
steam generator tube inspections at the Farley reactors in Alabama.
Three analyses were conducted. One was deterministic, and that said
inspect the tubes. One was a combination of deterministic and
probabilistic, that also said inspect the tubes. The third was a purely
probabilistic analysis that finally gave the utility the answer they
wanted -- don't inspect the steam generator tubes and operate for
another cycle.
Mr. Long told this committee that you can't, with risk
assessment, tell if it's a third of a year, you're fine, and if it's a
half a year, you're not fine. Risk assessments are just not that good.
Yet, this is precisely what the agency did.
He also said that you're getting yourself susceptible to
many more -- much more frequent transients.
I apologize for quoting your lines back to you, Dr. Powers,
but you basically called this regulation by religion and that we haven't
seen enough severe accidents and the progression of those severe
accidents into the steam generator to have a defensible basis for the
NRC's comfort.
According to the NRC, spontaneous tube ruptures have
occurred at a rate of approximately one every two years over the last 20
years. The Southern Company and NRC are gambling. They're betting that
Farley can operate for another fuel cycle without a steam generator tube
rupture. They're gambling that if the steam generator tube rupture does
occur, that no more than ten tubes will rupture.
They're betting that the operators will respond
appropriately so the coolant inventory will not be lost, causing a
meltdown of the core.
The NRC and the nuclear industry are gambling with the
public health and safety. If Farley is an example of how the NRC
intends to implement risk-informed regulation, then the nuclear industry
is going to melt down another reactor. You don't have enough history in
the industry to regulate based upon probabilistic risk assessments.
PRAs are a tool. They shouldn't be the basis of regulation.
I have an NRC document here, dated March 9, 1979. It reports that there
was less than one in 20 chance of a major accident occurring in the next
21 years and less than a one in two chance of a major accident occurring
in the next 400 reactor years.
It goes on to point out that even these probabilities
suggest unwarranted pessimism due to conservative analyses and lack of
engineering considerations. Less than three weeks later, Unit 2 at
Three Mile Island suffered a core melt, the worst accident in US
commercial nuclear power history.
DR. POWERS: That's not inconsistent with either one of the
previous estimates.
MR. RICCIO: I understand that.
DR. POWERS: Well, it looks like you don't.
MR. RICCIO: No. The reason I put it in there is to prove
that basically that your PRAs don't reflect reality. Less than 400
reactor years, one in two chance in less than 400 reactor years.
DR. POWERS: Yes. It can occur tomorrow.
DR. WALLIS: It doesn't tell you anything.
DR. POWERS: It can occur tomorrow, and still be okay.
MR. RICCIO: So it fits, then.
DR. POWERS: As far as you've presented, there is nothing to
say they're inconsistent.
MR. RICCIO: Okay.
DR. POWERS: I don't happen to disagree with your
probabilistics of one in 2,500 reactor years.
MR. RICCIO: It's two in 2,500.
DR. KRESS: It has to be a Bayesian prediction. So I'm not
sure it's right either.
MR. RICCIO: Well, we can figure that out.
DR. KRESS: You're basing your new prediction on --
MR. RICCIO: The fact is they melted down the reactor.
DR. APOSTOLAKIS: I think there is another element here. I
mean, Dr. Powers is right, it's consistent when you say the probability
is such and such and it happens. I think what Mr. Riccio is raising is
the element of surprise.
In other words, if it's one in 400 or whatever it is, you
really don't expect it to happen tomorrow.
DR. KRESS: No. That is a surprise.
DR. APOSTOLAKIS: That's really what he's saying.
MR. RICCIO: That was my point.
DR. APOSTOLAKIS: Not that it is inconsistent.
DR. POWERS: There is nothing to prevent it from happening
tomorrow.
DR. APOSTOLAKIS: No, of course not. There is a suspicion,
though, that maybe the number was not --
DR. POWERS: If you operate in a Bayesian sense --
DR. KRESS: That's what I said. If you did it Bayesian,
then you would just say, no, that number doesn't look right.
DR. POWERS: And somewhere earlier, he had numbers that came
out like two in 2,500, which doesn't sound to me like a terrible number
to have, terrible in the sense of wrong, because if I take --
DR. KRESS: It's one in ten-to-the-minus-three.
DR. POWERS: If I take roughly ten-to-the-minus-four for
operating events and roughly ten-to-the-minus-four for shutdown events
and roughly ten-to-the-minus-four for fire events and something for
seismic events, I come up with numbers like --
DR. KRESS: Yes, but that number wouldn't surprise you.
DR. POWERS: It wouldn't surprise me.
DR. APOSTOLAKIS: Why don't we let Mr. Riccio finish and
then we'll ask questions.
MR. RICCIO: Basically, I believe that NRC's and NEI's
attempts to risk-inform Part 50 are shortsighted and they will place the
public at greater risk. While I have grave doubts as to the industry's
ability to survive in a competitive electricity marketplace, I'm certain
that the industry cannot survive another meltdown at a US reactor.
Unfortunately, if NRC and NEI are successful in deregulating
nuclear safety standards based upon risk assessments, I believe more
meltdowns are likely to occur.
I thank the committee for its time and consideration of our
comments, and I'd be happy to answer any questions.
DR. APOSTOLAKIS: I just have one comment. You seem to have
a love-hate relationship with the NRC. You like what the AEOD did, but
then you criticize the agency for using risk information.
MR. RICCIO: I liked the AEOD's work. I think it was some
of the best work done in the agency, and, unfortunately, it's no longer
here.
DR. POWERS: Just as a potion of clarification. It's true
the AEOD is not here, but the function of acquiring and assimilating
data still exists.
MR. RICCIO: My understanding, though, is that for the new
oversight process, you're going to be using industry data rather than
what is generated out of this Commission.
DR. POWERS: Well, it's certainly going to take advantage of
the industry's database, because it was bigger and better than what they
have. But independent assessment of the data still exists in the
agency.
DR. APOSTOLAKIS: Yes, the same guys. The message I get
from you, Mr. Riccio, I mean, I can argue with some of the things you
said and that's not surprising.
MR. RICCIO: I'm sure you can.
DR. APOSTOLAKIS: I mean, would argue with everyone sitting
there.
MR. RICCIO: It feels like I'm doing dissertation.
DR. APOSTOLAKIS: But I think what you're telling us
basically is that you want to be assured -- well, you are putting in a
way -- you don't trust PRA results. Now, a side comment here. The
agency is not moving to a PRA-based regulatory system. It's informed,
you've heard those words.
Also, the question is not whether PRA, in an absolute sense,
presents the plant and what is happening there. The question is, is a
risk-informed system better than a purely deterministic system.
Of course, a purely deterministic system is being accused as
being overly burdensome and so on and you are bothered by the fact that
they are trying to remove the burden.
But let's not forget that for the last 20 years, PRA has
identified holes in that system and the staff was very willing to fill
them, the station blackout rule and so on. So it plays both ways,
really.
Now, the basic argument that PRA should be realistic for an
accident, or whatever the word is, representation of what is going on is
certainly true and the findings by AEOD and so on are disturbing, there
is no question about it. Of course, you have to evaluate them and all
that.
But you can't argue about that. If you use a tool, it has
to be a valid tool and, in fact, all the discussion on the performance
importance measures earlier had that motivation.
In fact, I believe that one of the reasons we have to make
sure that the importance measures have a sound basis, we all understand
the limitations, is public confidence. It's not just a matter of how
accurate PRA is.
MR. RICCIO: And my other concern goes to the fact that
basically I don't believe that the design basis is where the industry
believes it is, and I don't believe that the programs NRC has put in
place to force this industry to take a look at the problems with the
design basis are going to rectify that situation.
You're actually requiring that someone within the line of --
or at the utility fall on their sword to basically come up with
something that's going to basically put the operation of the reactor at
risk under this new regulatory program for preparing the design basis.
If you look at what just happened at Indian Point, those
folks swore under oath or affirmation that they had a program in place
that would detect and repair design basis problems, yet the reason the
AIT was sent up there is basically because of design basis problems.
DR. WALLIS: But you're taking on PRA and I'm very
surprised, because your argument on the final page here is you want to
reduce the probability of meltdowns. You want it to be less likely than
--
MR. RICCIO: The only reason I'm here.
DR. WALLIS: Well, what is PRA but a way of logically, with
the best tools you have, trying to predict whether or not -- what's the
probability of a meltdown. So if there is a better tool for assessing
what it is you're trying to bring about, please tell us. The PRA is the
tool we have to do exactly what you want to achieve.
MR. RICCIO: I believe it's a good tool, but I believe it's
taking on added significance in this deregulatory effort and,
unfortunately, I saw how it was used on Farley.
DR. WALLIS: What's better? Is there something better that
you know about that should be used?
MR. RICCIO: I think we should still be sticking with the
Part 50 regulations that exist.
DR. WALLIS: So you'd rather have the guess work
deterministic judgment type calls than the logical based on analysis
calls of PRA.
MR. RICCIO: I think you might get some -- I'm not going to
even respond to that one.
DR. WALLIS: Well --
MR. RICCIO: I think we have a regulatory system that has
said, up to this point, that has at least precluded this industry from
melting down another reactor in the last 20 years. You're shifting
regulatory horses in midstream and the optional nature of this entire
process is going to end up with a regulatory morass that is basically
what we've been trying to get out of.
You're going to have half the reactors operating under
regular Part 50, you're going to have the other half operating under a
new revised version of a PRA-informed Part 50, and if you look at -- I'm
sorry, but if you look at the Farley plant, I really believe that was a
bad move on the part of this agency.
And you're basically ignoring the DPO that says if you crack
as few as ten tubes, you're going to deplete your inventory. I just --
I have very serious misgivings about where this agency is heading with
this.
DR. WALLIS: Are you criticizing the way in which the PRA is
being used or are you criticizing -- when you say PRA should not be the
basis of regulation, it seems to me that to bring about your objective
of reducing the probability of core damage, there has to be a measure of
that and PRA is the measure of that.
So I can't see how logically you can say that PRA, per se,
is a bad thing. Now, you can criticize how it's used, but it intends to
do exactly what you want to happen.
MR. RICCIO: I'm not saying it's a bad thing. I'm saying
it's inaccurate information in some regards. It doesn't reflect reality
in some regards. One-times-ten-to-the-negative-six does not equal two
in 2,500.
I believe it's a good tool. I believe it's a tool that
could be used to tell the agency where to look. But when it becomes the
basis for regulation, the basis for not inspecting steam generator
tubes, for instance, I think it's being used improperly.
DR. APOSTOLAKIS: Can we move on now? Are there any other
questions on a different subject? Thank you very much.
MR. RICCIO: Thank you.
DR. APOSTOLAKIS: Mr. Chairman, we have a problem now. We
are already 18 minutes late. But we have not discussed option three at
all.
One idea is not to discuss it and just read what you have.
Another idea is to give three minutes, no more than three minutes to Tom
King to discuss the essence of it, which is on page four of the
viewgraphs you have, and the other one is just to recess right now.
MR. SIEBER: That would be option three.
DR. WALLIS: Which option do you recommend?
DR. APOSTOLAKIS: That's option four.
DR. POWERS: We are imposing on other speakers and we're not
going --
DR. APOSTOLAKIS: Well, Tom has agreed that just discussing
page four will give him an opportunity to give us his main message,
because a lot of it, as you see, is schedule and why are we doing it and
this and that. I don't think we need to get into that.
But if we put the basic structure of it on the screen and
the members have any questions, then it will give Tom an opportunity to
respond.
So shall we put him up there?
DR. POWERS: I don't think you can meet the schedule, no.
DR. APOSTOLAKIS: So what do we do?
DR. POWERS: I'm going to recess for ten minutes. Recess
till a quarter after.
[Recess.]
DR. POWERS: Let's come back into session. Our next
subject is a proposed resolution of the generic safety issue GSI-148,
smoke control and manual fire-fighting effectiveness. One of the two of
us, either John Barton or myself, is in charge of this, so I guess I'll
ask John to lead the way.
MR. BARTON: Thank you, Dana. The purpose of this session
is to hear presentations and hold discussions with representatives of
the staff regarding their proposed resolution of GSI-148, titled smoke
control and manual fire-fighting effectiveness.
A fire risk scoping study was initiated in 1987 in order to
identify the risk issues that were not previously addressed in PRAs.
GSI-148 was raised in SECY-89-170, fire risk scoping study, summary of
results and proposed staff actions.
GSI-148 was prioritized and classified as a licensing issue,
which means it's not being handled as a generic issue. The safety
significance will vary probably from plant to plant and it is unlikely
that any cost-effective generic resolution could not be achieved --
could be identified.
So, therefore, plant-specific reviews is the way the staff
is approaching this and some reviews have already been completed as part
of the IPEEE program, which is the way licensees are addressing this
issue.
The committee has requested us to prepare a report based on
this topic.
At this time, I will turn it over to Alan Rubin from the
staff, who will lead the discussion.
MR. RUBIN: Thank you. Before I begin, I think Tom King
wants to make some introductory remarks on this generic issue process
and where this generic issue fits in.
MR. BARTON: All right. Tom?
MR. KING: Yes. I guess there may be some question as to
why this was called a licensing issue, and licensing issues generally
don't come to the committee for review and requesting a letter before we
actually close them out.
DR. POWERS: Is it common for licensing issues to have RES
as the lead on them?
MR. KING: Now it is. In the past, licensing issues that
would be common, NRR would have the lead on them. But there aren't many
of those left. In fact, this may even be the last one. I haven't
checked lately. But basically, when this issue was looked at, it was
identified back, I think, in the early '90s, came out of the fire risk
scoping study.
It was given a number 148, was looked at, and it was
realized that the elements of this, most of them were being picked up
elsewhere, either covered by another generic issue or were going to be
dealt with in the IPEEE, and the thing -- because they were very
plant-specific issues.
And the thing the staff needed to do was develop guidance as
to, if this is in the IPEEE, how do we review the submittal and make
some judgment on whether 148 is adequately dealt with or not.
And since the staff was charged with developing the guidance
and that was what really was going to close the issue out, it was called
a licensing issue. So when we closed the issue out, I signed out a
package not too long ago that actually closed the issue out. What we
wanted to do was formally document what that guidance is and what we're
looking for in the individual IPEEE reviews to deal with 148.
What Alan is going to talk about is the elements of 148,
where they're being dealt with, and specifically the guidance that we're
using in looking at 148 as we go through the IPEEE process.
I know in the SERs that have been issued, each one
specifically addressed 148 to close it out and we'll continue to do that
on a plant-specific basis.
So that's all I wanted to say. With that, I'll let Alan
start the presentation.
MR. RUBIN: Start it? I think with your introductory
remarks, you finished it for me.
Good afternoon. My name is Alan Rubin, for those of you who
may not have heard from me before. I'm going to talk about the
resolution of Generic Issue 148, smoke control and manual fire-fighting
effectiveness.
The outline of my presentation, I'm going to cover the scope
of the generic issue in terms of its safety significance. I'll talk
about some of the background. You've heard a very brief summary before
I began my remarks. I'll talk about some of the significant documents
that have been completed that relate to this issue, going back to 1989
to the present.
I will talk about the closeout of the issue; also discuss
some ongoing research activities. These include the plant-specific
IPEEE reviews, as well as ongoing research on the effects of smoke.
Then I will reach some conclusions -- present some conclusions.
Let me first talk about the scope of the issue. An issue
was defined, Generic Issue 148. It covered four areas which potentially
could have an impact on plant risk. The first of those is that smoke
can reduce manual fire-fighting effectiveness. That's the title of the
issue that came out of the fire risk scoping study issues. An example
is that misdirected suppression could potentially damage equipment that
was not involved in the fire.
The second issue had to do with the effect of smoke and
potential damage that electronic equipment -- potential damage to
electronic equipment from smoke, that could either cause equipment to be
lost or potentially cause spurious response from the equipment.
The third area is can smoke hamper an operator's ability to
safely shut down the plant by causing evacuation of control centers.
For example, the control room in particular. Also, could smoke inhibit
operator actions elsewhere in the plant.
The fourth aspect of the issue was that smoke could
potentially initiate automatic fire protection system actuation in areas
away from the fire and potentially damage safety equipment.
I will get into each of the four of these, but let me just,
right up-front, tell you that item number one and three on smoke effect
on manual fire-fighting effectiveness and smoke's potential to inhibit
an operator's ability and abandon control centers is covered in the
plant-specific IPEEE program.
Item number two, the effect of smoke on electronic
equipment, is being covered in an ongoing research program. Item number
four, the actuation of fire protection systems, this part of Generic
Issue 148 was actually incorporated into a different generic issue that
covers automatic system actuation. That's a little broader than the
fire protection issue itself, but includes seismic actuation of fire
protection systems.
MR. BARTON: Alan, let me ask a question. You're trying to
close out this issue on a plant by plant basis, on IPEEE submittals, on
internal fires and how licensees are handling it.
One of the concerns with this issue is the effects of smoke
during manual fire-fighting and issue number two up there talks about
the impact of smoke on electric equipment, electronic equipment. You've
got a research project ongoing, which has not come to some conclusion, I
guess, at this point and it's still ongoing.
How can we close out this issue on a plant by plant basis
with something like this still hanging out there with unknowns? I guess
that's the problem I've got with the approach that you're trying to take
on closing this issue. That's an example.
MR. RUBIN: The reason that that aspect of this generic
issue was not covered in the IPEEE, getting ahead of the slides, but
let's take the question, because that is certainly a key question here,
is because of a lack of data, lack of available data to analyze and
understand and be able to do an assessment of the impact of smoke on
electric equipment.
There are ongoing research programs in that area, I'll get
into them, but I'll just briefly mention, one is on the effect of smoke
on I&C, digital I&C equipment. Another is the effect of smoke that's
being covered in the fire risk research program.
Part of that --
MR. BARTON: Is where?
MR. RUBIN: In the fire risk research assessment program.
You'll hear about that program in more detail in the future, and I'll
get into it in my slide.
MR. BARTON: Okay.
MR. RUBIN: But basically, of these four items, what I'm
saying is other than the effect of smoke on equipment, all of those
items are covered in the IPEEE program, in the plant-specific reviews.
Because there is lack of data, and it's a state-of-the-art
question, it's difficult to tell utilities to assess something when
there is not sufficient data and we wouldn't have the ability or
sufficient data to even review that.
So that's why this ongoing research program was undertaken,
in part, and also from user needs from NRR, as well. And if it turns
out that once we get that data from the research program and there is an
assessment of the potential risk impact, it's possible that this could
continue to be something that would need to be looked at on a generic
basis and perhaps another issue opened up.
But as far as 148, the basis for closing 148 is as I've
discussed.
DR. WALLIS: Is it only electronic equipment affected by
smoke?
MR. RUBIN: As part of the scope of this issue, that's what
is in Generic Issue 148. Is it only electronic equipment? Primarily.
I don't expect --
DR. WALLIS: It is primarily.
MR. RUBIN: Primarily, yes. I think that was brought up in
the fire --
DR. WALLIS: I guess that's the most obvious.
MR. RUBIN: Yes, the most obvious one.
DR. WALLIS: But there could be, depending on what's in the
smoke, effects on other things besides electronic equipment.
MR. RUBIN: And some of the issues with the effect of smoke,
it might not be an immediate impact. It might be corrosive effects that
could be over a longer period of time, for example, where smoke would
damage equipment and soot, so that it might not be an impairment of
equipment in a time where you're recovering from the fire.
So there are certain aspects of that issue that are
questions that are being addressed in the research program and once we
have the answers to those questions, we'll be able to better identify
the risk significance of that part of this generic issue.
It might help a little just to go over the background, how
we got to this point today and where we are, and I will just briefly
discuss this.
It was already mentioned that the Sandia report on fire risk
scoping study issues that was issued in 1989 identified six issues that
were potentially significant contributors to core damage frequency.
One of those was the manual fire-fighting effectiveness, and
that aspect of the fire risk scoping study issue was incorporated into a
piece of Generic Issue 148.
That included a time to respond to the fire, considering the
effect of smoke. One of the issues in the fire risk scoping study
that's of significance was the fire brigade effectiveness and the
importance of the fire brigade. It was seen that various plants, from
plant to plant, there were differences in the fire brigade training and
staffing and equipment that was available. That was part of this issue.
MR. BARTON: Has that been addressed in licensees'
submittals?
MR. RUBIN: Yes.
MR. BARTON: In enough detail that you can assess that the
training they've done and the scenarios that they've run adequately --
MR. RUBIN: Well, in other presentations, I've gone into
some of the aspects of the scope of the IPEEE reviews and the objective,
and we do not have the resources nor is it our intent to detail -- to do
the detailed review or verify or validate the calculations and analyses
that the licensees have done.
We're looking at the process, we're looking to see whether
there was a reasonableness and completeness of the licensee's IPEEE,
such that they could identify potential vulnerabilities and make fixes,
if necessary, to reduce those vulnerabilities.
DR. POWERS: Do licensees generally use your suggested
response times?
MR. RUBIN: If they don't, when we look over and -- again,
I'm getting ahead of myself a little bit, at least the presentation, but
that's okay. We've asked a number of requests for additional
information on these IPEEE reviews, and that's one of the slides later
on, particularly in the effect of smoke and fire-fighting.
And sometimes we ask questions because there is not complete
information or sufficient information in the IPEEEs themselves.
Sometimes we ask a question, related to your question, Dr. Powers, that
they've used optimistic assumptions in their analysis. They're stated
in the IPEEE and we just think they are optimistic; in fact, too
optimistic, and we've gone out and asked for responses to information.
An example of that would be related to this issue, some
IPEEEs have the assumption that the main control room would not be
abandoned. Manual fire-fighting would be effective in putting out a
fire in the main control room and it would not be abandoned.
Yet, we've seen on many licensees' other analyses that the
main control room is a significant risk contributor, one of the dominant
areas in fire risk. We've asked questions, what is the basis for that
assumption, was there optimistic credit for manual fire-fighting, where
are the detectors located in the main control room, are they in the
cabinets, are they not in the cabinets.
So we do go into that. Those are significant issues and we
ask questions and we try to get an adequate response, and I'll go into
some other examples of the RAIs that we've asked in this area.
We don't just necessarily take the original submittal and
say it's okay. We have actively asked RAIs and I think Fred Emerson,
who is sitting in the audience, from NEI, has seen these RAIs. We've
had some generic requests for information on the IPEEE program and we do
try to cover those aspects.
DR. POWERS: As I recall, I don't have the document here at
hand, so I can't pull it out and refresh my memory on which one it is,
as I recall, in some of the material sent to the licensees, there were
some recommended or suggested response times for the manual
fire-fighting crew, fire brigade.
And for the life of me, the best I could figure out, those
were somebody's wild guess. There was no technical basis for them,
other than somebody thought that they were good times.
MR. BARTON: YOu're talking about the curves?
DR. POWERS: Yes.
MR. RUBIN: I think the response curves that were included
in the guidance.
MR. BARTON: Typical probability of manual suppression for
various fires.
DR. POWERS: As far as I can tell, they were based on some
other computer code or some other analysis and they just adopted them.
MR. RUBIN: My understanding, it's based on some experience
data, but if I could turn to Nathan Siu to perhaps give a little more
insight on that aspect and those particular curves and recovery times.
MR. SIU: The IPEEEs varied in how they treated time to fire
suppression. Some of them did use historical database, which typically
they'd use in a generic manner. And we have quite a few events where we
have recorded suppression times.
I mean, we don't have many hundreds of events, but we have
several, tens to maybe a little over a hundred as of a little while ago,
and we might have more by now.
Others just simply looked at how quickly the fire brigade
responded to an alarm and that's typically the kind of situation where
we would ask a question, because obviously suppression involves more
than just responding to the alarm.
But certainly in the cases where they used those curves that
you've seen in the reports, those are far more than educated guesses.
Now, you can argue whether it's appropriate to use generic data for
plant-specific analysis, but that's --
MR. BARTON: That's the question I've got.
MR. RUBIN: And sometimes the licensees have used in their
analyses the time to extinguish a fire as the time for their fire
brigade or their response time from fire brigade training, and we've
asked questions, because the time to respond to the fire is not --
MR. BARTON: Does fire brigade training include going into
areas that are filled with smoke?
MR. RUBIN: In terms of typical fire brigade, I don't think
so. Steve, do you -- Steve, do you have a response? Steve, from NRR,
who is section chief in the fire --
MR. WEST: They typically would have some training that
requires that type of activity, where they do the hands-on training.
Not the classroom training, but the field training where they dress out,
could go to some fire academy and do training, go into smoke-filled
environments, use the equipment.
That's something NRR or the regions would normally verify
during an inspection.
DR. WALLIS: Well, if it's smoky enough, then they cannot go
in, can they?
MR. WEST: No. They put on the training breathing
apparatus.
MR. BARTON: They've all got FCBAs on.
DR. WALLIS: They still have to see.
MR. WEST: They feel.
DR. WALLIS: They feel? That's a pretty ineffective way of
doing anything.
MR. WEST: Trust me, fire-fighters do a lot of things you
may not believe or you may not feel comfortable with, but they do them.
In fact, one of the exercises is to take scuba gear, put it on, be fully
equipped, and then they put a blackout mask over the scuba mask and you
have to go into compartments and find things and bring them out safely.
So that's typical of your volunteer fire department down the
street and the fire brigades at nuclear power plants.
DR. WALLIS: But in a nuclear power plant, you're not
interested in going in and taking out things safely, really. You're
just protecting the key parts of the system.
MR. WEST: I believe they're also trained to rescue
personnel and operators that could be in the plants.
DR. WALLIS: That's important, sure, but who's controlling
response of the reactor systems and so on by feeling around?
DR. POWERS: The operators.
DR. WALLIS: The operators are still somewhere and they're
still operating.
MR. WEST: That's right.
DR. POWERS: The power brigade cannot include the operators
on duty at the time.
MR. RUBIN: Let me mention two other aspects related to the
fire brigade training, because that is an important part of the issue.
First is that there is an industry guidance document that EPRI put out,
called Fire Induced Vulnerability Evaluation 5 Methodology, which
includes in there guidelines on an adequate fire protection program --
I'm sorry -- fire brigade training, staffing and equipment.
That was put out for licensees to use to perform their IPEEE
analyses at the plant.
The second point is that NRR, as part of their fire
protection inspection program and triennial team inspections does
include actual review of the fire brigade drills as part of their normal
review process.
DR. WALLIS: It's very difficult to see how some of this is
related to CDF, these pictures of firemen feeling around in the dark
does not have any effect on CDF, does it?
MR. RUBIN: That level of detail would not necessarily be in
the CDF, but that's why they're looking at this response time.
MR. SIU: Let me try to respond to that. The way the fire
risk analysis incorporates this issue, it tries to model the component
damage as the outcome of a competition between the fire growth and the
fire suppression processes.
So fire suppression takes too long and you've got a fire
there that can damage a component, then it will do that, with obviously
some probability that the growth beats suppression.
So for those instances, and, again, I'll say this is done on
a generic basis, for those instances where, for whatever reason, fire
suppression was delayed, that factors into that curve that you saw. So
very long suppression times could very well be because there was a lot
of smoke, it could be because of other reasons.
I will also point out that the dominant risk contributors
typically are from scenarios that develop fairly quickly. So as long as
the response time is longer, then let's pick a number, 15-20 minutes, it
doesn't matter. If it's greatly delayed, you haven't beaten the fire.
MR. RUBIN: I won't dwell on this last bullet, but in 1989,
the staff issued a paper to the Commission that talked about how the
fire risk scoping studies would be addressed and basically said the
utilities would address these issues on a plant-specific basis and look
for vulnerabilities as part of their IPEEE program.
In 1992, the staff did a prioritization of Generic Issue 148
and prepared a memo and provided a copy of that to the ACRS, as well.
The conclusions of that prioritization were that the safety significance
of this issue would likely vary significantly from plant to plant.
In fact, it was very unlikely that we would be able to find
a cost-effective generic resolution for this issue. And plant-specific
reviews were going on on the IPEEE program.
But it was felt that there was a need to develop the
guidance for the staff review of the licensees' IPEEE submittals, and,
therefore, as Tom mentioned, this was classified as a licensing issue.
As a note, NUREG-0933, which documents the staff's status of
generic issues was updated in 1992 to reflect the prioritization memo
that was done to include the status of Generic Issue 148.
In 1992, the staff issued supplement four to Generic Letter
88-20, which was requesting licensees do an IPEEE evaluation, and it
also issued -- published a report, NUREG-1407, that provided
supplemental guidance for the conduct of the IPEEE itself.
Both of those documents specifically requested that
licensees address the fire risk scoping study issues.
Subsequent to that, the staff developed guidance for both
staff and contractor reviews for reviewing the IPEEE submittals, and
that includes Generic Issue 148. And when I get to the next slide, I'll
discuss some of that staff guidance.
DR. POWERS: When you asked the licensees to address fire
risk scoping study issues, you mean to address those issues other than
those that were to the effect that there was no data?
MR. RUBIN: I don't think it was that specific at the time.
The guidance that we have put out, though, for the staff reviews and
contractor reviews is explicit. It says that the data is inadequate for
looking at the effect of smoke on electronic equipment and, therefore,
that is not covered in the plant-specific reviews. But I don't think
there was that level of detail in the generic letter or NUREG-1407.
That's when we went into the details in the guidance and
that's why that was an important point.
So then recently, as Tom mentioned, in July of this year, we
issued a closeout memo of Generic Issue 148 and the conclusions for that
was that the review guidance has been developed for the reviews of the
Generic Issue 148. There were plant-specific reviews ongoing in the
IPEEE program and no further generic action was necessary.
MR. BARTON: In that document that Tom put out, it talks
about potential safety significance and there is some discussion on
smoke impacts, et cetera, et cetera.
But then the next section, which talks about staff review
guidance for the submittals, it doesn't address the smoke question, or
else I don't see it in here.
MR. RUBIN: It's not in there. That's probably, I'd say, an
oversight and we're clarifying that in our presentation today. It was
not explicitly discussed in there, but in terms of the prioritization of
the issue, the write-up in NUREG-0933 is explicit in there. It says
that the effect of smoke on equipment. It's just not information,
state-of-the-art is not sufficient to address that issue, and there is
an ongoing research program. That memo did not repeat that again.
DR. POWERS: It's just a little confusing. It says -- the
King to Thadani letter, it says no further generic action is needed, yet
there is generic action that's needed and apparently in place.
MR. RUBIN: It's the research program which is in place.
DR. WALLIS: So it's a re-definition of things or a
re-classification of things or what?
MR. RUBIN: It says closing out the generic issue, but with
a piece of it moved to the research program.
MR. KING: Clearly this is a follow-up item that has to be
done, continue the smoke testing, see what the risk significance is, see
if we need to do anything about it. I probably should have mentioned
that in my closeout letter. I didn't do that.
Whether it's significant enough we give it a new number and
then call it a new generic issue or not remains to be seen, but until we
get the research program done, we can't make that call.
If you're worried about it slipping through the cracks, I
think we can do something to make sure that doesn't happen. Maybe issue
a modified letter or something to make sure it's on the record that this
work still needs to be done.
But we didn't view holding up closeout of 148 as something
that we wanted to do and wait another couple of years till all the
research was done and analyzed.
MR. BARTON: Somehow, Tom, I think you need to make that
clear, because in your memo, you get the impression that, yes, smoke is
important, et cetera, but yet if I'm a reviewer looking at submittals, I
don't know what I would do about looking at licensees' submittals and
how they're addressing the smoke issue, other than the one that's in
research now, which is effects on electronic equipment.
I don't know what I'm supposed to do. Now, how can I close
out this issue, which I understand you're saying is I've kind of got
everything covered, because I should be looking at smoke-related issues
other than what's in research, and I'm also looking at the manual
fire-fighting that licensees are doing and they're telling me how
they're doing that in a submittal, based on I don't want to close out
this issue.
I think that's what you're saying. I don't think your memo
really conveys that message, at least not to me.
MR. KING: I agree. We ought to go on record that there's a
piece of this that's still ongoing. It's dependent upon a research
program and at some point we're going to have to come back and figure
out what does that research tell us and do we or do we not need to do
something.
DR. WALLIS: That is what a GI is, though, isn't it?
MR. KING: Not necessarily. I mean, a GI, you really need
to have enough information to say it's a safety issue and, therefore,
we're going to do something about it. I think at this point, until the
research is done, I don't think we can quantify the safety significance.
MR. RUBIN: And it gets prioritized as a high, medium, low,
or drop issue, in terms of safety significance or potential safety
significance.
DR. POWERS: Let me ask a little more about this research
program. The issue that came up with, gee, we have this smoke, smoke
goes up, and smoke goes out and it goes someplace other than where the
fire is and it can affect electronic equipment there.
Then in the IPEEE submittals and the background material
that was sent to the licensees, it says, gee, you can use the Comp Burn
code to do your analyses, but recognize that it has some shortcomings to
it; among those is it can't predict smoke transport.
Is that part of this research program that you've got here,
to do the smoke transport part?
MR. KING: Yes. There's two pieces. There is the testing
on equipment, and then I'll let Nathan talk about his program, which is
improving the fire modeling and how that's dealing with smoke.
MR. SIU: We do have a task and Alan, I guess it wasn't
going to be on this particular slide. There is a task, if you look in
the research plan, I think it's number 11, talks about fire modeling
limitations.
Where it talks about the limitations of the models, it
doesn't -- we don't have a component looking at developing better models
for generation and transport of smoke. Most of the work being done
right now looks at the back end, what's the effect of smoke on
equipment, and, also, and this is where we hope to get some of the risk
implications, what kind of smoke levels does it take to get these kinds
of effects.
If the smoke levels required are very, very high, in other
words, you have to be really close to the fire, then maybe it's not as
big of an issue as we're concerned about right now, but we don't know
that.
DR. POWERS: It seems to me that -- maybe this is not the
place to -- it would seem to me that if I was looking at smoke effects
on electronic equipment, there would be a time concentration curve; that
if I had relatively low levels of smoke, it might take a long time
before I manifested any effect of that smoke.
MR. SIU: True.
DR. POWERS: Whereas at high concentrations, that
manifestation may take place very quickly.
MR. SIU: That's true, but the other time scale, of course,
you have to think about is the reactor shutdown and as Alan pointed out,
things like corrosion, for example, we don't worry about so much because
that's well beyond the time scale of shutdown.
In the case of what you're talking about, it might be that
there's some long developing scenarios where that could be an effect,
and now we have to invoke the suppression process and whether that puts
the fire out before you get to those levels.
But, again, we don't know what levels we should be concerned
about at this point, I think.
DR. POWERS: So your strategy is you're going to find out
what concentration levels are and then if you need to, worry about smoke
transport.
MR. SIU: That's where we're heading right now on the fire
risk assessment part. Now, John Calvert might want to talk about the
digital I&C part, or, Alan, maybe you'd just continue on and get there.
DR. POWERS: Yes. I think we need to move right along on
this and try to get through things.
MR. RUBIN: This slide provides some additional background
information for some of the questions that we've already talked about,
but I, in fact, may have covered it already.
But I was going to discuss the four aspects of the generic
issue and how we consider that they're being addressed in the closeout
of Generic Issue 148.
Guidance has been developed for the first item, which goes
back to slide three, which is manual fire-fighting effectiveness. The
major aspects of that guidance was to look at the licensee's submittal
and see if it did a reasonable treatment or included a reasonable
analysis of the following questions; and if they didn't provide a
reasonable treatment in the original submittal, we asked for additional
information.
One of those was related to the licensee's analysis for
modeling of the plant fire brigade and their ability to respond to and
also extinguish the fires before significant damage occurs. Another was
how they treated smoke and how that could potentially hamper the
fire-fighters' effectiveness.
Another area was in the modeling of the fire brigade
response time, the question that came up earlier, as well as the time to
extinguish the fire. An important point is that the time to respond to
the fire is not equal to the time to extinguish the fire. We found that
some submittals included a fairly optimistic response and extinguishing
time in their analyses, and we sent some questions.
MR. BARTON: Alan, just so I can get a feel for this, how
many of these submittals have you reviewed and have satisfied yourself
that the licensee has adequately addressed the smoke issue in the
submittal?
MR. RUBIN: There are 70 IPEEE submittals total. We have
completed preliminary reviews of all of them. We've completed final
reviews, which means issuing the safety evaluation report, on 26.
MR. BARTON: So about a third.
MR. RUBIN: About a third. Of those 26 that have been
completed and we've written SERs, 22 of those we have found have
resolved on a plant-specific basis Generic Issue 148, with the exclusion
of the smoke impacts on electronic equipment.
DR. WALLIS: I have a question on --
MR. RUBIN: Most of the -- they have to go through the RAI
process and review the submittals, and most of the plant-specific
resolution or most of the plants have resolved this issue, and I'll get
into that a little bit later on as to how we've reached that conclusion,
what the logic is.
DR. WALLIS: I have a question, which may be generic, about
the whole review process. These licensees submit these things and then
you are asked --
MR. RUBIN: There are some of them that are pretty thick
things.
DR. WALLIS: Where if you were asked the question was this a
reasonable treatment and analysis of all these different things, how
much expertise does the reviewer have to have in this particular subject
or experience with fires to do this review? It seems to me that the NRC
has to make sure that its reviewers are savvy enough, experienced enough
that they stay ahead of what industry is going to throw at them in terms
of what appear to be reasonable analyses.
MR. RUBIN: That's a good point, and I --
DR. WALLIS: How do you assure that your reviewers are not
just people who look at it and say, yes, I saw the thing, that looks
okay, or are they really savvy enough to go in and say I know that they
ought to have investigated this and they didn't.
MR. RUBIN: I'll answer that question. I happen to make
presentations on the overall IPEEE program to some of the subcommittee
and full committee meetings and have gone into the process. But
primarily, we do have experts knowledgeable in PRA and fire areas.
First of all, we have contractors doing a preliminary review
in the fire areas. That's the Sandia National Laboratories. We have,
in addition to that, a senior review board, which includes experts from
both the staff and Research and NRR, as well as contractors, that have
detailed knowledge of many aspects, all aspects of the fire program.
In fact, for example, some of the data, the test data that
licensees have used to support some of their analyses are based on
Sandia tests. Well, we have Sandia on our senior review board, who have
actually been the ones conducting those tests. And if test data are
misinterpreted, we know that. I think we have a very strong team, a
review team in the fire area.
So I'm very pleased with the way things have been going in
terms of our ability to understand what's been done. But, again, on the
other hand, we don't review the details to try and verify the
quantitative numbers, for example, in the PRA, quantitative results.
DR. WALLIS: What assurance do you have that these
quantitative numbers are okay? Does someone do a sort of spot-check
here and there?
MR. RUBIN: If something is way out of whack or way out of
line with other plants, raises our eyebrows, if we see some things that
are very unusual in terms of either an omission of a significant area,
raises our eyebrows, we have some questions.
So there's a balance of knowledge and what information we
have from other submittals; is this plant similar, do they have a core
damage frequency that is either so high or so low that we need to really
pursue that further, and we've done that. We've looked at outliers for
plants, for example, that have come in with a ten-to-the-minus-ninth for
their original total core damage frequency from fires.
And on the other extreme, Quad Cities came in with a
five-times-ten-to-the-minus-three core damage frequency initially. So
we look at these outliers quite --
DR. WALLIS: Did they say wow when it was done?
MR. RUBIN: We said wow, they said wow. There's been a lot
going on in that area. We talked about that at ACRS. Most recently,
Commonwealth Edison resubmitted their IPEEE analysis for Quad Cities and
came in with about two orders of magnitude lower estimates, did a much
more detailed evaluation of the plant.
DR. POWERS: Alan, maybe you'd better progress. I'd like to
close this out at four.
MR. RUBIN: Let me close out. I think I touched on this
before, I'll just do this for completeness. Items two, three and four,
the effects of smoke are not covered in the IPEEE plant-specific
reviews, but there is a research program going on.
We do cover the main control room abandonment and the
operator recovery actions in the IPEEEs, and smoke initiation of the
automatic fire protection system is covered in a separate generic issue
and also a plant-specific basis in the IPEEEs.
MR. BARTON: Does that issue close out GSI-57?
MR. RUBIN: GSI-57 is closed out in 1993, and a memo was
sent to ACRS and ACRS agreed.
MR. BARTON: Okay. Thank you.
MR. RUBIN: I don't know if you want to get into this,
considering the amount of time, but this is actually an example of some
of the RAIs that we've seen if we needed additional information.
We've sent one or more of the questions listed here on this
slide to about 70 percent of the licensees, either because we didn't get
information, as I said, or we felt or our experts felt that the
assumptions were overly optimistic.
The first three questions were the most commonly asked, and
that related to the impact of smoke, as well as other effects, such as
heat and loss of lighting on human error probabilities.
We looked at -- we asked questions on the timing and
reliability of the fire brigade and operator recovery actions. We
looked if they were overly optimistic for manual suppression before
abandonment of the main control room, as well as the other issues listed
on this slide.
So the result was, in response to these RAIs, by the way,
typically, we'd get the additional information and sometimes there would
be some higher core damage frequency estimates after some re-analysis,
but typically there was not a major change in the total core damage
frequency from fires for the plant, and there was not a relative change
in what the dominant areas or sequences were from fires.
But at least we got the feel it was an adequate treatment,
that the plant could identify vulnerabilities, their process, licensing
process was capable of identifying potential vulnerabilities, and then
either no vulnerability was identified or if there were vulnerabilities,
then plant improvements were made to reduce that vulnerability. And
that's how the issue was resolved on a plant-specific basis.
I think we've touched on this question enough, what the
ongoing research program -- smoke effects on digital I&C, which is being
done in the Division of Engineering Technology. That work is nearing
completion. The fire modeling area, smoke effects on medium voltage
switch gear. Very important is this bullet on the risk implications,
past and current smoke experiments, and information including the
digital I&C experiments. That work is expected to be written up at the
end of November.
And the risk fire research program is also looking at the
experience from major fires, both in the US and abroad at nuclear
plants, and includes their effect of smoke and a report is expected at
the end of this year, at the end of 1999.
So let me conclude. I think I've said some of these things
before, maybe once or twice, but I think that's a good practice, to tell
you what you're going to say, say it, and say it again.
So the basis for the closeout of Generic Issue 148 was that
the IPEEE review guidance has been developed. Plant-specific reviews
are ongoing and there is no further generic action needed.
The ongoing activities are in the IPEEE plant-specific
reviews themselves. I've already responded to the question that we've
completed 26 of the reviews and have resolved the issue for most of
those plants. There is a research program going on looking at the
effect of smoke, and the NRR program of inspections, looking at the
adequacy of the fire brigade and fire drills.
In the future, once the IPEEE plant-specific are user
completed, we'll issue a final insights report and that report will
address the reviews of the plant-specific resolutions of Generic Issue
148, as well as other generic issues that are being covered in the IPEEE
program.
DR. WALLIS: These effects of smoke, there are all kinds of
smoke, it seems to me. So it's a pretty broad program. It depends on
what is burning and how it's burning.
MR. RUBIN: Fuel oil, cables, insulation.
MR. KING: It's cable insulation mostly, cables that are in
electrical cabinets or power cables. That is really what's been used
for the experiments.
DR. POWERS: There is, in fact, a fairly bold statement in
one of your documents that cables and lubricating oils are the
predominant heat, sources of heat in a fire within the plant.
I found that remarkable that you could make that statement,
because I think temporary trash piles and things like that would be a
more formidable danger than the smoke from cable insulation fires.
DR. KRESS: Are those generally there during short periods?
DR. POWERS: They're supposed to be controlled, but if you
do have a fire with them and there is likely -- they can put pretty
heavy heat loads on relative to burning cables; welds are probably
pretty high, too.
MR. RUBIN: Dr. Powers, in response to that, one of the
areas where there have been improvements as a result of the IPEEE
program has been --
DR. POWERS: Yes, transient control.
MR. RUBIN: Control of transient combustibles in an area
that turns out to be a high risk contributor. In fact, we've done some
site audits and we've seen some areas where the licensees have
instituted stricter controls on transient combustibles.
DR. POWERS: But even our own episodic visits to plants, they have gone
out of their way to assure us that as a result of their IPEEE walk-downs
and whatnot, they've had to institute greater controls on transient
combustibles. This is a problem, because there would be a lot of heat
load, and in unusual places, as well.
MR. BARTON: Any other questions of Alan? If not, thank
you, Alan and Tom.
MR. RUBIN: Thank you.
MR. BARTON: And I'll turn this back to Dana.
DR. POWERS: Before we leave the subject, I'd ask Fred
Emerson from NEI if he had any thoughts or comments on this.
MR. EMERSON: This is Fred Emerson, from NEI. We are
observing with interest the research program to see what effect smoke
has on electronic equipment.
I think one thing to keep in mind here is that manual
fire-fighting is only one piece of one element of defense-in-depth, and
that the controls that a plant puts in place to manage fire risk include
combustible controls, as was just mentioned, fire brigade effectiveness,
assuring the reliability of detectors and automatic sprinkler systems,
and taking action to limit the impact of damage following a fire, if one
were to get out of control.
So by balancing these elements of defense-in-depth,
certainly smoke is an important aspect, but there are other factors that
a plant tries to implement to make sure that the risk of fire is managed
successfully.
Thank you.
MR. BARTON: Thank you.
DR. POWERS: Thanks, Fred. Okay. I want to close that
issue and thank everyone very much. We can go off the transcript at
this point.
[Whereupon, at 4:07 p.m., the meeting was recessed, to reconvene at 8:30
a.m., Friday, October 1, 1999.]
Page Last Reviewed/Updated Tuesday, July 12, 2016