459th Meeting - February 3, 1999
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
459TH ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
U.S. Nuclear Regulatory Commission
2 White Flint North, Conf. Rm. 2B3
11545 Rockville Pike
Wednesday, February 3, 1999
The committee met, pursuant to notice, at 8:30 a.m.
DANA POWERS, Chairman, ACRS
GEORGE APOSTOLAKIS, Vice-Chairman, ACRS
WILLIAM J. SHACK, Member, ACRS
ROBERT E. UHRIG, Member, ACRS
MARIO V. BONACA, Member, ACRS
JOHN J. BARTON, Member, ACRS
ROBERT L. SEALE, Member, ACRS
GRAHAM B. WALLIS, Member, ACRS
THOMAS S. KRESS, Member, ACRS
MARIO H. FONTANA, Member, ACRS
DON W. MILLER, Member, ACRS
P R O C E E D I N G S
DR. POWERS: Good morning. The meeting will now come to
order. This is the first day of the 459th meeting of the Advisory
Committee on Reactor Safeguards. During today's meeting the Committee
will consider the following: (1) status of the proposed final version
of 10 CFR 50.59 (changes, tests and experiments); (2) proposed
improvements to NRC's inspection and assessment program; (3) proposed
ACRS reports. In addition, the Committee does have a meeting scheduled
with the NRC Commissioners between 1:00 and 2:30 and the Commissioners'
Conference Room, 1 White Flint North, to discuss items of mutual
This meeting is being conducted in accordance with
provisions of the Federal Advisory Committee Act. Dr. John T. Larkins
is the Designated Federal Official for the initial portion of the
We have received no written comments or requests for time to
make oral statements from members of the public during today's session.
A transcript of portions of the meeting is being kept and it
is requested that speakers use one of the microphones, identify
themselves and speak with sufficient clarity and volume so that they can
be readily heard.
Let me call the members' attention first of all to the
revisions of the schedule with the meeting with the Commission. What is
listed here are the presentation times and we have a Commission that is
extremely busy and we are going to try to hold these times very
strictly. There is approximately an equivalent amount of time
associated with each presentation for their questions and comments.
Members should have in front of them some items of interest.
I call your attention to, first, Chairman Jackson's speech considering
the continuation of the NRC mission, and also perpetuating a tradition
of excellence. I also call your attention to changes that have occurred
in the membership of the CRGR, and finally I call your attention to the
1999 Regulatory Information Conference agenda that is in the first week
I think members will find that to be of some interest and if
you are interested in attending there is a protocol for doing so. It
historically has proved to be of great interest.
I am intending to hold our agenda very tight and strict
today simply because we have a meeting with the Commission. In fact, I
am going to try to shave a little off this morning's sessions. After we
have completed the morning sessions we probably want to reassemble here
with lunch at the table so that we can deal with any last-minute
preparations for that meeting with the Commission.
Are there any points the members would like to raise during
this opening session?
DR. POWERS: Then I propose that we turn immediately to the
first item on our agenda, which is the status of the proposed final
revisions of 10 CFR 50.59, changes, tests and experiments, and John
Barton, I believe you are the Subcommittee Chairman there.
MR. BARTON: Thank you, Mr. Chairman.
The Committee this morning is prepared to discuss with the
Staff and hear the Staff's response to the 50.59 rulemaking since it has
just come back from -- since the public comment period has just ended.
As you will recall, the last discussion we had with the
Staff on the rulemaking was prior to the package going out for public
comment. Our last letter on this subject was to the EDO on the package
and the comments we had on that package that went out for public
At this time I would like to turn it over to Eileen -- do
you have the lead on this? -- Eileen McKenna, who will lead the
discussion for the Staff.
MS. McKENNA: Thank you, Mr. Barton.
My name is Eileen McKenna. I am with the NRR Staff. I hope
to be joined shortly by some of my management who I think looked at the
agenda and were looking at 8:45 so they will be here in just a few
As mentioned, I think our last meeting was in July. At that
time we were in the process of sending a proposed rule package to the
Commission, and I would like to pick up from there and tell you what
went on between July and now, what went into putting the notice out and
then the public comments we heard and then where we are and where we are
going with finalizing the rule package.
Just as a reminder, this rule really was intended to
preserve the licensing basis that was established through the initial
reviews, but also to clarify the requirements primarily through the
means of providing definitions of some of the terms that have been
subject to different interpretations and also, as noted, to allow some
movement off the so-called zero threshold for changes involving -- with
increases in probability or consequences, and the term that was used in
the Commission SRM was a minimal increase concept, so those were the
intentions of the rule.
This is not meant to be a more risk-informed type of rule
which I know there's been a lot of discussion about that. It's
recognized it may be something to go to in the future but it was not
something we were trying to accomplish with this particular package.
DR. POWERS: Well, you may not have been trying to
accomplish it with the package, but that doesn't restrict you from being
mindful of the volume of risk information that has come into the agency
over the last five or six years.
MS. McKENNA: Certainly, but unless there were more
extensive changes made to the regulatory approach it is difficult to
take full advantage of all that information if you are looking at the
context of the design basis type accidents that were part of the
original licensing basis as documented in the FSAR.
DR. POWERS: Right. I mean I guess what I am saying is that
there is not a religious discrimination that thou shalt not use and be
aware of risk information when you formulate this modest revision to
MS. McKENNA: Yes. As I say, it was not that we were --
exactly. It's just kind of how you do it in a way that --
DR. POWERS: Well, it would be useful as you go through your
presentation if you could tell us where you have taken into account the
risk information in thinking about things like definitions.
MS. McKENNA: Okay. We'll try to do that. I think the
other point is that for many of the kinds of changes that we are talking
about, it really doesn't come into play because if you are looking at
procedure changes or minor changes to systems, the assessments are
really more qualitative in terms that there is essentially no impact on
anything and therefore there is no risk impact, but it is not done in
any quantitative way of looking at risk as we consider it in severe
DR. POWERS: Well, I don't think -- I think it is fair to
say that there is a lot of use that can be made of the qualitative
aspects of risk analyses and I don't think risk analyses are confined to
just looking at severe accident space. I think they tell us a lot about
the equipment that is essential even under singlefold types of
MS. McKENNA: Okay. I won't spend a lot of time on this
next slide -- it's just background -- but it has a few datapoints. We
had the March, 1998 SRM which asked for the rulemaking on allowing
minimal increases, so the paper SECY-98-171 in July, providing that
proposed rule for the Commission to consider, received a September 25th,
1998 SRM from the Commission that approved going forward with
publication of a proposed rule with a number of comments and additions
that they wished to see in the notice.
I have listed a few of them here that they wanted to ask for
comment, for example on a wide range of options on margin. We put in a
question about whether the scope as presently defined as facility in the
FSAR should be maintained and there were some other issues that I will
elaborate on in the next slide.
This SMR -- excuse me -- let me say we published the
proposed rule in October.
DR. APOSTOLAKIS: Eileen --
MS. McKENNA: Yes?
DR. APOSTOLAKIS: -- I am trying to understand what the
thrust of this is.
We are getting around with this SECY the issue of, the
problem of not having explicit quantitative probabilities in the SAR, in
the licensing basis, and not having therefore a definition of minimal
changes by saying if we were to license this facility again or today and
there was this change, would that affect our judgment as to the
acceptability of the facility.
In other words, it is stated in here in the SECY that the
basis for deciding what is minimal is qualitative and that when people
reviewed, when the Staff reviewed the original license they made some
qualitative judgments regarding probabilities.
First of all, I wonder whether that is true. Did people
actually do that or they just followed the regulations?
Second, is the industry happy with this and their staff is
happy with this? In other words, they feel this is going to work?
MS. McKENNA: Let me try to answer that in two parts. The
first part that you asked, as to whether during initial licensing to
what extent the probabilities were considered, and I think that was
considered in a relative sense, that there were -- there's a spectrum of
events that was postulated to occur and that for those events that were
considered to be more likely the expectation was that the outcomes of
those events be more acceptable -- for instance that we know no fuel
damage for more frequent events as an example and that the other events,
for instance, double-ended guillotine LOCA, was a less likely event. It
would be acceptable to have some degree of damage so to that extent
things were, probabilities of the events were taken into account.
The second part, about where people are on probabilities and
minimal -- I was going to get to that a little later, but we can talk
about it now -- I think given the sense that people generally do want to
continue to look at these things in a qualitative way that the minimal
being something large enough that you could see it and touch it perhaps
has a little bit of discomfort because once you move off the well, we
are not sure whether it increased or not to yes, it did increase, but we
don't know how much is okay before it is not okay, there is some
discomfort about that in a qualitative sense, and I think the general
sense of the industry is perhaps that is more than they want to take on
at this point, that just having the -- as long as we can't really tell
whether it changed that's okay, may be good enough for the purposes.
DR. APOSTOLAKIS: So "minimal" then would be different for
different accidents and different components, depending on the original
assessment or qualitative assessment of probability or is it in
general -- in other words, if one of the original events that was judged
to be reasonably likely -- let's say not very likely but reasonably --
if a change affects that, then you would allow a larger change in the
probability because the probability is already relatively high than say
for a double-ended guillotine break, which is a very low probability
event, in which case -- in other words, is the concept of minimal
defined with respect to the probability of the event?
MS. McKENNA: We did not try to do it that way just because
of the relative lack, if you will, of precision on the assessments of
the probability in the first place.
I think that in theory that would be the case, but in actual
practice it would be more difficult to undertake, so I think our present
thinking is to not to try to do any quasi-quantitative --
DR. APOSTOLAKIS: Right.
MS. McKENNA: -- approach and to continue the -- I think we
used the word "attributes" at an earlier meeting, the more qualitative,
that you are still meeting the design requirements for that particular
system, and if that is the case there is no -- not an increase in the
probability of failure, rather than trying to step too far off into
minimal concepts for probability when we don't really have a basis to do
MR. BARTON: So in the final rule your definition of
"minimal" is going to be there or not there or --
MS. McKENNA: It is going to be this qualitative type -- I
think we have given some information on the proposed rule and we got a
number of comments on ways to improve that and I think that is what we
would continue to use as a qualitative base to judge when there is a
need to get the approval.
I want to introduce people at the side table. We have Scott
Newberry, who is now our Deputy Director for the Division of Reactor
Program Management and Frank Akstulewicz, who is the Acting Branch Chief
for Generic Issues and Environmental Projects Branch, and they may
contributed to the conversation as we proceed.
DR. WALLIS: I picked up on what you said a minute or so
ago. You used the term "We don't know if it changed or not." Are you
claiming this is a better state to be in than trying to evaluate
MS. McKENNA: Well, I think it is something that is more
easily dealt with --
DR. WALLIS: I think it sounds retrograde to say that
ignorance is better than trying to figure it out.
MS. McKENNA: I think it allows room for engineering
DR. WALLIS: But that is detestable. That is the worst
possible way of deciding things.
I am sorry -- I have broken my vow of silence already.
DR. POWERS: Professor Wallis, the vow of silence was
forbidden by the Chair and you will remember that from the last meeting.
DR. MILLER: Professor Wallis and I may not agree on that
statement of engineering judgment.
My question is do you believe that minimal will move to a
state where we'll define it quantitatively?
MS. McKENNA: I'm sorry, I couldn't quite hear your
DR. MILLER: Do you think we will move to a situation where
will define minimal quantitatively or do you think it will always be
qualitative, dependent upon engineering judgment?
MS. McKENNA: Well, I think at least in the present term
that it would be qualitative. There was some interest in perhaps
tackling a more quantitative approach, although I think in general those
who are interested in that would not want to be looking at the
individual probabilities of the events but a more quantitative approach
that perhaps uses some other kind of criteria, getting more into a
combination of the probabilities and consequences, for instance, as
opposed to trying to apply quantitatively to probability of a particular
DR. WALLIS: So you're going to use the judgment of the
policeman as to how far the car was going rather than actually trying to
MS. McKENNA: Well, again, I think the point is you're
trying to look at the incrementals change, if you will, of the foot on
the accelerator and --
DR. WALLIS: Why don't you just measure it?
DR. MILLER: Well, it has been a tradition, for the first 30
years, or whatever it is, this term "minimal" is not used, but there was
the use of engineering judgment very successfully.
MS. McKENNA: Yes. And I think --
DR. MILLER: And throughout the use on 50.59 up till about
two years ago.
MS. McKENNA: Um-hum.
DR. MILLER: We never defined "minimal" but implicitly it
MS. McKENNA: I think that's true.
DR. MILLER: And it's been very successful.
MS. McKENNA: If you look at the industry guidance
documents, I think they do embrace that kind of concept where it's so
small a change, a change that you really can't tell whether or not it
really changed, that that's not a change. And we would still think that
that would clearly meet a minimal-increase standard. And that's
generally what people have been using.
DR. APOSTOLAKIS: Eileen --
MS. McKENNA: Yes.
DR. APOSTOLAKIS: I was wondering whether if you dropped the
word "probability" your life would be easier. In other words, say the
change is so small that the licensing basis is not affected, and get out
of the probability business.
Now as you had in your earlier slide, you want to preserve
the integrity --
MS. McKENNA: Right.
DR. APOSTOLAKIS: Of the licensing basis, okay? Which was
deterministic. There are some -- there are of course also this document
here talks about the conservatisms and so on.
MS. McKENNA: Right. Correct.
DR. APOSTOLAKIS: So we know that there are a lot of
conservatisms all over the place. So I am preserving the integrity of
the licensing basis even if I allow some small changes here and there.
If you put it that way, then perhaps all this discussion would not take
place. I think the use of the word "probability" in this document is a
red flag, because it was never quantified.
DR. KRESS: And it will be difficult.
DR. APOSTOLAKIS: And it will be difficult to quantify. So
DR. KRESS: I agree with you.
DR. APOSTOLAKIS: I wonder whether you can just go through
page by page, line by line, and just cross out "probability" and don't
try to justify that we had a qualitative estimate and so on.
Look, this was an engineering judgment. We licensed the
facility, a lot of them are operating successfully. It worked, okay?
And it's very conservative. When you do that, you make judgments all
over the place, okay? And nobody says that that was the way and there
is no other way, right? When you select the parameter values of the
ranges in this worst case. There is a lot of room for changes.
So the spirit of 50.59, is the proposed change, test, or
experiment going to affect significantly the integrity of the licensing
basis or not? And this will be as judgmental as the original decision
of the acceptability was. Don't put the word "probability" in the way.
Then you will not get all this.
MS. McKENNA: It's an interesting thought. I mean, I agree
with you that the intention of the rule is to do exactly that to
preserve, to look for that, but it's how you define that in a way that
everyone can understand it in the same way and reach the same decision.
DR. APOSTOLAKIS: It's three lines, Eileen.
MS. McKENNA: Yes.
DR. APOSTOLAKIS: We want to preserve the integrity of the
licensing basis, a lot of judgments there, and we will not require prior
approval of any changes that do not affect that integrity.
DR. KRESS: You will have to go on and define attributes or
something that a person -- the licensee can say what changes now qualify
me to say that this doesn't affect --
DR. APOSTOLAKIS: I didn't see any attributes in this
DR. KRESS: But you're going to have to have something. I
mean, that -- you're just changing the words there, George.
DR. APOSTOLAKIS: Okay. All right. So you will say --
DR. KRESS: There will have to be more.
I like your approach, but there has to be --
DR. APOSTOLAKIS: Yes.
DR. KRESS: Some definition criteria or something that the
licensee can actually use to say he meets that --
DR. APOSTOLAKIS: Okay. So let's take the original 50.59,
the three criteria, preserving the integrity of the licensing basis
means -- and you something on initiating on accidents, but you don't use
the word "probability," you don't say that the probability may be
increased. There was a judgment made at the time, right?
DR. KRESS: What do you say?
DR. APOSTOLAKIS: I don't know. They know better --
DR. MILLER: The judgment was -- their judgment was the
DR. APOSTOLAKIS: No.
DR. MILLER: In a nonquantitative sense --
DR. APOSTOLAKIS: No. The judgment was that this was
DR. WALLIS: Well, the judgment was made by --
DR. APOSTOLAKIS: No undue risk to public health and safety.
DR. WALLIS: People who have now left the Agency, and no one
knows what the basis of their judgment was, how can you go back and --
DR. APOSTOLAKIS: I'm sure the ladies and gentlemen of the
staff can come up with the right words in three minutes.
MR. BONACA: One comment I have. Although there wasn't
specifically a lot of calculations done quantitatively, the design
level, the vendors that designed the plants used a lot of insights in
probability, whatever the tools may have been at that time, to design
these systems and these plants. And I think what happens with the
probability is that anytime you make a change, it may affect
redundancies, separation, those criteria which has an NGDC which then
you have to interpret in terms of have you affected the probability.
Now that's where the judgment comes in. Assume that you eliminate, for
example, diversity in instrumentation for some reason, that would be a
USQ, and you can show -- you assume, you assume that that would result
into an increasing probability.
DR. APOSTOLAKIS: But why do I have to bring the probability
at all into this? What you just said, if you eliminate diversity, you
will have a USQ. Put it that way.
MR. BONACA: And that's a possibility.
DR. KRESS: Those are some of the attributes I had in mind.
DR. APOSTOLAKIS: Yes.
DR. KRESS: Diversity, redundancy, things related --
DR. APOSTOLAKIS: Right.
DR. KRESS: To defense in depth. If you affected those,
then you have an USQ. That's the kind of things you have to expand on.
DR. APOSTOLAKIS: But it's still judgment, though. It's
still engineering judgment.
DR. KRESS: Oh, it's absolutely judgment. Yes.
DR. APOSTOLAKIS: You're staying within -- and the question
you are asking is, "Is this still acceptable?" Am I preserving the -- I
like these words -- am I preserving the integrity of the licensing basis
without saying I am judging that the probability did not change, what,
more than an insignificant amount, therefore it's okay? I don't think
you need that detour through probability space. And then it's clear
that this refers to the basis of acceptability. That's how they put it
in the SECY.
DR. KRESS: What you do with that, and it's -- now some of
those changes that wouldn't meet those attributes and therefore not be
allowed, could be changes that have minimal change in probability,
minimal change. You're just -- too bad, you're going to -- this is
where you've got to draw a line.
DR. APOSTOLAKIS: It's too bad.
DR. KRESS: Yes.
DR. APOSTOLAKIS: Exactly.
DR. KRESS: Yes.
DR. APOSTOLAKIS: And that's something that we have not made
clear -- occurred to me last night actually.
DR. SEALE: We've gone --
DR. APOSTOLAKIS: When we go to risk-informing 50.59, we are
not going to preserve the class of problems to which 50.59 applies. We
will expand it --
DR. SEALE: Yes.
DR. APOSTOLAKIS: Significantly. So people who complain
about the additional analysis now will have a cost-benefit problem in
front of them. Do you want to have this additional flexibility? Then
you have to do something.
DR. SEALE: We've gone through an awful lot of agony here
lately over what the role of defense in depth is as we look to
risk-informed regulation, and I think making the case the way you've
suggested, George, which identifies these various defense-in-depth ideas
like redundancy and multiple systems and so forth and not building a
case which really can't be supported for probabilistic arguments at that
time puts those two issues in the proper perspective with each other and
I think will help us if we can get that idea across as we look into what
we do with defense in depth in a risk-informed regulatory climate as go
forward. So that's a useful thing to point out because it also properly
defines the issue for our considerations down the road.
DR. APOSTOLAKIS: So what you're saying, Bob, is --
DR. SEALE: You're right.
DR. APOSTOLAKIS: That what I said was useful.
DR. SEALE: Very useful. Very useful. Because for other
reasons than just this issue, because it does put those two ideas in
juxtaposition to each other.
DR. POWERS: Please go ahead.
MR. BONACA: On the other hand, I'm saying that from a
perspective of the significance of 50.59, the fact of addressing
probability and consequences, which is fundamentally risk, you know,
it's pretty enlightened, I think, as a general regulation, and also it
provides an opportunity for introduction of risk-informed regulation in
fact if we were flexible about how to use it. The moment in which you
begin to eliminate from these rule terms as probability and
consequences, which really qualitatively were at the foundation of the
whole design of these plants, we are eliminating opportunity, it seems
to me. Just a thought.
DR. APOSTOLAKIS: No. But the staff is already working on
making something like 50.59 risk-informed, and they don't need
opportunities to inject risk information there. They are starting, you
know, by considering a number of options. So this here, the rule we're
talking about here is something that is needed urgently, right? Because
the plants out there need it. And what we're trying to do is rephrase
certain things that are already there and clarify a few things.
But our objective is not to give opportunities to use risk
information in this context. I mean risk is coming later in a different
rule so if the word "probability" creates so many problems, it seems to
me if you can drop it completely and go back --
DR. SHACK: It creates problems for you, George. I am not
sure it creates problems for anybody else.
DR. MILLER: What Bill is saying, I agree that does it
create problems for those who really use it on a daily basis.
DR. POWERS: I don't think it really does.
DR. MILLER: I think that's --
MR. BARTON: For those that use it --
DR. MILLER: Risk insights are already being used.
DR. APOSTOLAKIS: How?
DR. MILLER: Just their judgment on use of PRA insights.
MR. BONACA: By the way, there were other foundations on a
number of cycles and components and so even if we know much better than
whatever they could come up with at the time, even the tocology, it
really forces licensee to stay within certain anticipated transient
rather -- which are really related to the number of cycles that you have
on components and things of that kind, so there is a framework and a
structure but you got to be careful, you know, in my mind, not to upset
without a lot of thought about what is going to happen once you remove
it because ultimately it paid off pretty well insofar as having safe
plants out there.
So I am only saying --
MR. BARTON: Think hard before you take the word out.
DR. POWERS: I guess I am very interested in Professor
Apostolakis' suggestion because my belief is that the word "probability"
is causing difficulties, not because of what probability is and how it
has been used but the fact that we have gone from an era where
probability was looked upon in a very qualitative sense to an era now
where it is used in a very highly refined and quantitative sense, and
you get people looking at issues where probability can be assessed at
such microscopic detail that in fact it would have been glossed over in
the past and assumed no change.
The difficulty you get into then as a designer or an
engineer is a qualitative sense of the probability but a concern that
the regulator is operating with a much more refined sense of
probability. It may be that indeed you don't want to eliminate the term
but you want to get its sense across by coming in with saying things
explicitly like the attributes of loss of redundancy, loss of the
ability to deliver safety functions in a diverse mechanism is some
better definition of what you mean by probability on this.
DR. KRESS: Let me comment on that too.
One of the reasons that we shy away from using the words
"probability" and "risk" going all the way to the end of that, is
because in this 50.59 space there doesn't seem to be a really good way
of using our only tool, PRA, to calculate changes in probability or
changes in risk. That is why I said one of the attributes one might use
to determine -- what we really are interested in is not allowing changes
to things that are important to safety. I hate to use that word but if
we define "important to safety" to mean those things that contribute a
certain amount to the consequence, to the risk, if we define it that
way, then we are risk-informed, and the question is how do you know
which those are. Well, you can find the ones that contribute to risk.
I mean that part PRAs can do, and the things that are left, the one
minus that, are the things you can then say those are 50.59 space -- we
don't have to deal with those.
If you are not touching or bothering these things that are
important to safety, unless you are improving them -- you always allow
improvement -- then you have defined what you mean by minimal change
with an attribute. It is an attribute. These things do not belong to
that set of things that are important to safety.
DR. APOSTOLAKIS: Which the Commission does in this document
exactly what you are saying. On page 26, probability of equipment
malfunction -- the Commission believes that the probability of
malfunction is more than minimally increased if a new failure mode as
likely as existing modes is introduced.
Then later on they say the probability of malfunction of
equipment important to safety previously evaluated in the FSAR is no
more than minimally increased if design basis assumptions and
requirements are still satisfied. That is close --
DR. KRESS: That is close to what we have.
DR. APOSTOLAKIS: And say if you introduce new failure
modes, don't do 50.59, right? And I don't have to use the word
See, here is an attribute of the kind that everybody seems
to like, me included, which makes sense to me, and then it says further
on because we want to preserve the integrity -- beautiful words -- the
probability of malfunction of equipment important to safety is no more
than minimally increased if design basis assumptions and requirements
are still satisfied. Now that I understand.
DR. KRESS: That is a troublesome phrase for me though.
DR. APOSTOLAKIS: Well, that is what it's saying.
DR. WALLIS: Mr. Chairman, I would like to know how far
along the Staff is with resolving this issue.
DR. APOSTOLAKIS: I don't think they are trying to resolve
this issue. Are you?
DR. WALLIS: Are you? I thought we were going to hear a
presentation now instead of a discussion among the members.
MS. McKENNA: Okay. Yes. I will get to that where we are
in a little bit, if you will allow me to just pick everybody up back to
where we were.
I was talking about in September the SRM, just briefly from
what we had from the Commission on what they wanted to see in the notice
that went out for comment, and so we did solicit comment on a wide range
of options on margin.
We asked for ideas on options to refine the guidance that we
had given on minimal and also the options we had given with respect to
consequences, and the Commission -- as I say, we had included
definitions and the Commission asked the question as to whether there
was a need for the definitions, and also ideas about the particular
definitions and whether they could be improved in any way.
The Commission also asked the Staff to solicit comment as to
whether there was a need for a definition of the term "accident." It
does appear in a couple of the criteria with respect to accidents
previously evaluated and also accidents of a different type and I
believe the Commission's interest was recognizing that in the context of
the FSAR you are looking at the so-called design basis accidents as
opposed to the full spectrum of accidents that might be considered in
other contexts, so we did put a section in the notice about accidents
and asked whether there was a need for such a definition.
The SRM established a date of February 19th for the final
rule, which we believed was extremely ambitious given that the comment
period would not be closing until December, and just noted that we had
proposed that we continue to apply enforcement discretion for instances
that were of low significance as things went on.
So this is what we changed in the notice compared to what
you saw in the SECY paper. We added the discussion on accident, added a
larger section on margin that included a number of different approaches,
the one being what we had originally offered in the paper with respect
to input assumptions that supported analyses for tech specs.
We also offered the possibility as to whether margin as a
separate criteria could be deleted on the basis that the other criteria,
other regulatory requirements and tech specs really would provide the
envelope of things that you really need to consider, and we also in the
context of other ways of going at margins put proposals of looking at
various results of analyses or identifying particular parameters that
may be of interest and then judging to what degree they could be changed
and still be done without approval, whether that is a minimal type of
change or where there is some particular limit that has been established
for a particular design parameter, whether that would be the point at
which you would judge when there is a need for approval.
DR. MILLER: Eileen, are you going to discuss in some detail
where you are on the issue of margin?
MS. McKENNA: Yes. Yes, I will.
DR. MILLER: I haven't looked ahead, I guess --
MS. McKENNA: I'll try to do that.
MR. BARTON: Before you move that slide, I don't recall, on
the discussion in the definition of "accident" did you try to redefine
"accident" or are you -- or is the industry's definition in 9607
MS. McKENNA: I think the general sen se is that it is
acceptable as a definition of accident previously evaluated. The
question of accident of a different type has always been a little bit
more problematic, but I think it does -- kind of the balance between
what you consider to be accidents and what you consider to be
malfunctions, they kind of fall in either of the bins, but I don't think
in general there was much disagreement about that, which is one of the
reasons why we really hadn't proposed anything on that in the first
I think this slide I have more or less covered including the
range of options that were included in the notice. I mentioned I think
DR. APOSTOLAKIS: Are you going to discuss the public
comments at all?
MS. McKENNA: Yes, that's the very next slide, so if you
want me to, I can --
DR. WALLIS: Are you going to discuss your conclusions?
MR. BARTON: That's all right, keep going. You're doing all
MS. McKENNA: Okay.
DR. WALLIS: Are you going to discuss your conclusions
following the --
MS. McKENNA: Well, as we advertised, I believe, for this
briefing that this really is meant to be a status. We have not
completed our review at this point. We do plan to come back to the
Committee at a later meeting when that has been done. But we did want
to give you the opportunity of hearing where we are at this point and
what we were hearing from the comments and where we may be going with
it. But I can't today tell you exactly how everything's going to come
Here's the status. We received comment letters. There were
58 submittals totaling about 300 pages of comments, and indicated there
commenters, as you might expect, were largely from the power reactor
licensees, certain organizations, NEI, some law firms that represent a
number of the utilities, some of the vendors, both the NSSS type of
vendors but also some of the vendors for dry-cask storage, because, as
you may recall, this also was applicable to Part 72 for the independent
spent-fuel storage facilities. And then there were some letters from
individuals. Nothing that was identified as, if you will, a public
DR. WALLIS: It seems to me the only noninterested party in
this whole discussion is the ACRS.
DR. POWERS: Well, disinterested as opposed to
DR. WALLIS: I suppose disinterested is a better term. I
stand corrected. But the only one that doesn't have some personal
interest in tweaking the regulations for their own benefit seems to be
MS. McKENNA: Well, I think it reflects that the rule is
something that is used by these utilities on a regular basis and that
changes to it obviously impact them the most, and so they were obviously
looking for opportunities to make the process more efficient from their
own purposes, so I'm not at all surprised that that was the spectrum of
DR. POWERS: I notice that you also got comments asking NRC
to consider an equivalent rule of maybe expanding the scope of this one
to be applied in other areas such as the transport packaging --
MS. McKENNA: Yes. Yes, we did.
DR. POWERS: Have you given that any thought, or have you
just set that aside as something in the future?
MS. McKENNA: No, I think we have been giving that thought.
There are some complicating factors, if you will, on the transportation.
There are some questions that IAEA standards and DOT standards, we need
to look at closely to make sure that we're compatible with those.
DR. POWERS: Do you really have to have compatibility?
MS. McKENNA: Well, I think it's just something that, you
know, I'm not really -- the Spent Fuel Program Office has primarily
been, but we have had discussions on this. I think that they are giving
that thought, but are not prepared to go wholesale on Part 71 at this
point. I think they're giving the hardest thought to the question of
the dual-purpose casks where they fall under both Part 71, Part 72, and
I think they believe that's appropriate to do. It's just kind of the
timing and how we do it.
More broadly, I think that given the range of type of
transportation issues that there are, they're looking at does it make
sense to do it across the board or focus it on, say, fuel transport. So
that's still under consideration, but, you know, it has not been just
DR. POWERS: Good.
MS. McKENNA: Let's go back to basically the nature of the
comments that we received. And I tried to group them into a number of
areas based on what were the things that we asked for or the things that
were of most interest to those who filed the comments.
The first one I listed there was of course margin. We
probably got the most number of comments on margin, and a considerable
spectrum of views as well. We had those that suggested that the
approach in the old NSAC 9607 on using acceptance limits was the way to
go. We had a number that embraced the proposal to delete it as a
criterion, that it was kind of more trouble than it was worth in terms
of what it was really going to give you. We received a proposal from
NEI to I guess I'll say reformulate the idea of using limits to judge
when there is a reduction in margin but not to -- try to get away from
the words "margin" and "reduction," because they are things that have
given us difficulty in the past when you really identify well, you know,
margin from what to what kind of kind of questions.
DR. APOSTOLAKIS: So the theme seems to be drop words,
eliminate words, eliminate probability --
MS. McKENNA: Or use different words, anyway.
DR. APOSTOLAKIS: Use different words.
MS. McKENNA: At least in this area I think that there was a
strong sense that the words presently there, margin of safety as defined
in the basis for any tech spec, have too many threads that are not
understandable and that it's better to kind of get more directly to what
it was really trying to do.
DR. MILLER: I find it interesting that there are as many
comments supporting deletion, of course --
MS. McKENNA: Yes.
DR. MILLER: Even from the Commission. On the other hand,
NEI seemed to have put a red flag on that one and say gee, that could
introduce some holes, so to speak --
MS. McKENNA: Yes. I think --
DR. MILLER: If you introduced another approach.
MS. McKENNA: Right, and part of that is a recognition I
think that all tech specs are not created equal, and that, you know --
so that some plants there may not be any holes, others there might be,
and that having an additional criterion that would perhaps test some of
these other things may be useful. So I think that was the thinking
DR. MILLER: So the staff has not anyway coalesced on
thinking in those areas, or --
MS. McKENNA: Not totally. As I say, we have this proposal
from NEI which obviously we're giving serious consideration because it
does reflect an overall view from the industry, you know. In their
process they did circulate their proposal among their members, and so it
does have a degree of support from the users, shall we say.
At this point we are doing that same kind of circulation
among the staff to see are there holes, are there things that we think
it might allow through the net that shouldn't get through? Is there a
better way to define it? And one of our goals certainly is to have
something that is understandable and can be consistently used from place
to place. And I think their proposal may offer those advantages I think
the staff is looking at to make sure that it's not too narrowly focused.
DR. POWERS: Professor Seale.
DR. SEALE: Eileen --
MS. McKENNA: Yes.
DR. POWERS: When we started, when we opened this can of
worms about a year-and-a-half I guess ago I guess --
DR. MILLER: April of '97.
DR. POWERS: Yes.
DR. MILLER: Almost two years ago.
DR. SEALE: We heard that NSAC 9607 was a process or gave a
process that many utilities had used successfully in screening their
50.59 applications, and where that had been followed by the utilities
there didn't seem to be any residuum of problems. It was only when they
didn't follow that kind of process that they ran into difficulty.
Does the staff still feel that that was a proper assessment
of the situation at that time?
MS. McKENNA: I might just modify slightly what you said.
DR. SEALE: All right.
MS. McKENNA: I think following the process kind of led the
licensees through the right questions --
DR. SEALE: Certainly.
MS. McKENNA: And attributes, if you will, the kind of final
answer whether it was a yes or no. I think there were probably a few
cases where they might have gone to yes, I can do it, whereas we might
have said no, under the rule as written, you cannot, but that the
significance of those areas of debate was probably relatively low.
DR. SEALE: Um-hum.
MR. BARTON: Bob, as you remember, that was with the NSAC
125, and then the industry committed in their revised 9607, they got 80
percent of the utility to sign up --
DR. SEALE: Right.
MR. BARTON: They would all sign up. I believe they're all
using the revised 9607 --
DR. SEALE: Right.
MR. BARTON: Process at this time. Which brings me to the
question of, and I haven't seen it on here, your work on the regguide
that will tie the 9607 with --
MS. McKENNA: Okay. I'll get to that --
MR. BARTON: Is it coming?
MS. McKENNA: In a moment.
MR. BARTON: If it's coming, it's fine.
MS. McKENNA: Yes.
MR. BARTON: Fine.
MS. McKENNA: I think it's kind of --
MR. BARTON: Okay.
MS. McKENNA: This last bullet, guidance development.
MR. BARTON: All right. Go ahead. Just continue where you
were then. If you'll address it, that's fine.
MS. McKENNA: Okay. So as I started to say, on margin that
we are considering the full set of comments. We're looking at this NEI
proposal, which I don't know if any of you had a chance to look at --
see it, but in essence they would offer, instead of the margin criteria,
a criterion that states that prior approval is required if as a result
of the particular change the design-basis limit for -- directly related
to integrity of fuel-clad reactor coolant system pressure boundary or
containment boundary would be exceeded or altered. So that's the
language that they came up with.
MR. BARTON: That's the NEI's proposal.
MS. McKENNA: That's their proposal.
DR. KRESS: The crux of that is would be exceeded by whose
determination, and --
MS. McKENNA: Well, I think the crux of it is this concept
of design-basis limits that there is a limit established through either
regulation or code or whatever that that particular parameter must
DR. KRESS: Sure.
MS. McKENNA: And then as long as as a result of this change
that limit is still met --
DR. KRESS: I know, but in order to make that determination,
somebody has to make a calculation.
MS. McKENNA: Correct.
DR. KRESS: And I'm asking whose calculation will that be --
MS. McKENNA: Well, that's why you saw on this list the
question on margins, the question of methods.
DR. KRESS: Yes.
MS. McKENNA: That's an issue that we are also still
wrestling with, because there was a method that was used originally in
the FSAR that was reviewed by the staff, and if they're doing a change
and if they use the same method and you still meet the limit, I think
that gives the staff more comfort that, you know, in terms of preserving
the licensing basis as opposed to I'm changing something and I'm
changing methods, so then --
DR. KRESS: So you have 15, 20 different methods out there
of determining, depending on the plant --
MS. McKENNA: Yes.
DR. KRESS: And the limit, determining this, and you're
saying you just are going to let each one of them use their own method
and see if they exceed this limit?
MS. McKENNA: Oh, I think what we're saying is that they
should use the method that was used before --
DR. KRESS: For the SAR.
MS. McKENNA: Which the staff has seen as opposed to just
using any method.
DR. KRESS: But my understanding is the staff didn't really
approve that method, they just approved the value that was calculated --
MS. McKENNA: I'd say there's a range. Some cases there
were topicals or other things, the methods were approved --
DR. KRESS: If there's more than one or two exceptions, then
the general case is they did not approve the method.
MS. McKENNA: Right. I think what we're -- like I say, what
we're trying to do again is to gauge the effect of the change, and if
you leave the methods alone, then you can see what the effect of the
change is and then determine whether that change is acceptable.
DR. KRESS: Yes, but you don't know where the absolute value
is. The change is relevant to the -- what amount of change is
acceptable is -- the limit may have already been exceeded, but you don't
know that. You probably are assuming that, but my question is, if you
went with that particular option, don't get me wrong, I like the option
MS. McKENNA: Okay.
DR. KRESS: So all I'm saying is that there is a need now to
go back and look at all these methods and say okay, we really ought to
approve the method as being conservative, or we have to approve the
level of uncertainty in that method is low enough that I can have a
certain level of confidence that my limit has not been exceeded. That's
the only rational way to do that, and it's a good way to do it. I like
the proposal, but I think you have to go back and do something in order
to permit it.
MR. BONACA: A question I have for you is --
MS. McKENNA: Yes.
MR. BONACA: These methods were not best estimate.
MS. McKENNA: That's correct. They were generally
MR. BONACA: They had very forced assumptions --
MS. McKENNA: Yes.
MR. BONACA: Like 20 percent decay heat --
MS. McKENNA: Certainly, yes. Penalty factors and
uncertainties and things would apply.
MR. BONACA: And so what you really want to preserve is the
commitments in the methods to those conservatisms, okay? And by doing
so, the only unknown is whether or not you have assumed everything
stacked in a certain direction and really have a scenario where you
should think about the opposite way, and that is the only -- but I'm
saying that this is not best-estimate calculation.
MS. McKENNA: That's absolutely correct; yes.
DR. KRESS: But even conservative calculations have
uncertainty in both directions.
MR. BONACA: I agree with that. That's why I said --
DR. KRESS: So you can't really say just because it's
conservative, I still have sufficient margins. You have to really look
at the uncertainties, even in a conservative calculation.
MS. McKENNA: But I think also that in establishing whatever
those limits are there is recognition of taking into account
uncertainties and giving yourself margin, if you will, to some -- the
conditions that you don't want to be in, that you establish your limits
at a place that takes some of those things into account.
DR. KRESS: It's another conservative on the other end.
MS. McKENNA: Yes.
DR. KRESS: That's all right too.
MS. McKENNA: Sure. But the question of methods, as I said,
is something we are giving --
DR. SEALE: Cascading conservatism.
MS. McKENNA: Giving a lot of thought to.
Let me go back -- from the comments, I think the question of
minimal we have talked about to a certain degree. I think as I
mentioned we got a lot of comments about the guidance we had given in
the notice -- some of the points you mentioned about the language, about
no more than minimally increased if the design basis requirement for
instance are met. I think we had a number of people say, well, if the
design basis requirements are met, you know, there is no increase.
I think we were trying to give a -- clearly there was no
more than a minimal if these are met, so that was what we were trying to
accomplish with that and so there were comments along those lines of
perhaps ways we can improve those qualitative thoughts, but again I want
to emphasize that people were concerned that the way we wrote it made it
suggest that we were suggesting quantitative analysis of probability and
certainly that's not the expectation, and we don't really anticipate
that in general that is what is going to be what happens.
There were a few that I think might be interested in a way
that they could take more advantage, shall we say, of PRA information,
but that is -- it is a little harder to weave together with the criteria
based on the FSAR analyses. They don't quite fit in terms of -- as delta
CDF or some such criteria as opposed to a consequence of a design basis
accident, so I am not sure we can do that in the context of the current
language. That may be something that would have to wait to a later
The other part on this minimal is the area of consequences
and Staff had offered a couple of ideas, such as looking at the
difference between the regulatory limits -- for instance, Part 100 --
and the current value and allowing some percentage change as a means of
minimal, and I think there were still those who felt that they ought to
be able to go up to the limits from a risk-informed perspective, that
meeting the Part 100 is still adequately protective but in general
people were willing to go with an approach to the limits by using some
percentage of the remaining margin, shall we say, as a way of making
sure you did not get up to those limits.
The last part of that particular one was on this question of
cumulative effects and whether that needed to be documented and reported
in the FSAR because of, for instance, the issue of probabilities and
whether -- how qualitative or quantitative that could be.
There was a lot of concern that that might require too much
effort and burden that really wasn't necessary and that for the more
quantitative issues such as consequences or some of these other factors
that the existing language where if they made a change in the
calculations and the results were affected then that would be reflected
in the FSAR, that that was sufficient. It wasn't necessary to put
additional language in 50.71(e). I think at this point we are inclined
to accept that, that comment.
The next one there is screening. That has to do with the
definition of change. We had many comments that were looking for
language in the rule that would more easily facilitate a screening
process that if it is a change that is kind of at the level of detail
that it couldn't possibly affect anything, it ought to be able to be
screened out very simply without having to do an evaluation against the
We are looking at some language that would make that a
little easier to accomplish but at the same time make sure that where
appropriate it does look at the criteria and not be just, well, this
doesn't affect anything and I'm done -- without really understanding
what the effect of the change is.
Additional clarifications on the other definitions -- I
think I mentioned the Part 72. You mentioned it, Dr. Powers, I believe
the Part 71. Some of the other comments on Part 72 were really aimed at
making the language in 50.59 and Part 72 even more similar. There are
some differences that exist now and it is a goal of the Staff to try to
do as much as we can to align them because we recognize the overlap in
those who use those parts, so we are looking very hard --
MR. BARTON: Eileen, on the clarification definition --
MS. McKENNA: Yes --
MR. BARTON: -- what do you mean by that?
MS. McKENNA: Okay. We provide definitions of what the
facility has described in the FSAR or procedures as described in the
MR. BARTON: Right, but as I recall there was a lot of
comment back on -- the thought was that the new definitions greatly
expanded the scope of the rule. I remember --
MS. McKENNA: There were some that thought that because of
things like the analysis.
MR. BARTON: Do you agree with that and intend to change the
definitions based on that?
MS. McKENNA: I guess I don't see that it greatly expanded
it. I think that it may have -- again, depending on how people
interpreted it in the past -- that it may be an expansion to a certain
degree, but we believe it is an appropriate expansion and that that in
combination with the language I talked about on change I think may get
us to the right set of information.
So we do plan some changes to the definitions but not to the
extent that I think perhaps you might have been suggesting.
MR. BARTON: Okay.
MS. McKENNA: The last couple of bullets here on enforcement
policy -- I think there were questions about how we would look at
changes that, evaluations that were made a year ago once the rule is in
place or during the first few months after the rule goes into place
while people are getting their programs in order, so we are looking at
the interplay of that with what we should propose as the effective date.
On the one hand, and obviously I think people would like to
have the rule be made effective as soon as possible, but on the other
hand we may not have all the guidance developed and agreed to, so we may
have a little bit of a period where we would say proceed to implement
and we would, you know, continue to exercise discretion if there is
question of their procedures didn't quite catch up yet.
DR. SEALE: Could I ask another question?
MS. McKENNA: Yes.
DR. SEALE: Again, two years ago when the wheels came off it
was I think the result of the response to a questioning of what the
words "zero increase" had as their meaning --
MS. McKENNA: Maybe increase language, yes. Yes.
DR. SEALE: And that was the result of an interpretation by
a particular subset of the legal profession, namely the Office of the
Now I notice here that we have had quite a few comments back
from law firms, but they of course represent people with other positions
on this issue, and I am curious as to whether or not you have confidence
that the product you are coming up with is not going to run into a stone
wall when it gets back to the general counsel after you put all these
modifications into it.
Is your modification of the concept of minimal increase
going to survive that kind of review?
MS. McKENNA: Well, obviously the Office of General
DR. SEALE: Will speak for itself.
MS. McKENNA: -- was involved. Yes, they will speak for
themselves, but I will say they were certainly involved in the
development of the proposed rule and are involved now in looking at the
changes that we are suggesting as a result of the comments.
That is probably the best I can answer.
DR. SEALE: So there is light at the end of the tunnel?
MS. McKENNA: I hope so, yes.
DR. SEALE: You just don't know if it is a gorilla holding
MS. McKENNA: That's right. Okay. I think we are on the
question of guidance.
As you know, the 96.07 document is out there. We have had
previous Commission direction to attempt to -- they would like to see us
endorse that document or some modification of the document if we can.
I think NEI has indicated they are willing to modify the
document to conform to the rule as it ultimately gets finalized but
obviously that will take some degree of time to make the appropriate
changes once the Commission agrees on some language that they would like
MR. BARTON: Will that go on in parallel with the rule?
MS. McKENNA: Yes, and that is one of the reasons why I
mentioned that we may in terms of the timing of implementation as to
whether we delay implementation until the guidance is ready or allow it
to be effective with, say, a full implementation in a year to allow the
guidance to proceed. That is kind of our current thinking -- to say
make it immediate --
MR. BARTON: Get the rule out and then the guidance would
MS. McKENNA: Yes, with some -- with a period to achieve
full implementation while we proceed with the guidance.
MR. BARTON: -- with the guidance, okay.
MS. McKENNA: That is an issue I'll get to in a moment,
where we are in terms of getting back to the Commission.
MR. BARTON: I understand what your thinking is on that.
MS. McKENNA: Yes. That is what we have on the comments.
We are still trying to wade through them all, and make sure we captured
DR. APOSTOLAKIS: Would you remind me what Part 72 is?
MS. McKENNA: Part 72 -- requirement for licensing of
independent spent fuel storage facilities and monitor retrievable
storage, essentially dry cask type storage facilities.
They presently have a section 72.48, which is virtually
identical to 50.59.
We are in the process of going through the comments. I
mentioned that we had a large volume of them. The comments are on a
number of different topics.
What we did was go through each of the letters and identify
in a particular comment what aspect of the rule that it applied to and
then kind of group together the comments on particular issues and to get
to an overall view of, okay, in this area commenters either generally
agreed or agreed with some degree of suggestion or changes or in certain
areas there was four or five different themes we were hearing to try to
help us get our arms around the comments and the issues.
DR. WALLIS: I am concerned by this that you don't just get
tossed hither and thither by all these comments, that you have some
basis for a rational decision. I don't see what it is.
Earlier you seemed to think vagueness was better than
clarify and I am not sure that is a good criterion.
I think simplicity is certainly desirable --
MS. McKENNA: Simplicity --
DR. WALLIS: I would like for instance Commissioner Diaz'
approach where you actually state something simple and then try and use
that to assess all these comments and I don't see that in your
approach -- unless you have some logical train of thought, you won't be
able to analyze these comments.
The big question behind it all is what is the effect of
doing this on nuclear safety and I don't see how you are going to assess
MS. McKENNA: Well, I think --
DR. WALLIS: Whatever you do it's not just responding to all
these inputs. You have got to say, look, rationally looking at all this
stuff this is the way to go and I am going to defend that position on
MS. McKENNA: I think we tried to lay that out in the
proposed rule as to what we thought the rule was trying to accomplish
and how the criteria and the definitions would allow that to happen and
we are still trying to use that as the template, you know, if we had
comments to say, well, this, as you say, this part really isn't clear.
Somebody might interpret the words we use to mean a slightly different
thing. Then perhaps we need to change those words or provide a little
more explanation of the words but we still were trying to operate --
DR. WALLIS: I think you should.
MS. McKENNA: -- operate within this is a process for
licensees to look at their changes that they are making, how does it
affect their information presented in the FSAR, and gauge how much of a
change it is and therefore whether they can make that on their own or
they need to have the NRC be involved.
DR. WALLIS: I think you also need to assert some
principles -- this is why we made our decision -- and stand by them and
not get too lost in the details.
MS. McKENNA: Yes, I would agree with that and as I say I
think it goes back to the comment about preserving the licensing basis
is really what the ultimate objective is and --
DR. WALLIS: That sounds good. I am not quite sure if it
MS. McKENNA: Well, that's why we didn't try to write that
as the criterion in and of itself, but that is the intent and that in
terms of gauging whether what we are doing is on the right track, I
think that is what we look to as what does it -- does the language lead
things in that direction, will the changes be looked at appropriately,
will the right ones come to the NRC.
DR. WALLIS: Well, when you get interviewed by television
reporters saying what is the effect of this legislation, you are going
to say we have reduced the burden on industry without sacrificing any of
the safety of the plants, or something like that.
MS. McKENNA: That is a good statement.
DR. WALLIS: Really assertive statements that make sense --
we have simplified the regulations so that they are more understandable.
DR. APOSTOLAKIS: I don't know that we have simplified them.
Would you go that far?
DR. SHACK: Clarified?
MS. McKENNA: Clarified --
DR. APOSTOLAKIS: Clarified is a little bit --
MS. McKENNA: -- is a better word, yes. I mean sometimes
you sacrifice simplicity for clarity because you explain more what you
mean but yes, clarify might be a better word.
DR. APOSTOLAKIS: Isn't part of the problem here, maybe the
problem, the fact that in the SAR and the traditional deterministic
licensing process we don't have a metric that reflects the inpact or
contains the impact of all these thousands of decisions and numerous
other things in one place, so now you have to worry about all these
decisions independently, and what you are trying to do with 50.59 in its
current incarnation is to make sure that all these decisions or none of
these decisions is affected in a serious way.
All this stems from the fact that you don't have a common
method, namely the equivalent of CDF perhaps at a lower level, so it
will be sensitive, although that is an issue that we need to discuss.
You don't have that common thing that says, oh, gee, everything I have
done to the plant now results in this number, and I don't want to change
that number by much. Okay?
So this is really the heart of the problem here, that you
are trying to deal with a lot of different things, different decision,
different parameters, design decisions individually. Then you have
You say, well, the original decision was based on judgment
to some extent, so the change will be based on judgment. Isn't that
MR. BARTON: Yes, but what kind of metric would you use,
because CDF I don't think is the right metric, because --
DR. SHACK: This is the champion of the integrated
decision-making process and the enemy of bright lines.
DR. APOSTOLAKIS: No, no, no --
DR. APOSTOLAKIS: -- no, no, no. Wait, wait, wait, wait,
I would not make my decision only based on CDF, for example,
but for example I can take a system, just one system, and take its
unavailability. Now I am way down there now, okay, and take the
unavailability of the system and say with 50.59 I will allow everything
that will not change the unavailability by more than such-and-such.
DR. MILLER: Is there any reason you can't do that?
DR. KRESS: Of a system important to safety.
DR. MILLER: Is there any reason you can't do that under the
DR. APOSTOLAKIS: Yes. It is not allowed. There is no
place to put the unavailability anywhere. They will take you back to
the licensing basis. Unavailability was not part of the licensing
MR. BONACA: George is correct. In fact, to the point where
if you perform an evaluation of probability based on hard data, on the
PRA data, and then you say that shows that there is no increasing
probability, oftentimes you are against the regulation because the
regulation says the original SER was granted based on engineering
DR. APOSTOLAKIS: Yes.
MR. BONACA: And so even if you show me this analytically,
you haven't concluded that, and that particularly where you have
commitments supporting the original decisions which were to do with
diversity, operation and redundance, so that is really where truly you
really are discouraged from using any PRA in the current environment.
I would also like to say that from what I have seen, current
PRA, Level 1, using CDF, can be extremely successful for safety
evaluations in fact if you use it because what is important is the
engineering analysis or the discussion that tells you which way you have
gone and the soundness of it.
I mean you are not limiting yourself to the bottom line that
you can do that, but the discussion is very informed typically and very
credible, so there is a place for PRA, no question about it.
DR. APOSTOLAKIS: But right now there isn't.
MR. BONACA: Right now there isn't.
DR. APOSTOLAKIS: You have to change something more
fundamental, because if I bring up the issue of unavailability, it was
not in the original part of the integrity -- part of the integrity? How
did you put it? I have to memorize it -- preserve the integrity of the
licensing basis, so in that sense I think we should drop all references
to risk and probability and acknowledge this was a judgment, and if we
manage -- I mean we granted the license so I don't see -- you know,
preserving that integrity should be on the same basis, judgmental, and
stop thinking about other things.
DR. WALLIS: So integrity of licensing basis is another way
of saying the judgment of the NRC?
DR. APOSTOLAKIS: Yes. Yes.
DR. WALLIS: We have got to protect the ability of the NRC
to make qualitative judgments.
DR. APOSTOLAKIS: Well, no. Preserve the quality of the
judgment that was made at the time the license was granted.
MS. McKENNA: Right.
DR. WALLIS: Well, I think then you are reading the minds of
someone who did something 30 years ago and you can't do that.
DR. APOSTOLAKIS: Evidently they can.
DR. POWERS: And I don't think we can turn back the clock
either. I think we have to acknowledge that we have, what, 3000 reactor
years of operating experience and we have a new technology that we have
exercised broadly that has brought to us information on things like
reliability that simply cannot be ignored and in fact are now part of
the engineering judgment process.
DR. KRESS: The problem of preserving the licensing basis is
that each plant has its own licensing basis so that means something
different for every plant. Not all parts of the licensing basis have
any relevance to real risk metrics anyway -- so I don't see that as a
good goal. I mean I don't even think --
DR. APOSTOLAKIS: What, preserve?
DR. KRESS: Yes. I don't even think that is a good goal to
have, frankly. We're kind of with it to some extent --
DR. APOSTOLAKIS: We are stuck with it, yes, for the time
DR. KRESS: -- but I think in terms of risk-informing 50
that we don't want to use that --
DR. APOSTOLAKIS: Oh, no -- no.
DR. KRESS: 50.59 is part of 50, so why should I not even
abandon it for 50.59 also?
DR. APOSTOLAKIS: But there is a short term and a long term.
DR. KRESS: Yes, in the short term.
DR. SEALE: This is an interim rule.
DR. KRESS: You are still talking about short term and there
maybe it has some relevance in the short term, but in the long term I
would want to abandon that as a goal.
DR. APOSTOLAKIS: I am a little bit perplexed by a statement
Mario made earlier, that the word "probability" is in fact needed. Why?
MR. BONACA: No, I agree with you that typically it is there
to preserve certain commitments, as you said before, and I am intrigued
by your recommendations, okay? I think that you are correct. Maybe we
would have a paramount change to the verbiage and what I want to try to
say is that probability was used by non-probability experts for a long
time in designing these plants and they looked at it in terms of, you
know, what is an anticipated transient, one that will happen with
So you design certain components to meet certain criteria
and so on and so forth, so there was a loose application of the word
DR. APOSTOLAKIS: Right.
MR. BONACA: -- that went so deep into the design of these
plants that at this stage in attempting to achieve an agreement on words
between the industry and the regulator, it doesn't pay -- it would take
I believe a long time to go deep and to modify it.
You know, conceptually I totally agree with you, George,
DR. APOSTOLAKIS: What I understand you are saying, Mario,
is that, and maybe I am wrong, is the people who put together the
regulations did that.
MR. BONACA: Well, actually it was the designers in many
DR. APOSTOLAKIS: Okay.
MR. BONACA: The standard, the ANSI standard.
DR. APOSTOLAKIS: But then, if I get an application from
Palo Verde I can review the proposal to grant the license without any
reference to probability. Can I do that? Because I am using now what
is in the books.
Now the guys who put them in the books used qualitative
judgments, but I do not have to do that because the book says when it
comes to pressure make sure this happens, when it comes to temperature
this is allowed, so in that sense the word "probability" does not appear
in the licensing basis. Is that an accurate understanding?
MS. McKENNA: I guess I would say that, however you count
them, there are three or seven criteria. Not all the criteria would
apply to all changes or issues, that some cases the thing you are doing
it only has something that affects, you know, a dose thing. It has
nothing to do with probabilities or anything, so I would say that that's
DR. APOSTOLAKIS: So Eileen, do you think it is feasible to
go back and see whether the word "probability" can be dropped?
MS. McKENNA: I think it is something that could be
considered on the timeframes that we are talking about. I think it
perhaps would be an additional complication that would get in the way of
our achieving where we're going to go.
The other comment I wanted to make --
DR. APOSTOLAKIS: We could go line by line and cross it
MS. McKENNA: The other comment I wanted to make on this --
DR. APOSTOLAKIS: Use the delete button.
MS. McKENNA: -- this topic where we are referring to such
attributes as diversity and redundancy and things, I think in some cases
those things are embedded in other requirements -- for instance, GDC and
things like that, that even if your 50.59 criteria allowed you to
consider those in this context, the need to satisfy the underlying
regulation may prevent you from doing that anyway.
That is always something that has to be considered in
looking at a change as to whether you still meet the body of
regulations, not just whether this change I need, I need to ask your
permission on, but if it is something that would removed redundancy
where it is required even if they asked our permission we would say no.
DR. KRESS: George, you will find the word "probability" in
the licensing basis.
DR. APOSTOLAKIS: I will?
DR. KRESS: Yes, you will. It has to do with the
determination of the design basis accidents and their frequencies and
you will definitely find it as part of the licensing basis.
DR. POWERS: But it is important to recognize that the scale
that was being used at the time broke it down into not decades but
duo-decades. That is, it was in factors of 100 in which there was
likely, less likely, and --
MS. McKENNA: Much less likely.
DR. POWERS: -- and beyond the scale of human reason, and I
mean that seems like a crude scale to many of us here, but that scale is
still an operational scale if you design a facility for the Department
That is why I think, yes, there is a confusion, George, when
the -- in the modern time when we use the probability we are used to
thinking about the differences between 4 times 10 to the minus 4 and 2
times 10 to the minus 4 -- I mean a much, much narrower scale is being
used now and it does cause confusion in that respect. Okay.
DR. APOSTOLAKIS: Yes.
MS. McKENNA: Just a couple more bullets here, just in terms
of schedule and approach.
As I mentioned earlier, we originally had a date of February
19th for the final rule and we went back to the Commission in December
and said we really can't do justice to this and give you a final rule in
What we instead offered to say -- well, we realized there
are certain issues that we really -- the Commission has had some input
on and that there is a diversity of views perhaps among the Commission.
Perhaps it would be more effective if we came back to the Commission
with some of those particular issues and allowed the Commission to have
the benefit of what we learned from the comments and the current
thinking and give us some feedback on where we are before the Staff
expended the time and effort to put together a final rule package with
all its accompanying accoutrements and review process for something that
is kind of missing the mark as to where the Commission wanted us to go.
So that is our current plan of attack is to provide a paper.
Presently the date is in fact still the February 19th date but it is a
little different kind of paper than a final rule. What it would be
instead is taking those issues I had listed on the slide before, at
least most of these -- I am not sure every one will be in there -- and
say this is what we have learned from the comments, this is what we
would presently recommend -- please give us your feedback, Commission,
as to whether you agree or you have other ideas and that once we get
that feedback, then we can do a better job of preparing a final rule,
come back to the Committee with the final rule recommendations.
You will obviously see the February paper and have the
benefit of the Staff's thinking on those issues and then, say, when we
get some Commission feedback then go ahead and try to put together the
final package, make sure all our stakeholders within the agency are
We have been trying to do that but with the timeframes
available to make sure that between us and NMSS and the regions and OGC
and all these others that we are all on board. It just takes time to
make sure everybody understands, thinks the same way and that we haven't
missed something in the process.
DR. WALLIS: I don't get the sense that you are on a
convergent path to resolving everything.
MS. McKENNA: I would say probably not everything. I think
on most of these issues we are on a convergent path.
DR. WALLIS: I looked at the blizzard of comments and then I
looked at the comments from the Commissioners and it seems to me there's
a great deal of work to be done to figure out what is the right
resolution of the various points of view.
MS. McKENNA: I think in a couple of specific areas that's
true. I think in most of them that we are close. Margin is probably
the one where there is the biggest stumbling block.
MR. BARTON: Margin of safety is the biggest issue, biggest
MS. McKENNA: Yes.
DR. POWERS: I mean in all honesty, we have been close a lot
of times. I mean it seems to me that beginning with NSEC 125 and going
to 96.07 the contention has always been rather close except for a couple
MR. BARTON: We are further apart now than ever.
DR. POWERS: I think we are too.
DR. SHACK: Yes, but this time you get to rewrite the
guidance and the rule at the same time.
MS. McKENNA: And make them match.
DR. SHACK: And make them match.
DR. POWERS: I think that just adds degrees of freedom in a
nonconvergent algorithm here.
Let me ask you about your final bullet on this slide.
MS. McKENNA: Yes.
DR. POWERS: Many seem to have questioned the resolve that
the Staff has to go to Phase 2.
Can you offer a testimonial on your resolve here?
MS. McKENNA: Well, I think one of the points in terms of
future changes, that really kind of came on two fronts. One is the
question of scope, as to whether FSAR is the right set or whether that
should be current licensing basis or some risk-informed subset of
information in the licensing basis, and the other I think is a question
of criteria, whether you have some different criteria that allow more
consideration of, say, delta CDF or other risk metric as a decision
criteria for moving forward.
What I think we have suggested in the paper on
risk-informing Part 50 in general, which I think we still believe is
probably a more important thing to do before you try to take on the
specific of 50.59 -- scope has been put forward as one of the first
things to take on, and I think that those kind of considerations on
scope will then shed light on what the scope of 50.59 is.
Just in the fact that if this is the scope of things that they are -- I
think the words were used "regulatory treatment," shall we say. Well,
50.59 is a means of regulatory treatment, and that they will naturally
kind of come together. At the time that you redefine that scope, then
it may be necessary and useful to also give further thought to the
question of criteria.
DR. POWERS: If I was a suspicious type, I would say very
good strategy, admire the strategy, create a blizzard of additional
questions, and I can delay this thing forever. I mean, come in and say
oh, well, we'll just think about the scope, we'll think about the
applicability, we'll expand it here, create a lot of additional
questions, and I can avoid having to make Phase 2.
MS. McKENNA: I guess we look at it as making sure that
whatever we do in 50.59 is consistent with the underlying set of
requirements in Part 50 that we're trying to judge the changes to, and
if that means that it has to wait until some of those other things get
settled, I think we feel that that's perhaps a better plan of attack
than trying to move forward on something that may not match up well with
how the rest of things will come out.
MR. BARTON: Well, if you think you're going to a risk
approach to Part 50, where would 50.59 come in your pecking order?
Pretty far down the list, I would guess.
MS. McKENNA: It's probably not the first, but it's not the
last, I guess is how I would --
MR. BARTON: It's pretty far down the list. Yes. Okay. I
don't know, it's just my perception of thinking where you're going to go
with the risk-informed approach to Part 50.
MS. McKENNA: Well, I think in terms of, you know, you've
got to decide what your objectives are of the risk-informing Part 50 and
how you think risk-informing 50 is not who's going to benefit you --
MR. BARTON: Right.
MS. McKENNA: Considering the kinds of changes that we're
looking at now, it may or may not help people to have to look at them in
a risk-informed manner. You know, there may be -- I think it was
mentioned earlier that there is a, you know, that if you're going to use
a risk-informed approach, then you're going to have to have an analysis
base or other kinds of things in order to support that, and whether you
want to do that or not as a particular licensee, you know, is obviously
DR. APOSTOLAKIS: No, but the benefits, though, must be
commensurate with the effort.
MS. McKENNA: Yes. Yes.
DR. APOSTOLAKIS: So you're not going to worry about little
changes here and there any more. I mean, I'll give you an example. All
changes that do not change the CDF by more than 10 to the minus 5 will
not be reviewed. Now you're going to see people doing a lot of
calculations, because the benefit is tremendous. So that's something
that we have missed in the past.
MS. McKENNA: Um-hum.
DR. APOSTOLAKIS: We thought, you know, we'll make it
risk-informed, but it will be the same 50.59. Well, it won't be. So
maybe it will be worth it, worthwhile doing these extra analysis.
MS. McKENNA: Yes. I think the other thing we have to
decide is the view of a degree of hybrid of terms of a CDF and any other
criteria that you would want to apply, you know, certainly in looking to
Reg Guide 1174 as our risk-informed approach, there was not just CDF,
there are other factors that are in there, and if you were doing it in a
50.59 context, how you would apply those other factors would have to be
So I think that's -- the last bullet I think we just covered
in terms of the potential future changes.
DR. POWERS: Let me say that I think if we're going to make
progress, thrust really has to be to narrow and refine the focus and not
broaden it, or you will simply bollix yourself up with inalterable
questions. Maybe indeed the strategy is start at the top with 50,
because otherwise you can always find some reason that the scope has to
be broadened because of some other rule within the system.
MS. McKENNA: Well, I think some of the meetings we had on
the risk-informed options brought that to light, that even if you
changed 50.59 and risk-informed it and didn't change other requirements,
how far you can go? You may run up against these other regulations, you
know, that okay, we want to make this change under 50.59 but we can't
because it's the safety-related definition that gets in our way or
whatever. I mean -- so that's another reason why looking at in the more
wholistic way we think is the way to go.
MR. BARTON: Any other questions for Eileen at this time?
DR. SEALE: Are we going to write a letter?
MR. BARTON: Not at this time. This was the information or
status briefing on where they were with respect to feedback on the rule
they laid out for public comment.
Now our schedule on this issue, we're far from being through
with this ourselves, at the March meeting will will review the
Commission paper that Eileen talked about that they owe to the
Commission in the middle of this month which ought to talk about
reconciliation, public comments, and proposed positions for the final
rule. They're going to get some feedback from the Commission on that,
and they owe a final rule on -- I think at that time we need to say
something to the Commission, in the March meeting, because the final
rule schedule is April, the end of April, the next time we will get a
shot at this. Or we may not even get a shot, I guess. In the May
meeting we'll review the proposed final rule, and I guess if we've got
some problems with it then, we can say something. But I think we've got
to get our oar in the water to the Commission in the March meeting.
DR. POWERS: Well, needless to say, we get a little oral
communication in a few hours.
DR. SEALE: In a few hours.
DR. POWERS: But I think a more definitive statement is
going to be made in our March meeting. And we might want to take some
time in this meeting to think about what the length and the breadth of
that letter ought to look like, because it probably merits some advance
planning in anticipation of the kind of information that we're likely to
get out of -- at the final briefing.
MR. BARTON: If there are no other comments, Mr. Chairman,
I'll turn the meeting back over to you.
DR. POWERS: Thank you. We are now scheduled for a break.
John's given me back 15 minutes. I appreciate that. So I will have us
break until 10:20.
DR. POWERS: Gentlemen, I want to come back to our session
here. We're a little early for our speakers on this session, but I
thought it would be worthwhile to begin the session with a little
internal discussion on where we thought we should be going with the
improvements on the inspection and assessment program and even where we
thought we were going on 10 CFR 50.59.
And so John, if you can give us some introductions --
MR. BARTON: Well, on the assessment program, we had a
subcommittee meeting a week or so ago, at which time we were briefed on
what was in the Commission paper SECY-99-007. At that time we had
Members raise several questions for the staff. The questions that were
raised at that time are in your package under Tab 3 under the status
report on page 5, were the issues that we had raised with the staff at
The staff is prepared today to go give us an overview of the
status of the process, and we should get into our issues that we raised
in the subcommittee meeting as laid out on page 5 of Tab 3 and try to
get answers to those questions.
We do owe a letter to the Commission on the integrated
assessment process coming out of this meeting. I guess the least we've
heard on the overall plan was details -- in the overall process was
details in the transition plan. We heard some on the assessment process
last time, had some questions on that.
DR. POWERS: My understanding is the Commission's been
briefed on this.
MR. BARTON: It may have. I don't --
DR. POWERS: I think I've gotten a communication from Mr.
Markley on questions they posed in that briefing process.
MR. BARTON: Oh, yes. Yes, yes, yes. I've got too many
walls in the area today.
DR. POWERS: Yes, too many pieces of paper in front of you.
MR. BARTON: Too many pieces of paper.
There was a Commission briefing on January 20, I believe.
There are also questions from the Commissioners that came out of that
briefing. I don't know if you've got copies of those. I have a copy of
that. We will also question the staff on some of those issues.
DR. POWERS: Do we anticipate in our meeting with the
Commission today that they would have questions on this?
MR. BARTON: It's not on the agenda, but if they should ask,
you know, where are we, I guess we could answer that. I don't know that
we have the Committee, you know, agreement on, you know, our position on
where we are at the inspection assessment process, but we could answer
DR. POWERS: We can answer status questions.
MR. BARTON: Yes, that we should be able to do, definitely.
DR. POWERS: Okay. And our intention is indeed to produce
MR. BARTON: Produce a letter.
DR. POWERS: A letter, and this issue would then be resolved
as far as we're concerned? This is one that has recurring -- recurs for
MR. BARTON: Yes, I think as far as we're concerned, Dana,
this is the final action we need to do, is just put out a letter on what
we believe, you know, what our position is on the overall assessment --
integrated assessment process.
DR. POWERS: Okay.
MR. BARTON: I'm not anticipating any other meetings or
briefings on this. Although there are a lot of open, you know, open
items in the plan. It's going to be out there for trial. There are
going to be pilots. And there may be an opportunity down the road as
this plan gets fully implemented over the next year or two to have
further discussions with the staff. I don't know.
DR. SEALE: Wouldn't we expect --
MR. BARTON: There's nothing planned at this time on that.
DR. SEALE: Wouldn't we expect to hear something about the
results of the pilots --
MR. BARTON: Yes.
DR. SEALE: And maybe we want to stake that out.
DR. POWERS: Well, I guess that gives me an idea of what our
overall strategy here is. My right clock tells me that the time has
come to move back into this session, and the left clock tells me we may
be a couple of minutes ahead.
DR. SHACK: It's 10:26.
MR. BARTON: We need to synchronize the clocks.
DR. POWERS: I would hope that we can synchronize clocks one
of these days.
So, John, if you could go ahead and introduce the speakers
for this, and give us the appropriate background on this issue.
MR. BARTON: Thank you, Mr. Chairman.
The purpose of this meeting is to continue the Committee's
discussions in review of the proposed improvements to the inspection and
assessment program with the staff, including initiatives related to
development of a risk-based inspection program performance indicators.
In preparing for our letter on this subject, we've asked the staff to
today give us an overall picture as to, you know, where they are, what
are the open issues, and also I would anticipate that Members' issues
and questions that were raised in the January 26 subcommittee meeting
would get again addressed to the staff today.
DR. POWERS: And I guess I would appreciate it if you think
that any of the questions that the Commission posed to you in your
briefing remain unanswered or were left unanswered if you have responses
to those that you'd like to share with us, that would be useful.
I got the impression from my digest of the questions that
most of them were answered by either more inspection or we'll tell you
after the pilots, but -- both of which are appropriate responses.
MR. GILLESPIE: Yes, that's true.
MR. BARTON: So at this time I'll turn it over to -- Frank,
are you going to take the lead on this?
MR. GILLESPIE: Yes. Let me give you a current status on
some documents that will probably be coming out in the next couple of
weeks, because we're continuing to move.
First, the selection of the pilot plants. We're working
with NEI and we've pretty much narrowed down the eight pilots. NEI is
working with the people at those pilots. One of our criteria was
hopefully that it was someone participating as one of the utilities
participating in kind of the equivalent utility group to us so that
there was actually some knowledge within that group so that we weren't
starting with a plant that's been totally uninvolved and asked to be a
volunteer and not know what they're getting into.
We're not going to name those plants until a couple of
things are done. Once we basically say shake hands with the industry,
we're going to need to inform the Commission. We've kind of taken on a
self-imposed requirement. We'll then call the State representative that
works with us from that State to let them know, offer a briefing to the
State representatives and State people so they understand what we're
doing, particularly some States like Illinois, New Jersey, Pennsylvania,
which are very active in this area. So we've got some steps to do in
the next couple of weeks before we see the names coming out in a press
DR. POWERS: I guess I understand your reluctance to name
the plants, because of course there's slips between the cup and the lip
here that are always potential, but let me understand better criteria
MR. GILLESPIE: Okay.
DR. POWERS: There's been a lot of work in the severe
accident and probabilistic risk assessment space on defining
representative plants, and they, you know, you can -- all plants are
different, so you never get a perfect representation. But I think they
generally feel they've got a broad representation. The defect in their
strategy was they didn't ever have a replicate. There was no measure of
the experimental error in these pilots. Are you getting what you think
is a broadly representative group of plants, and have you made
provisions for measuring your experimental error in the pilots?
MR. GILLESPIE: The answer is we hope we've gotten a broad
representation, although limiting ourselves to eight. There's P's,
there's B's, and our perspective I'll say of performance in talking to
one utility which had multiple plants, we told them which plant we'd
like, and one of our criteria was some things happen to have -- there
needs to be some things going on we can count. Taking a plant that has
every indicator in the green zone, the licensee response zone, would not
be a measure of what we're trying to measure.
So we had a mix of plants. And this utility said but we'd
like you to take this plant over here, not that plant over there. Well,
this plant over here wouldn't attempt to measure what we're trying to
measure to try to get it at these kind of differences. So that's the
tentative negotiation that's kind of taking place right now.
DR. POWERS: I think what you told me is that the things
that we often think of in terms of broadly representative, is it a P, is
it a B --
MR. GILLESPIE: Those things are covered.
DR. POWERS: Subatmospheric containment, is it an ice
MR. GILLESPIE: Yes.
DR. POWERS: That's not nearly so important as that you get
a range of performance.
MR. GILLESPIE: That's important to us, that we exercise the
system, because if the evidence was that everything was a null set, then
we certainly don't have a system -- we haven't exercised anything. So
the focus here is really on those plants, not that there won't be some
excellent performers in the eight, but clearly we're looking for a broad
range to exercise the system to ask the question are we seeing what we
would expect to see?
The other element we're going to have in doing a true pilot
is let's take a normal two-unit site that has three residents. And I
already said we're not picking the absolute best performer in the
country, so it's going to have the third resident.
Under this program, it's not envisioned the third resident
has anything to do as part of the pilot. So we'll be needing to devise
a method of using the third resident almost as a way of the independent
eyes. He's going to have to have a different routine to try to keep him
from influencing the norm, yet it gives us an opportunity to have a guy
on site who can kind of follow around and say did these guys go deep
enough in challenges. So we've got some challenges in designing the
pilot just in the environment we're going to be in.
MR. JOHNSON: Frank, can I --
MR. GILLESPIE: Sure.
MR. JOHNSON: Just to add a couple of other things that we
thought about, that are less important than the things that Frank
mentioned, but they also went into it. And that was -- one of them was
we wanted to pick plants, or NEI suggested plants who were members of
their task group who have been working with the process all along, the
thought being if those plants are chosen, then they have an
understanding of where we're trying to go, and so we're not starting
from ground zero when we try to get the pilot plants up to speed. So
that was something of a sort of a consideration.
And the other thing was we found a couple of cases where
while there would have been a plant that NEI would have suggested, we
thought it wasn't a good idea, but because, for example, we had a
resident who was turning over and so we didn't want a lot of internal
NRC things in terms of changes in staff or whatever to impact the pilot.
So those were a couple of other considerations.
MR. GILLESPIE: Yes.
MR. JOHNSON: Okay, the other element is, and this was a big
question from the Commission, if there was a major academic or
philosophic hole in our package, it was the scale to measure inspection
results. Anyone can calculate a number in a model, given you know the
piece of equipment doesn't work. That's relatively straightforward.
But how do you deal with the subjective results? We do have kind of
about the fourth revision of a draft that we're now testing some past
results through, so the staff is right now exercising a first draft of
They do not feel comfortable necessarily issuing it and
putting their name on it until they've run some more examples through,
so we're taking some examples from Waterford, I think it was, D.C. Cook,
some Millstone examples of things that were actually found and running
them through what is a two or two-and-a-half-stage screening process
with a first set of questions that are attempting to get rid of the
trivial, and then a little more in depth, if you violated an LCO, for
example, was it one day, was it seven days, was it 30 days over, to give
some perspective -- it's risk-informed, not risk-based -- but some
perspective of importance to violating the LCO is not the end of the
world if your shutdown happened to take an hour longer and it's seven
days in one hour. So that's kind of stage 2 screening.
Stage 3 of screening is you get the SRA from the region
involved because it's something that does require a little more in-depth
perspective, calculational knowledge, sense of uncertainties. So we're
exercising that now.
I would guess that we'd see a draft coming out also in about
the next two to three weeks, because this was really a major point that
we promised we'd show up with in March. And they're exercising it with
one major, major criterion, and that's to have a high probability of not
having a false negative. So the screening process doesn't have to be
perfect relative to a high-risk event as long as the high-risk event
doesn't get screened out. So if we allow some low-risk events through
the filter, that's okay, and that's the kind of perspective they're
trying to test it to to make sure that a high-risk issue would not
artificially get screened out too early. So it's kind of a lopsided
test. We're allowing deficiencies in one direction and trying to assure
no deficiencies in the other direction.
DR. POWERS: Well, that's an interesting part of your
program, and one I think is pretty well thought out. I'm still
concerned about what I call the experimental error in your pilots, and
the way an experiment should measure experimental error is you do a
replicate. Do you have that?
MR. GILLESPIE: No, not a replicate in a true sense.
There's multiple P's and multiple B's, but there's different performance
-- there's different known performance. I mean, deliberately picked for
different performance. Right now I wouldn't want to say that there's --
there's a good comment, that if we had two P's and they were looked at
as approximately the same performance, comparing the relative findings
from the two, to see if the nature of what was being seen is consistent.
DR. POWERS: You're going to draw conclusions, and --
MR. GILLESPIE: Yes. I can only say at this point we hadn't
gotten that far in the thinking of the comparisons, but it's a good
point. I understand it. And I think we need to look at what we can do
with kind of a small sample.
DR. POWERS: Um-hum.
MR. GILLESPIE: I think in Statistics 101 at one point if
your sample was at least 15, you're okay, but if you're less than 15,
you're in trouble.
DR. POWERS: There are a lot of things, many of them -- I
don't want to draw this analogy too far, because it can be carried to a
MR. GILLESPIE: No, but --
DR. POWERS: But it does seem to me --
MR. GILLESPIE: A qualitative comparison, though, could
easily be done from that perspective.
DR. POWERS: I think you just need to be aware that when you
draw conclusions out of these pilots that there's liable to be an
experimental error. Now you may not be able to do it on this plant
versus this plant. Maybe you can identify that experimental error on
this finding versus this other finding --
MR. GILLESPIE: Um-hum.
DR. POWERS: We found the same thing at two plants, so there
must not be any error in that one, but we didn't find it at another one,
and should have. Maybe there's an error there. Maybe you can get some
capture of how definitively to draw conclusions from your pilots.
MR. GILLESPIE: Yes, and this is what I talked about this
extra resident. We have to be kind of artful in how we use him, because
all plants are operated differently.
DR. POWERS: Yes.
MR. GILLESPIE: And it's unlikely that we'll find exactly
the same error at one plant or another. In fact, if we do, we have to
ask the question do we have a generic problem.
DR. POWERS: [Laughs.]
MR. GILLESPIE: So I'm hoping that if the error is in theory
randomly distributed among the plants, we would test is the severity of
what we call something at one plant consistent with the severity at the
other plant. Clearly that's one of the things we're exercising is our
own criteria. If you apply the criteria, does it make sense across
But part of my thought process here is, we have to have some
outside view in because if the inspector -- if we have him looking at
the wrong things or at a disproportionate number of things, do we want
the other inspector who's not part of the process to look at something
that we're not looking at, one of those things that we threw out because
we had PIs to confirm that not looking at it is the right thing to do.
That we won't get by a plant to plant comparison, because
the people at each plant will be looking basically at fundamentally the
same issues, the same high risk components.
The question would be have we lost anything by not looking
at support systems enough, so should we take this extraordinary resource
which is located on the site and have him looking at the things that we
actually threw out, to not look at as much, to confirm that what we said
didn't need to get looked at was the right thing so we have some design
of experiment work that we still need to do on what are we trying to
prove and how are we going to get some essence, the support that we made
right or wrong decisions.
DR. POWERS: What you want to do is say all my uncontrolled
variables indeed are behaving randomly.
MR. GILLESPIE: Yes, yes.
DR. POWERS: But again I caution you about drawing the
analogy too strongly.
MR. BARTON: Are you always going to have the luxury of an
MR. GILLESPIE: No.
MR. BARTON: What are you going to do in that case?
MR. GILLESPIE: Oh, during the pilots we will.
MR. BARTON: During the pilots you will?
MR. GILLESPIE: Yes. I am going to anticipate during the
pilots we will.
MR. BARTON: Okay -- if there is not one assigned to the
site, there will be one to do the pilot?
MR. GILLESPIE: Yes. If there is not one at the site, we
are going to have to do some external monitoring to answer those kind of
You know, if there is any concern when you -- the public
concern -- let me switch to public confidence. The program is smaller.
How do you know you threw out the right stuff and kept in the right
stuff? That is a fundamental question that we are going to have to
If it means putting some extra people to confirm that at a
site that doesn't have the extra resident right now, we are going to
have to address it. We have to do the effort anyway -- because that is
a concern and that is a specific public concern that we are going to
have to look at.
So there's two documents. I'll be happy as soon as we are
ready to relieve them in the next two weeks to send copies over here and
Mike can distribute them, but the second document, I think, is the key
one and that is the measure of what you do with inspection results. I
probably feel more comfortable with what the Staff's done than the Staff
does right now. They want to test it out a little more on some real
results and kind of prepare a report on it.
The reason I feel more comfortable is right now we have no
scale, and probably having a scale someone could take shots at is better
than no scale at all, so that is the reason I feel comfortable we are
With that, Mike, you were going to try to touch upon some
questions on the assessment process?
MR. JOHNSON: Yes, I was going to try to touch upon some
questions on the assessment process. I guess what we have handed out is
the slide package that we handed out last time and I wasn't going to go
over the slides. In fact, how would you like me to proceed? Would it
be best for you to ask questions or me to --
DR. POWERS: I would just --
MR. BARTON: Why don't you just go on. We will jump in and
interrupt you like we usually do.
DR. POWERS: Yes.
DR. APOSTOLAKIS: Weren't there some questions raised at the
DR. SEALE: Yes --
DR. APOSTOLAKIS: Why don't you jump through them.
DR. POWERS: Well, I would say just go through the
presentation and emphasize those points that are a response to the
MR. GILLESPIE: We are kind of at a disadvantage because we
didn't get the questions.
DR. APOSTOLAKIS: No, I mean during the subcommittee
meeting? Oh, you mean the Commissioners' questions?
MR. GILLESPIE: No. Whatever is on Tab 3 of Enclosure 5.
DR. APOSTOLAKIS: Tab 3, page 5 is the --
DR. POWERS: Oh -- our apologies.
MR. GILLESPIE: How can we answer them? That's why I felt a
Go ahead and go down the questions and just go through those
first and then we'll see if we have any residual --
DR. APOSTOLAKIS: These are the Commissioners' questions?
DR. SHACK: No, no, no -- our questions.
DR. POWERS: There are relatively few of them anyway. I
will be glad to share my copy with you.
MR. BOEHNERT: I've got it.
DR. POWERS: Okay, just taking the questions, why a full
year pilot was not proposed -- and the answer there is as we stated it.
We really didn't want to do a full year pilot. We in fact went with a
six-month pilot out of consideration for the schedule that we were being
asked to meet.
But I mean it wouldn't be the first time that somebody has
gone to the Commission and said your schedule is unreasonable.
MR. GILLESPIE: Well, we did that. In fact, the original
schedule was a three-month pilot and the program fully implemented in
October, so what we did is we balanced -- well, we would like a year --
which would be a full cycle --
DR. POWERS: That is the reason. You would get a --
MR. GILLESPIE: We said, well, we could do a credible job in
six months given that even in that first year there's going to be bugs
we are going to have to work out and have a feedback loop, and as Mike
said I think in his last presentation, at the end of the first year
there is a major reassessment effort that is going to have to take place
to make adjustments.
So with that knowledge that there could be a major
readjustment after the first year, we gave that also to the Commission.
At the Commission presentation we said with that knowledge we could
probably do this in six months but we are going to recognize that it is
not the perfect system in January, it's the best we can do with working
at what we can work out in six months, and there is nothing much more in
depth behind it than that compromise.
MR. JOHNSON: Yes. We have done some things to try to
buttress that, if you will. We for example are going to ask plants to
report PIs every month as opposed to every quarter for the pilots. We
are going to ask them to report, to go back historically and report some
data even before the start of the pilot and I think we are talking about
two years but we haven't finalized that so we will have some PIs leading
up to the start of that six-month period.
In addition to that, we are going to try to look at some
additional plants beyond the eight plants that we are talking about for
a pilot to get some PIs on, so we are trying to, even though it is a
six-month period of time, we are trying to exercise the entire process.
We have tried to artificially raise the amount of
information we have about those plants and the frequency with which we
get that information such that we can get the most out of the pilot, but
we really are sort of confined to the schedule. If we want to get to a
January 2000 ready to implement for all plants we really did need to try
to limit it to a six-month pilot.
Frank, do you want to start? Why don't you take the second
DR. POWERS: I mean you told us the advantages. Have you
lined out explicitly what you think the disadvantages of not going the
full year are?
MR. JOHNSON: Sure. We have talked about them.
One of the disadvantages is for example we are talking about
trying to exercise the entire risk-informed baseline inspection program
that we know is going to be over a year cycle in a six month period of
time. That is going to cause us to do something artificial.
Either we are going to have to spread out pieces of the
risk-informed baseline inspection at the various pilot plants so that we
have got full coverage of the procedures, but maybe not coverage at all
the plants, or that causes you to do increased inspection at those pilot
plants to cram it all into a six-month period of time and we know plants
are not going to be wild about that, so there really would be a lot of
advantages and we recognized a lot of advantages to doing the 12-month.
DR. POWERS: All of this, of course, the aggregation of all
these problems that have been identified here in the lat five minutes is
of course to increase the uncertainty you have on any conclusion that
you arrive at.
MR. JOHNSON: Right.
MR. GILLESPIE: Now let me take it on the positive side. We
have actually attempted to provide a structure to apply against
inspection and assessment, which replaces a system right now which is
very, very, very, very subjective.
It is not clear to me that even if we don't have the perfect
system, we are proposing a system that is better than what we have, and
from that perspective the loss of momentum from waiting a whole year,
what would the incremental improvement be from six months to a year?
It is not clear to me that we are not going to be 80 percent
to where we should be in the incremental -- we'll be at 20 percent
improvement and that we can't really deal with that actually better on
an industry-wide basis. It's going to be a hassle because it's going to
be more exceptions and we are going to get more comments but we will be
exercising the system so that is kind of a trade-off.
I think if we extended this a whole year, the idea of having
a risk-informed oversight process overall would lose so much momentum
that it may be two years. I think that is a real danger, so we need to
keep a certain effort going and a certain commitment on the industry
part and our part.
DR. POWERS: Well, I think I would be most concerned about
losing momentum on this team that has done so well here, that, you know,
you guys will burn out eventually.
MR. GILLESPIE: Well, I think some of the guys we have
shifted, moved around, because --
DR. POWERS: Quite frankly, the hard part of your work is
ahead of you.
MR. GILLESPIE: Yes.
MR. BARTON: None of these pilot plants are scheduled for a
refueling outage during this pilot program I take it?
MR. JOHNSON: No, I don't -- in fact, we would like it if --
I haven't looked. Alan Madison, the Task Force leader, has looked at
that. We actually want a plant in shutdown. We want to see what
happens and so we think there is a pretty good chance based on the
population of plants we picked, but yes, we are looking for a plant in
MR. BARTON: Now if several of the eight plants get
unfortunately into a forced shutdown, what does that do to your pilot
program? Does that impact it or --
MR. GILLESPIE: Oh, no, that's great.
MR. BARTON: Okay.
MR. GILLESPIE: I shouldn't say it's great until we force
somebody to shut down, but it means something happened and the question
is would the PIs in the inspection process have picked it up before the
So that would not be a bad occurrence --
MR. BARTON: It would give you a different look but you
won't be able to track the month-to-month PIs?
MR. GILLESPIE: No, because once they are shut down, they
are down and then you go into a different process and this is one of the
holes we currently have is we promised that we would try to develop in
some sense some shutdown PIs.
MR. BARTON: That was my next question. Do you have an
operating program and a shutdown program?
MR. GILLESPIE: Yes, we are going to need a set of shutdown
PIs that go to multiple water sources, availability of ultimately heat
sink, multiple power sources being -- a kind of configuration set of PIs
that could easily be --
MR. BARTON: But that is not yet developed?
MR. GILLESPIE: No, and that is something we are committed
to doing in the next six weeks, and that is one that is going to be one
of the harder ones because of Staff availability with the right
MR. BARTON: Okay.
MR. GILLESPIE: By the way, if we don't, it's not the end of
the world because we revert back to what we are doing today.
DR. APOSTOLAKIS: Sure.
MR. GILLESPIE: So it is not a total void.
MR. BARTON: I just wondered how it would impact the whole
pilot program, since it is a relatively short period to begin with,
MR. GILLESPIE: Mike -- you are being silent on me, Mike?
MR. JOHNSON: Yes, I am.
MR. GILLESPIE: I was anticipating a question too.
MR. JOHNSON: That's no accident.
MR. JOHNSON: There was a question about the distinction
between an oversight program and a regulated activity or process and we
talked about that quite a bit, and George, that was your question.
Would you restate the question or give some context and we
will try again?
DR. APOSTOLAKIS: Okay. If you look at the document the
Staff has prepared and you go to Appendix H, where the performance
indicators are selected, the general approach is to look at the number
of plants over the industry, across the industry, and say look at
unavailability, unavailability of a particular system or the initiating
events and so on and have a histogram, which I understand was supplied
by NEI to you, of how well each plant performed on that particular
metric, so I have now, say, 35 inputs on initiating events.
Then the threshold for action is selected such that about 95
percent of the plants have a frequency that is less than the threshold
and five percent are above, so now the question is -- okay. That
creates a problem in my mind.
The problem is the following. An inspection program is a
quality control program really. You want to make sure if we borrow the
terminology from manufacturing that you have a process that is producing
something and the process is acceptable.
What you want to do with inspections is to make sure that
the process has not changed, okay, with time, so every Monday or every
first Monday of the month you take a sample of 10 items. You do your
measurements and you declare, yes, it has not changed. A change might
be a shift in the mean or increase in the variance and so on.
So the critical item here is or the point is that all you
want to do with the inspection process is to make sure that the way the
plant was remains -- that it is the same way in the future.
If I select the lower 95 percent, the threshold as the 95th
percentile of the distribution of plant to plant variability, then what
happens to those five percent plants that are above the threshold. I am
not now making sure with my inspections that even though their frequency
was higher, it remains there, it has not gone up, because if it goes
down of course it's nice.
So I am not really -- I am deviating now from the idea of
quality control because I am saying the threshold is at, say, three when
there's a number of plants that were at five, seven, and eight. What is
going to happen to those plants?
Are they going to be forced to come down to three, in which
case this is not quality control anymore. This is now regulation, okay?
See, conceptually now the process has changes. Instead of
making sure that whoever had five remained at that level or improved,
now we are telling them that five was not good enough to begin with.
That is not inspection anymore. We are beyond inspection now. You are
telling them what to do.
The answer we got last time was that while these are
thresholds that will act as a red flag that will tell us something is
going on and that we'll look into it and take it from there and you will
make sure that this will not become an additional regulatory action, but
this is really my concern.
It seems to me it is a fundamental conceptual point -- what
is the purpose of inspection? The purpose of inspection in my mind is
this plant is operating now. It has an unavailability for the RHR
system of .025. I want to make sure through my inspection program that
it does not become .06 -- but it is none of my business. I don't have
the authority to tell them .025 is too high, you should reduce it.
Somebody else should have that authority, not the inspection program.
That was the concern.
MR. GILLESPIE: Go ahead, start.
MR. BARANOWSKY: There's a lot of stuff there, George.
We could pick this threshold anywhere. We have to pick the
point in any QA process where we trip our flag, if you will, and we ask
ourselves are things getting out of whack here to the point where we
have to take some action?
In this case, we have picked it at the 95 percent for the
green to white band. You could raise the same question, by the way,
with the white to yellow band. There is not a regulatory requirement
that would keep the temporary change in core damage frequency to less
than 10 to the minus 5, for instance, so these were just points that we
picked that we thought made some sense for us to in a graded manner
increase our observation of licensees to be sure that these indications
that went beyond those points were nothing more than maybe aberrations
and that in fact the licensee would be taking appropriate actions to
maintain a level of performance that is deemed to be acceptable and
doesn't require any NRC directed activities toward them.
So these are really observation points and no action would
be taken unless there was an observation that the trend was there and
that there was a regulatory requirement that needed to be satisfied, so
that we weren't necessarily talking about forcing certain availability
or trip frequency requirements on licensees. What we're trying to find
out is if these indications in the first threshold was to get the
earliest one possible were indicative or problems that could lead to
more significant changes in performance -- because our whole philosophy
on this thing is to take small steps first and keep everything operating
in the regime where we don't have major problems develop and major
regulatory responses by both the licensee and the NRC.
MR. GILLESPIE: George, let me get to what I think your
point is and that is the possibility of an unattended result. We are
going to publish these reports. It is going to show someone who's
busted a threshold and while what Pat said is exactly right on how we
would react, there would be a published report that shows a threshold
broken, and that has spinoff potential just as SALP scores got used in
financial markets and other things, so there is a possibility of an
unintended result, and this is something after at least the first
year -- I don't think we are going to have enough information from the
pilots, but your articulation of what the program is trying to achieve
One of the things that they did when they came up with the
95 percent was they actually looked back at -- Pat, I don't remember the
years -- '91 through '95?
MR. BARANOWSKY: I thought it was '92 to --
MR. BARTON: '92 to '95 or '92 to '96 or '93 to '97 --
something like that.
MR. GILLESPIE: I forget the years, and really when we are
setting a threshold it is sustaining that level of safety, so you are
absolutely right. The problem then is saying are you continuing to
operate at that level of safety?
If there is something inherent in a design which would cause
someone to be outside or it's correspondingly made up for someplace else
because of a specific design difference -- maybe it's multiple diesels
or some other specific design difference, I think we would have to
commit to looking at that after the first year, to step back and say
what are the unintended results, because it is a distinct possibility.
On the other hand, it may be a non-problem because we picked
the baseline year, you might say, to -- I hate to say we have defined
safety, but from the oversight perspective we have defined at least a
threshold of safety as an indicator in '92 through '96.
It may be that those plants today -- it may be 100 percent
if we picked today, and so I don't know. I have a hard time -- I don't
want to assume five percent of the plants are outside or inside. We need
some experience and if that means it needs to be customized because of
plant-specific engineering differences, I think we have to be
open-minded enough after about the first year to say part of the
feedback is now let's start customizing these things plant to plant to
plant to plant, where it makes sense, and where there's an engineering
difference or some -- you need some concrete foundation for it, not just
that they modified three procedures.
I think we have to be open-minded enough to say that that is
a distinct possibility.
DR. APOSTOLAKIS: I realize that --
MR. GILLESPIE: It is the unintended effect I think you were
DR. APOSTOLAKIS: Exactly, and I wanted to -- I think it is
interesting though that, you know, we have to discuss this conceptual
problem of what is the intent of an inspection program, because another
question might be -- see, fundamentally the more I think about it, the
more I am leaning towards the point of view that this really should be
DR. KRESS: Should be a percent change of the plants --
DR. APOSTOLAKIS: You should have some sort of a
plant-specific threshold. You can't do that right now, of course.
MR. GILLESPIE: I think we are very open to that comment,
DR. APOSTOLAKIS: Because let me give you the other example.
I don't know if we all have this thick document --
MR. BARTON: It's the SECY paper.
DR. APOSTOLAKIS: Yes, the SECY paper, page -21.
MR. GILLESPIE: I am not sure if we brought it.
DR. APOSTOLAKIS: This is the BWR RHR system unavailability,
so the threshold is set at .015 and there are one, two, three, four,
five, six, seven, eight plants that are above it, okay? But there are
also several plants that have an unavailability for that system that is
significantly lower, so let's take Plant Number 33 -- 33, okay? That
has an unavailability which is what? -- .003? Something like that, 3
times 10 to the minus 3, right?
MR. GILLESPIE: Yes.
DR. APOSTOLAKIS: Three times 10 to the minus 3.
Your threshold is at 1.5 10 to the minus 2, so there is a
significant gap. Is the intent of the inspection program to make sure
that Plant Number 33 will not have an unavailability greater than three
10 to the minus 3 or to make sure that its unavailability stays below
In other words, are you giving now a license to that plant
to increase the unavailability by about an order of --
MR. GILLESPIE: No. No, no, no. George -- here is one of
DR. APOSTOLAKIS: So it is really plant-specific.
MR. GILLESPIE: And you have hit one of the limitations of
what we have done. We have picked a limited number of systems so the
idea of trading one system against the other in this indication
process -- it is not complete. It is not complete.
Now Steve Mayes's work in now Research I guess as of two
weeks ago, looking down longer term, would collect the reliability data
which would be more complete. I don't want to say it is perfect. I am
just kind of roughly familiar with where Steve is going with it, so in
the long-term the agency has a program in place which would get you that
sense of a profile across all the major safety systems versus just
picking four and saying okay, these four are kind of indicative of what
is happening to all of them, and it's a basic limitation.
It is not that we are giving those guys license to operate
DR. APOSTOLAKIS: All right.
MR. GILLESPIE: That is not the intent.
DR. APOSTOLAKIS: I know that is not the intent, but I
MR. GILLESPIE: So I think Steve is on a two -- maybe a two
MR. BARANOWSKY: I just left a meeting --
MR. GILLESPIE: Help me out. You are letting me drown.
MR. BARANOWSKY: Whether we should have plant-specific or
peer group specific thresholds for these things and we are looking into
those as to what is practical.
When you go to plant-specific it gets a little bit hard
because you don't have enough data to work with to know what the
practical bounds are on it so we have some statistical issues to deal
with. Believe me, we couldn't handle them in the few months that we did
this over here, but we know that we would likely want to see some
differences when you see Plant -- what is it? -- 78, for instance is
almost zero unavailability, so there's quite a big spread here in terms
of orders of magnitude.
I wouldn't disagree with that but from a practical point of
view, this is what we came up with for today.
DR. APOSTOLAKIS: I understand that and I am not
particularly criticizing this. I am just raising the point that
eventually they have to be plant-specific, the thresholds. Now when
eventually is I do not know.
MR. BARANOWSKY: Or peer group or something but we are going
to come back here with that program that Frank was talking about in the
not-too-distant future, a few months.
DR. APOSTOLAKIS: But this is not very novel, though. My
understanding is that in the maintenance rule it was each licensee that
set the thresholds, the criteria, so this is not unreasonable. I mean
you give them general guidance and then each licensee comes back and
says yes, for me the threshold for RHR unavailability is this.
I mean you don't have to do every little detail because that
is a burden, of course, but right now I agree. As long as we all
understand that this is the first step, that there are these limitations
and that the intent of an inspection program is not really to tell the
licensees what to do. It is just to confirm that what is out there,
what we thought was out there is indeed there.
MR. BONACA: Yes, and I would like to add one thing. Your
point is well-taken, George, from a perspective of really the foundation
to monitoring a plant is trending and not absolute comparison with
something that may be unrelated.
What I mean by that is that if you have a very low
unavailability and you trend higher and higher, even if you don't meet
the threshold it would have to be a concern that you have the trend.
MR. GILLESPIE: Yes, and it is important in this process.
This is the regulatory process. The intention was that these thresholds
for our intervention would be set at a point where the licensee in fact
has the opportunity to trend without our involvement, so in fact they
are kind of looser for the majority of the plants, because the intention
wasn't for us to get involved with the small incremental change in
unavailability or on the first scram.
The intent was that the licensee has to have enough freedom
to actually have something randomly occur -- we are looking for a
systematic flaw, something that is systemic that goes across, and we are
actually now saying we are recognizing random flaws exist. You should
be able to trend those random flaws and correct them before they get so
systematic to cross this boundary, and so the philosophy is there.
Something George said -- there's more than plant-specific
here. In the future just going plant-specific with these PIs is not
enough. It has to also be complete.
These are four safety systems, not all the safety systems,
and I am going to expect that when someone crosses a boundary the first
thing they are going to do is come in with a system that we are not
tracking and say but see how good this one is, so in risk space we are
really not there 'cause this is better than our prior assumption.
It is going to be interesting to see how we have to deal
with that. I expect that there will be some plants that will come in
with that argument.
DR. APOSTOLAKIS: Yes.
MR. GILLESPIE: So there is a sense of plant-specific but
there -- or peer group specific -- the completeness. We have to work
towards those two goals.
DR. KRESS: Excuse me. I would, before we get off this
point, I would like to express just a little bit of difference in
opinion with George's concept.
George's concept is a good one that the purpose of your
inspection is to go in and ensure that the licensee has maintained its
licensing basis in the sense of performance, and that it is a way to do
that. I think your inspection program has more than that as a
I think there are plants out there with varying levels of
safety, if you will, and that those that are at a level of safety that
is not very good, let's say, need more inspection, need more attention,
and the part of the inspection program is to perhaps identify those
kinds of plants, so there is an element of absolutism.
You were talking about relative change, trending, for
example, versus the absolute level. I think you need a little bit of
both in there and so I would say that to have absolute thresholds that
some plants are outside of is -- probably should be part of the system
also, as well as your concept.
DR. APOSTOLAKIS: But what to do about it is not part of
DR. KRESS: Oh, that's right. You're just identifying it.
DR. APOSTOLAKIS: Yes. Just identifying it.
DR. KRESS: You are not necessarily identifying when the
licensee is changing. You are identifying his relative status to the
other plants, and I think that is important too.
DR. APOSTOLAKIS: I think it is important, yes.
DR. KRESS: Okay.
DR. APOSTOLAKIS: But my point is that when you set the
thresholds, one way is to do what the Staff did, given the time
pressures they had. You look at the plant to plant variability and pick
the 95th, approximately the 95th percentile.
In an ideal world though, okay, I would try to have
thresholds that form a coherent whole, which comes back to what Frank
was saying, that maybe they have a high RHR system unavailability
because they have something else somewhere else that compensates for
that. Ultimately what matters is the accident sequences.
Ultimately it is the LERF, the CDF and those things, because
that is the real thing, so in an ideal world again you would have a
coherent set of these criteria and by looking at those sequences say
yes, those guys have a high rate of these initiators but look at what
else they have, right?
And then you want to make sure that that is preserved, that
is the inspection program's --
MR. BONACA: And what it is really pointing out is
preserving the function rather than the specific piece of equipment that
you are using.
DR. APOSTOLAKIS: That's right. That's right.
MR. BONACA: Because, especially the older plants,
typically, were not symmetric as much as the newer plants. They have
multiple systems to make up water, for example, for high pressure. And
so if you only look at one component, it may be under design by the
standards, but then for a PRA, you see that you have plenty of
over-compensation from other. So, to some degree, right now, going to
look at just equipment components' performance rather than bigger
functions like, I don't know, high pressure injection, okay, you may
tend to over-penalize some of the units, particularly the older ones.
DR. KRESS: Yes, but that is why they had this matrix.
MR. BONACA: In degree.
DR. KRESS: You know, it is two out of three, or three out
SPEAKER: Yeah, and that was to temper --
DR. KRESS: And that is an attempt to take care of that.
MR. JOHNSON: Yeah, let me -- in fact, if I can remind us,
we are talking about a system that tries to seize on performance
indicators, or inspections used like performance indicators, as a first
sign as to whether we ought to engage. The actions that we take are
very much going to need to be plant-specific, as is indicated by the
action matrix. You know, something that when a plant -- plant A crosses
a threshold, there may be factors, as you are going through this matrix,
and you decide that you are going to do some inspection to follow-up on
that, that lead you to believe that the licensee is on top of it, that
there are other things that are going on that make this threshold not be
maybe as applicable for that plant.
So, I mean there's flexibility within the matrix. The
actions are very much going to be based on what is going on at the
plant. And in any -- and in all cases, as Pat pointed out, the actions
are going to be tied to our ability as the regulator to take action
based on some regulatory requirement. It is very much our notion -- in
fact, when you look at the action matrix, the actions that we talk about
are regulatory actions. It is not -- we are not relying on some
pressure or influence on licensees to perform with respect to the
DR. APOSTOLAKIS: When -- I mean the pilots will last six
months, the inspection six months?
MR. JOHNSON: Yes. Yes.
DR. APOSTOLAKIS: And then what? At some point you have to
issue something, right?
MR. JOHNSON: Well, the approach is that we will have
success criteria for the pilot that we provide to the Commission, that
we will do the pilot plant -- the pilot plant activity for six months.
Near the end of that pilot plant activity, we will go back and look at
the results against those success criteria and, absent an indication
that we haven't been able to meet those success criteria, and we expect
to make some changes to our procedures and processes based on the pilot,
but, absent any indication that we have had a pilot, we plan to then
proceed with full implementation for all plants, and that should happen
in January 2000.
DR. APOSTOLAKIS: Now, two questions. Can you include in
the eight plants, I think, eight pilots, one or two which are in the 5
percent area? So that it will be interesting to see how we handle
those. And, second, can you -- do you have the time now to really think
seriously about the plant-specific nature of the thresholds and this
coherent system that I mentioned earlier, or is that a refinement that
has to come later, or it will have to wait until we see the lessons
learned from the pilots?
MR. GILLESPIE: No, I think in parallel.
DR. APOSTOLAKIS: Parallel.
MR. GILLESPIE: I mean the key is going to be -- and this is
kind of, I will say, a new union between the inspection program and
research. And so we are going to have long-term things. And Pat, you
are coming next month, you said?
MR. BARANOWSKY: Yes.
MR. GILLESPIE: So you are going to see us sending user
requests over to research to deal with the one year, two year timeframe,
where we need a stronger basis for doing things.
DR. APOSTOLAKIS: I see.
MR. GILLESPIE: And the focus in NRR will be kind of on
tactical, the tactical day-to-day, how do we keep it going and
incrementally improve it. So, yeah, we are going to be walking
hand-in-hand, so it will be in parallel.
They are not going to wait for us to finish the pilots to
continue doing what they are doing on more risk based indicators.
DR. APOSTOLAKIS: So how about the other question of whether
one or two of the eight plants will be --
MR. DAVIS: Dave?
DR. APOSTOLAKIS: Please come to the microphone and identify
MR. GAMBERONI: This is Dave Gamberoni of NRR. The pilot --
right now we are still finalizing the selections, but we have chosen
plants across the full range of performance. We have plants which we
believe should definitely exceed the thresholds.
DR. APOSTOLAKIS: Okay. So then it will be interesting to
MR. GILLESPIE: Yeah. I mean that was one of the things I
MR. BARTON: They are not all green plants.
MR. GILLESPIE: They are not all green plants. I mean the
ideal industry selection would be you get all green plants and you say,
DR. APOSTOLAKIS: Yeah.
MR. GILLESPIE: And that wouldn't test the system.
DR. APOSTOLAKIS: Thank you. I think we covered the third
bullet as well.
MR. GILLESPIE: Right. Okay. Good.
DR. APOSTOLAKIS: Then --
MR. BARTON: The third bullet is covered.
DR. APOSTOLAKIS: The fourth then. Yeah, this -- I don't
remember asking the question, but it is a good question. I don't mind
having my name. I don't think that as a committee, at least I haven't,
have spent -- have scrutinized those little fault tree type diagrams you
have in the report, and the logic there is not always transparent.
MR. BARANOWSKY: But it is scrutable.
DR. APOSTOLAKIS: Now, of course, when you want something,
you can't find it. Oh, here, maybe I am lucky. No, I am not. Do you
-- can you help me here, Pat?
MR. BARANOWSKY: Which one are you looking for/
DR. APOSTOLAKIS: Any one.
MR. BARTON: H --
DR. APOSTOLAKIS: Where you have -- no, no, it is not in --
MR. BARTON: He is not looking at the PIs.
MR. BARANOWSKY: Appendix II-II?
DR. APOSTOLAKIS: II-II, let's see --
MR. BARANOWSKY: Appendix 2, I guess that is appendix.
DR. APOSTOLAKIS: II-II -- two -- two-A.
MR. BARANOWSKY: It is Roman numeral two-dash-two.
MR. BARTON: What appendix, where are you?
MR. BARANOWSKY: It is one of these --
DR. APOSTOLAKIS: A, B, C.
MR. BARANOWSKY: One of these charts.
DR. APOSTOLAKIS: Yeah, one of those. Yeah. We have A, B,
C. How come you have Roman II?
MR. BARANOWSKY: I think that is appendix --
DR. APOSTOLAKIS: Oh, attachment.
MR. BARANOWSKY: Attachment 2.
DR. APOSTOLAKIS: Attachment 2.
MR. BARANOWSKY: Yeah. No, it is Attachment 3, Appendix 2.
DR. APOSTOLAKIS: Attachment 3, Appendix 2.
MR. BARANOWSKY: Page -- yeah, that has got to be it.
DR. APOSTOLAKIS: Roman II, you are right, now we found the
MR. BARANOWSKY: And page 2 or 1, or whatever one you want
to talk about. I think that is the chart you are looking for. It is
about two-thirds of the way back in the document.
DR. APOSTOLAKIS: Okay. Which page?
MR. BARANOWSKY: I am looking at Roman numeral II-II.
DR. APOSTOLAKIS: II-II, I think we found it.
MR. BARANOWSKY: See, that one is on mitigating systems. Is
that -- do you have that?
DR. APOSTOLAKIS: Yes, on mitigating systems. The one
before was on initiating events, right?
MR. BARANOWSKY: Yes.
DR. APOSTOLAKIS: See, I am sure there is logic here, but if
I look at it without talking to you, --
MR. BARANOWSKY: Right.
DR. APOSTOLAKIS: What do I see here, mitigating systems.
That's the top level, then there is design, protection against external
events, configuration control, equipment performance, procedure quality,
MR. BARANOWSKY: These attributes were primarily the ones
that were brain-stormed out in the workshop that we had in September of
'98, end of September 1998.
DR. APOSTOLAKIS: Now, for a mitigating system itself, you
will have the unavailability as the metric, right, as a performance
indicator? We just discussed the RHR unavailability, didn't we?
MR. BARANOWSKY: Right. What we said was for mitigating
systems, that unavailability -- or availability, reliability and
capability were the three performance attributes that we were concerned
DR. APOSTOLAKIS: Now, isn't capability part of reliability?
MR. BARANOWSKY: Well, we went through that argument, too.
And some people might like to say it is, some don't. It depends on how
you want to define it.
DR. APOSTOLAKIS: Well, it says reliability is the
probability of successful operation for a period of time.
MR. BARANOWSKY: I guess it is a matter of --
DR. APOSTOLAKIS: If you are incapable of operating, how can
you have successful operation?
MR. BARANOWSKY: I think that is from a specification point
of view. For instance, if someone didn't have the right specification,
and the system was tested out as showing that it could always operate,
and that it was available based on record-keeping, but that, in fact,
during an accident, its functional capability would not be adequate
because of specification, that's what the capability part meant.
MR. GILLESPIE: It gets at the heart of the design.
MR. BARANOWSKY: But I mean I do the same thing, when I am
doing an analysis of reliability, I say if the equipment is not capable,
it is not reliable.
DR. APOSTOLAKIS: Sure.
MR. BARANOWSKY: I can't count on it. Reliability is can I
count on it. I can't count on it.
MR. GILLESPIE: Yeah, one of the problems with we have here
is when you are collecting data on reliability or unavailability, it
doesn't get at the essence of what happens if it was misdesigned or
modified and, therefore, its capability to provide the necessary
function has been degraded unknowingly. It is very reliable but it only
provides half the flow needed to do what needs to be done. So --
DR. APOSTOLAKIS: I wouldn't call it reliable.
MR. GILLESPIE: Right. Right. But, see, the number -- but
the number gets reported, the statistic would still be there, yet, you
are still open to the design aspect.
DR. APOSTOLAKIS: Let me understand that. Which statistic?
I mean one statistic is the availability, it will start. That doesn't
tell you whether it supplies the actual flow.
MR. BARANOWSKY: Correct.
DR. APOSTOLAKIS: But then in the reliability calculation,
shouldn't you be looking at the actual flow?
MR. BARANOWSKY: Yes. I don't disagree with that at all.
In fact, we would -- we would take something that was designed
improperly and we would say it is not capable, and, therefore, it is not
reliable for that function. We chose to split the things up because we
were dealing with 90 percent of people who are inspection-oriented, not
reliability analysts, and that's the way they talk and think, and so
this jargon is based on that.
DR. KRESS: There is a problem with that, though, George.
DR. APOSTOLAKIS: There is no problem with it?
DR. KRESS: There is a problem. Reliability normally is a
spectrum of probabilities on whether it works or not. Capability is
almost a delta function. Is it or not capable of doing it? And you
have -- they are really two different animals, and it makes some sense
to treat them separately, I think.
DR. APOSTOLAKIS: But reliability implies capability. The
DR. KRESS: Yeah, but if it is incapable of doing it, then
your reliability is zero. That's what I am saying, it is a delta
function, and you need to treat it differently than normal reliability,
which is a spectrum. So it makes sense to treat them differently. I
don't know how you do it in a PRA, but it makes --
MR. BARANOWSKY: Actually, the way we are doing it here is
DR. APOSTOLAKIS: Let me grant you that point.
MR. BARANOWSKY: We were going to treat it as being
unavailable if it was incapable.
DR. APOSTOLAKIS: Okay.
DR. KRESS: Okay. That would be the way. Zero
DR. APOSTOLAKIS: My point is this, is the logic here the
way I think it is. You have the mitigating system at the top. You have
a PI, right. You have the unavailability metric, or the reliability
metric. Then to go to the next level, I thought the thinking was what
is it that this performance indicator does not cover? Therefore, I have
to worry about it in addition to the numerical value, right.
MR. BARANOWSKY: Yes.
DR. APOSTOLAKIS: So you think that all of these things,
design, protection against external events, configuration control,
equipment performance, procedure quality and human performance are not
covered by the unavailability?
MR. BARANOWSKY: No, that's not quite right. What we said
was, what are the attributes that the performance indicator or the
inspection program need to cover in order to give us assurance that we
really do know the reliability, availability and capability of the
MR. JOHNSON: Right. What is all --
MR. BARANOWSKY: We need to know these things. We couldn't
identify things beyond this that we thought we needed to know.
DR. APOSTOLAKIS: Okay. So, now, some of these are covered
by the PRA and some are not.
MR. BARANOWSKY: Right.
DR. APOSTOLAKIS: And this -- where am I going to see that,
not in the figure?
MR. BARANOWSKY: You will see it in the prior Appendix 2
where the discussion of each of the cornerstones is made. You will see
this is covered by inspection, this is covered by performance
DR. APOSTOLAKIS: But let's go on a little bit. It would
seem to me that the PRA would be a major guiding force here in
determining what needs to be looked at, right? Not just the judgment of
MR. BARANOWSKY: Right.
DR. APOSTOLAKIS: I mean the judgment of people is the
ultimate thing, but we have to structure that judgment. So, for
example, you say procedure quality and human performance. Why do you
single out the procedure quality? I mean human performance, it seems to
me, is what you are interested in. And I don't know that procedure
quality is the most important thing when it comes to human performance.
MR. BARANOWSKY: Okay. Again, as I said.
DR. APOSTOLAKIS: So there have been some judgments there,
some decisions already made based on perhaps the experience of people,
but not guided by the quantitative and structured approach of a PRA.
MR. BARANOWSKY: Okay. I can tell you this first line was
initially derived based on the experience of the group I described at
the performance assessment workshop. We then brought some different
folks together with a little bit more PRA background, and the results of
PRAs, like the IPEs, for instance, in which we had summaries of the
important contributors to different types of plant designs associated
with different mitigating systems. And that is where we brought in
additional understanding from that insight as to what aspects of these
things, design and procedures and so forth, were most important to pay
attention to, and see whether the performance indicators or the
inspection program covered it.
That's the -- basically, that was done for every one of
DR. APOSTOLAKIS: But do you at some point plant to go back
to the fault tree handbook or some PRA and look at how they derived the
unavailability and unreliability of the system and ask yourself, have
these six boxes really captured what the calculations do? And see --
MR. BARANOWSKY: Well, I think the --
DR. APOSTOLAKIS: I mean get some additional guidance
perhaps what dominates. I mean, as you know, when you calculate the
system unavailability, there's a bunch of terms there.
MR. BARANOWSKY: Right.
DR. APOSTOLAKIS: Unavailability because the system is down
due to tests, or this, and this and that.
MR. BARANOWSKY: That is why we have two risk information
matrices. The first one is the generic one, which was the primary tool
used to derive this structure you see here. Then there is the second
one which is plant-specific, which says in addition to this, what do I
know about this specific plant? And, therefore, we bring in the
plant-specific PRA. That is brought in in that factor.
The performance indicators, I don't believe need the
plant-specific PRA the way they are set up here. If we were to have
indicators that were more relevant to differences in design and
operation, then we would have to have some additional flexibility in the
structure of the indicator models. That's the kind of stuff we are
working on in the future, and that would allow us to refine our whole
inspection program a little bit better, but we are not there yet.
DR. APOSTOLAKIS: Well, ultimately, I would like to see an
argument that says, look, here is an expression for the system
unavailability and unreliability. Each term represents this, so if the
number is below the threshold, we are covered there. What is left is
something else that we believe, as experienced engineers, that is
important and it is not covered by this, we will add an extra box. So
that argument has not been made completely.
MR. BARANOWSKY: Well, actually, we did that kind of
thinking but we didn't jot it down for every single one of these things.
DR. APOSTOLAKIS: But that is something that eventually will
make me happy.
MR. GILLESPIE: I think, George --
DR. APOSTOLAKIS: And I know you will make me happy.
MR. BARANOWSKY: I am going to make you happy.
MR. GILLESPIE: One of the ways we are trying to get at a
piece of what you just said is this document we should have out in
another couple of weeks on looking at what you find. And what I mean by
that is the old term "regulatory significance." If you have -- six in
the past, if you had six to eight findings of procedural noncompliance
in some small period of time, several months, the licensee might get a
letter that says you have got a programmatic failure. This is
regulatory significance. It might not be safety significant.
Part of the screening process for putting together --
because this deals with that always unknown beta factor that kind of
goes in the front of all the terms. And one of the questions I asked
the staff to do, not as -- as part of the backup to attack this problem,
was, given the information that we would get from this program, how many
-- I don't know how they are going to do this -- but how many things
like Appendix B violations would have to occur before it becomes safety
significant? Is it six, three? Some of the things we have seen in the
past. Or is it 700? I mean it is a phenomenal number?
From some work I was involved in about five years ago, I
have the feeling it is a very high number.
MR. BARTON: I would agree.
MR. GILLESPIE: When I was with the Regulatory Review Group,
we did kind of some peer groups and saying -- how much could the QA
system degrade before it would actually show up in safety?
MR. BARTON: Well, there are so many elements in Appendix B
and sub-elements, et cetera, it would be a high number before you would
MR. GILLESPIE: Philosophically, when do you get to the
point where these would not show up in the indications, would not show
up in inspection, so they are very subtle, yet would occur to cause the
configuration of the facility so that it would occur in an overlapping
manner, so the configuration of the facility at some time is unsafe, or
past some threshold?
The staff is thinking about it. It is almost an
unanswerable question, but we needed to start thinking that way because
we are going to have to have a firm basis for saying -- there's, to pick
the easy one, there is a maintenance backlog of 1200 items at this
plant. You know, one might say is that safety significant? 1200 is a
big number. All plants count them differently. It may not be.
MR. BARTON: What are the man-hours, what are the systems?
What are --
MR. GILLESPIE: What are the systems? And do you -- so we
are at least thinking that way now. And we almost have to do this to
deal with the other question that you haven't asked about, that we do
have an IOU as the staff in March, that's enforcement. We have to mesh
enforcement into this whole scheme, because what comes out of
assessment, if you are going to take action, be it an order, a notice of
violation -- when do you turn it over to a licensee? When do you send
out a notice of violation? And that is the whole scale in Mike's
metrics. So we have to have a sound basis for why we are throwing out
also what we are throwing out, or you might say turning it over to the
I think it gets us to the essence of the trivial factors,
the factors that really might not matter, but yet they are requirements,
and you still have to comply. But how we dispose of them is our choice.
So what we are asking, the question, is an integral hole that starts to
get at those insignificant factors. And I have one view on how I think
it will come out, but I need to let the team mull on that one for a
while. And so they are trying to give that some thought this week as we
kind of approach. How do you enforce differently than you have enforced
before? This is a very significant change in philosophy on enforcement
for these kinds of items. So we are trying to get at that.
The last point I would like to make is plain English and
public readability. We have been working with Public Affairs and Bill
Beecher's staff, and they are about -- they have put a pamphlet
together, and we have gone through several drafts of it. And we are
going to be putting a pamphlet together that takes this 400 page paper
down into something that is more readable for the general public. And
Public Affairs was very concerned that the right message wasn't getting
out. We eliminated SALP as an institution. Gee, the NRC is backing
off. Yet, under this process, there will be more information available
in a more timely way than we have ever had before for everybody, if we
can put it in context.
So Public Affairs has kindly stepped up and teamed with us
to put a pamphlet together. I talked to Victor Driggs in Public Affairs
this morning. They are hoping to go to the printer late this week. So
that all of us who are going out talking to people will have 100
pamphlets in our briefcases that are written in plain English. So that
was one of the items that we discussed here, and it is the kind of thing
for reporters, League of Women Voters, public around site.
Actually, I hate to say it, but it is probably better
written than our paper, relative to understanding.
DR. POWERS: When you go on your evangelical missions --
MR. GILLESPIE: That's what we like.
DR. POWERS: I wondered if you speak to the downsides as
well as the upsides.
MR. GILLESPIE: Yeah. And, actually, we have been trying to
-- we are trying to be very balanced with that. We are -- and here's
one of the -- the big downside is we are going to be looking at less.
We may have the feeling we are looking more focused, and we are looking
at what is important, but we are clearly looking at less. And I think
we have -- that has a negative connotation to it.
DR. POWERS: Yes.
MR. GILLESPIE: But we may be looking at less, but because
we are getting PIs in and more information that directly relate to
operations, we will have -- I think we will have more knowledge.
DR. APOSTOLAKIS: What do you mean looking? What does the
MR. GILLESPIE: Well, I think the general public, and I try
every once in a while to put myself in the place of that person who kind
of just lives around the site and just reads the paper once in a while
and sees something, I think the general public generally sees that if
the inspectors aren't there as many hours, then they are not finding as
much information. This whole concept of performance indicators, I don't
think -- I think it is foreign to a member of the general public.
MR. BONACA: That is why, you know, I asked that question a
week ago, and I don't see -- I still believe that you are proposing a
new approach, but you are not summarizing anywhere what the differences
are and results insofar as the areas that we are not -- you will not
cover. And I may feel comfortable about your assertion that you looked
at it, and you didn't keep the record, but the fact is I would like to
know what you are not going to look at. And probably many other people
will ask the same question.
DR. APOSTOLAKIS: But I think we need to interpret and
explain the word "look." What does look mean? If you look at the
performance indicator, are you looking? Or looking means physical
inspection of an area?
MR. GILLESPIE: It depends on who you are. And this is the
quandary we are in. I believe that if you are a member of that general
public population --
DR. APOSTOLAKIS: They don't consider it.
MR. GILLESPIE: -- look is an inspector on the site looking
at the pump, looking at the worker doing the work, looking -- being
DR. APOSTOLAKIS: Right. And I would then make sure that
this pamphlet sent a message that the information that becomes available
to the agency from the inspection program is at least as good as it used
to be, and that the information does not come by looking at physical
MR. GILLESPIE: It probably -- it does not -- it tries to
say that. It doesn't go into a lot of detail, because it is up here a
little bit. But it is an important point, George, and that is --
DR. APOSTOLAKIS: Yeah, but don't just --
MR. GILLESPIE: No, I am not dismissing it, because this is
DR. POWERS: Evangelical.
MR. GILLESPIE: Yeah, this is our mission. And it is
DR. APOSTOLAKIS: Yeah, but I mean if you focus the
attention on the information that the agency gets rather than looking
physically at areas, you are going to go a long way towards explaining
MR. GILLESPIE: And that's the attempt. Yeah. And that is
the attempt, and that is our focus -- is what information do you need to
say the plant is safe? How we get it -- but this pamphlet doesn't go
that far. It is not that good yet. The next printing -- the next
printing -- we are evolving, we are not perfect here, we are evolving.
DR. POWERS: I was very enthusiastic and excited about your
discussion that you presented at the subcommittee meeting on how you
respond -- how you would approach this, given that the average member of
the public probably has a limited understanding of the concept of
cut-sets and other arcane vocabulary used by gentlemen of the
DR. APOSTOLAKIS: That's why it should be taught in high
DR. POWERS: And I felt that your thrust that you had in
those comments of pointing out how much more you get by this approach
than you got on the other approach was an awfully attractive beginning
of a dialogue in this area.
MR. GILLESPIE: And, in fact, the day after we had that
meeting here, Mike -- I think it was the day after, Mike and I, Al
Madison and Corny were in Bill Beecher's office with the entire OPA
staff, including a regional rep on the phone. We were running exactly
this concept, and it is funny, the guy who wrote this was the regional
rep in Region III. So Bill actually had someone write it who was
separated from us.
DR. POWERS: Okay.
MR. GILLESPIE: It is interesting the way he did it. And
knew some -- there were some things, that he read the paper, and a smart
guy, and was totally kind of off-track. And so we -- I think this -- it
is a first try at coming out on how to explain it. It was immediately
following that, so that went right from us to OPA and they, by design,
had someone do it was totally ignorant of what was going on, other than
hearing about it long distance.
So I think it is a real good product as a first go-around.
So we are trying to get smarter and evolve. And the technical aspects
are easy, it is all this other stuff that is much harder.
What message do you give to whom when. Mike is laughing.
DR. POWERS: Well, I mean the fact is that you are charged
with the responsibility of protecting the health and safety and they
would like to know how good you're doing it.
MR. GILLESPIE: And that's right -- and we're damn good.
DR. POWERS: So far.
MR. GILLESPIE: So far. But now we are approaching the
point as we get into the pilots where we are going from the paper to the
application in a very rapid way and that is why we are going to go out
and talk to the states.
I don't know that this agency has ever taken on a project of
this size that is industry-wide, basically from a concept in one January
to full implementation in the next January, and we are rapidly moving
from pieces of paper which are easy to throw around -- was it Rickover
once who said that all paper submarines are on schedule and on time?
MR. GILLESPIE: So I think we are now going to test the
budget, the schedule, the application, the design and the theory very
rapidly and, you know, June is only four months away. We are now in
February. We will deliver the piece of paper in March -- we promised
the Commission -- and it will be a damn good piece of paper. This is
what I told Sam Collins. He's nervous about March. I said I am not
nervous about March. I am nervous about June.
MR. BARTON: We are rapidly running out of time. I think
there may be some more members' questions.
There was an issue during subcommittee, a comment made
regarding executive override and the value in this and I don't know
whether the member who asked has gotten a satisfactory response.
George, I think it was you?
DR. APOSTOLAKIS: No, it wasn't.
MR. BARTON: Okay -- then I don't know who asked it.
Whoever asked it if you are satisfied with the response --
MR. BARTON: Transition team makeup -- the question is what
is the seniormost member in the region on that team? Is it going to be
a senior representative in each region?
MR. GILLESPIE: Let me hit that. The actual workers on the
team are going to be generally Branch Chief level people from the
regions. We have now put together, and Sam will be signing out in the
next day or so, calling it now an Executive Council, which will likely
be the Deputy Regional Administrators --
MR. BARTON: -- will be part of that team?
MR. GILLESPIE: No, what they are going to be is not a
steering committee -- a murder board. We expect that when deliverables
come out we will pull them together. We will not be part of that. It
will be basically a four-person panel that will appoint a chair and the
expectation is that that chair will sit at the Commission table with us
so that we can get, you might say, direct raw feedback into the system
from the regions as each major product comes out and at the end, so that
memo is being put together now.
It is going to be interesting to see who among the regional
deputy RAs gets appointed to be chair, but knowing in advance he is
going to be sitting at the Commission table with us we will probably his
focus significantly on the products and where they are coming from.
MR. BARTON: Good.
DR. MILLER: Is that one way you are going to get some
indication of regional variability?
MR. GILLESPIE: Yes. Yes, and difference in regional
One of the reasons we couldn't capture a deputy RA -- you
know the regions went through a lot of turmoil with the loss of managers
and stuff we recently had. Virtually 50 percent of everyone seems to be
new in whatever position they are in, and there are still a number of
holes, and to pull a deputy RA in full-time at this point for four
months would have been disruptive to keeping the program going, and we
still have to keep the program going, so this is the way we are going to
It is going to be interesting to see who fights to be
chairman of this group.
DR. SHACK: Do you want volunteers?
MR. GILLESPIE: Well, the idea is they have to pick among
themselves so however they want to do it. No, we are going to leave it
The intent is that we shouldn't influence it. It should be
an independent group because they are the implementers and we shouldn't
dominate those -- their critiques.
MR. MADISON: Alan Madison here. Just so we have the
terminology right, we already have an Executive Council. I was reminded
of that this morning, so we will probably retitle it as an Executive
MR. BARTON: Okay. Instead of Transition Team or --
MR. MADISON: We will still have a Transition Task Force but
this will be an Executive Forum that will provide as Frank discussed
some oversight for us.
MR. BARTON: Okay, Alan, thank you.
DR. SEALE: Could I ask, this eight pilot candidates that
you have, do you have any -- are a pair of those run by the same
operating company and in different regions?
MR. GILLESPIE: The intent is two per region because part of
this is --
DR. SEALE: Yes, I know.
MR. GILLESPIE: -- is a training exercise also.
MR. JOHNSON: Dave, would you come to the mike, or Alan?
MR. MADISON: That is not the case. We didn't look at that
as one of the criteria for selection.
DR. SEALE: It will still be interesting to see down the
road and actually I think you ought to solicit from operating companies
whether they perceive an even enforcement or handling of this from the
MR. GILLESPIE: Yes -- we have taken that comment back. The
other interesting piece is I hope we have an operating company that
has -- for six months it has a plant under the old system and the new
system and one of the feedbacks we might ask is how did you see the
difference knowing that there's some artificialities about a pilot, but
that will give us a comparison point.
MR. BARTON: The enforcement plan -- there's discussion,
enforcement plan, violations, severity levels, et cetera. I don't find
anything regarding actions that lead to CALs or more severe actions. I
think there was a question in SRM in November '98, and I was looking
through the enforcement thing and couldn't find it. Is it addressed at
How do you get from violations to severity levels down to
CALs or shutdown orders or whatever? I don't see that in here. Did I
MR. JOHNSON: Well, I don't know if you missed it.
The action matrix just talks about the fact that we would
take increased action including orders of 50.54(f)s and that kind of
stuff. Many of the actions are considered enforcement sanctions and so
the process we would use to take those actions, if we were going to
issue an order, it would be the process that we use today to issue an
It's just that this action matrix, because of a plant's
performance it would drive you into that consideration and then you
enter that process to get the order.
MR. BARTON: Okay. I had a general comment.
Last time we talked about the package was put together by
several groups, teams, et cetera. We talked about the acronyms, need
for a list of acronyms. Also, I don't know what your intention is to
proofread the document but you will find areas where it just kind of
disappears into Never-Never-Land. I don't know how many there is but
there's an example in Attachment 4 on page 5 is one example where the
thing just doesn't make sense because something is missing, so it looks
like it really hasn't had a good proofreading since it was put together.
Does it all hold together as a document? I think you need to look at
that. Have you got that, Mike? Attachment 4, top of page 5.
MR. JOHNSON: Yes, I have that.
MR. BARTON: It just --
MR. JOHNSON: Yes, thanks.
MR. BARTON: There is something missing there -- at least in
my SECY paper there is.
DR. POWERS: Well, if that one isn't there are several other
places that it looks like the cutting and pasting may have overlapped a
little bit or something.
MR. BARTON: And you need to look at that.
MR. JOHNSON: Right, and I think we tried to make the point
last time that we didn't, we didn't see ourselves refining this
document. This was the communication vehicle to get it to the
Commission and then to get it out, but we do see ourselves, as Frank
indicated, in going with -- to put out some plain language information
about the process but also to then take this document and capture it
into the implementing procedures and so those implementing procedures
will get the kind of going over and we will be able to make sure that we
have corrected the things here.
For example, the Attachment 4, I would see, speaking for the
Transition Task Force a little bit, but I would see Attachment 4 as very
easily going into what would be a management directive that replaces
today's SALP management directive, for example.
MR. BARTON: Do you have any insights or inputs into the
policy issues at this point? They are going to have to get addressed
sooner or later.
MR. GILLESPIE: We have had discussions, including all the
way up through talking to Bill Travers at the EDO level with Frank
Miraglia and Pat Norry on the organizational aspects that could be
affected and trade-offs between generalists and specialists in
N-Plus-One, so -- and actually Paul Bird was there, so we have got the
Human Resources people involved, so at this point it is -- they are
being looked at. Nothing has been resolved but there is a recognition
that in general they are kind of the right issues that we have to look
There's organizational impacts from this kind of change.
Now how we accommodate them and move forward has not been decided.
MR. JOHNSON: Incidentally, I'll add, you know there are, in
fact we didn't draw out all the policy issues and put them in the
Commission paper but there are many more policy issues and policy issues
with a small "p" that we will have to take on and we are trying to deal
with those every day.
For example, we don't talk a lot about the allegation
program or how that fits in the oversight process, but we certainly need
to come to grips with that, and we are working on it as we go through
this transition period.
MR. BARTON: Do any other members have any other questions
at this point?
I have run out of my notes and I think we have basically
covered most of the -- I think we have covered all of the issues that
came out of the subcommittee meeting and most of the issues that came
out of the comments in the Commission briefing.
I don't know what kind of letter we are going to put out
based on the last two meetings we have had with you, but we will get a
letter out from this meeting.
No other comments at this point?
MR. BARTON: Dana, I will turn it back to you.
DR. POWERS: Thank you. My perception is it will be a very
positive letter, by the way. The team has done an awfully good job.
MR. BARTON: I don't think that's the problem. It's going
to be the details to support the statement that says it is a positive
DR. POWERS: That will be an interesting discussion.
DR. SEALE: The nitty and gritty.
DR. POWERS: I want to recess this session, ask members to
collect whatever they want for lunch and come back so we can have a
little pre-discussion as needed on our meeting with the Commission which
will take place at one o'clock, so I think we should probably count on
making our migration over there about twenty 'til 1:00.
[Whereupon, at 11:58 a.m., the meeting was recessed, to
reconvene at 8:30 a.m., Thursday, February 4, 1999.]
Page Last Reviewed/Updated Tuesday, July 12, 2016