468th Meeting - December 2, 1999
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
***
MEETING: 468TH ADVISORY COMMITTEE ON
REACTOR SAFEGUARDS
U.S. NRC
Two White Flint North, Room T2-B3
11545 Rockville Pike
Rockville, MD
Tuesday, December 2, 1999
The committee met, pursuant to notice, at 8:30 a.m.
MEMBERS PRESENT:
DANA A. POWERS, Chairman, ACRS
GEORGE APOSTOLAKIS, Vice-Chairman, ACRS
THOMAS S. KRESS, ACRS Member
MARIO V. BONACA, ACRS Member
JOHN J. BARTON, ACRS Member
ROBERT E. UHRIG, ACRS Member
WILLIAM J. SHACK, ACRS Member
JOHN D. SIEBER, ACRS Member
ROBERT L. SEALE, ACRS Member
GRAHAM B. WALLIS, ACRS Member. P R O C E E D I N G S
[8:30 a.m.]
DR. POWERS: The meeting will now come to order. This is
the first day of the 468th meeting of the Advisory Committee on Reactor
Safeguards. During today's meeting, the committee will consider the
following: proposed final amendment to 10 CFR 50.55a regarding
elimination of the 120-month ISI and IST programs update requirements;
low power and shutdown operation risk insights report; license renewal
application for Calvert Cliffs Nuclear Power Plant and the associated
safety evaluation report.
We will also discuss future ACRS activities; reconciliation
of ACRS comments and recommendations; and hear a report of the Planning
and Procedure Subcommittee. We will spend substantial time working on
the proposed ACRS report.
The meeting is being conducted in accordance with the
provisions of the Federal Advisory Committee Act. Dr. John T. Larkins
is the designated Federal official for the initial portion of the
meeting. We have received no written statements or requests for time to
make oral statements from members of the public regarding today's
session. A transcript of portions of the meeting is being kept, 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.
I have a note of sad news for the members. Jay Carroll,
former member of the committee, is ill and apparently has been ill for a
couple of weeks; is in intensive care now, apparently with some form of
ammonia -- pneumonia.
Other members have comments that they'd like to make at the
opening of the meeting?
[No response.]
DR. POWERS: Seeing none, we will turn to the first item of
business. Dr. Shack, I believe that is our work on proposed final
amendments to 10 CFR 50.55a regarding the elimination of the 120-month
ISI and IST program updates.
DR. SHACK: Okay; as the members know, there is currently a
requirement in 10 CFR 50.55a for licensees to update their ISI and IST
programs every 10 years or 120 months to essentially the latest edition
of the code that's endorsed in 10 CFR 50.55a. There was a proposed rule
change last spring that would eliminate this update. At that time, we
wrote a letter saying that we thought the update should be retained.
The staff has then gone on to hold a workshop to get input from
stakeholders on whether the update requirements should be kept or not,
and they have prepared a draft Secy paper with some options concerning
the update.
One of the options, of course, is to eliminate the update,
and then, there are a number of possible choices for a baseline
inspection. With the elimination of the mandatory update, licensees
would have an opportunity to voluntarily update. The staff would retain
the option to mandate updates if they could make a case through the
50.109, a backfit rule criterion. One of the current elements is that
the updating is mandatory and does not have to pass a 50.109 criteria.
One of the other options prepared by the staff essentially
maintains the update. The third option is actually available to
utilities now, which is to propose an alternative that is acceptable.
What the staff is proposing there is to come up with a more specific set
of criteria to assure that the treatment of those proposed alternatives
is more uniform and more predictable.
I see Gil Millman here. We're waiting for Tom Scarborough
for the staff. We're a little bit early, ahead of schedule. We had a
subcommittee meeting yesterday. We had presentations by the staff, by
the ASME Code. The code essentially went through a fairly thorough
review of the changes over the past 10 years so one could get a flavor
of just how significant the changes in the requirements were over a
10-year period, and they classified the requirements in terms of their
impact on safety, on industry standards, and again, a fair number of the
changes are basically editorial or reporting formats.
DR. POWERS: Well, it seems that we had a difference of
opinion on that. I think your statement that a fair number of just
cosmetic changes or editorial -- everybody agreed to that, but it seemed
to be a difference of opinion that may be borne of different time
periods that they were looking on how many changes were important to
safety or safety-related. And some of that seemed to bear down on to
just defining what safety was.
DR. SHACK: And explaining it.
DR. POWERS: If we get some clarification of that over the
course of today's presentations, that would really be useful to me to
understand better. Because sooner or later, we have to put this in some
sort of a framework of whether it's in a defense-in-depth framework or
in a risk framework but somewhere in a framework to base an opinion on,
and we've taken a preliminary opinion on this, which we sent to the
commission, and we said not to jump on immediately getting rid of this
120-month, and that preceded this period of discussion that we've had on
it, and there are a lot of people with -- that are impacted by decisions
one way or the other, so we're going to have to have some fairly
substantial justification, I think, for whatever; if we retain -- adhere
to our past view or take a different approach, it's going to have to be
justified in some framework. And defense-in-depth and risk seem to be
the two easy frameworks to work with here.
DR. SHACK: Yes; you know, I think the comparisons that we
did see were, of course, on different time periods. One was a 1989-1992
change; the other was essentially a 10-year change, but I agree that the
question of what important to safety was was left undefined.
DR. POWERS: And the other issue that we need to understand
a little better, I think, is there are those who say, gee, the 1989
issue -- version should be the baseline version whatever you do, and
there are other people, including the staff, who say that the 1995
version should be the baseline version, and I guess I need some
framework to understand how you make decisions in those regards, because
it seems to me that you have two decisions that you potentially could
make: to retain the 120-month update or not, and if you -- whatever you
do, you have to have some code of record that everybody adheres to or
can adhere to, and this question of 1989 versus 1995 --
DR. SHACK: 1995 or 1998.
DR. POWERS: Or presumably 1998 as well; I mean, I don't
have a good framework for making that decision. It seemed to me that a
statement was made that people were contending that in the 1989, they
had reached a climactic point of where they had made all of the real
corrections to the code that had substance to them, but there were flaws
in it, and so, you could use that and just correct those flaws. Other
people say newer is better seemed to be the argument.
DR. SHACK: Right; and again, I think we heard from the
utility or licensee representative from Entergy who had long experience
on the code. He had been on the code himself 20 years or so; supporting
the dropping of the update. There was some discussion as to whether
this would decrease participation in the code, and again, there are
essentially various opinions of whether, in fact, one is reducing the
burden by keeping -- eliminating the update or reducing the burden by
continuing the update.
DR. POWERS: The arguments on participation in the ASME
effort -- you're right; we heard two different stories there, equally
fervently preached, but I think it's neither here nor there for us; that
that's another issue outside of our domain. So I'm not sure -- I mean,
it's interesting what people's opinions are, but I don't know that we
can help one way or another on that. It is interesting that the -- we
get these diametrically opposed opinions, neither one of which is
defended with any definitive data for it. It's all perception.
DR. SHACK: I see Mr. Scarborough has arrived and Mr.
Wessman.
Mr. Wessman, would you like to start us off?
MR. WESSMAN: Yes, sir; I apologize. We must have
misunderstood the schedule, because we thought we were starting at 9:45.
DR. POWERS: 8:45.
MR. WESSMAN: Yes; and so, we were perhaps a few minutes
late in getting here.
Yes; my name is Dick Wessman. I'm deputy director of the
Division of Engineering. With me, on my right, is Gene Embro, chief of
the mechanical branch and at the table in front of you is Tom
Scarborough and David Terao. Tom will do most of the presentation, and
we will do it slightly briefer than we did yesterday because of the
press of time.
Before he starts, I will repeat a couple of my comments from
yesterday for the benefit of the members that were not with us. As you
know, the regulations require an update every 10 years, every 120
months, to licensees' ISI and IST programs to incorporate the latest
version of the ASME code that's endorsed by the regulations. As you may
also recall, recently, in September, we, after a long period and a lot
of work, updated Part 55a of the regulations to endorse the 1995 version
of the code with 1996 addenda with certain limitations, and this brings
the code requirements from the 1989 code up to the 1995 version.
We met with the ACRS in April and discussed the concept of
replacing the required 120-month update with a baseline and then
allowing licensees to voluntarily update as time moved forward, and at
that time, we had proposed the 1989 version of the code as the baseline,
because that was what was on the books at the time, and many of the
licensees were conforming with that version of the code, and at that
time also, there had been some healthy public comment and healthy
comment and discussion with the staff, and the ACRS suggested that --
and wrote to us, in fact, disagreed with the concept of the 1989
baseline and recognized that hey, we have much later versions out there,
but the ACRS, as they wrote us, also indicated that the concept of the
later baseline, with a voluntary update, seemed like a reasonable
approach if updates took the code as a whole.
There have been complex views; there have been public
discussion on this, and we had a public workshop on it, and on balance,
we think the proposal that we will discuss with you today represents the
best balance of a recommendation to eliminate some of the burden on the
licensees and utilities of the periodic update and puts in place as a
baseline the 1995-1996 version of the code and allows future voluntary
updates.
The presentation that we have does represent a consensus of
the work by NRR and research as well as our regional offices, and so,
we're glad to be back with the ACRS to finish up this discussion. I
guess I would point out also that most of the players and participants
that you see on the staff that have worked on this were all members of
the ASME, and we've been involved in the ASME process, some of us for a
number of years, and so, in that sense, we've worked as a liaison and a
representative of the agency on code committee activities and I think
have a solid appreciation for the ASME process.
With that, let me turn it over to Tom Scarborough, who will
take us through the presentation.
DR. SHACK: Dick, the question came up yesterday about how
licensees choosing the risk-informed inspection programs would be
affected by the updating requirements. Is Tom going to address that or
--
MR. WESSMAN: Yes, sir, he is, and he'll pick it up in the
middle of his discussion, but we did research that, and we'll make it
slightly clearer to you, sir.
MR. SCARBOROUGH: Good morning; my name is Tom Scarborough,
and we're going to finish up our discussion of the 120-month update
issue. I've revamped the slides somewhat to try to focus on some of the
more topical issues that we might want to talk about this morning, and
I've included some backup slides, which I probably won't go over
directly, but they're there for your information.
Just to give you a brief overview that Dick has mentioned in
terms of where we came from and what we've done so far on this issue, in
April we met with you to talk about the concept of the ISI IST update
requirements. In April, later in April, we issued a proposed rule,
which would have replaced the 120-month update requirement with a
voluntary updating provision with a proposed baseline of 1989 edition.
In May of this year, we had a public workshop, and we had about 60
participants from various areas of nuclear power and various
stakeholders.
In June, the commission asked us to redirect the effort
somewhat in terms of separating the then-ongoing work to endorse the
1995 code and to deal with the 120-month update issue separately, and we
did that. Our public comment period ended in June also, and we received
about 40 comment letters. We did finish the 1995 code endorsement in
September, and that's on the books now. That takes effect November 22.
And where we are as of today is we're preparing a commission
paper to deliver to the commission in January which summarizes the
public comments; provides options and recommendations regarding the
update requirement.
Okay; in terms of the public comment areas that were listed
in the Federal Register notice of the proposed rule, these are the
typical areas that we ask for comment regarding: potential effect on
safety; the selection of the proper baseline; the benefits and
hardships; effects on licensees submittals; the range; the potential
effect on risk informed initiatives, and maybe I'll jump right to that
question, since it came up. In terms of the risk-informed initiatives,
the alternative that is approved in the safety evaluation regarding the
risk-informed IST and ISI programs is for the entire life of the plant.
You know, they don't have to resubmit and re-request an alternative as a
risk-informed program. But they do have to review at the latest code
that's been endorsed in the regulations at the end of their 120-month
interval and determine if there's any significant changes regarding
anything they've committed to specifically; for example, if they
committed to a test according to the code in a certain area, and that
test has changed; and they're not able to do that, or they want to
propose some alternative to that, then, they would send in a relief
request for that, but if there wasn't any significant change in that
code area, they would just review, update their program to the latest
version of the code and keep going.
They would not have to resubmit any request for the
risk-informed program. That's approved for life, but they just have to
monitor the code to make sure there's no significant changes that might
affect how they do their program in terms of a method that they might
have committed to.
In terms that we also asked for comments about the effects
on states and the application also of portions of ASME Code in terms of
what we call cherry-picking in terms of should licensees apply entire
editions of the code when they voluntarily update, or would they be
allowed to select portions of that code. And then, we received a number
of miscellaneous comments.
In the back of your package, you'll find a more itemized
list of the individuals and areas, different organizations we received
comments from and also examples of the public comments themselves that
we talked in detail about yesterday, but suffice it to say that there
was significant comment on both sides of the question and some very
detailed comments in terms of their proposals that we had.
So after we reviewed those public comments, we derived three
basic options with some permutations among the options. The first
option would be to replace the 120-month update requirement with a
baseline and allow voluntary updating to a later NRC-approved code
edition. It would have to be endorsed in the regulations unless a
baseline was revised, and that would be applying the 10 CFR 50.109
criteria, and in terms of the initial baseline, we had three sub-options
in terms of which baseline to select. Option 1a would be to apply
basically the baseline that was proposed in the proposed rule, with the
1989 edition for ISI and IST; the 1992 edition for metal and concrete
containments; and the 1995 edition for Appendix 8 on qualification of
ultrasonic personnel is basically the proposed rule.
Option 1b would be to apply the 1995 edition with the 1996
addenda as incorporated by reference and currently in the regulations.
That was completed in September. Option 1c would be to apply a later
version, and typically, we talk about the 1998 edition, because that's
the one that we currently have under review at this time.
Option 2 would be to retain the current 120-month update;
basically, the current approach, the approach that we're following now.
The third option would be a somewhat permutation of the option 2. It
would be retain the regulatory requirement for 120-month updating, but
we would develop guidance for plant-specific alternatives, and we would
make that -- that's a much more structured approach for that.
So those are the options that we came up with, and then, we
went through a process of looking at the advantages and disadvantages of
the options, but also later, we had more detail on how they matched up
to the strategic goals of the commission as well. But under option 1 --
this ought to give you an idea of what that would entail, and I wanted
to make sure I made this clear, maybe more clear than I did yesterday,
is that we would continue to review future code editions under option 1.
That would not change. And we would incorporate those new editions into
the regulations for voluntary use.
So that would continue on as normal, as we are doing now.
We would also evaluate code improvements for backfit implementation, and
that would involve the aspect of have cumulative changes occurred in the
code over time such that it would be appropriate to revise the baseline,
so that licensees would implement a new baseline, that they would have a
new baseline to come up to, or we would look at, as we do now, for
specific provisions of the code and determine whether or not specific
backfits are appropriate as we do now and as we did with Appendix 8.
Were there certain aspects of the code that should be backfit on a
specific provision basis?
DR. SHACK: What backfit criteria do you think would be
invoked to do this? I mean, Appendix 8 was done under a compliance
backfit. What would you envision would be the route for future changes?
MR. SCARBOROUGH: I think there could be for all three
possible backfit paths could be used. For example, you could, if you
had a specific provision that would, like Appendix 8, that was felt to
be a compliance issue, that could be backfit through that approach, and
you could have a backfit added on to the sort of baseline, specific
provision like Appendix 8 might be, or you could process it through a
cost-benefit analysis and determine that the improvements of the code
over time, cumulatively, both qualitative and quantitatively, built up
to such a point that it would be appropriate to re-baseline everyone,
and the schedule for that might be a less aggressive schedule than if it
was a compliance or something where we might feel that it was
appropriate to do this on a more rapid basis, but you could do this
through that approach, too.
And then, adequate safety, of course, would be something
that would be considered to be very important and should be done very
quickly. So I think we're going to continue to monitor the changes;
we're going to do it much more structured, in a structured fashion than
we have in the past, to see what the changes are and evaluate those
changes in terms of both looking at sort of the risk changes and also
more on a qualitative basis: is there an improvement in the code
techniques and methods that as they accumulate, it would be appropriate
to rebaseline everyone.
So I see us looking at it from all of the possible backfit
criteria to see which one best fits or if there is a need to rebaseline.
So, of course, the licensees could voluntarily implement new
code editions as they come out, as they're incorporated by reference in
the regulations, and under option one, they would be allowed to
voluntarily implement an entire edition or addenda without NRC approval
or without prior NRC approval. Now, we would require that they have
prior approval for using portions, and that's intended to make sure that
there isn't a cherry picking process and to make sure that the
interrelated requirements are imposed. Now, we might go through and
pick certain areas which would be appropriate to implement on a specific
basis like ASME Code Case O&M-1 or Appendix 2 on check valve condition
monitoring, so we might endorse certain aspects of the code that would
be appropriate for individual implementation without the entire code
edition, but for the most part, if they wanted to use portions, they
would have to ask for that through a request.
And also, we would continue to participate in the code
process. That would not change in that respect, and once a licensee did
select a new code edition or addenda, that would be their code of record
that they would be responsible for meeting. So that's a summary of
option 1, but I wanted to make sure to indicate that we were intending
to continue our endorsement process for the ASME code.
MR. WESSMAN: And I would point out that we expect to be
more timely in our endorsement process. This has come up a couple of
times, and, of course, the industry has chastised us for the amount of
time that it took for us to finally move from 1989 to 1995. We know we
didn't do a good job, and our intent is to be more timely and do better
as we look at future versions of the code.
DR. POWERS: And my understanding is that we have a future
version, the 1998, isn't it?
MR. WESSMAN: The 1998 version, and we are just started on
working on that right now, and Tom and Dave have the responsibility for
that work in the mechanical branch.
DR. POWERS: So it is clearly going to be something like the
year 2000 or 2001 before that is endorsed by the regulations?
MR. WESSMAN: That would be the timing, and because of where
we are, it's likely to be the midpoint of 2001 before that is in final
form because of the process of CRGR and the ACRS and public comment and
this sort of activity and then go through the cycle again for a final
rulemaking.
The estimate is because of the various wickets that one must
go through in the process that it can take from a nominal 14 to 18
months for all of these things to happen from the time that we start
looking at it and we go out for public comment on the proposed rule and
then finish up the final work on the final rule. It is, unfortunately,
a lengthy and somewhat tedious process.
DR. POWERS: Well, what I am thinking is that it looks to me
like 1998, whereas it might be better than 10 years, still reflects some
substantial delay, and I'm wondering is -- it looks like you're behind
to where you can't catch up, because you've got a finite amount of
staff, and you've got these things coming off the presses every 3 years
whether you're ready or not. Have you given thought to skipping one
cycle to catch up?
MR. WESSMAN: Well, no, we really haven't, and I think our
intent is to try to get on with 1998, not wait until 2001 and skip that
aspect. And some of the traps that we're in are external to the agency.
The Administrative Procedures Act requires a certain amount of these
things that we do on this that just plain takes this amount of time.
You know, it's unfortunate and not wholly satisfying, and we've looked
at are there things that we can do with the OGC, and we're still
exploring that as far as endorsement of the code cases, but that
deliberative process is tedious, but certainly, part of the thing that
contributed to the situation that finally ended in September with the
endorsement of the 1995 code is that time went on, and then, we decided
oh, well, gee, yes, the next version of the code is almost on us, or
there was a debate about some portions of the content; let's wait and
see what the code then looks like and work on the next one.
We can't do that. We have to be more ambitious on that.
But the vision that we have is, of course, that if we go forward with
the proposal that we are discussing with you today, the 1995 baseline
will be in place, and along will come the staff endorsement with
limitations, if that's appropriate, the 1998 version of the code, for
voluntary implementation by licensees, so they choose.
MR. SCARBOROUGH: Okay; regarding the advantages of option
1, we may have a reduction in burden for the licensees through option 1
by the replacement of the mandatory updating process, but there were
public comments regarding possible increase in relief requests, to use
portions of future codes or code cases and such, so that might minimize
the savings that they see.
Significant safety improvements would be imposed through the
50.109 process. And that would be consistent with other new regulatory
requirements. So we consider that to be an advantage.
There would be a provision for -- provisions for
rebaselining and specific backfits would continue to emphasize the
importance of the code, and where there was a rebaselining, and there
was an update review, that would help identify program weaknesses, which
has been pointed out as an advantage of the update process.
Option 1a with the 1989 code might allow more burden
reduction for licensees, because most plants are currently using the
1989 code. 1b would apply to most recent endorsed code improvements in
the regulations, and 1c would apply something even more recent that we
haven't completed our review yet, such as 1998 or something beyond that.
DR. SHACK: Now, a licensee coming up in 2002, when its
10-year update was scheduled, could choose the 1998 version.
MR. SCARBOROUGH: Yes, and if this was in place at the time,
if they chose the entire 1998, they could do it, you know, without our
approval, without our prior approval, you know, as through the process.
So, they could voluntarily pick that, and that would be their code of
record. So they would be allowed to do that.
DR. SEALE: But if they wanted to cherry-pick, they would
have to wait until you had evaluated and taken your position on the
latest version.
MR. SCARBOROUGH: Well, if they cherry-picked, they would
have to come in for prior approval. If they select the entire code
edition that's already incorporated by reference in the regulations,
then, they can do it without our prior approval. But if they start to
cherry-pick, then, they would have to come in for prior approval.
DR. SHACK: I guess what I was asking is if my 10-year
required update was in 2004, if you picked 1b, I wouldn't have to update
to 1995 and then come in and voluntarily adopt 1998; I would just update
to 1998 and be done with it.
MR. SCARBOROUGH: You could, or you could do 1995, of
course. You could just --
DR. SHACK: Okay.
MR. SCARBOROUGH: -- pick 1995 and do 1995, or if you wanted
to, you could jump right to 1998.
DR. SHACK: It would be my choice.
MR. SCARBOROUGH: Exactly; it could be their choice, and I
think there are a couple of licensees who indicated that they liked the
1998, and so, I would envision that they probably would be pointing
toward that direction when they do their updating. They could do that
once we finish our endorsement of it.
In terms of the disadvantages of option 1, it does remove
the historical exclusion of 50.109 -- of the updating process from
50.109, and that's a disadvantage in the sense that historically, we
have not required them to go through that process of looking at the
criteria of 109 and such, and that was something that would be removed.
In terms of additional resources, we would need -- we would need
something on the order that would bring us up to about one FTE just for
sort of maintaining the code editions because of the need to go through
now and look at the changes to see if the changes were significant --
sufficiently significant that you revise the baseline. So that would be
a new burden for the staff to have.
DR. SEALE: Excuse me.
MR. SCARBOROUGH: Sure.
DR. SEALE: That's the timely evaluation, because in the
past, you've evaluated them; it's just that you've been perhaps not as
prompt as you would like to be.
MR. SCARBOROUGH: Right; and also, now, with option one, we
would be looking at the changes a little more specifically to see if
this change reflects a need to rebaseline, something we really haven't
done in the past. So that's a little bit more additional work we would
have to do.
In terms of the ongoing staff activities, you know, we have
a lot going on with Part 50 entirely in terms of risk-informing it.
There's risk-informed ISI IST programs and such, and so, there's a lot
of work going on with Part 50 itself, so the overall approach in a few
years might change, so that's something that might affect the final
conclusion once we get down the road a few years.
There might be an assumption of a reduction in importance of
the code, but we intend to try to mitigate that by indicating that we
will continue to more promptly endorse the latest code versions.
There would be additional licensee burden to do at least one
additional updating to come up to the 1995 or 1998 edition, so they
would not receive -- that would be some additional burden there. We
wouldn't have the sort of typical 120-month update program review that
has been helpful in identifying weaknesses in the programs, and in terms
of potential inconsistencies, the State of Illinois was concerned that
there might be a situation where a state might endorse a more recent
version of the code rather than what the staff was using for a baseline,
and so, there was that potential there. We did receive some feedback
from OGC that that would basically be resolved by Federal preemption in
that respect.
But we do intend to continue the review, and where we see
significant changes, we plan to rebaseline, so hopefully, that would
mitigate that effect as well.
One other aspect that was mentioned by the State of Illinois
was that the multiple code editions, such as we laid out in the option
1a with 1989, 1992, 1995 code editions all sort of mixed into one
baseline, they indicated that it could be confusing to their inspectors,
and they were concerned about that. We didn't receive specific comments
on a proposal for a 1998 or some other baseline edition, but we did
receive quite a few comments on the baseline itself, so we felt we
received sufficient information in there to try to decide what the right
baseline would be or the best baseline would be.
Just thinking of the ISI portion of this, again, we've been
through the risk-informed ISI and the GSI-190 resolution, where it's
very difficult to justify ISI, at least for piping systems on a sort of
core damage frequency basis. It doesn't seem to have much impact. But
really, the justification for ISI is basically a defense-in-depth
argument.
Have you done defense-in-depth arguments through the 50.109
process? Does it really work with defense-in-depth arguments?
MR. SCARBOROUGH: I think Appendix 8, you could say that
what that was, sort of that process, because there wasn't -- you
couldn't quantify the concern in that respect in terms of the ultrasonic
examination personnel being unable to fully recognize flaws in the
piping, so that was a case where we didn't have -- we didn't have the
ability to do a PRA-type of analysis so this would be the increase in
the risk, in fact, but it was more on a qualitative basis in terms of
analyzing the fact that there were significant inaccuracies in their
qualification that led the staff to be able to make a successful
argument to the CRGR that there was a need to have this as a backfit.
And so, I would imagine there would be similar arguments if
there are significant changes to the code down the road where we would
make that type of argument. We would try to use risk insights as best
we could, but recognizing that it may be difficult to do that for an
entire code but try to pick out maybe portions of the code, revisions of
the code, to show that yes, this is an improvement in safety on a
risk-insight basis.
But I would imagine it would be more qualitative in nature,
and we have checked with CRGR in terms of their charter, and it does
have the provisions for that sort of review of a qualitative and
quantitative argument with them, and in our meetings with them in the
past, they have recognized that and have been receptive to that, so we
think there is a success path there. We recognize it would be difficult
if we just -- our only tool was a risk CDF increase of some type, so
because of that, we've discussed that with CRGR, and they recognize that
we could make a more qualitative argument, and that's probably where the
more greatest thrust of our argument will be if we can get to a point
where we're going to rebaseline.
DR. SEALE: Yes, but, Tom, you are talking a very special
situation in that case, because you have a lot of performance-based
assessments of the capability of these inspectors; that is, the people
who interpret the data from the tests to judge their capability. You
may not know what the risk is, but you know what -- you have a pretty
good idea of the failure rates and so forth, and it seems to me that's a
very different kettle of fish than coming in with something where you
really don't have the performance assessment either; you don't have the
risk assessment, but you don't have performance results either, so here
you are, you're truly in never-never land now and what-if space.
MR. SCARBOROUGH: Yes; I agree. I think what we tried to do
was go back to what was the basis for the change? Why was the change
proposed? Was it a concern that was raised regarding weakness in the
test methodology, and that led to a change in the methodology? For
example, motor-operated valves and stroke time tests; it's recognized
that that's not a fully informative type test, and there are efforts
from ASME O&M 1 Code Case to improve that to make it more diagnostic.
So what we might have to do is go back to what was the test
method before and what did we change and why did we do it and make the
argument that there was a concern with that method and make it more
along that line. It wouldn't have the advantage of having, you know, so
many failures of the personnel identifying the flaws, but we might be
able to make an argument that the test method was inadequate, and I
think we can do it pretty well with the stroke time testing for MOVs and
to go to something that's more definitive in terms of the type of the
method; but yes, I agree, it's going to be -- it's not going to be an
easy task to perform.
MR. TERAO: If I may add, Tom, excuse me, I think what we
are trying to say here and maybe more concisely is that when we evaluate
a code and look at the improvements to safety, it is not going to be
solely on a risk quantitative perspective. We will look at
defense-in-depth and other qualitative factors as far as improvements to
safety. It's not solely on CDF.
DR. APOSTOLAKIS: I hear the words defense-in-depth a lot.
Somebody explain that to me, please, how it enters here? What is
defense-in-depth in this case?
MR. TERAO: Defense-in-depth in this case would be
maintaining the principal radiation boundaries if the core itself, of
course, if the pressure integrity of the vessel and the piping and the
containment itself. So when we talk about in-service inspection, what
we are basically looking at is the pressure integrity of the vessel and
the piping and the containment pressure integrity.
DR. APOSTOLAKIS: Well, the existence of these barriers is
defense-in-depth.
MR. TERAO: Yes.
DR. APOSTOLAKIS: Now, doing other things, is that part of
defense-in-depth? I mean, I don't understand what defense-in-depth
means in this case. Are you trying to -- I mean, the risk space, I
would say, you know, I have already very low probabilities of failure,
and I want to make sure that they stay low. I don't know what
defense-in-depth is.
DR. POWERS: Well, I think I would agree that the process
and the procedures one goes through to ensure integrity of the primary
piping system and the reactor pressure boundaries is an element of the
defense-in-depth, and it comes in through the independence and greater
degree of conservatism that you have in defining defense-in-depth. So I
think I would -- I would concede that these particular requirements that
we're dealing with here are part of the defense-in-depth concept.
DR. KRESS: I would view it in the form of a principle in
the sense that you put on defense-in-depth because of uncertainties in
your risk evaluation. And what you are trying to do is put enough
defense-in-depth to make that uncertainty acceptable.
DR. APOSTOLAKIS: But where does this uncertainty come from?
DR. KRESS: It comes from not knowing -- well, there are two
sources of uncertainty. One of them is in the numbers themselves, the
risk numbers, but there is also uncertainty in how well your
defense-in-depth provisions operate, how well they're defense-in-depth.
So if you want -- if they're there to reduce uncertainty, then, one way
to be sure that uncertainty is acceptable is to look at your
defense-in-depth provisions and reduce the uncertainties in those and
their capabilities, and you do that -- one way to reduce that
uncertainty there is to inspect and test.
DR. APOSTOLAKIS: So, but the root cause of the uncertainty
is what? That there may be another mechanism that I'm not aware of?
DR. KRESS: Yes, yes, absolutely, yes.
DR. POWERS: And very definitely; nor, do you know what the
magnitude of the insult may be.
DR. KRESS: That is right; you don't have good numbers for
the insults and the forces and the vibrations.
DR. BONACA: And quantitatively, I think it comes in in the
assumed frequencies, for example, of small breaks and large breaks, et
cetera. There are assumptions that you put, in particular, in the PRA
model that may be challenged insofar as if you have aging taking place,
and you'll inspect to determine when or how you have -- all the
uncertainty, all those values are significant.
DR. APOSTOLAKIS: So this is really the rationalist view.
It's not the defense-in-depth view line.
DR. POWERS: It is the rationalist view. It is part of
defense-in-depth as seen by the rationalists.
DR. APOSTOLAKIS: Ha!
DR. KRESS: It is also a structuralist view.
DR. BONACA: Yes, I think so, too.
DR. KRESS: It's a little bit of both.
DR. BONACA: I mean, we have evidence of, in fact,
irradiation happening over a 10-year period, so the quality of the
inspectors to determine, to detect the degradations is --
DR. APOSTOLAKIS: I would find an argument that says, you
know, there are uncertainties; for example, we have been surprised in
the past and so on, and we want to make sure that this is not going to
happen again; much more convincing than saying in the name of
defense-in-depth, do it. And I think words and statements at this stage
of the game when we are shifting to a new regulatory system are very
important. I mean, I am convinced, you know, if you give me an argument
in terms of uncertainty and, you know, it makes sense to me.
MR. TERAO: I would just like to add one more thought is
that when you get down to the basics, what is the purpose for the ASME
boiler and pressure vessel code? It is a code to ensure the pressure
integrity of the vessels, meaning the piping and the containment.
That's what it is. It ensures the pressure integrity of the components.
So when we evaluate the acceptability of the boiler and pressure vessel
code, what we are looking at is the integrity of the piping and the
containment from the defense-in-depth standpoint.
MR. SCARBOROUGH: Regarding option 2, basically to maintain
the current approach of the 120-month update, some of the advantages
would be that licensee event reports have revealed that numerous program
deficiencies have been found during the update reviews, and we talked a
little bit yesterday about the significance of those, and some of those
did result in requirements for a notice of enforcement discretion, so
some of them were rather significant in that respect, and we also heard
from an individual that many times, by finding those early, they are
able to conduct the examinations before the 120-month update time ran
out, and so, they were able to mitigate the consequences of missing that
examination that may have been missed.
As part of that, the programs do diverge over time, and
that's part of the advantage of the program update. It does -- the
updates do help safety by incorporating new experience and techniques.
It would retain public confidence, you know, in terms of the inclusion
of safety-significant information in future code revisions, and it
responds efficiently through the ASME code process to address emerging
issues.
A disadvantage is that by retaining the 120-month update
requirement is that it doesn't reflect a current effort to improve our
justification for imposing new requirements in terms of such as applied
through the 50.109 process.
Regarding option 3, there, we would retain the regulatory
requirement for the 120-month update, but we would develop guidance on a
more plant-specific basis for alternatives to the 120-month update
requirement through 50.55(a)(3)(i), which is acceptable level of quality
and safety, and we would prepare guidance for making those decisions,
such as including considerations of operating life and safety
significance of the changes since the previous update.
The advantages of option 3 are that it uses an existing
regulatory process, so we retain the 120-month update in the
regulations, so we minimize the potential for reduced code
participation, and we would reduce the potential inconsistency concern
regarding state and Federal requirements to a more plant-specific issue
if a particular plant was allowed to, as an alternative not to update;
then, that would be more of a plant-specific issue.
Regarding disadvantages, the licensees would have to justify
this alternative, and that could be rather extensive in terms of going
through the entire code to look at all of the sections. Further, this
process might not include public participation, and that could be a
concern to some, and we would need resources to develop this guidance
for the acceptance of the other alternatives. So that's how we came
down with the advantages and disadvantages of the --
DR. BONACA: Just a question.
MR. SCARBOROUGH: Yes, sir.
DR. BONACA: There would be an option 4 which you haven't
looked at, which is why don't you do the update less frequently? For
example, since the rate of changes is decreased, and that's the basis
for the request not to update, why not do it every 18 months -- 180
months, 15 years? I mean, that will relieve the burden but also will
keep some order that is appealing in the 120-month update? I just would
like to know if you thought about that option.
MR. SCARBOROUGH: There was some discussion of it, but we
thought it was sort of similar to what we're doing now, and we were sort
of considering should we take a more -- a maybe more drastic approach to
the 120-month update. I think over time, I think, you know, that has
been expanded out, I think from -- you know, it was initially much
shorter, and now, it's up to 120 months. It has gone longer. So that
would be a possibility. We did talk about it, but we thought that we
would sort of focus on these particular options; in terms of the
possibility of doing that, certainly. I think that's something that we
could talk about and see what, if that might be an option we would like
to propose.
DR. KRESS: Wouldn't that require a rulemaking?
MR. SCARBOROUGH: Yes.
DR. KRESS: You would have to change the rules.
MR. SCARBOROUGH: We would have to do a rulemaking on that.
DR. KRESS: Not only that; you don't have a good basis for
deciding whether that would be 180 months or 200 months or 50 months
even.
DR. BONACA: But there is nothing sacred about 120 months.
DR. KRESS: There is something; it's on the books.
[Laughter.]
DR. BONACA: I understand that.
DR. KRESS: It makes a big difference.
MR. TERAO: Well, actually, there is something -- there is a
correlation between, of course, the 120-month update requirement and the
10-year intervals established by the ASME code. The in-service
inspection and in-service testing codes still have separate requirements
for 10-year intervals, and it just happened to correspond with the
120-month update requirement. So, in other words, for example, when the
licensee performs inspections of the vessels and piping, the program is
for 10 years.
So if we were to say update every 15 years, that would cause
some perturbations in this program. They may have to change the ASME
code to have 15-year programs. So it's more complicated than just
selecting a date or a time period.
MR. SCARBOROUGH: Okay; we not only looked at the options
through the advantages and disadvantages sort of test, but we also went
back to the strategic goals and applied those to the various options to
see how they stacked up against each other that way, in terms of
maintaining safety, increasing public confidence, reducing unnecessary
burden and making the NRC activities and decisions more effective,
efficient and realistic.
So what I'd like to do is walk through in summary fashion
how did these stack up against each other regarding these goals? With
respect to maintaining safety, we feel that each of the options will
maintain safety, in the sense that they may require updating of the ISI
and IST programs, but the criteria would depend on the option; for
example, the option 1 would raise the bar for when you would be required
to update to issues where we've revised the baseline to incorporate
safety significant information. So we've sort of raised it up to a more
50.109 test.
Option 2 would continue the current process of its
automatically updated, and option 3 would be on a plant-specific basis
of whether or not to update or not. We did find that there have been
improvements to the code, and these are talked about quite a bit
yesterday, between 1989 and 1995, such as the OM code issuance and the
weld techniques and the comprehensive pump test.
So while these can't be explicitly quantified, there
definitely is an improvement that you can see in terms of the quality of
the provisions in the code, and we wanted to recognize that. Regardless
of which option we select, we'd continue to review future code editions
for endorsement. As part of that, we would continue the importance of
the ASME code, and where we did have an update, rebaselining of some
type, there would be the means to assess program adequacy as part of the
update process. So we felt all of the options would satisfactorily meet
the goal of maintaining safety.
In terms of increasing public confidence, there are some
permutations here that I would like to go through. One is that the
option 1 could increase confidence by applying all new requirements
consistently. The 50.109 tests would be applied across the board. For
option 1a, which is the 1989 code edition baseline, there could be a
negative perception because of the age of that baseline, just from a
perception point of view; it's an 1989 code. It's 10 years old already.
From options 1b and 1c could have a negative perception if
it wasn't apparent that there was a need to update; if it was just done
frivolously; if there wasn't a need that was demonstrated, it could have
a negative perception. Option 2 could increase confidence by continuing
to maintain the latest ISI and IST improvements in the program as they
go along, but if it involves unnecessary burden, it could have a
decrease in confidence.
Option 3 could have a negative effect if it was perceived
that we were just avoiding public participation in the process, so there
are some permutations there. We do feel that all of the options are
consistent with the National Technology Transfer and Advancement Act,
because we will continue to review and endorse the latest versions of
the code, and we will, whichever option we pick, we will establish a new
baseline periodically based on the criteria that we have selected for
that particular option.
And as we've mentioned before, in terms of public
confidence, all of the public commentors, regardless of which argument
they took in terms of the 120-month update, all of them argued strongly
that confidence could be most directly increased by more prompt
endorsement of the future code editions, and we got that message loud
and clear.
In terms of reducing unnecessary regulatory burden, the
licensees indicated that they could save $1 million to $1.5 million
every 10 years if their programs were not required to be updated.
However, those savings would be offset to some extent by the cost of
submitting relief requests; to apply portions of future code editions
and code cases and also updating if a new baseline was established. So
it wasn't quite clear exactly what the savings would be, but there would
be some competing factors there.
But we did feel that option 1 provided the greatest
flexibility to minimize their burden. If the licensee felt that a
future code that had been endorsed for voluntary updating in the
regulations was an advantage from a cost resource benefit, then, they
would update. They could have that option. If they decided that it
wasn't, and they could stick with the current baseline, they could
continue that, and it would be when we decided that the baseline needed
to be revised that they would be brought up to the next level.
Regarding the efficiency and effectiveness of NRC decisions,
if there were no program updates, there would be no relief requests
associated with those. However, the licensees are going to continue to
submit relief requests to use portions of codes and code cases, so
basically, we don't feel there's a resource savings, and there might be
some small increase in resource requirements for the staff to develop
under option one for the review of the future code editions for updating
the baselines. There might be some small increase there.
Option 1 removes the historical exclusion of 50.109 from the
updating process, but it does make it consistent with the other new
requirements for operating plants, so that does make it more consistent
with the other approaches and maybe more realistic from that point of
view. We would mandate significant code improvements. That would
continue. And we would have to do this additional process of review to
look for a new baseline. Option 1a, as a small part, we would have to
go back and look at option 1a to see what other aspects of the 1989 code
might need to be part of the baseline, but that's something that we
would have to do under that option.
I have a few more comments here on the activities on this
next slide. We are encouraging the use of risk-informed ISI and IST
programs by the licensees, so we have a major initiative under Part 50
to make it more effective in a risk-informed environment, and we're
emphasizing that a standard committee such as ASME code that needs to be
responsive to industry and regulatory issues. So we have a number of
activities ongoing; we're also evaluating the ISI IST process for errors
in improvement. So down the road, within a few years, there might be a
different approach for this whole process, so this whole area is in a
state of improvement that may cause some changes.
We don't think there is going to be a significant effect on
the range of code editions applied, because each program now is unique.
I mean, each plant has their own program; they have their own relief
requests, and they're unique as they are now, so we don't think that
there would be a significant change in that. The Illinois Department of
Nuclear Safety was concerned about the multiple editions in the
baseline, and so, we recognize that that's a concern that there might
need -- indicate a need to revise the baseline if we start to see
multiple editions, a lot of backfits tagged on to a baseline might
indicate a need to revise the baseline.
We're going to continue to participate in the code process,
so we don't think there will be any adverse impact on the risk-informed
initiatives under the code. And finally, regarding any state NRC and
consistency and requirements, we feel that would be resolved by Federal
preemption, but we do plan to try to make sure that as the code makes
significant changes that there would be a review to determine if the
person needs to rebaseline.
So that's how we stacked up against the four goals, and
based on our review of the public comments, the advantages and
disadvantages of each one of those and looking at the goals, the bottom
line that we came up with was that no particular option has an
overwhelming advantage over the other options regarding the goals. We
feel that option 1b would reasonably combine the strategic goals, and we
would recommend that the 120-month update requirement be replaced with
voluntary updating unless the baseline was revised, and we would revise
it using the 50.109 criteria.
Our basis for selecting the 1995 code edition is essentially
twofold: one is that it is incorporated by reference in the regulations
today with a requirement that licensees update to that edition at their
next interval and also the improvements that have been identified in the
code since 1989 through the staff review and the public comments, those
can't be quantified, but they can be looked at and definitely have an
increase in the capability and the provisions for ISI and IST
requirements.
DR. POWERS: These are essentially the same arguments you
made at the subcommittee meeting on why the 1995 versus the 1989, and we
subsequently had another speaker who argued that really, the 1989 was
the one to go with as the baseline and not the 1995. His argument
seemed to be that gee, everybody is there already, save maybe for four
or five plants, and so, if you're looking for burden reduction while at
the same time taking advantage of the big changes that have occurred in
the code, the 1989 was the issue to go with.
Have you given these arguments any thought?
MR. SCARBOROUGH: Yes; we talked about that yesterday
afternoon, and the selection of the 1995 code is -- was not based on any
one particular goal or any one particular aspect of it. We looked at
what was -- in terms of safety, what was the improvement in safety? Was
there an improvement in safety going from 1989 to 1995? And our answer
is yes, there is some improvement there because of the amount of changes
that were identified through our review and the public comments.
MR. BARTON: Let me ask you something. You said improvement
in safety. I'll take your word for that. Now, let's assume that we're
in the regulatory environment here where this 1995 would have been
voluntary. Could the staff make a case for backfit on those safety
improvements that you've identified in the 1995 code over 1989?
MR. SCARBOROUGH: We'd have to go back and look at each one
individually. I think there are some aspects that are significantly
different. I think the comprehensive pump testing would be an aspect
that you would have to go back and determine, you know, would that be a
backfit type? When 1989 was endorsed, and when we went to 1995, because
of the fact that the 120-month update process was in place, we didn't --
we were not concerned about areas that were being picked up eventually.
We were only really concerned about areas that need to be done on a more
prompt basis than the 10-year.
So our vision was is it okay for the plants to keep doing
what they're doing for the -- until they finish their next interval,
which is sort of a different question than what it would be if this is
what they're going to do forever, and that's what we would have to go
back and look at, and that would be a significant challenge to go back
and look up all of the changes, and I think we've heard about some of
those yesterday that were made to see if any of those in particular
raise up to the level that would be a backfit requirement, but we
haven't tried to do that.
MR. BARTON: Okay.
MR. SCARBOROUGH: I think what we're trying to
do --
MR. BARTON: I just asked that because you said the reason
you want for 1995 backfit basically that there were some improvements in
safety.
MR. SCARBOROUGH: Right.
MR. BARTON: Well, I thought if I were under a regime where
licensees could voluntarily take that 1995, could you really -- how
strong would you feel about safety improvements, or could you really
want a backfit case?
MR. WESSMAN: I guess if I could butt in on you a little
bit, Tom, I think that each of these incremental improvements between
1989 and 1995, if you look at it singularly, whether it's the flaw
evaluation process or the comprehensive pump test and some of these
other things, it's real hard to quantify and say aha, that makes it
safer. It may make it more efficient; it may recognize a different
technology and allow gaining a little bit more of information, but I
think it would be very difficult to quantify and say aha, those three
things definitely added safety benefit.
I think it's generally accepted by the code community and
the industry and the staff that some of these methodologies are better
methodologies. The one area between 1989 and 1995 that we recognized
was a substantive safety benefit and we applied the backfit test to was
the work under Appendix 8, and that test was applied, and it was used to
accelerate the implementation of the provisions of Appendix 8 for
in-service inspection as opposed to let the licensees just pick that up
when they do their next 120-month update.
That was the one stand-out of the changes between 1989 and
1995, but it gets much more difficult to try to quantify, and to some
degree, that lends support to some of the arguments by some of the
speakers you heard yesterday that said gee, leave it at 1989. Clearly,
we have a lot of plants out there under the 1989 code right now, and if
we felt that they were unsafe, the staff would have to take action, and
we would. We can't reach that conclusion. Can we say there are things
that are better as far as the inspection and test processes in 1995 when
compared to 1989? Yes, I think we can, but I don't think we can go out
and say licensees, you better go do it now.
MR. TERAO: And also --
DR. SEALE: Excuse me; as I understand it, though, if you
had some new code case that came down the road that you were going to
apply the individual aspect assessment to, the cost-benefit measure
would have to be applied to each individual possible change and not to
the aggregate of all changes that were contained within that code
package, since you only can buy -- or you only pick and choose on an
individual basis.
MR. WESSMAN: No, I think we could look at it as an
aggregate, and you see when the time came to say like the hypothetical
situation in 2010, whenever it is, there is an aggregate change, and we
would have to see whether that argument can, in the aggregate of those
incremental improvements be made either on a cost-benefit or a
qualitative, somewhat qualitative overall improvement in overall plant
safety.
DR. SEALE: Well, in that event, then, you would require
that the utilities make all of the changes and not just certain ones of
them.
MR. WESSMAN: Yes, sir; that would, of course, establish a
new baseline. And then, at that point, the utilities would be bound by
that new baseline.
DR. SEALE: So that's essentially sidestepping the idea of
picking only those aspects that would be of particular use to your
particular plant.
MR. BARTON: But that baseline has to pass the test.
MR. WESSMAN: Yes.
MR. BARTON: Yes.
DR. SEALE: But if you don't change the
baseline --
MR. WESSMAN: No, but so long as we are periodically
endorsing for voluntary use and so on in timely manner, and so long as
we are endorsing code cases for use by the utilities in timely manner,
then, these incremental provisions, if a utility says I see the benefit
of doing that particular type of test using this new technology, they
can choose and go ahead and use it.
DR. SEALE: It smacks a little bit of the idea that if you
divide your risk assessment up among enough sequences, you'll never find
a dominant one that's above 10-5 or whatever the number may be, and
that's, you know, at what point have you sliced and diced the problem to
the point where it is not a problem specifically any more?
MR. SCARBOROUGH: Yes; we recognize that's going to be a
concern, and that's why we emphasized in terms of our discussion of this
and in the position papers that, you know, we will look at cumulative
changes over time, because, yes, there is that potential there if you
slice things up so small, you never do anything, so we do want to try to
make sure that when we do review each new code edition for endorsement
in the regulations, that we discuss what are the cumulative changes and
what is the overall safety effect of that, and it might raise to the
level of changing the baseline.
DR. KRESS: What is the meaning of the -- when you say
endorse in the regulation? What does that mean? That's not a
requirement; it shows up where? In a reg guide or something? That says
we would endorse that --
DR. POWERS: That would be in the regulation itself here.
DR. KRESS: You would actually put that in the regulation?
MR. SCARBOROUGH: Yes; but what I envision is that we would
continue to incorporate by reference into the regulation 50.55a --
MR. BARTON: They don't have to comply with it.
MR. SCARBOROUGH: Right; there would be a baseline provision
that would be incorporated by reference, and that's why I'm using the
word endorsement, because it's more of a voluntary word.
DR. KRESS: So it's not a requirement, even though it's in
the recommendations.
MR. SCARBOROUGH: Right; they could use it pre-approved.
It's already pre-approved if they use the whole thing. So that's how
we're preapproving it, but the baseline might be like 1995, whereas, an
endorsed one might be 1998 or incorporated by reference for voluntary
use. I don't know if we would use the word endorsement. It might be
incorporated by reference for voluntary use or similar words like that.
But that's what we would do.
So there are actually the sort of two sets there. There
would be a baseline that would be the required; they have to meet, and
then, there would be later versions that had been incorporated by
reference for voluntary use.
DR. KRESS: Let me ask another question. Anybody can feel
free to answer it. And that is do you honestly believe a meaningful
backfit cost-benefit can be made for this particular issue? I mean,
where the benefit has to do with the changes in CDF, say; now, LERF II
because you're doing containment? Do you actually think those things
can be quantified in a cost-benefit analysis to decide on a regulatory
analysis backfit?
MR. SCARBOROUGH: Let me make sure I understand. Are you
talking about the process for changing from the automatic 120-month date
to the baseline and that whole process?
DR. KRESS: To a baseline and then a rebaseline.
MR. SCARBOROUGH: And then a rebaseline.
DR. KRESS: Yes.
MR. SCARBOROUGH: Oh; I think it would be difficult unless
we had some major event or provision change in the code. I think it
would be very difficult to do it on a pure risk basis, and I think that
gives me a lot of qualitative analysis in that, and that's why we have
talked to CRGR about that and that there is going to be a path, a
success path, to rebaseline based on a more qualitative and combined
with quantitative where possible.
DR. KRESS: I guess that is my question. Is there a success
path for that?
MR. SCARBOROUGH: I think we have, because there has been a
lot of concern about that: is there a success path? And I think there
is.
MR. WESSMAN: Wally wants to chime in if he could.
MR. NORRIS: Wally Norris, Research.
I think that's an important aspect that needs to be
considered. The last two amendments to 50.55a each took 7 years, and
about half of that was because of this judgment call that has to be
made; the considerable discussion, and it's very difficult to make.
MR. SCARBOROUGH: I just want to add that, you know, talking
about the 1989 versus 1995, sort of follow up and just finish your
question very briefly. We did look at all the goals in terms of
selecting which, whether 1989 or 1995, you know, maintaining safety,
public confidence; we talked about 1989. It's a very old code, even
though, you know, in certain cases, it may not be that significantly
different from 1995, it's very old from a public perception point of
view.
Burden, we do recognize that there will be a need to update
one more time, so we recognize that. In terms of efficiency, the 1995
code is already endorsed in the regulations. It's there. We don't have
to go back to 1989 and try to decide, okay, if we were going to baseline
to 1989, do we need to make other changes, or are we going to have a lot
of relief requests, because 1989, there have been a lot of relief
requests associated with that code. 1995 solves a lot of those.
So we looked at the whole gamut of the question. We just
didn't focus on one aspect of the goals. We tried to look at all of
them in terms of our final recommendation.
MR. BARTON: The bottom line on your option 1b, as a
licensee, I never have to update my program unless it's a reduction to
burden for me; then, I can voluntarily take on some changes in the new
code. But any increased requirement, I wouldn't take on unless you
forced it on me through backfit, and you'd have to prove the backfit
case. Otherwise, I'd never change my program for additional testing or
inspections.
MR. SCARBOROUGH: I mean, there are two ways we can impose
something that we felt was significant. We could do it through a
specific provision backfit, like we've always been able to do, like
Appendix 8, or we could do it through a cumulative effect where we
rebaseline, and we bring everybody up to new baseline. So there is the
option for us to do that, but we'd have to make that case. No longer
would it be automatic for rebaselining. Right now, we're going to
rebaseline every couple of years when we endorse a new code. But now,
we would raise that bar so that it would be a more -- consistent with
all of the other new requirements that we impose under the 50.109 test.
MR. BARTON: Okay; thank you.
DR. SHACK: Unless there is a burning question, I think we
will have to stop here. We're a little bit over schedule already, so I
thank the presenters for helping clarify some important points that we
didn't address yesterday in the subcommittee meeting.
MR. SCARBOROUGH: Thank you.
DR. SHACK: Mr. Eisenberg?
MR. EISENBERG: Well, thank you very much. I'm Jerry
Eisenberg, director of nuclear codes and standards at ASME. With me is
Owen -- to my far right, Owen Hedden, current chairman of subcommittee
for nuclear in-service inspection at ASME; John Ferguson, who is the
current vice-president of nuclear codes and standards; and to my left,
James Perry, who is the past vice-president, nuclear codes and standards
with 40 years' experience in the nuclear industry, 11 of which as a
vice-president of a nuclear utility.
Today, we'd like to summarize the points made in some of the
ASME letters to the commission and to the chairman regarding this
rulemaking; kind of outline the important code changes over the last 10
years and provide a basis for supporting the retention of the 120-month
update, and with that, I'd like to turn it over to Mr. Perry.
MR. PERRY: Well, thank you very much. It's certainly a
pleasure to have the opportunity to present our view to the ACRS
committee. Just for those who may not be familiar with the documents
we're talking about, there is a code of the Section 11 of the ASME code
back in 1971. This happens to be the 1998 one, and I think that, you
know, it reflects changes and increases and I think a much better code
than what we had before, and that's what we hope to show.
Next slide, please. Now, the ASME codes are really first
and foremost safety codes, and they're intended to protect the health
and safety of the public, so that's our main thrust. But since this
proposed change to the regulations, the changes are based on burden
reduction, that's what we're going to address, that aspect. But I
didn't want to lose sight of the fact that first and foremost, our
responsibility is safety and not burden reduction per se.
We say that the benefits outweigh the costs of the update,
and let me explain that. What we mean there is that using the ISI
program as an example, based on information feedback that we received
from seven utility representatives on the committee, they estimated the
average cost of the update for ISI 10-year is around $200,000, and we
can argue about what's right and what's wrong. My feeling is that that
is a cost incurred in one year, but it also is going to benefit you over
a 10-year period. So I think that the cost per year really comes to
about $20,000.
Now, to put it in perspective, what was in the proposal was
the 1989 as the base, and so, what we're saying is the one-time added
cost significantly applies to this total. For example, when the NRC
mandates IWE IWO being implemented, that's a brand new requirement that
utilities never were required to meet in the past as part of ISI, so
that's a significant increase. Likewise, Appendix 8 is very, very
complex and comprehensive and involves a significant change to the
in-service inspection program. So when I treat those two, which we're
saying are mandated anyway, that's the bulk of what it costs to update
the code, but when you do that, you're only doing that to 1989, and I'm
glad to hear that we're looking at, since the code on record at the
moment that applies is 1995, that we're looking at that one as a base.
In addition, the cost incurred by a utility for the review
fees for exemptions and relief requests are significant. So, for
example, if you baseline to the 1989, many of the changes and
improvements that utilities really need and want are in the later codes
or in code cases that are not in the 1989 yet, and so, they must go
forward to the NRC and request exemptions and relief requests in order
to do that in a program.
Now, my understanding is if you look at code cases
themselves, asking for approval of a code case, depending on the nature
of the code case itself and whether it's one utility applying for this
or it's a group of utilities, that cost per code case could be anywhere
from about $15,000 to hundreds of thousands of dollars, so it's no small
thing. So that's an added cost and burden not only to the utility but
to the NRC to respond to all of those. If you impose the later code,
where that is all covered, it minimizes the exemptions and relief
requests that have to be asked for.
One other key point: when you update like we have in the
past, every 10 years, that 10-year interval, you really focus, utility
focuses on an evaluation of the entire program where you identify
deficiencies, and it forms a basis for making corrections and
enhancements as you look at the changes necessary to do it, and that is
a powerful thing, and I can speak from personal experience at the
utility I was in: we found major problems when we did that and paid the
price because we didn't pay attention.
Now, I submit if you don't require continued 120-month
update, that is going away. That is not going to be done.
DR. KRESS: But excuse me.
MR. PERRY: Yes?
DR. KRESS: I guess I'm addressing this to the staff behind
me.
MR. PERRY: Yes.
DR. KRESS: When you make a regulatory analysis,
cost-benefit, do you include this particular aspect as a safety benefit,
or is there any possible way you can do that? Because it's uncovering
things that are unknown deficiencies, and I don't know how you can
predict they're going to happen and to what extent in the future, but
does that play any role in the cost-benefit of a backfit?
MR. WESSMAN: This is Dick Wessman from the Division of
Engineering. At least looking backwards, I don't think that we can say
it has so far in the past. I guess I would also observe that to some
degree, some of these things that the utilities identify at this 10-year
interval situation, it may be a little bit dependent on the individual
utility and how carefully they manage their programs over the 10 years.
If they did a good job, they shouldn't find a bunch just because they
decided to carefully scrutinize their program at a 10-year point. They
should do their modifications properly; they should reflect on the code
implications, and if a new check valve is installed, and it falls under
the provisions of the code, it should be added to the program at that
time.
I think that's about all I can say on that.
MR. PERRY: I would like to comment on that point, because I
think it's significant. The NRC, in their presentation yesterday,
referred to some LERs that relate to this problem that were reported by
utilities, and I'll give you one example of one when -- of the utility
that I was in when I was a VP.
We were at the end of the 10-year interval. We were going
into a refueling outage. We had a small group that was managing ISI
controlled by the vice-president of operation without proper checks and
balances between engineering and QA. We used a consulting firm to
implement the ISI. What we found at the last minute during this outage,
many instances occurred when the welds that were supposed to have been
UTed were not, and what the supplier said was he wrote a report; can't
do it because this is B-31 plant, and it hasn't been ground, and that
was thrown in the file; never evaluated.
Now, that was a significant impact. I'm not saying all
utilities do this. But I think the point we need to be careful of is
those utilities that participate in the code, that keep up and pay
attention to these details, I'm not worried about those. I'm worried
about the ones who don't get involved, who don't look at what's
happening, who don't pay attention, and those are the ones more inclined
to have the problems. And if you're forced to go into the update and
reevaluate this, you find things that you've been doing that have
degraded over a period of time, and I submit right now when we talk
about more pressure on utilities to be competitive, changing in
organization structure and responsibilities, the risks are greater today
than they were in the past.
I would like to conclude on this chart by saying that, you
know, it's a difference between short-term focus and long-term focus.
The short-term, which I think from a utility executive's point of view
might be what does it cost to upgrade and change this ISI or IST
program? And if we're allowed to not be required to do it, I can save
money by doing it. It's attractive. And I think I would say that, you
know, that's something you might want to do.
But from our perspective, I think we're looking at
long-term. We're looking at the 10-year interval. We're looking the
not just the up-front administrative costs to make the update, but what
are the benefits that you derive by doing that, by implementing these
changes over a 10-year period? And we think that the net result of that
is that the benefits outweigh the costs when you look at it in that
perspective.
Next chart. We see that the ASME codes are living
documents, and what we mean by that is that changes result from new or
improved inspections and tests; materials and design methodologies, and
really, they reflect the lessons learned over the 30 years of nuclear
experience and really are very responsive to the user feedback. In
addition, our codes are moving from the prescribed repetitive
inspections and tests to more risk-informed and performance-based
approaches, and even if you apply the risk-based techniques in the code
cases, and pilots are doing that, ultimately, you know, they tell you
how to separate out between the risk-significant and less significant
items, but ultimately, you come back to the code with respect to what
criteria you have to meet, what tests you have to meet, what ways in
which you document it and what are the qualifications of people.
DR. KRESS: Let me ask you a question.
MR. PERRY: So we're moving the code in that direction,
recognizing that's coming.
Yes, sir?
DR. KRESS: Yes; let me ask a question about that. I
envision a risk-informed performance-based code --
MR. PERRY: Yes.
DR. KRESS: -- that would have in it, then, some guidance on
the required frequency of updates because that is generally one of the
things you change when you risk-inform things. So the code itself might
address the appropriate frequency based on some sort of risk
considerations and based on some performances.
MR. PERRY: Yes.
DR. KRESS: If that were the case, and if NRC endorsed that
code, would that not automatically change this 120-month thing by
itself?
MR. PERRY: Well, I think the way I see it, it's the other
way around. The current code requires 120 months update.
DR. KRESS: Yes.
MR. PERRY: And, as I think the staff eloquently said
earlier, that's tied also to the interval in which you take 10 years to
do all the necessary inspections before you repeat or all of the
necessary tests before you repeat. So I don't think the code would want
to get in the business of saying, you know, to preempt the NRC, here's
how often you have to update. I think what we concentrate on is what
must you do, what methods do you use, what criteria do you use and how
often you do it maybe, and that's where the risk-based changes the
frequency and concentrates on risk-significant ones in terms of maybe
better methods and techniques and, like, UT in Appendix 8 to be able to
find these flaws and defects.
DR. KRESS: It is conceivable that that 120-month might get
altered in the code itself when you do a risk-significant and
performance-based --
MR. PERRY: Well, it could be, yes.
Okay; so, what we're also saying is numerous changes have
occurred since the 1989 edition that we feel improve safety; improve the
industry standards that we use; reduce the burden and respond to
inquiries and user feedbacks, and to be a little bit more specific, I
would like to now look at the next chart, and I will be talking -- the
summary from this chart, but I will also be calling your attention to
one of the handouts we provided you, which headed at the top says
important: Section 11 subgroup noninstructive exam code changes and
code cases.
What we're trying to show on this chart, and this is a new
piece of data that was not in our letters of submittal to the staff on
this revision, we have tried to summarize here the changes that have
been made since the 1989 edition through the 1999 addenda for the
Section 11, and we've grouped it according to the way Section 11 is
organized; in other words, subgroup on noninstructive examination;
subgroup on water-cooled systems; subgroup on repair, replacements and
modifications; and subgroup on liquid metal coolant systems.
So we have a total here of 255 changes during this 10-year
period. Further, the committee has looked at these changes and really,
they put it into two large groupings. One group, they say these are the
important changes; and then, I think the balance were less important.
So of the important changes, their total is like 80, 80 of the 255 are
considered by the committee to be important. And what they have done is
also looked at each change and said what is this change -- what is the
benefit related to? What is the primary benefit?
And so, on the left-hand side, they talk about a change
category for the changes that are incorporated in the code. First, they
talk about improved safety; second one is improved industry standard.
Third is the reduction in radiation exposure personnel; and fourth is
reduced requirements; and last, if it's a maintenance one. Now, when
you look at the handout, you will find that if you look at page 1, for
example, if you just look at item one for a second, here's one of the
changes that relates to what NRC has mandated, which relates to Appendix
8 on performance demonstration of ultrasonic exams.
But in the column in the middle on purpose and benefits,
you'll see that they say this change of benefits include safety
improvements, safety improvements from the point of view of improved
flaw detection and sizing of reactor vessel underclad reaches and piping
systems; safety improvements in terms of improved confidence that the
non-instructive exam contributions to failure probability calculations.
It also minimizes personnel exposure by appropriate detection and flaw
sizing approaches the first time, and this saves money as well.
Now, if you look at the classification on the right, the
first one that's listed, which they call primary, is improved safety.
The second one has to do with improved industry standard; the third one
is reduced radiation exposure. The table that I show you on the board
there and the view graph only looks at the primary one, but we can't
ignore the secondary ones.
At this point, I would like to draw your attention to page 6
of the handout, and there, we have two notes. The first note
summarizes, on the left, all of the changes in the E area over this
period to get to 1989, which matches what we show here. But on the
right, in addition to what's incorporated in the code, we show the code
cases and code revisions, which I'm not taking credit for here, and the
reason I don't is those are alternatives to what's in the code, and
they're optional to be used, but they also benefit people if they use
them.
More importantly, if you look further down on page 6 under
note two, you have a standardize and define what we mean by improved
safety, and what we mean by that are the action items that have obvious
effects on plant safety, such as improving the assurance of the pressure
boundary integrity. That's what we're all about: pressure boundary
integrity, or the other condition is reduced core damage frequency as
determined from the risk approach.
So it may not be the same definition as the NRC staff uses
in terms of is it safety significant, but in our view, since this is a
safety code, these are important considerations, and I submit the bar
under which we say this one is listed as safety improvement is pretty
high. You have other ones in there that I think relate to safety but
don't meet that bar, so we identify that as improved industry standards.
Likewise, as you go down the page, we define all the other terms as
well.
Now, if I look at the -- incidentally, if you look at the
handout, the first six pages really relate to the And to the NDE
portion. If you're looking at water cooled systems, you'll find that on
pages 7 through 9, and they have the same summary of the number of
changes and code cases and definitions. 10 through 14 cover the repair
and replacement, and then, the liquid metal coolant systems are at pages
15 and 16.
Just to summarize, the important safety items, there are 10
of those, and all 10 of those are listed in this attachment, and they're
listed as important to safety; in other words, all 10 are considered
important by the committee, and they're included in the tabulation. The
ones dealing with the improved industry standards, there are 124 total.
The committee has determined 55 of those, almost half of them, are
important ones, and so, those are listed and justified in terms of the
enhancements.
Under the reduced radiation exposure, there is only one, and
that one is listed as important, and I mentioned earlier there are many
others that show reduced radiation exposure, which, in my view, relates
to safety, because what we want to do is protect the health and safety
of the public, and certainly one of the public members are operators who
run the plants and do the maintenance to satisfy ALARA.
Under the reduced requirements, 13 of the 29 are listed as
important, and on the maintenance, one of the 91 are listed.
I'd now like to shift to the O&M code. A similar analysis,
not quite as detailed as this one, was done by the O&M committee, and
there, it's identified really the change categories are pretty much the
same as what we had before, and they have their definitions on page 17
of the handout, and what we have listed on page 17 really are the ones
that they considered important, and that's the three dealing with the
improved safety, and I think some of that was already mentioned by Dick
Wessman with respect to pump -- comprehensive pump tests was one in the
pump area; in the valve area, it's really the condition monitoring and
in-service exercising for check valves, and the third one really has to
do with the dynamic restraints in terms of the service life monitoring
of these. So those are the three that are considered important.
The 14 dealing with the improved industry standards are also
identified here as well, and the ones dealing with reduced radiation
exposure are listed at least by topic of the change also. What I would
like to call your attention to is that in addition to the code changes
in the O&M code from the 1990 edition to present, there were, I think,
nine code cases, O&M code cases, and most noteworthy of these is the one
dealing with O&M-1 on motor operated valves. A significant effort went
into putting out the document during the last 10-year period, and
incidentally, just to show you the impact, it has been endorsed by the
NRC in Reg Guide 96-05, so that one, I think, is a significant
contribution to the industry in terms of making sure those motor
operated valves are going to continue to operate when we need them.
Three other ones that are noteworthy are the risk-informed
related in-service testing group. That's O&M 3, 4 and 7, and one of
those relates to the method by which you classify the components as
safety significant or less safety significant. The others, I think, are
more specific to pumps and valves, and another one is coming up that's
specific to snubbers.
In conclusion -- last chart -- keeping the 120-month update
maintains a stable system which works. It provides an in-graded
approach to safety improvement and burden reduction. Maintain the
update process; use option 2. It works well.
Thank you very much.
DR. KRESS: Thank you.
MR. PERRY: I'll be glad to answer any questions you may
have.
DR. SHACK: I think you've made your case very well. Thank
you very much, Mr. Chairman.
DR. POWERS: You're welcome.
I think it will, because we've had a chance to get some
responses to the presentations and what not, and one of the contentions
that came up was 1989 versus 1995. You've had an opinion on that.
You've heard the staff's opinion. I can hear what your view on that
opinion is. Similarly, we had differences of opinion on how many
changes had occurred and whether they were safety significant or not.
You've had a chance to hear contentions on that and an interesting
rebuttal on that.
Kurt, did you want to say anything?
MR. COZENS: First of all, I'm Kurt Cozens with the Nuclear
Energy Institute. I'm the senior project manager responsible for this
issue, managing it on behalf of the nuclear utility industry.
Steve Lewis, who was with us yesterday, is not here today
and obviously will not be making further comments. I did very much
appreciate his presentation yesterday to bring a plant-specific
perspective on the issue. I wasn't initially going to make comments to
this panel today, because as I sat back listening to the other
presentations, it sounded like all of the issues that we had indeed
raised yesterday had been somewhat debated around the table.
But at the request here, I'll go ahead and make a few
comments and just highlight a note or two, and I'll try to be very
brief, just a few minutes. First of all, the industry does support the
elimination of the 120-month update. This is documented through
correspondence we've had with the utilities and feedback we've had. I
believe that in many cases, the utilities have written directly to the
commission supporting the change.
We do have concern, however, with the choice of option 1b
versus 1a, which was the 1989 versus the 1995 edition as the baseline
code. We look at this issue as looking at the primary function of the
NRC, which is to assure that public health and safety is maintained, and
I think it's so well characterized just listening to the last
presentation and hearing it basically two days in a row is that the ASME
code is identified between 1989 to 1998 that 255 changes, as they
characterize them, that were important were made.
But only 80 of those were really important in their
characterization as they documented it. If I understand correctly what
they said, that they subdivided it, although we kind of put this in the
larger bin, 80 of those were important. If you do the ratioing, that
approximately means that two-thirds of the changes that they said were
important maybe weren't real important. I think that's the problem that
the industry is dealing with, that when we go to the automatic update,
we need to deal with the whole of the baggage that comes with that, some
of which really don't have a lot of impact on assuring public health and
safety.
We also are concerned that there are two definitions of
safety, one that the NRC staff might use and then one that might be more
general nature that says yes, this is better technology; this is an
improvement. But in the regulated environment that we are in, we need
to assure that not only is this change an improvement of technology,
things that are nice to have, but that it's really essential that it be
mandated on licensees to ensure that public health and safety is
maintained.
I believe that should indeed be the threshold. I believe
staff's proposal is indeed moving the regulatory process in that manner,
where a baseline would be established, and then, the same guidance that
applies to all other regulations, the 50.109 backfitting regulation,
would apply then to changes in the future. Admittedly, there may be
some adjustments that the staff might consider doing on how they
actually implement that, but they're a little bit different than they've
done before; and that's something probably for a future discussion.
But yet, it means that this requirement just has the same
level of importance, the same level of management, the same level of
control that all other requirements under Part 50 have. So we think
that's an important change that is really balanced and appropriate at
this point in time.
We believe that the 1998 -- excuse me, the 1989 edition --
of the ISI/IST is appropriate as a baseline edition versus the 1995-1996
as proposed in the option 1b because even if you go to the option 1b,
you will have 10 more years of licensees living with the 1989. So you
ask yourself if they are safe enough to operate, and the commission has
deemed that they are safe to operate, why are you mandating this change
if you haven't demonstrated that the safety is necessary there? Even
the staff admits that at this point in time, it's been a subjective
evaluation, not a more rigorous evaluation, on why the 1995 would be
adopted.
We believe that it is just a logic question. If you go back
to the basic principles that the staff went through, the strategic goals
of the NRC staff, and one of them was to eliminate unnecessary burdens
on licensees. So if this rule is promulgated, and the 120-month update
is eliminated, you'll say oh, that's nice. But however, we're going to
do it one more time; 10 years, everybody will have to do it one more
time, and the question I would have to ask is the safety significance
there to ensure that -- to justify imposing new requirements on them
when they haven't even been evaluated?
And so, that really completes my statement. Any questions?
[No response.]
MR. COZENS: Thank you.
DR. SHACK: Thank you very much, Kurt.
Mr. Chairman, I'll hand it back to you slightly late.
DR. APOSTOLAKIS: We're going to take a break now until
10:30.
[Recess.]
DR. POWERS: Let's come back into session.
We're going to turn now to a topic that the committee has
expressed some marginal interest in --
[Laughter.]
DR. POWERS: -- dealing with low-power shutdown risk. And I
will turn to Professor Apostolakis, and he will remind us what it is
that the committee had some interest in this subject for and where we're
going now.
DR. APOSTOLAKIS: Okay; thank you.
Well, the committee has written at least one letter where it
expressed its concern that the tools that are available for assessing
the risks from low-power and shutdown operations are not up to the level
that we would like them to be, certainly not if you compare them with
the tools available for assessing the risk from power operations, and
there is evidence that these operations may, in fact, create levels of
risk that are comparable to those from power operations.
The staff has prepared a report, which is called Low-Power
and Shutdown Risk: A Perspectives Report. We had a subcommittee
meeting on November 18, and we had a presentation by the staff on this
report. We had a version of that report at that time. I understand
there is a new version now which I just saw today; maybe -- I don't know
how different it is.
And there were several questions that the subcommittee
members raised at that time, and I hope that the staff will address them
today. I'm not sure it's worth repeating them. One interesting
question was if the risk is so high, how come we haven't seen any
events, serious events.
DR. POWERS: The contention was the other way around, that
you have lots of events, but they never seem to progress to damaging
fuel.
DR. APOSTOLAKIS: That's what I meant, the damage.
DR. POWERS: And it suggests that the part that I think PRA
techniques have a hard time accommodating is unproceduralized
intervention actions. It's very difficult for the PRA to address those.
But maybe in shutdown, inherently, there are facile ways to always
intervene --
DR. APOSTOLAKIS: Yes.
DR. POWERS: -- detect and intervene that aren't
proceduralized. And so, maybe you never get to core meltdown,
discounting those things like blind dilution events that have very short
time scale on them. I mean, that seems like a possibility to me, but
maybe our risk assessments have just been unduly pessimistic in these
areas.
As long as I'm talking, I might as well keep going here and
ask some other questions that come to mind when I think about shutdown.
What we do see among the utilities is a great deal of attention being
paid to outages and a lot of effort going to planning these outages so
that things are done very efficiently, very smoothly and presumably very
safely. What I want to know is what kind of tools do our inspectors and
senior reactor analysts in the regions have to assure themselves that
indeed, the detailed plans that licensees follow during outages are
moving not only in an efficient direction but a direction of safe
operations?
DR. APOSTOLAKIS: So shall we have the staff start the
presentation?
Mr. Cunningham?
MR. CUNNINGHAM: Okay; good morning; I'm Mark Cunningham
from the Office of Research. The bulk of the presentation will be made
today by Erasmia Lois and Mary Drouin of the branch. I wanted to just
have a few introductory comments.
First of all, as Dr. Apostolakis noted, we had a
subcommittee meeting a few weeks ago. The presentation that we'll make
today is we've tried to tune a bit relative to the subcommittee meeting
to reflect some of the comments we've received. Probably one of the
bigger changes is where I think we've gotten to the point of better
describing how well existing tools can be used in risk-informed
regulation activities and what we see as outstanding issues.
The report that we provided to you a few weeks ago is also
being changed. I think we talked at the subcommittee meeting that this
is due to the commission by the end of this month.
DR. APOSTOLAKIS: So we did not, then, receive a new
version.
MR. CUNNINGHAM: I'm not quite sure what you received. The
principal change that's being made to the report kind of in parallel
with this presentation is at the chapter 5, which is the kind of
analysis and recommendations part, laying out our plans for the next
couple of years in this area. You'll see at the end of this
presentation, in effect, what will be the substance of that new chapter
5.
DR. APOSTOLAKIS: When is this paper due to the commission?
MR. CUNNINGHAM: This is due to the commission on December
31.
DR. APOSTOLAKIS: So basically, this committee will not have
an opportunity to see the final version before it goes out.
MR. CUNNINGHAM: Well, one thing we can discuss today --
well, let me back up. At the subcommittee meeting, we requested a
letter, and again, we continue to request a letter from the committee on
this report. Recognizing that the committee has a version that's
several weeks old or almost a month old now and that we are making
considerable changes, one possibility is that we could provide a new
version of the report and a draft of the commission paper sometime next
week for the committee to look at.
DR. POWERS: There is no opportunity on our agenda for the
committee to look at it as a whole.
MR. CUNNINGHAM: That's --
DR. APOSTOLAKIS: So we will have to discuss this after your
presentation.
MR. CUNNINGHAM: That is right.
DR. APOSTOLAKIS: It looks like it will have to be deferred.
MR. CUNNINGHAM: Our intention was that the presentation
today would provide the substance of what the report would say and what
the commission paper would say. The paper itself I think will be
primarily just a summary of the report focusing on what we see as the
recommendations for future work.
DR. POWERS: Painful experience has led to the committee
making a pretty iron-clad rule that we do not send up comments on things
on the come.
MR. CUNNINGHAM: Yes.
DR. POWERS: That too often, we've had heartfelt promises
made, this is what the paper will say, and it did not, and so, the
committee just really doesn't do that.
MR. CUNNINGHAM: Yes; okay; obviously, that's the
committee's decision to make.
DR. POWERS: Okay.
MR. CUNNINGHAM: Okay; with that, let's go into the
presentation.
Erasmia, I believe you're going to -- okay; one of these
people to my right will make a presentation.
[Laughter.]
MS. DROUIN: Actually, it's going to kind of flow this way.
MR. CUNNINGHAM: Okay.
MS. DROUIN: I'm just going to give some highlights of what
Erasmia will be talking about.
DR. APOSTOLAKIS: Who are you?
DR. POWERS: You didn't attend our subcommittee meeting, so
we don't really know who you are.
[Laughter.]
DR. APOSTOLAKIS: Yes.
MS. DROUIN: I am Mary Drouin with the Office of Research.
I apologize for not being here at the subcommittee meeting; however, I
was very involved in low-power shutdown at the time, attending an ANS
meeting.
DR. APOSTOLAKIS: So it was not that you were in France
again.
MS. DROUIN: No, I was not in France.
[Laughter.]
MS. DROUIN: I was working with the project team on ANS on
what their intentions are on trying to do in regards to the standard on
low-power shutdown.
DR. APOSTOLAKIS: That's a mystery to me, by the way, how
you can have a standard without having the method.
DR. POWERS: Your obviously limited appreciation of these
things, I suppose.
DR. APOSTOLAKIS: But I'm sure Mary knows what is happening.
DR. POWERS: I'm still struggling with whether being in
France is like being in dutch or not.
[Laughter.]
DR. KRESS: Or being in limbo.
MR. CUNNINGHAM: One of the issues that will come up
sometime today is that it may be worthwhile for the committee to hear
from the ANS.
DR. APOSTOLAKIS: Oh, it will definitely be worthwhile --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: -- Mr. Cunningham.
MS. DROUIN: Yes.
DR. APOSTOLAKIS: And you can rest assured --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: -- we will be --
MR. CUNNINGHAM: -- in the area of shutdown and --
DR. APOSTOLAKIS: -- very interested in that.
MR. CUNNINGHAM: And in this area as well as the standard
that they are developing in seismic.
DR. APOSTOLAKIS: But anytime you feel, Mary, that the ANS
work is ready to be debated here, feel free to ask for time.
MR. CUNNINGHAM: Okay.
MS. DROUIN: But again, I'm just one member on the project
team. The appropriate person to contact there would be the chairman of
the project team, which is Bob Budnitz.
DR. APOSTOLAKIS: Yes; we're not going to contact him.
We're just telling you that we are very interested.
MS. DROUIN: Yes.
MR. CUNNINGHAM: We can presume that if Bob Budnitz
volunteers to come to talk to the committee, you would be open to those
-- to that overture.
DR. APOSTOLAKIS: It will be a pleasure to have him.
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: Dr. Budnitz here.
Okay; Mary.
MS. DROUIN: Okay; today, we're going to just quickly, you
know, revisit, you know, the purpose of, you know, why we're here and
why we're looking at low-power shutdown, and in looking at low-power
shutdown what approach that we are taking in trying to take the
insights, the models, et cetera, and how these get factored into, you
know, the decision making process for our risk-informed activities.
We're doing this, of course, because of the significance we
feel is associated with low-power shutdown, and we will quickly revisit
that. We're going to go over the methods and tools that are currently
available; what those benefits are and how we feel we can build upon
these methods and tools to get where we think we need to be.
DR. POWERS: When you look at the world's methods and tools,
do you look at it and say, okay, here's what they have, and that's what
I want to have? Or do you look at it in the sense of I know what we
need; now, do they have it?
MS. DROUIN: The one you just said. This is what we need.
DR. POWERS: So do you have a document somewhere that I can
look at that says here's how well we have to be able to calculate
low-power and shutdown risks?
MS. DROUIN: Whether it addresses the question to your
satisfaction -- hopefully, it does, and hopefully, when we finish up
with the insights report, that was one of the things that we went
through and laid out the different approaches for incorporating
low-power shutdown into the decision making process. Whether you were
going to go via this qualitative, non-PRA approach that we call it
versus a plant-specific PRA approach, what is it that you need? And
given what you need, here are the methods and tools that are out there,
and how do they match up?
DR. POWERS: I guess I'm intrigued here that there are -- I
hear people all the time saying gee, what can we do to minimally treat
low-power and shutdown risks? And they said maybe we can use
qualitative techniques. And these things all surprise me, having
developed a wonderful, wonderfully quantitative technology for
operational risks and now wanting to couple it with something that's
qualitative, and it always surprises me. Is there a reason for wanting
to do that?
MS. DROUIN: Well, I think that you want to try to take
advantage of what's out there and where you can use it without
necessarily maybe going and doing, you know, a full scope PRA. You
know, is that truly needed for every application that you're going to
have?
DR. POWERS: Another way of thinking about it is we haven't
had it for any applications.
MR. CUNNINGHAM: The industry has made a considerable
investment to apply some techniques to manage their outage risk, and a
lot of those techniques are qualitative. I think that we would be
remiss if we didn't at least consider whether or not those techniques
could have applicability in our interest --
MS. DROUIN: Right.
MR. CUNNINGHAM: -- our regulatory interests, and that's
part of what we'll get to in a little while.
MS. DROUIN: Yes.
DR. APOSTOLAKIS: It really depends on the use.
MR. CUNNINGHAM: Yes, it depends on the application.
MS. DROUIN: Absolutely; it depends on the application.
MR. CUNNINGHAM: And as we'll see, we've got a number of
potential applications here within the agency.
MS. DROUIN: Okay; and then, the last two things we really
want to put the focus on is, you know, what is our plan? What are the
tasks that we envision over the next 2 years and, you know, what the
schedule is? And even though, as Mark said, we don't have a commission
paper written at this point, we do intend on taking the slides and
essentially writing a commission paper around these last slides that
focus in on the plan, the task and schedule so that you will have at
this point hopefully a good feel for what we plan on putting in our
commission paper.
DR. APOSTOLAKIS: Now, as you know, some commissioners have
expressed doubt that we need to do any extensive work in this area. Is
one of the purposes of submitting this paper to convince the commission
that we should go ahead with this?
MR. CUNNINGHAM: In effect, yes, that right now, we have
funds available in FY 2000 for work in this area; the commission has
basically said we'll consider funding this work in future years based on
what the staff tells us in this and probably other forums.
DR. APOSTOLAKIS: Okay.
MR. CUNNINGHAM: So yes.
MS. DROUIN: Okay; next slide. Again, I don't want to spend
a whole lot of time here.
DR. APOSTOLAKIS: What kind of font is this?
MS. DROUIN: What kind of font?
DR. APOSTOLAKIS: It's strange.
DR. POWERS: One that bleeds.
[Laughter.]
DR. APOSTOLAKIS: One that bleeds?
MS. DROUIN: This is not the font; this was the printer.
DR. APOSTOLAKIS: This was the printer?
MS. DROUIN: The printer, and it was bleeding on the
transparencies.
MR. CUNNINGHAM: Or the fuzzy logic.
[Laughter.]
MS. DROUIN: But it's actually a Times Roman font.
DR. APOSTOLAKIS: Okay.
MS. DROUIN: But the printer was unfortunately bleeding, and
we didn't have time to fix that problem this morning. So we apologize
for that. It's not bled on your piece of paper, though.
DR. APOSTOLAKIS: No.
DR. POWERS: Already, we find a difference between our
handout and what's presented.
[Laughter.]
MS. DROUIN: The point that I want to make on this slide is
that the uses of where low-power shutdown, you know, plays a part in the
regulatory activities, you know, is quite extensive. You know, it goes
way beyond just reg guide 1.174 and changes to the licensing basis. You
know, we're embarking on Part 50; low-power shutdown is going to play a
role there. The plant performance work in terms of inspection and
enforcement; looking at events analysis; performance indicators;
low-power shutdown, the risk associated with low-power shutdown is
playing a role in all of this, and so, our understanding of low-power
shutdown; what is the risk-significance of it; you know, what are the
important design features, et cetera, are all things that we need to
know in moving forward in these different areas.
DR. POWERS: I guess I'm stunned by this slide.
MS. DROUIN: Stunned by this slide?
DR. POWERS: Nowhere up there does it say and 58 percent of
our augmented inspection teams are sent out to investigate events that
occur during shutdown and low-power operations.
MS. DROUIN: I guess I'm not understanding your comment,
Dana.
DR. APOSTOLAKIS: He is giving you more ammunition, Mary.
DR. POWERS: It looks to me like the only reason you're
pursuing low-power shutdown is because of a reg guide and some
aspirations to risk-inform the regulations, when, in fact, it seems to
me that in years past, when we had an AEOD, we had them coming in
telling us that 58 percent of our augmented inspection teams are sent
out to look at incidents that occur during shutdown and low-power
operations.
MS. DROUIN: I might not have those exact words, but NRC
oversight, that encompasses that; that thought is in there.
DR. APOSTOLAKIS: Make it clearer, though.
DR. POWERS: But, I mean, what could be more important to
the agency? Addressing issues that occur at the plant or refining which
tools you have to evaluate applications under 1.174?
MR. CUNNINGHAM: I think what we're trying to say here is
there is a variety of important applications here. A front line one, if
you will, is what we talk about as the oversight program, and you
alluded to it earlier, I believe, that licensees are out using some sort
of qualitative methods to decide, to manage their outages. Our side of
it is, well, what kind of tools do our SRAs or our residents need to
have in order to be able to effectively review that information or make
-- pass judgments on the information the licensees provide? So that's
kind of implicit in the third bullet there, that because we see the
importance of shutdown risk in terms of numbers of AITs; in terms of
risk studies, that sort of thing, we see a need to be able to make sure
that the people in the oversight program have adequate tools.
DR. APOSTOLAKIS: Also, would the transition risk be
somewhat there? You know, the transition mode? You said last time that
you thought that was one of the areas where we don't really have any
tools.
MR. CUNNINGHAM: Yes.
MS. DROUIN: That is correct.
In supporting all of the various activities and looking at
the previous slide, that slide was not meant to say these are the
activities. The point was that we do have a whole variety of them, and
it wasn't even meant to be a prioritized list. But in supporting all
the different activities, in looking at it globally, how we plan to, you
know, approach the work in terms of what we need to be doing on
low-power shutdown, the first one, you know, is looking at collecting
the information; you know that's out there.
We've already embarked upon this particular activity quite
extensively over the last 6 months, you know, by having the workshop.
We've been doing a lot of literature search. We've gone out on some
plant visits. We don't think we're complete in this area, even though
we have done quite a bit of activity there. The same thing on the next
bullet, in terms of looking at the methods and tools. We've made a
major investment in terms of gathering information there on the various
tools and methods and getting an understanding.
DR. POWERS: How do you evaluate the adequacy of current
methods and tools? Do you have some idea of how accurately one thing is
to be done? How completely, how comprehensively? Or do you say well, I
don't have that, but I have what I have for operational stuff, and I can
compare it to them? How does the adequacy determination fit?
MS. DROUIN: I think it's a mixture of both, and it's also
looking at here, what our needs are for the different applications, and
are these tools and methods providing that?
DR. POWERS: Do I have a list of these needs somewhere that
I can look at?
MS. DROUIN: We took a first shot in the report; again,
whether they're all listed there; whether you agree with them; whether
you think they're sufficiently explained, you know, is another question.
DR. WALLIS: Isn't one way to evaluate the adequacy to
actually use them and see if they work?
MS. DROUIN: Absolutely.
DR. WALLIS: Is that what you're doing, too?
MS. DROUIN: At this point, in terms of actually getting
some of these tools -- now, when I talk about the tools, in getting
them, I would be talking more about like, you know, using ORAM or EOS or
some of those. Right now, I don't think that was one of the things that
we were looking at in terms of these software tools.
DR. WALLIS: Well, sometimes, there is a difference between
what the tool says it will do and what it actually will do.
MS. DROUIN: Right.
DR. POWERS: Surely not.
[Laughter.]
DR. WALLIS: The concept of a hammer is quite simple when
you look at it, but using it is another matter. There are hammers and
hammers, and some are better than others.
MR. CUNNINGHAM: The tools that Mary talked about, EOS and
ORAM and things, were designed for a very different purpose, and again,
we want to build on what they have, but I think kind of implicit in what
we're saying here is that for our purposes, we don't think right now
they're completely sufficient, because we have a very different need
from that.
DR. WALLIS: I was just wondering if adequacy is the right
word. I mean, you can drive a tack with a sledgehammer adequately, but
it may not be appropriate.
MR. CUNNINGHAM: Appropriate.
DR. WALLIS: And the best use, the best design for the
purpose.
MS. DROUIN: Yes; I mean, a better word might be whether
they're sufficient to be given, because we were not embarking on a
program to go and, in a sense, do a quality assurance on these, but are
they giving us the products and the results of what we need? So, I
agree; you know, adequacy may not be the appropriate word there.
And as part of the approach, as I said, you know, ANS has
embarked on their effort to develop a standard. It's in the very early
stages, but we're going to be working very closely with them in this
effort.
DR. POWERS: Isn't that remarkable, that you can develop a
standard and yet, to my knowledge, we haven't seen the kind of effort on
estimating risk during low-power and shutdown operations that preceded
the development of a standard for PRA during operations? Which is also
going on just now, but it's bunkered with 25 years of people trying to
develop the technologies in a quantitative sense.
MS. DROUIN: Right.
MR. CUNNINGHAM: So the issue may not be so much as the ANS
work a little early is the development of a standard for full-power
operations a little bit late in the development.
MS. DROUIN: Because I personally feel the ANS -- that this
is the time to be doing it.
DR. POWERS: Good point.
MR. CUNNINGHAM: But I think it's also fair that the ANS one
will probably have more -- it will, of necessity, be a little softer, if
you will.
MS. DROUIN: Oh, absolutely --
MR. CUNNINGHAM: Than ASME, just because of the state of
technology of shutdown risk.
DR. APOSTOLAKIS: We'll have to wait and see.
DR. POWERS: So maybe one can successfully look upon the ANS
effort as an intent to defend the adequacy.
MS. DROUIN: Yes.
MR. CUNNINGHAM: It's an important parallel effort to help
us define what we need to do; very much so.
MS. DROUIN: In the tasks that we're going to be getting
into and that we will get into more detail in the presentation, you
know, develop and prioritize and implement the task, they're going to
involve three areas; you know, guidance, analytical work and new
methods, and we'll get into the specifics of these, because in terms of
the prioritization, we've done part of that to determine what we want to
be doing in the next 6 months; what we want to be doing over the next
year; what we want to be doing over the next 2 years. And so, we will
get more into that during the presentation.
DR. WALLIS: Do you have a list of achievements you wish to
come out with at the end of the work?
MS. DROUIN: I think the answer to that is yes, if I
understand your question, so I would ask to bear with us until we get to
that part of the presentation.
DR. APOSTOLAKIS: What is the fundamental -- I mean, if one
asked you how's the PSA or what would be the important differences
between a PRA for power operations and a PRA for all other nodes? What
are the key issues that, in your opinion, would make one more
challenging than the other, perhaps? Or how is it --
MR. CUNNINGHAM: Again, if you'll bear with us, I think
we'll come back to that in a little while.
DR. APOSTOLAKIS: So you do have something on that.
MR. CUNNINGHAM: Yes.
MS. DROUIN: Yes.
MR. CUNNINGHAM: We'll come back to that.
DR. APOSTOLAKIS: Okay; go ahead.
MS. DROUIN: I'll let Erasmia pick up at this point.
MS. LOIS: It's a very tough one. I guess before we had any
subcommittees, we were saying that the shutdown risk used to be
comparably to low-power; here, we say it's significant, and we present
some results in terms of fractions.
The results from the foreign studies are reported as
fractions, and therefore, we would, you know, most of them, we don't
know what is the actual CDF, or how did they come up with these
percentages, but the point here is that shutdown risk is not negligible,
and therefore, it should be considered.
DR. APOSTOLAKIS: Now, has anybody on the second or third
column done work, either U.S. industry or foreign, that goes beyond what
the NRC sponsored at Sandia and Brookhaven?
MS. LOIS: I guess the Goesgen PRA, for example --
DR. APOSTOLAKIS: Which one, for example?
MS. LOIS: Goesgen, Goesgen.
DR. POWERS: The Swiss, the Swiss studies.
DR. APOSTOLAKIS: Switzerland?
MS. LOIS: It's more complete in the sense that they also
include startup risk, which is something that we haven't done at the
NRC, but on the other hand, it may not be as complete because I don't
think they did a level three analysis, which was done at the NRC, so I
guess all of them have various points are more complete or less
complete.
DR. APOSTOLAKIS: Was this done by American studies?
MS. LOIS: It's a PLG study.
The Dutch PRA has done a more comprehensive HRA analysis
according to my knowledge, but there are differences in the studies.
DR. POWERS: What I know is that those studies were not a
look at the entirety of low-power shutdown. They looked at one node of
operation. Is the apparent discrepancy between US plants as far as
their LPSD CDF as a factor of full power CDF and the foreign ones just a
matter of scope?
MS. LOIS: I don't think so, because one, the industry PRA
studies in the domestic industry are more shutdown studies. But then,
if you compare the NRC studies and the foreign studies, they may be,
some of them, more complete in some sense than some other ones. For
example, the NRC studies, there's an analysis of the modes and plan
operational states before they analyze specific modes or states. So
doing a screening analysis, one could argue that did not disconsider it.
DR. UHRIG: Are these values instantaneous values, or are
these integrated values over a period of time?
MS. LOIS: These are integrated per year. I mean, this is
percentages on comparing full power and shutdown on a per year basis.
DR. UHRIG: And that assumes a certain number of shutdowns
per year or one per year?
MS. LOIS: If it's one every 15 months, it would be average
to a per year basis.
DR. UHRIG: Yes.
DR. APOSTOLAKIS: I guess I don't understand; look. You
calculate -- I mean, one of these studies calculates a core damage
frequency during a particular mode. Now, that is a frequency, so it has
nothing to do with how long you are in that state, right?
MR. CUNNINGHAM: It's a conditional -- yes, a conditional --
DR. APOSTOLAKIS: But is that the number you compare with
the core damage frequency from power operations? In other words, if the
plant was in this mode for the whole year, then, the core damage
frequency would be this and its 50 percent of the power core damage. Is
that what you are comparing?
MS. LOIS: My understanding is that it also said these
studies, we have these fractions. We got these numbers from
publications. And therefore --
DR. APOSTOLAKIS: Oh.
MS. LOIS: -- we cannot say how these numbers are derived,
okay?
DR. WALLIS: Doesn't the publication define how they're
derived?
MS. LOIS: Not really.
DR. WALLIS: It's a very simple matter.
MS. LOIS: They're very high level publications, and they --
DR. APOSTOLAKIS: Because that would be an important
consideration.
DR. POWERS: Well, I think I know exactly what the ones on
our Brookhaven and Sandia studies --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: So what are they?
MS. LOIS: Yes.
DR. POWERS: They calculate the risk during the shutdown,
and they divide it by the necessary factors so that it's averaged over a
year.
DR. APOSTOLAKIS: Oh.
MS. LOIS: That is correct.
DR. APOSTOLAKIS: Averaged meaning -- so in other words, if
I'm in shutdown for 50 days, they will multiply by 50 and divide by 365?
DR. POWERS: Yes.
DR. APOSTOLAKIS: Okay; so it's a normal --
MS. LOIS: Yes.
DR. APOSTOLAKIS: But I have another question that perhaps
my colleagues with utility experience can help a little bit here. When
we say that the plant is in mode X, does that mean that there is a
number of different configurations that go with that mode?
MS. DROUIN: Possibly, yes.
DR. APOSTOLAKIS: Okay; because -- so just by saying that we
are studying, you know, mid-loop operations, that may mean many
different things, depending on which components are out of service and
so on, and I wonder whether that is actually part of the risk
assessments here, because one thing that has been of concern to this
committee not only for low-power shutdown but also the maintenance rule
and so on is how do you control these different configurations? How do
you know that you are actually evaluating a risk numbers that make sense
and reflect the actual configuration?
So I wondered to what extent the configurations themselves
are part of the calculations? Now, Mark wants to say something and then
Dr. Bonaca.
MR. CUNNINGHAM: I was going to say in the two NRC studies,
I think what we found was, as I recall, is just saying it's in mode four
or mode five was not sufficient.
MS. DROUIN: That's right.
MR. CUNNINGHAM: In effect, we had to subdivide them and
define what we called operating states, and there were many more
operating states than modes because of this, of the phenomenon that
you're talking about, that within a mode, you can have different
configurations. So our operating states were defined more by
configurations in the plant, so they subdivided the modes. And I think
that's fairly typical.
DR. SEALE: It's also typically dependent upon whether it's
a BWR or PWR, is it not?
MR. CUNNINGHAM: Well, that's the first place. And then,
again, within BWRs have different set of modes in effect than Ps.
DR. BONACA: The point I wanted to make is the most striking
difference; for example, if you decide --
DR. SEALE: Could you speak up a little bit more?
DR. BONACA: If you decide to offload the whole core versus
not to offload the whole core just for the same mode, you have
significantly different risk. That's just an example of how significant
the risk; that's the decision alone.
DR. APOSTOLAKIS: And this is entirely up to the utility?
DR. BONACA: Yes.
MR. BARTON: They are limited by what's in the specs.
They're limited by what configurations are in the tech specs.
DR. APOSTOLAKIS: I guess my question is when the analysts
do this analysis, how confident can they be that they are really
analyzing real configurations? Do they have to speculate? Or something
random can happen, you know, 3 and a half years from now that will
create a different configuration? Is that possible?
DR. SEALE: Likely.
DR. APOSTOLAKIS: From a different configuration?
DR. SEALE: Almost guaranteed.
DR. APOSTOLAKIS: Wait, wait, wait; this is a very serious
point, then, if it's guaranteed.
DR. SEALE: Of course it is.
DR. APOSTOLAKIS: Yes?
MR. CUNNINGHAM: I was going to say again, from the purposes
of the NRC studies, what we did to define the plant operating states was
to look at the historical record of outages in those plants. That by no
means assures that the next outage would be, if you will, covered by the
ones we've defined so far.
DR. APOSTOLAKIS: And that is certainly a very reasonable
thing to do.
DR. KRESS: That's the only recourse you have.
DR. APOSTOLAKIS: But still, the question is well, okay,
let's look -- say you look at the historical record, as you should, and
you have, say, five outages in the recent past, and you find that in
each one of them, there were different components.
DR. KRESS: Oh, absolutely.
DR. APOSTOLAKIS: Now, what do you do? Do you analyze each
one separately? Is that the plant operational state?
MS. LOIS: I guess -- can I say something?
DR. APOSTOLAKIS: Yes, sure.
MS. LOIS: Going back to the tools, right now, the industry
has the capability to do a mini-PRA for each one, plant operational
state, if you will. So, if you -- as you transit from one state to the
other, then, you know exactly what equipment you have available. And
therefore, you can reconfigure your model and come up with a conditional
CDF or boiling frequency, whatever it is, for that specific
configuration.
And you do a predictive analysis here, and for example, in
South Texas is one of the most, you know, the most advanced plants, they
are using this capability to determine what would be the optimal
schedule with respect to, you know, their schedule as well as economics
and safety.
DR. APOSTOLAKIS: So they are using their risk monitor to do
this?
MS. LOIS: That is right.
MR. CUNNINGHAM: Yes.
DR. KRESS: You can't really do that for the purposes of
risk informing regulations, because that's specific known configurations
that you have there. George, you've hit -- you've put your finger on
the exact problem that exists with doing low-power and shutdown risks,
and that is if you really want to know the risk, you have to know the
risk for all future configurations.
DR. APOSTOLAKIS: Yes, exactly.
DR. KRESS: And there's no methodology in the PRAs to do
that at the moment.
DR. APOSTOLAKIS: Yes, yes.
DR. KRESS: Unless you use past experience and project it
into the future some way.
DR. APOSTOLAKIS: Or if Erasmia, you know, the point Erasmia
just made, if a plant has done this analysis and says look, in the
future, we will use these configurations for these modes --
DR. KRESS: Yes, but that leaves out the unplanned
shutdowns.
DR. POWERS: Yes.
DR. APOSTOLAKIS: That's right; or some need
for --
DR. KRESS: It even leaves out the planned shutdowns,
because you never, never go with the way you project.
DR. APOSTOLAKIS: So seems to me --
DR. POWERS: This is the question here: you've hit on a
conundrum that you've spoken to publicly and privately on for some time
about this and even advocated a resolution on this, and a very
imaginative resolution it is that I kind of like, but let me ask this
question: if I look at the history, and I say what has happened is that
vertically, the configurations I get there and what will happen in the
future are closely related, how big of an error do you think I make by
doing that? And is that error large or small in comparison to the error
inherent in estimating the risks of a particular configuration?
DR. KRESS: Yes; I don't know; I would guess that the error
you get by projecting into the future is going to be small compared to
the error you get by just calculating the risk of a given configuration
in the first place.
DR. POWERS: Even if comparable?
DR. KRESS: Even if comparable.
DR. POWERS: It argues to get on with it --
DR. KRESS: Yes.
DR. POWERS: -- and not hang up on this difficulty of
projection.
DR. KRESS: Right; and I, of course, don't know the real
answer.
DR. BONACA: One thing I'd like to point out: is something
happening there that you need to at least have a PRA person during the
shutdown to review the changing configuration that takes place? Because
the configuration change is clearly due to emerging issues. I mean, you
have some component that else and is not supposed to; you have to change
the schedule. So that change takes place all the time, and on the other
hand, there is, you know, all nine evaluations that take place now, and
it is more and more frequent. I'm just saying that because it affects
really the, you know, core damage frequency.
DR. POWERS: I think what you're saying is fear not, Tom
Kress, that feasible configurations that could potentially have high
risk won't ever occur, because we've got a guard watching the shop.
DR. BONACA: Where they're using that, yes.
MR. BARTON: That's right.
DR. BONACA: That is correct.
DR. KRESS: There may be some credibility to that.
DR. POWERS: I mean, what they are saying is your tail from
your Monte Carlo distribution can't be too big. I mean, there's always
the chance that the guy fell asleep on the job and didn't notice, or
somebody fooled him or something like that, so it couldn't go to zero
out there. You can't clip the distribution. It's just that they're not
very big out there.
MR. BARTON: That's right.
DR. KRESS: That's right; yes, it would be nice to include
all those tails in your decision process.
DR. POWERS: Sure, because your decision process is very
likely to be dictated by the tails no matter how small they are.
DR. KRESS: Exactly.
DR. BONACA: By the way, on the positive perspective, once
the operators -- I mean, the guy in charge of operations -- gets used to
having the PRA person helping him, it's incredible how they want that
person there. I mean, it's one of the most wanted resources right now.
DR. POWERS: One of the joys of visiting plants both at
Brown's Ferry and at Susquehanna, we got exactly that word in very
tangible fashion; that indeed, this fellow, he who knows is consulted
regularly and enthusiastically by people.
DR. BONACA: Yes.
DR. POWERS: It's just flat real helpful to them, and they
appreciate that help.
DR. BONACA: And there feedback is actually lack of sleep.
DR. POWERS: Yes.
DR. BONACA: Continuous sleep --
DR. POWERS: Yes.
DR. BONACA: -- cycle.
DR. APOSTOLAKIS: But the point is, though, that this
process may itself fail in controlling these very risky configurations,
and if I am to do a predictive analysis, I have to have somehow an
allowance for that.
DR. POWERS: That's the long tails.
DR. APOSTOLAKIS: And I would -- I think this is a most
serious issue here of model uncertainty.
MS. DROUIN: Absolutely; I was just going to say that.
DR. APOSTOLAKIS: And it's comparable, I would say, to the
big problem they have on the other side of the house, where they really
don't know whether 6,000 years from now, they will have transport of
radionuclides through porous media or something else. I mean, once
you've determined the medium, they know what to do. They have very good
codes. But we don't know the medium, and that's an uncertainty that I
don't know can ever be resolved, and it seems to me we are going that
way now here, even though --
DR. POWERS: I think that's just not the case.
DR. APOSTOLAKIS: Oh, I think it is.
DR. POWERS: That just what Miles says is you've got a
watchdog on this; yes, he can fail, but it is not like every conceivable
configuration is equally likely to occur.
DR. APOSTOLAKIS: No, no, but the media there are not
equally likely either. But it makes a hell of a difference in the
calculations.
DR. KRESS: But my point is you can use past experience over
the fleet of plants to develop some probability or likelihood of
occurrence of a given configuration. And that is a Bayesian --
DR. APOSTOLAKIS: It wasn't a good Bayesian scale.
DR. KRESS: Oh, okay.
DR. SEALE: Not only that, but you also have the fact that
there are certain things that stack the deck, and an emergency shutdown,
where you have the highest possible after heat burden, and perhaps the
cause of the shutdown is the -- is multiple system -- mitigating system
failure and so on, those things clearly have to be in your assessment.
Now, those are the things where you really have the problem.
DR. APOSTOLAKIS: Yes.
DR. BONACA: One thing that is important is that there has
been, in fact, now more and more of a separation between strategic
issues and tactical issues during the shutdowns, which is -- a strategic
issue, I mean, there are certain configurations that you want to stay
away from, and that's really -- so you plan for it, and you don't get
into those.
DR. SEALE: Yes.
DR. BONACA: Or you try to really or to isolate them, like,
for example, at South Texas, you know, the issue of focusing all of
their resources on a midloop operation. And there are issues that are
considered tactical measures, and that's really where the PRA analyst is
involved in evaluating those changes.
DR. APOSTOLAKIS: But is every utility doing this?
MR. BARTON: Just about that I know of.
DR. BONACA: Now more and more.
DR. SHACK: But still, I mean, it's not the planned
configuration that's likely to cause the problem.
DR. APOSTOLAKIS: Right.
DR. SHACK: I mean, you know, if you were in the planned
configuration, things are probably quite reasonable. It's when the guy
--
MR. BARTON: It's the human error.
DR. APOSTOLAKIS: It's the human error.
MR. BARTON: Yes.
DR. APOSTOLAKIS: We have examples up there right now.
DR. SEALE: Yes, but you don't want to run all the safety
systems off the same bus.
DR. SHACK: I mean, it's not that you have to look at it
historically the outage history; you have to somehow estimate that
chance of getting those other configurations.
DR. WALLIS: If you'd have a sort of a picture, which would
say here's the core damage probability versus time; it goes along like
this in normal operation, and when you get into certain configurations,
it goes around. It looks like sort of the --
MR. CUNNINGHAM: Right.
DR. WALLIS: -- the silhouette of New York or something. It
has bumps and so on.
[Laughter.]
DR. WALLIS: And you evaluate all of those configurations,
and then, you've got to look at the probability of getting into those
configurations.
MR. CUNNINGHAM: That's right.
DR. WALLIS: If you had a figure like this, it would sort of
explain it all, and it might help the commission better to understand,
at least to talk about it; it's much more difficult.
MR. CUNNINGHAM: We have such figures. We don't have
anything here; but that's --
DR. APOSTOLAKIS: I don't think we need to discuss six; go
to seven.
MS. DROUIN: Okay.
DR. KRESS: George, the other approach might be to recognize
that if the -- if indeed it's true that the risk to low-power shutdown
is only equivalent to power operation, that's like instead of 1 x 104,
you're at 2 x 104, and you might be able to accommodate that range of
variability in your risk-informed measurement without always having to
calculate low-power and shutdown risk.
DR. POWERS: Well, I think we do, but I think that's not --
DR. KRESS: Yes.
DR. POWERS: -- that's not the issue here. And I think this
is a big mistake people make is yes, if -- supposedly, it's comparable.
Suppose it's twice? You have increased the CDF by 30 percent, 200
percent? I didn't believe the number that accurately to begin with. It
makes no difference to me. Where I think it makes a difference is when
we use these results to then decide on how do we categorize equipment?
How do we categorize --
DR. KRESS: Risk importance of equipment.
DR. POWERS: -- risk importance of things?
DR. KRESS: I think you're right.
DR. POWERS: And the likelihood that you're going to make
decisions that says ah, this system is not so very important, and I will
give it less attention in my maintenance program when it's absolutely
crucial in shutdown, during normal operational events; it's nothing --
DR. KRESS: I would fully agree with that.
DR. POWERS: That's the danger --
DR. KRESS: Yes.
DR. POWERS: -- that you have as we move toward using risk
as a guidance on how we treat these plants. Right now, I'm not very
concerned, because I don't think that risk guidance is so powerful. But
as we progress on, I think it becomes more and more, and for every
system that you say gets a lot of attention, there are 20 that are going
to get very little attention.
DR. UHRIG: But it is being used, but it is being used
today. I was told by the licensing manager of a utility that they had
used their risk meter to determine that it was better to take an
emergency diesel generator out of service while they were still at power
than to have it out during the outage.
Sure.
DR. POWERS: Ipso facto true, yes.
MR. BARTON: So?
DR. POWERS: I mean, it's very -- there are lots of these
things that are surprising, and you say very obvious once I think about
it, but you never think about it until you have a tool that's --
DR. KRESS: Well, once again, if you knew -- had some
estimate of the likelihood of given configurations based on past
experience, you could incorporate that into importance measures pretty
easily without having --
DR. APOSTOLAKIS: Right, if you had that.
DR. KRESS: If you had that.
DR. APOSTOLAKIS: Yes.
DR. KRESS: Which is a database which I think is needed, and
in fact, I think we said that in one of our letter.
DR. APOSTOLAKIS: Okay, Erasmia.
MS. LOIS: I guess the point of this slide is that the
industry has developed a capability to -- do you want me to --
DR. APOSTOLAKIS: I suspect there is going to be more
discussion later. I think we are convinced this is true.
MS. LOIS: Okay.
DR. APOSTOLAKIS: Let's go to the next.
MS. LOIS: Okay.
DR. APOSTOLAKIS: Unless you want to make a big point; yes,
that's something that we should discuss.
MS. LOIS: With this chart here, we try to put in some kind
of a order or fashion what we need, what are our needs and how we can go
about them, and the top box shows the values, applications that we could
use, where we could use information or risk insights from low-power
shutdown status, and given that there is a variety of needs for the NRC,
I guess what we need is to have insights regarding the significance of
shutdown risk, and that encompasses not just planned outages but the
various other modes: transition risk, unplanned outages, et cetera.
The insights regarding the contributors of shutdown risk,
would it be the plant design? For example, we do have an indication
that fire makes a difference, during shutdown makes a difference if --
regarding from your plant design. And the significance of the different
plant configurations and the discussion we just had within planned,
specific mode, shutdown mode, how many different planned operational
states and which ones may be more important to be analyzed.
DR. APOSTOLAKIS: This doesn't seem, though, to be an
approach for determining what you need. It really is a first step in
that approach, is it not?
MR. CUNNINGHAM: Yes, yes, it's the first step.
DR. APOSTOLAKIS: These are the needs, these are the
ultimate uses.
MS. LOIS: This is what you would like to have. You would
like to have --
DR. APOSTOLAKIS: Tools that would allow you to do this,
yes.
MR. CUNNINGHAM: Right. This is the first step in the
process.
DR. APOSTOLAKIS: Right.
MR. CUNNINGHAM: A first step in the process.
DR. APOSTOLAKIS: Right.
MS. LOIS: And in order to get this information, which if we
have, we can use in the regulatory applications, we can either go to
using the quantitative PRA model or a qualitative approach.
And if we look at the PRA plant specific we call it --
DR. APOSTOLAKIS: This is for shutdown mode?
MS. LOIS: For shutdown.
MR. CUNNINGHAM: Yes.
MS. LOIS: Now, this chart is trying to capture the idea
what is the current status of the art and what we need to do, and if you
look at the -- from a scope perspective, which encompasses different
modes, different planned operational steps, et cetera, we think that we
need methods for doing screening analysis.
DR. APOSTOLAKIS: Is it really a screening? Or what Dr.
Kress mentioned, you know, maybe trying to get the likelihood that the
POSes will occur? I mean, screening implies that you are leaving things
out. That's not --
MS. DROUIN: I think it's a mixture of both, because it's a
screening -- here, we're saying your plant operational states but also
your different operational modes. Do you need to go out and model every
mode? Or is there some kind of screening technique that you could use
to say, well, I don't really have to go and model this particular mode;
is there some way to look at some of these things that haven't been done
to justify whether or not I really need to do them?
DR. KRESS: That's probably like cutting off sequences at a
certain frequency level.
DR. APOSTOLAKIS: Yes, but the big difference is that there,
you are doing it after you have quantified.
DR. KRESS: I know.
DR. APOSTOLAKIS: And here, it's a priori, but on the other
hand, I think what Mary says makes sense, I mean; you don't have to
quantify everything before you decide that something is unimportant.
DR. KRESS: That's right.
DR. POWERS: I think that it would be more comfortable
deciding when I didn't need to quantify after I had quantified.
DR. KRESS: Yes.
DR. POWERS: At least once.
DR. KRESS: That's more comfortable to me.
[Laughter.]
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: As an example, the screening analysis that
was done for the NRC studies a few years ago was a semi-quantitative
evaluation of all of the planned operating states, and in that
circumstance, it turned out that there was a few of them using that
method that jumped way up, and there were a bunch of others that seemed
to be trivial by comparison, and that led us to focus on it.
DR. POWERS: Of course, the fundamental problem is one of
those trivial ones is the one where the Wolf Creek draindown event was
right in the middle of, and it has a huge ASP number associated with it.
The problem, I think, with screening now is I don't
understand enough about shutdown to come in and say yes, we really --
POS4 is a no, never mind and 5 is a big deal.
MR. CUNNINGHAM: Yes; there's an element, I guess, in this
of what do you do first? That's valuable. It's also valuable to be
working on the issue we've talked about a couple of times in the Wolf
Creek example: human reliability; human performance in these is
probably less certain, and so, you may have to -- another topic we get
to is our ability to model human performance in these.
DR. APOSTOLAKIS: Why don't you replace screening by
prioritizing, huh? Would that do it? That's what Mark just said.
MS. DROUIN: To me, screening -- I guess my interpretation
of the word screening, that includes that, but we can bring that out
explicitly.
DR. APOSTOLAKIS: No, because prioritizing means I will do
this first, and then, I will look at the others. Screening implies, you
know, I will only look at these, because the others have been eliminated
due to some reason.
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: So prioritizing makes more sense to me.
DR. WALLIS: For each POS, there is a corresponding human
operational state. There are certain demands being made on the human --
MR. CUNNINGHAM: Yes.
DR. WALLIS: -- which can be defined, I would think.
MR. CUNNINGHAM: Yes.
DR. WALLIS: But there are clearly certain times when
they're under more pressure --
MR. CUNNINGHAM: Yes.
DR. WALLIS: -- to make decisions.
MR. CUNNINGHAM: That is correct; and so, that, in a
screening or a prioritization, we should be able to reflect that.
DR. WALLIS: Right.
MR. CUNNINGHAM: I suspect one of our issues on this list is
that our ability to model that human performance isn't up to snuff
relative to other things.
DR. WALLIS: But it's not hopeless. I mean, you can start
to describe what it is they have to do.
MR. CUNNINGHAM: Right.
MS. DROUIN: But I don't think, George, just a
prioritization. I think there is a screening aspect to it.
DR. APOSTOLAKIS: I understand that, but that --
MS. DROUIN: Okay.
DR. APOSTOLAKIS: -- may come a little later for purposes of
--
MS. DROUIN: And that's why, you know, if you look on the
slide coupled with it, you know, is we have the words there further
study.
DR. APOSTOLAKIS: I understand that.
MS. DROUIN: Which is getting to some of Dana's concerns.
MS. LOIS: And we do further study before we develop the
guidance in terms of what we do here first.
DR. APOSTOLAKIS: Go on to initiating events.
MS. LOIS: Regarding initiating events, our feeling is that
the methods for identifying initiating events are good enough, but we do
need to develop guidance to tell people how they will do it for
low-power and shutdown analysis.
DR. APOSTOLAKIS: Well, what do you mean by initiating
events? Do you mean the standard nuclear stuff like loss of coolant
capability and that stuff?
MS. LOIS: And those that are shutdown specific; for
example, loss of shutdown cooling, which is not --
DR. APOSTOLAKIS: Right, right; but I guess I disagree with
that conclusion, because if you look again at Wolf Creek, there was a
series of human actions that led to the draindown event. So are the
methods for studying these actions mature enough for us to say that we
don't need to do any work on this? And during those modes, it seems to
me these are much more important than during power?
MR. CUNNINGHAM: Getting at the issue of human-induced --
DR. APOSTOLAKIS: Yes, initiators.
MR. CUNNINGHAM: -- initiators and things.
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: And that, I don't think we had that in mind
when we said things were sufficient. That may be something that's --
DR. APOSTOLAKIS: Yes, see, that's why I'm bringing it up.
MS. DROUIN: To me, the human part is under the HRA part, so
I think it's just how we have divided up this pie versus how you would.
DR. APOSTOLAKIS: So what you're saying, Mary -- what you're
saying is that yes, there is a problem with the humans but --
MS. DROUIN: Absolutely.
DR. APOSTOLAKIS: -- all they can do is create one of the
initiating events that I have already identified.
MR. CUNNINGHAM: Right.
MS. DROUIN: Yes.
MR. CUNNINGHAM: Right.
DR. APOSTOLAKIS: And I tend to agree with that, but
somehow, that has to become clear.
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: Because I think for those modes, the
probability that humans --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: -- will do that is much higher than --
MS. DROUIN: Absolutely, absolutely.
DR. APOSTOLAKIS: Okay; then, we agree.
MS. DROUIN: Absolutely.
DR. KRESS: Yes, but then, you've got -- if you have an
initiating event identified, you have to have a frequency associated
with it, and the tendency would be to use the old frequency --
DR. APOSTOLAKIS: Well, they just say I have --
MS. DROUIN: And we did that on purpose. This was not the
quantification part.
MS. LOIS: So, then, the quantification part is the next
bullet, where we think that we need to do a further study to really
comprehend the evolution of events during shutdown.
DR. APOSTOLAKIS: Now, methods for sequence delineation
sufficient; now, again, I'm not saying that I disagree, but I'm trying
to understand what you mean. It seems to me that the major new element
here, in addition to what we said earlier about the configurations, that
is not present during power operations is the fact that you have a
decaying decay heat.
MS. DROUIN: Right.
DR. APOSTOLAKIS: And things are much more time dependent,
so when we identify sequences, are you talking about time dependent
sequences? Are we really in a position to claim that we are doing this
well enough?
MS. DROUIN: When we talk about the methods for sequence
delineation, this one, you know, was hard to capture the thought on a
slide without having to write several paragraphs.
DR. APOSTOLAKIS: So give us a paragraph.
MS. DROUIN: We give you a paragraph. The accident sequence
development really, you know, in my mind kind of almost gets to the
heart and soul of your PRA, because it defines your accidents. In order
to define those accidents, you know, understanding, you know, what your
success criteria is, how the plant responds to those events, once you
have that, we have the software, so to speak, of how to delineate that.
It's a very narrow part when we say that part's sufficient.
DR. APOSTOLAKIS: So what you mean then -- let me put it in
different words and see if I understand it --
MS. DROUIN: Okay.
DR. APOSTOLAKIS: -- is that you can identify the sequence
in the sense; in other words, I have to lose coolant.
MS. DROUIN: Yes.
DR. APOSTOLAKIS: I have to do this; I have to do that, and
then, I have damage.
MS. DROUIN: Right.
DR. APOSTOLAKIS: But how these things occur in time is not
something you can do, right?
MS. DROUIN: That is correct.
MS. LOIS: And that's the next bullet.
MS. DROUIN: I'm just saying we don't need to go out and be
developing the --
DR. APOSTOLAKIS: Not just the response.
MS. DROUIN: -- software to do it.
DR. APOSTOLAKIS: I don't care about the software. Let's
not talk about software right now. It's the basic understanding. But
the question is, though, whether putting time on that sequence, on the
events that I have already identified, the minimal cut set, in other
words, is just a simple thing to do, or it's much more involved, and in
fact, the presence of time may create new ways that the sequence may
evolve. What you thought was a sequence, now maybe somebody intervenes
or something happens, and all of a sudden, you have 10 different ways it
can go.
I don't know; I don't know, and I'm sure --
DR. KRESS: I think you're absolutely right, George, and, in
fact, I think you have to treat power as an uncertainty parameter in
your calculation and incorporate it in that mode somewhat.
DR. APOSTOLAKIS: So maybe -- yes, I understand the need for
putting a paragraph there, but somehow, you have to put a word or two to
explain that there is this caveat there that we don't know what happens
when things start evolving in time.
MS. DROUIN: Right; and that's what those next two bullets
were trying to get to; you know, your understanding of your accident
response is insufficient.
DR. APOSTOLAKIS: Yes.
DR. POWERS: The next one that you have, success criteria,
one that particularly fascinates me in low-power shutdown, how do you
know you've won the game?
DR. KRESS: It will be a function of time, too. That's the
other thing: success criteria in this mode --
DR. APOSTOLAKIS: Right.
DR. KRESS: -- will be a function of time.
DR. APOSTOLAKIS: The response rate is a function of time.
DR. KRESS: Yes.
DR. APOSTOLAKIS: Everything is a function of time.
DR. KRESS: And that's going to be an important
consideration when you try to put together this.
MR. CUNNINGHAM: I think at the subcommittee meeting, we
talked about the issue of modeling transition risk. That's an issue, I
think, that remains open --
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: -- in the documents where the dynamic
nature of the evolutions and the timing may become even more critical in
some of those transitions. People, I think, tend to agree we don't
model that very well at all.
DR. POWERS: Is anyone coming along and making this a
simpler job?
[Laughter.]
DR. POWERS: Everything we talk about just makes it more or
more complicated.
DR. WALLIS: Yes; I'd like to picture this. I mean, you
could do this at a very crude level, and it may be that there are a few
things that dominate, and that's all you need, or you could do it with
tremendous detail, endless job, you know. How do you decide where to be
in the spectrum?
MR. CUNNINGHAM: That's the screening part that we had
talked about earlier that we need to develop.
DR. WALLIS: I don't think it's so easy. I think you just
have to do it. You have to do it at different levels and see what the
payoff is.
MR. CUNNINGHAM: Again, we have done it at a couple of
levels. I don't know that we'd use those in examples as the way to do
it in the future, but it gives us some perspective on it.
DR. WALLIS: Does this tell you that you need lots of
detail, or does it tell you that you can get rid of a lot by screening?
DR. APOSTOLAKIS: I would offer a suggestion here that may
help also Dr. Wallis. I think you need to strengthen this slide by
changing the third column development of needs and actually identifying
some of the issues we have been discussing; for example, when you say
methods for sequence delineation sufficient, then, under developmental
needs, you might say something to the effect of that the time element is
something that is needed to be incorporated and then maybe address the
question from Dr. Wallis.
I mean, how are we going to do that? Maybe we'll start with
a static sequence then do some variations in time, and as we go, we are
getting insights, and somewhere there, we draw the line and say we
understand, because in power operations, for example, when you lose
off-site power, right, that's one of the few cases where time comes into
the picture: are you going to recover off-site power before something
bad happens, like a station blackout? If you have the diesels back and
so on.
But the treatment of time is trivial there.
DR. BONACA: Yes.
DR. APOSTOLAKIS: It's a simple calculation that this time
is shorter than this other time.
DR. BONACA: Yes.
DR. APOSTOLAKIS: And people have found this to be good
enough.
So I think some sort of guidance under the third column,
which I think will attract a lot of attention from the commission, is
needed. It will help your case. And then, you go back to the screening
issue. We discussed it a lot. Maybe you use different words, and so
what do we need to do? Because right now, I don't think the only thing
that is needed when it comes to sequence delineation is guidance. I
mean, what is the HRA part? Right there, you may want to say except
that when the humans initiate this, we really don't know how to do it.
So now, people very clearly see what the needs are and why
research needs to be done and most importantly, why they have to give
you the money to do it.
[Laughter.]
DR. WALLIS: And what's the payoff?
DR. APOSTOLAKIS: If you just say I need to do further
study, given what the commissioners have stated in public --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: I mean, you're not going to get it.
DR. WALLIS: Give us some numbers very early on.
MS. DROUIN: I just want to say in a little bit of defense,
we're not turning in the slides to the commission.
DR. APOSTOLAKIS: No, but when you write the report, too --
MS. DROUIN: When we write the report --
DR. APOSTOLAKIS: -- bear in mind these comments.
MS. DROUIN: Absolutely, and I agree with that.
DR. APOSTOLAKIS: But won't you make a presentation to the
commission?
MR. CUNNINGHAM: Not necessarily.
MS. DROUIN: No.
DR. APOSTOLAKIS: Okay; if you don't, you don't, but what
I'm saying is that this third column could be even stronger.
MS. DROUIN: And I agree with you. I mean, the details,
hopefully, and I do appreciate the comments, and we do have some
details, and as we write it, we will ensure that those details are
there.
DR. WALLIS: I'm looking at the bottom line. You gave us
the -- grand gulf 0.5, sorry 0.1. Now, really, this is gone to 0.5 plus
or minus 0.3 or something.
Now, are you doing work in order to try to find out if it's
0.6 or 0.51 or 0.513? What kind of precision do you want as an answer
to this problem? You probably already know that it's not a trivial
problem; I mean, the numbers like 0.5, 0.41 indicate it's significant,
but how much more do you need to know before someone makes some
decision?
MS. LOIS: Can I say something?
DR. APOSTOLAKIS: Yes.
MS. LOIS: I guess the -- one of our concerns is that we had
done studies on a limited number of plants that may not be, a)
representative design --
DR. APOSTOLAKIS: End states.
MS. LOIS: -- even for BWR and PWR, and the kinds of
applications we showed, which is 1.174, Part 50, these are applications
that are across plants, and therefore, we need information from all
different -- as a minimum, all different types of plants, and therefore
--
DR. APOSTOLAKIS: And modes of operation.
MS. LOIS: Yes; these are just starting points in our
understanding.
Now, what we tried to capture here is that we take advantage
of other studies done domestically and internationally, and therefore,
where we feel confident that we can develop guidance for our review or
for the industry application, we do that, and where we need -- that we
need to do additional work to either clarify or explore more, these are
the areas where we're going to concentrate from a research point of
view.
DR. WALLIS: Well, you see, I have problems with things like
developmental needs further study without some metric of where you're
trying to get to. And that's why I'm encouraging you to be more
specific about it.
DR. APOSTOLAKIS: It's not just the metric, though; I think
it's also the why. If I look at the experience --
MS. LOIS: Yes.
DR. APOSTOLAKIS: -- and I see people creating all these
parts and creating initiating events, I want to understand what that
means.
DR. WALLIS: Well, what questions are you asking, and how
much precision do you need in your answers?
DR. APOSTOLAKIS: Right, right; it's not just the
risk-informed -- well, actually, risk-informed is --
DR. WALLIS: Yes.
DR. KRESS: And I hate to bring up the subject, but I don't
see you mentioning uncertainties anywhere, and I suspect if shutdown and
low-power risks are the same magnitude as full power --
MS. DROUIN: Can we get to the next page, Tom?
DR. APOSTOLAKIS: No, no.
DR. KRESS: -- I think it will dominate the uncertainties.
DR. APOSTOLAKIS: Let's go back to 9. Let's go back to 9.
MS. DROUIN: We get into uncertainties on the next page. We
can stay with 9.
DR. APOSTOLAKIS: We wouldn't dare to get into that without
--
[Laughter.]
DR. APOSTOLAKIS: Systems analysis; methods for system
evaluation sufficient. I don't know if -- does everyone agree?
DR. KRESS: I see it.
DR. APOSTOLAKIS: Don't we have the same problem with the
sequences here, that things may be time-dependent? And most
importantly, what you mean by systems analysis is if I give you the
system, then, you can analyze it.
MS. DROUIN: Right.
DR. APOSTOLAKIS: But given the problems with the
configurations we're talking about, I don't know how sufficient the
methods are. I mean, if I tell you yes, this is the system you will
analyze, but now, you know, three and a half years from now, maybe this
component will not be there, or it will be out for some reason. Then,
are the methods sufficient? No; we're coming back to your idea about
likelihoods, Monte Carlo, Las Vegas, you know.
[Laughter.]
DR. APOSTOLAKIS: So I'm not sure that -- what I'm saying
there is that this is too strong.
DR. WALLIS: Those methods should apply across the board.
MS. DROUIN: The methods should apply.
DR. APOSTOLAKIS: Oh, Newton's Law; of course, excuse me,
conservation of mass, oh, yes; I don't think we need new principles.
What do we need by methods?
MR. CUNNINGHAM: Okay; this was more, I think, in the vein
of your first part of it, which is given a configuration, can we analyze
--
DR. APOSTOLAKIS: Exactly.
MR. CUNNINGHAM: -- a complicated system with fault trees
and things like this. That was the intent of this.
MS. DROUIN: Right.
MR. CUNNINGHAM: But you're right.
DR. APOSTOLAKIS: Yes; it's Newton's Law. I mean, it does
apply.
DR. KRESS: I just don't think you can take a -- any kind of
current PRA we have and do a -- and use the same method that it uses to
do a low-power shutdown risk. I don't see it.
DR. APOSTOLAKIS: I would put on the third column there, so
that you wouldn't get comments like that, the configuration issue needs
to be resolved and the time issue.
MS. DROUIN: Absolutely.
DR. APOSTOLAKIS: What if I have to do fault tree analysis
that is time-dependent? I'm not saying we cannot do it, but it's not
something we do routinely, right?
MR. CUNNINGHAM: In effect, what you're defining is a couple
of overarching issues, if you will --
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: -- to use that term.
DR. APOSTOLAKIS: That's right; you like that there, right?
MR. CUNNINGHAM: I like that word, yes.
DR. APOSTOLAKIS: Overarching.
MR. CUNNINGHAM: Overarching principles.
DR. APOSTOLAKIS: Yes.
DR. POWERS: They can't be overarching principles so --
DR. APOSTOLAKIS: At least not in here.
[Laughter.]
MS. DROUIN: But I think that's a major point, because
that's how -- you know, that's one that we didn't explicitly put on that
slide, and, you know, shame on us, but, you know, the time dependent one
to me is not a systems analysis question, because it cuts across the
whole PRA. It's an issue in and of itself.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: Sure; but, I mean, systems analysis is a
subset.
MR. CUNNINGHAM: Yes; we haven't captured that.
DR. APOSTOLAKIS: I mean, anyway, I think we agree.
MS. DROUIN: Yes.
DR. APOSTOLAKIS: It's just a matter of what to put in the
third column.
MR. CUNNINGHAM: Yes.
MS. DROUIN: Yes.
DR. APOSTOLAKIS: If you put those key words
there --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: Then, you are taking away a lot of the
steam from my questions.
MR. CUNNINGHAM: Okay.
DR. POWERS: It seems to me that we want to be cautious
about overindulging on this. If I said, okay, I'm going to analyze a
plant during normal operation, you would say oh, no, no, no, because at
the beginning and ending of the period, you are going to turn it down,
and some part of that is going to be a part of your PRA assessment, and
you've got to take into account the time dependence of things and
complexity there. It's clear that we don't do that, that exactly, when
we do an operational state. We don't worry about the fact that on the
last day of power operations, the guy is sitting there thinking -- the
operator is sitting there thinking tomorrow morning, we're going to shut
down, and so, I've got to do a lot of things. We don't take a -- it's a
nick.
And so, I think it's easy to overdramatize the magnitude of
the problems, especially when you haven't done it.
MR. CUNNINGHAM: Yes.
MS. LOIS: On the other hand, we do want to make the point
that we're building on existing technology, and that's where we came off
to say sufficient guidance development, at least from the perspective of
the principles or the mechanisms to be used here. That's the intent of
this.
DR. APOSTOLAKIS: Anyway, these comments are not for the
general --
MS. LOIS: Is this clear, or should I move to the next?
DR. APOSTOLAKIS: I think you've got the spirit now, didn't
you?
MS. LOIS: Yes.
DR. APOSTOLAKIS: We don't need to discuss every single one
--
MR. CUNNINGHAM: No.
DR. APOSTOLAKIS: -- unless a member has a comment.
What is the word uncertainty?
DR. KRESS: I'd like to read ahead.
DR. POWERS: No, I was particularly excited in one of your
draft documents, and I don't know, I can't tell you which version I was
last looking at, but I seem to have gotten three or four of these, each
one of which is different. But, you know, fire is not something that we
address during normal operations as a cause or initiating event from
normal operation PRAs probably because it isn't all that huge of a
contributor, maybe.
MS. DROUIN: Some people would argue that.
DR. POWERS: But, now, low-power shutdown, it is a -- and is
the evidence such that one would say that the risk of having a fire
higher during low-power and shutdown operations? I mean, a significant
fire; is it higher, or is it lower? Or you don't know?
MS. DROUIN: Depending on what studies, you get different
answers, and that was just an oversight on our part not to have it on
the slide, because we do have it in the report as an issue to be looked
at for low-power shutdown. You cannot exclude fire or flood.
MR. CUNNINGHAM: Again, I might go back and say this may not
-- this is a place where shutdown risk analyses may be ahead of full
power in the sense that fire is kind of built in, at least from our
studies, as an internal initiator, and it's just there, and it's not
treated in some special way, and it may be that, at least if I had
anything to do with it, I would argue that the full power ought to do it
too; that just routinely, a full power analysis ought to have a fire
initiator in it.
DR. APOSTOLAKIS: Now, what if an earthquake hits, and you
have -- you're in midloop operations? What does that do to you?
DR. KRESS: That happens at such a frequency that you can
forget it.
DR. APOSTOLAKIS: Oh, it's one of those incredible
hypothetical events.
DR. KRESS: Yes.
DR. APOSTOLAKIS: So what saves you is the fraction of time
you are in that mode.
DR. KRESS: Yes, sir.
DR. POWERS: So what you're telling me, most -- and a
risk-smart activity is to shorten down the --
DR. KRESS: As short as you can, yes.
[Laughter.]
DR. APOSTOLAKIS: Rush it.
DR. KRESS: Yes, rush it.
DR. APOSTOLAKIS: That's a pretty good argument.
DR. KRESS: Yes; somewhere, I think it peaks and goes the
other way.
MR. CUNNINGHAM: There's an optimum there someplace, isn't
there?
DR. WALLIS: Well, that's the question: are you going to
develop methods which help determine optimal like that? And you can say
after awhile that they've gone too far in cutting down the time of
shutdown?
DR. POWERS: I think one of the biggest uses I can imagine
of a capability to routinely do low-power shutdown is going to be to
make tradeoff decisions on whether to do online maintenance or wait
until the shutdown.
DR. APOSTOLAKIS: Yes.
DR. POWERS: I think that's going to be the biggest utility;
I bet you we get some consciousness-raising experiences once we have a
good capability.
DR. APOSTOLAKIS: Yes; go ahead. What are we talking about?
MS. LOIS: This is the qualitative or non-PRA approach, and
I guess what we try to say here is that the industry has been using a
qualitative approach in addition to 1.174 provides for qualitative
arguments, and therefore, we're going to look into that.
Regarding the NUMARC guidelines, the problem we say is the
inconsistency, and therefore, we feel that we have to develop guidance;
and also, we have -- we are thinking that it may not be the only
qualitative approach, and therefore, we have to look into other types of
qualitative approaches.
DR. APOSTOLAKIS: But I mean your goal here is to argue that
we need to do more on the risk assessment part of these modes. Why do
you care about configuration management as practiced by the utilities
now?
MR. CUNNINGHAM: Again, we don't -- well, there is an
indirect issue there, but in terms of assuring that they do it safely,
but our issue here is that they have these methods out there, and
they're using them, and if there's a way we can find to use those
methods with some adaptation in issues that are important to us, then,
there is a benefit to the fact that they're already out there, and we
don't have a new infrastructure to build or anything.
DR. APOSTOLAKIS: Their objectives are different.
MR. CUNNINGHAM: Their objectives are different, correct.
DR. APOSTOLAKIS: You know, they're into defense-in-depth,
and they want to make sure that they have alternate ways of doing
things.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: That may help you a little bit in the
effort to bound the possible configurations.
DR. KRESS: And not only that; they input the configurations
by hand.
DR. APOSTOLAKIS: Yes.
DR. KRESS: The known configurations.
DR. APOSTOLAKIS: Okay; so, in that sense, this is not that
interesting.
DR. SIEBER: Could you explain a little bit what you mean by
the inconsistently applied?
MS. LOIS: On that one, at least according to the
information that we gathered during the summer, we were told that the
utilities have the capability to determine themselves what are the reds,
the oranges --
DR. SIEBER: Right.
MS. LOIS: -- and the greens.
DR. SIEBER: That's true.
MS. LOIS: And also, they judge themselves whether or not
they are conforming with the guidelines. Therefore, there is no uniform
acceptance of what it does; when a red is red and how red it is. So I
guess we need to strengthen those areas.
DR. APOSTOLAKIS: That's not your job.
MR. CUNNINGHAM: That's right; our job is to -- right.
DR. APOSTOLAKIS: I mean, the NRC must do that. Are you
also considering the quality of these configuration management programs?
No.
MR. CUNNINGHAM: No, we're not. Again, the way we would get
at these issues is the inspectors have to go out there and --
MS. LOIS: Right.
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: -- watch what's going on, and our idea is
that you've got tools and --
DR. APOSTOLAKIS: That's right.
MR. CUNNINGHAM: -- mechanisms in place to help the
inspectors.
DR. APOSTOLAKIS: I think we should go to number 12.
[Pause.]
MR. CUNNINGHAM: What we have tried to do in this is kind of
summarize the things you've heard or would have heard in the various
slides to try and lay out which do we think should we tackle in FY 2000;
what do we think in the general sense we want to attack in 2001 and
2002.
DR. APOSTOLAKIS: I guess I don't understand what guidance
development means.
MS. DROUIN: I think that one of the things I was saying to
Mark is that one of the problems here is, I think, how we are defining
and using the term method and the term guidance versus yours, and I
think a lot of your definition of method is we're putting that under
guidance.
DR. APOSTOLAKIS: Ah.
MS. DROUIN: And so, we're going to have to be very careful
when we go back to make it very clear what we mean by these terms.
DR. APOSTOLAKIS: Let me tell you what I would do if I had
the unpleasant task that you have.
[Laughter.]
DR. APOSTOLAKIS: I would say that in the immediate future,
near future, 2001, what's this 20?
MS. DROUIN: That's a typo. That should say 2000.
DR. APOSTOLAKIS: Oh, okay.
DR. POWERS: It's like a millistone.
[Laughter.]
MR. CUNNINGHAM: In a millistone, yes.
DR. APOSTOLAKIS: So the first thing that I would do, if I
were you -- bear with me -- is I would look at the two or three major
differences between these modes and power operations, the elements that
would make a big difference in the assessment that we have identified
here, the configurations, the time and the human error that leads to
initiating events.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: These are the ones I can think of right
now.
DR. KRESS: Those are the major three.
DR. APOSTOLAKIS: And I would spend a few months, several
months, thinking about what to do about them.
MS. DROUIN: Absolutely.
DR. KRESS: And with the time, I mean, the time dependent
configurations --
DR. APOSTOLAKIS: Yes.
DR. KRESS: -- and time dependent success criteria.
MS. DROUIN: And, George, on the slide --
DR. APOSTOLAKIS: A nice list of what the issues are; some
ideas of what to do. I would start with that, and I would be prepared,
in fact, to argue very strongly that unless I have some answer to these,
I really cannot do anything.
MS. DROUIN: And we are in total agreement, and that's what,
when you come down under analytical work, where we talk about further
studies, and you will see down there, you know, 2000, that's our first
thing we want to do.
DR. APOSTOLAKIS: Okay; I guess again, I suggest that
instead of using these general terms like analytical work, methods --
MR. CUNNINGHAM: Yes.
MS. DROUIN: Yes.
DR. APOSTOLAKIS: Go right to the heart of it and say this
is it, both in the report and in the view graph. You never know where
you are going to use it. And you identify those three areas or maybe
more; I don't know, and say that this is not simply a matter of taking
methods that are widely used now in this area. There are these
fundamental differences.
MR. CUNNINGHAM: Yes.
DR. KRESS: They have to be accommodated.
DR. APOSTOLAKIS: That somehow, we have to come up -- first
of all, understand better their impact as we
were --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: -- just saying.
MS. DROUIN: Right.
DR. APOSTOLAKIS: You know, it's not just in the sequence;
the success criteria may change and so on. And then, propose solutions
or research programs to resolve these issues, having in mind Dr. Wallis'
guidance there.
DR. WALLIS: Oh, I have some other things I want to say.
DR. APOSTOLAKIS: Why do I want to do this? How important
is it? I mean, we have the example from power operations with the
offsite power, a very trivial calculation; takes into account time. You
can do more sophisticated calculations, but you're not really changing
anything.
DR. WALLIS: I would like to see discussion of who is going
to use the results of this work; what kind of results they need; what do
they need first; which ones are most valuable; how precise your
arguments need to be and so on, and I don't see this. I see a very
interesting study, a very interesting analysis that's going to probably
be useful to somebody. But I don't see -- I think you need some input
from people who are going to use it.
What are the questions that people have to make decisions
for asking? How will you help them?
MR. CUNNINGHAM: We have a set of tables back in the office
that we didn't really present here that go into, okay, if an application
-- the people who are going to be reviewing license amendments under reg
guide 1.174 have a set of needs for this; what are they? The people who
are going to be risk-informing the special treatment requirements have a
set of needs for this, and we're working that problem. We just thought
if we put the table up here, we would never get past that page 1, if you
will. That has to be done clearly, and they are different.
DR. WALLIS: This is development. There's nothing
deliverable here that by some date, we're going to get some particular
thing, and that's going to be valuable, because we've done it.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: I think it's important to emphasize to the
commission that there are these three major elements here that make this
kind of analysis very different from the power operations analysis. I
mean, just to talk about analytical work --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: -- I mean, this is what we normally do.
But I think you should go straight there and say look: there is an
issue of timing; there is an issue of configuration; there is an issue
of human errors, and if you don't believe me, here is what has happened
in the last 20 years, and I think that, if you start with that, you can
also answer the earlier question from Dr. Wallis --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: -- that you are not really trying to see
whether the 0.5 is 0.53.
MS. DROUIN: Right.
MR. CUNNINGHAM: Right.
DR. APOSTOLAKIS: The concern is that by using relatively
crude methods that have not really addressed these three elements to our
satisfaction -- we don't know that -- we are getting this high number.
So what's going to happen if we do it more rigorously?
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: I don't know what's going to happen. The
0.5 may become 5; I have no idea whether it won't.
DR. KRESS: It won't, and along with that, these elements
that you talked about, these three, will change your perception of the
uncertainties.
DR. APOSTOLAKIS: Of course.
DR. KRESS: And somehow, when you just do the 1.174 process,
we talk about using mean values for the metrics. I know -- I'm not sure
mean values are all the -- I mean, the mean values are going to shift on
you if you're actually doing an uncertainty.
DR. APOSTOLAKIS: Well --
DR. KRESS: And I'm not sure whether the mean value is an
appropriate way to use uncertainties.
DR. APOSTOLAKIS: If you remember, there is a nice paragraph
that Dr. Perry wrote, as I remember, or maybe three paragraphs that
advises people how to handle model uncertainty in the context of 1.174.
If I take his advice -- well, this is the agency's advice -- I really
have to start playing sensitivity games.
DR. KRESS: Yes, yes.
DR. APOSTOLAKIS: Okay; tell me how you're going to do that
with all of these configurations.
DR. KRESS: That was the point I was trying to bring in:
the perception of how to do sensitivity uncertainty changes.
DR. POWERS: I don't think we can overestimate the magnitude
of effort to pursue your Monte Carlo techniques, because I am reminded
that the attraction of Monte Carlo techniques is it takes so few
samplings to get things to converge pretty well.
DR. KRESS: I cannot agree more.
DR. POWERS: And, in fact, just looking at people, they're
using Monte Carlo techniques to look at reaction kinetics for complex
systems.
DR. KRESS: Oh, yes.
DR. POWERS: And I'm stunned how completely you can get
those things to converge.
DR. APOSTOLAKIS: From which distributions would you sample
to do your Monte Carlo evaluations?
DR. KRESS: You'd have to have a database.
DR. APOSTOLAKIS: Your model uncertainty; that's the
fundamental question.
DR. POWERS: Another truism is not to overestimate the
difficulty of getting an adequate distribution for those, because the
distributions are about the same.
DR. APOSTOLAKIS: Well --
DR. POWERS: In the Monte Carlo sampling system, all
high-entropy distributions are about the same. It really doesn't make a
difference.
DR. KRESS: That is about right.
DR. APOSTOLAKIS: I'm skeptical.
Is there anything else that you ladies and gentlemen are
dying to tell us?
MR. CUNNINGHAM: To come back to the beginning --
DR. APOSTOLAKIS: Oh, yes; the letter.
MR. CUNNINGHAM: The letter. We -- the report that you have
in some form will go to the commission at the end of this month --
DR. APOSTOLAKIS: Right.
MR. CUNNINGHAM: -- with the paper. It will have the types
of recommendations we have here, probably a little more solidified and a
little bit more precise to reflect the types of discussions we've had
here. I'm presuming that this is a subject the committee will still be
interested in hearing about, so there may be more subcommittee meetings
coming up over the next X months at some interval that you can decide.
DR. APOSTOLAKIS: Now, when you send that document to the
commission, what action do you expect them to take?
MR. CUNNINGHAM: At this point, I don't think there's a
specific action. What I think the intention will be is that this is
where we're going in shutdown risk analysis. And then, we have to, in
part of the budget process, then, we'll have to revisit -- which will be
a separate action, we will have to revisit. And this is kind of laying
out our intentions that we will fund -- our intention is to fund this
work in 2001-2002 and things.
DR. POWERS: I'm wondering if the right vehicle for us to
comment on this at this early interim stage -- because I think it's so
difficult for us to comment on the document you actually send to the
commission -- I wonder if it isn't appropriate for us to comment on the
topical area via our research report.
MR. CUNNINGHAM: That's another way to do it, yes.
MS. DROUIN: When does your research report come out?
DR. POWERS: Well, we're going to --
DR. SIEBER: Tomorrow morning.
DR. POWERS: We are breaking a leg, and the back of
certainly the chairman and perhaps his members to have a draft by the
end of this meeting that you will admire over the holidays and finalize
in February, at the February meeting.
MR. CUNNINGHAM: I think that's a perfectly legitimate
vehicle to deal with this.
DR. APOSTOLAKIS: Or we can write a letter in February.
MR. CUNNINGHAM: You can write a separate letter.
DR. KRESS: It may get a little lost in the research report
if --
DR. APOSTOLAKIS: No, if the research report comes out in
February --
DR. KRESS: We can do both.
DR. APOSTOLAKIS: We could write a letter.
DR. KRESS: We could do both.
DR. APOSTOLAKIS: If you have your document, I will know
what I'm talking about.
DR. POWERS: You won't have time to do it then. I have the
February agenda fully booked.
DR. APOSTOLAKIS: Well, and I think the research report is
going to be a little crowded right now.
Anyway, I don't -- I mean, our preliminary assessment, the
committee will have to discuss it -- is that we cannot really write a
letter now, because we don't have the final document. So what vehicle
we're going to use to let the commission know our views is something to
be decided.
MR. CUNNINGHAM: Yes, that's fine.
DR. APOSTOLAKIS: So anything else on this subject by -- no?
MR. CUNNINGHAM: I don't think so.
DR. APOSTOLAKIS: Members?
MS. DROUIN: Once the commission paper comes out, you can
then write a letter.
DR. APOSTOLAKIS: Unfortunately, we are not meeting in
January.
MR. CUNNINGHAM: In January.
DR. APOSTOLAKIS: So, that's the thing. Yes, after you have
the document, of course, we will.
MR. MARKLEY: Mark, are you issuing this thing for public
comment as the draft NUREG?
MR. CUNNINGHAM: No, I don't think we had in mind that we
would put it out for public comment. Again, this is our plan at this
point for proceeding over the next few years.
DR. POWERS: Well, I think that -- I mean, we have other
vehicles to advise the commission than a formal report, and we may avail
of those; we may look at the final document and decide to do something
specific in February. I think it's -- I think the important thing is
just keep us informed.
MR. CUNNINGHAM: Okay.
DR. POWERS: And we'll do our best to keep you informed.
MR. CUNNINGHAM: Okay; again, at some point in the next few
months, it may be appropriate to have a subcommittee meeting where we
revisit and go into more detail the last few slides of this and then
invite Dr. Budnitz and/or somebody to come in and talk about the ANS
part.
DR. APOSTOLAKIS: Certainly.
DR. KRESS: Yes.
DR. APOSTOLAKIS: That will be a pleasure.
MR. CUNNINGHAM: One of the issues that we have here in
defining our work is how do we ensure that we're not duplicating what
the ANS is doing and taking advantage of what ANS is doing? And that
makes it a little difficult to shape our program, because the two of
them are being developed in parallel.
DR. POWERS: If I wanted to have a better explanation of
ORAM from your staff, who on your staff would I talk to?
MR. CUNNINGHAM: Erasmia.
DR. APOSTOLAKIS: Any other comments from anyone?
[No response.]
DR. APOSTOLAKIS: Hearing none, back to you, Mr. Chairman.
DR. POWERS: And I will recess us until 10 minutes after
1:00.
[Whereupon, at 12:10 p.m., the meeting was recessed, to
reconvene at 1:10 p.m., this same day.]. A F T E R N O O N S E S S I O N
[1:10 p.m.]
DR. POWERS: Let's come back into session.
Welcome, gentlemen. We're turning to one of our more
pleasant activities, I think. We're back into the license renewal area,
and Mario, I think you're going to lead us through this.
DR. BONACA: Okay; just a few notes. This is the fourth
ACRS meeting concerning the Calvert Cliffs license renewal application
and the related SER. During the session, we will hear presentations
from represents of BG&E and the NRC staff. I would like to point out
that BG&E submitted its license renewal application for Calvert Cliffs
unit I and II on April 8, 1998. The staff issued a safety evaluation
report in March of this year. The ACRS reviewed the application and the
SER in April and issued an interim letter in May. The staff issued a
final SER on November 16, 1999, which includes the resolution of open
and confirmatory items for the SER.
Our subcommittee held a meeting on November 18 to review the
resolution of the open and confirmatory items. Today, we have asked
BG&E to come and make us a presentation on the plan components that
would limit continued operation of Calvert Cliffs and on the processes
they will use to ensure compliance with the commitments contained in the
license renewal application.
We also asked the staff to respond to certain specific
issues that were discussed at our November 18 meeting. With that, I
will turn the meeting over to Mr. Doroshuk of the NRC.
MR. DOROSHUK: Thank you, Dr. Bonaca.
DR. POWERS: Which is new and exciting for us.
MR. DOROSHUK: Thank you.
DR. POWERS: Explain yourself a little better.
MR. DOROSHUK: Good afternoon, Mr. Chairman and other
members of the Advisory Committee on Reactor Safeguards. My name is
Bart Doroshuk, and I'm with Constellation Nuclear Services. As you
know, I did work for BG&E for 10 years, and I was the project director
for license renewal up until this summer. BG&E did create a new
subsidiary to provide assistance to other utilities going through the
license renewal process, so I am responsible for interfacing with other
utilities on the same process we're discussing here.
For the purposes of representation, I have been authorized
to speak on behalf of the BG&E license renewal project team. We are one
company working together for the same overall company. On my right is
Craig Sly, senior engineer from the nuclear regulatory matters section
and the acting director of that section. He will be speaking about
regulatory commitment management. And to my right is Dr. Gryczkowski,
who is from the nuclear engineering unit at Calvert Cliffs, who will
assist us in fielding questions regarding nuclear fuel issues that you
were discussing perhaps at the last subcommittee meeting.
As Dr. Bonaca pointed out, we've been asked to talk about
three items for you, and that is the discussion of life limiting, if you
will, components that we may have looked at during our tenure
evaluation; to discuss the one-time inspection activities and regulatory
commitment management. What I would like to do is start out and make
those two presentations on the life limiting issues and one-time
inspections, and then, Craig will follow up with the third presentation.
For the purpose of -- when we heard this discussion at the
subcommittee, we -- I observed some terminology that we use differently
at the plant and others do at their utilities, and that is there is an
end-of-life and there is an end-of-license, and we believe at Baltimore
Gas and Electric that the end of a life of any facility is driven
primarily by economics, and that will dictate when a particular
generating facility would be retired. In the case of Calvert Cliffs,
the nuclear plant, we recognize there is one standard on all of our
evaluations; I would like to point out up front that we did not seek to
reduce any existing safety margins when we evaluated Calvert Cliffs for
the period of extended operation, and none of our findings rely on
reducing any of the current licensing basis safety margins to show
public health and safety during the period of extended operations.
So from that perspective, that is a very important standard
that I think not only Calvert Cliffs has to live up to but the rest of
the industry. Certainly, if there are cases where licensees do seek
exemptions or changes to the safety margins, those are done on
plant-specific, individual license amendment basis and are reviewed
separately.
With respect to the end-of-license, we believe that that is
a calendar period. We believe that the license renewal option gives us
an option to operate out to that period. So there are, in our view,
distinctly different definitions.
What I would like to do today is discuss the major
refurbishment plans and concepts for Calvert Cliffs from that broad
perspective, and hopefully, we can touch on any limiting areas. As a
result of our review, as a result of our submittals, Baltimore Gas and
Electric and Calvert Cliffs found no major refurbishments to the plant
necessary to support operation solely during the period of extended
operation. We had a few cables identified, sets of cables we will be
replacing. There are EQ equipment that we will continue to change out,
but from the purpose of going beyond year 40, we identified no large
component replacements or limitations.
Nor during our reviews of the plant and in our application,
our application did not request any exemptions, any exclusions or any
special considerations with respect to RCLB or the current regulations
to support our submittal which requested the renewed license. So we
believe that our application, the technical findings don't identify any
technical limitations to the plant, and we also believe our technical
findings, which the staff has reviewed now, do not encroach upon the
established safety margins in our plant, in our license.
However, we do manage the life cycle of our plant, and we do
have engineering organizations that look out over the operating horizon
and develop the modification profiles for various equipment systems and
components and structures on an ongoing basis to address two different
items: number one, to ensure that we continue to meet safety
requirements established by the NRC; and two, to address reliability
requirements we impose upon ourself to meet the needs of the customer
that we serve.
And the following lists, we took out, I took out of the
current look-ahead for 5 years or 6 years or so, and these are
conceptual projects that we took a cut on. There are hundreds of
things, but we cut off at $1 million, and we said let's just look at the
ones that are above $1 million, since we're really talking about
economics, and let's see what -- how they fan out over the next several
years. So what you're going to get here is a presentation or an
overview of the photographs of what we think we'll be doing to the plant
for various reasons over the next 5 or 6 years. Some of them are very
conceptual, and so, when you see replace unit I pressurizer, it's
because in 1989, we had a pressurizer issue that shut us down for
awhile, and we've got that on our radar screen to make sure that we
continue to watch do we have the problem addressed? So we track large
items like that because it makes sense to us.
So you can see in 1999, replacing reactor coolant pump
rotating assemblies; large circuit breaker replacements, motor control
center bucket replacements; in the year 2000, we are considering
replacing the refueling machine because of reliability issues and the
fact that it's a very difficult machine to perform maintenance on. You
can't perform maintenance on it when you're operating, because you don't
have access to it. Auxiliary building roof considerations, self-site
power, minimum load and peaking power project with respect to our loads,
electrical loads, control room habitability improvement projects, main
generator voltage regulator, replace switch gear HVAC compressors.
Some of these are driven by our continued compliance. You
can see control room habitability being one, where we are continually
trying to make sure that we are meeting expectations of the NRC and
their regulations. Some of them are clearly addressed by reliability
issues, and certainly, the HVAC compressors would be a good example.
MR. BARTON: The circuit breakers for KV13, are they
selective, or is that all of them?
MR. DOROSHUK: I don't know the details of that particular
one.
MR. BARTON: MCC buckets, all of them or just safety-related
or what?
MR. DOROSHUK: Those are the safety-related ones.
MR. BARTON: Safety-related?
MR. DOROSHUK: Right.
MR. BARTON: So those are safety-related?
MR. DOROSHUK: Right.
MR. BARTON: Both of them?
MR. DOROSHUK: Looking out to 2001, you can see we do have
-- and we do have at this point -- an active steam generator replacement
project scheduled to replace the generators in 01 and 02 on both units;
looking at main generator staters, in-core instrumentation; ICIs; salt
water valve replacements and electronic processing equipment, and you go
into 2002 and go down that list and see there is a range of -- the
further you get out, the more conceptual you begin to get because the --
some of the problems we're just monitoring as a monitoring; you can see
the pressurizer there.
That certainly is not a planned replacement at this point,
however, we do put these markers in our strategic plans to make sure
we're not caught by surprise from a budgeting or an economic
perspective.
DR. SEALE: Could I ask a couple of questions?
MR. DOROSHUK: Yes, sir.
DR. SEALE: From what you've said, I take it that these
items are place holder --
MR. DOROSHUK: Some of them are.
DR. SEALE: -- in a sense.
Now, but do you -- do I read this to mean that you have set
aside the resources in your financial program to cover all of these
items or something of comparable cost but greater priority that might
arise?
MR. DOROSHUK: Yes; we have a -- the planning horizon for
Calvert Cliffs, we looked at during the license renewal effort, and the
analysis we did assumed between a $20 million and $30 million a year
capital improvement program. In fact, the conservative one was a $30
million profile for over 36 years, and if you multiply that out, that's
$1 billion of investment above and beyond the normal maintenance.
So those are the types of book ends we put around our
planning on an annualized basis, and of course over the entire period,
it adds up to be quite a lot.
DR. SEALE: Yes.
MR. DOROSHUK: And on an annual basis, we will add up and
budget and determine on an annual basis what our actual expenditures
are, but they range between $20 million and $30 million a year.
DR. UHRIG: I notice you have a power uprate scheduled in
2002. How large is this? Is it 5 percent
or --
MR. DOROSHUK: That would be -- I think it's a 10 percent
theoretical upgrade. We had a project in the early nineties that we ran
in the life cycle management organization that looked at a 10 percent
power uprate as a, you know, potential for Calvert Cliffs when the steam
generator degradation increased, and we knew we were replacing steam
generators; when we were looking at the license renewal application, we
felt that was in conflict technically and from a regulatory perspective
with those efforts, and so, what we did is we took it out beyond the
horizon.
Now, as we go forward and replace generators, we're building
in --
DR. UHRIG: Capacity?
MR. DOROSHUK: -- capacity, so that there is a potential,
there is an option, and of course, when you get out to that time period,
you look at is it cheaper for me to buy the 100 megawatts from
Pennsylvania than to go through the $100 million or $200 million upgrade
to increase the power? And that's really what the tradeoff is: can I
buy it cheaper than really putting that stress on the plant.
DR. SHACK: That secondary pipe replacement, that's erosion
corrosion --
MR. DOROSHUK: Yes.
DR. SHACK: -- type thing?
MR. DOROSHUK: Yes.
DR. BONACA: Is that the $38 million a year forward pretty
much consistent with the expenditures you have seen in the past, say, 5
years?
MR. DOROSHUK: They were down around $20 million to $25
million in actuality, but as a book end, we looked at that as an
assumption.
DR. BONACA: It's in the same --
MR. DOROSHUK: Yes.
DR. BONACA: -- order of magnitude.
MR. DOROSHUK: Yes. I mean, we always wish we could spend
more.
And then, of course, as you look out at 2003, 2004 and 2005,
you can see some of the longer term issues that we are looking at.
DR. UHRIG: Well, now, some years will be greater; for
instance, your steam generator replacement --
MR. DOROSHUK: Sure.
DR. UHRIG: -- is a major --
MR. DOROSHUK: That's a separate capital project altogether;
not included in that core life cycle expenditure.
DR. UHRIG: Okay.
MR. DOROSHUK: We believe that we continually, proactively
manage the life cycle at Calvert Cliffs so that all safety requirements
are met and/or exceeded and that the reliability goals are maintained
irrespective of time and life. We have not identified any technical
limitations on the 60-year license.
We believe Calvert Cliffs can continue to safely and
reliably generate electricity during the current and renewed license
period as a result of our evaluations.
DR. BONACA: A question I have.
MR. DOROSHUK: Yes, sir?
DR. BONACA: Would you replace the steam generators if you
didn't go to license renewal?
MR. DOROSHUK: We looked at the steam generators as a
stand-alone project, and it was coincidental that it happened when we
were doing license renewal. In fact, the actual inspection, the
degradation began to increase in 1995, which was 5 years after we
started working with the industry to do license renewal.
If there are no more questions on life-limiting issues --
DR. BONACA: Well, I have one.
DR. POWERS: What is the life-limiting issue for the plant?
You're getting a license extension that will take you to 60 years. What
prevents you from getting another one to take you to 80?
MR. DOROSHUK: The life-limiting issue is whether or not we
can sell Calvert Cliffs energy competitively.
DR. POWERS: That's not a good answer.
MR. DOROSHUK: That's not a good answer?
DR. POWERS: No.
MR. DOROSHUK: Okay.
DR. POWERS: I was looking for something specific on what it
is about the plant that will -- you finally say yes, I've finally got to
quit.
MR. DOROSHUK: Well, we haven't identified -- I mean,
everybody looks at the vessel and says okay, where do you stand on
pressurized thermal shock? And we have submitted our package, and it's
been approved for at least 60 years. We believe it goes way beyond
that. We haven't identified any new aging that has required enormous
increases in operation and maintenance or replacements. So it's hard to
say.
I happen to think that maybe it may not be Calvert Cliffs,
but for another plant, it might be an environmental issue. But at least
in our case --
DR. POWERS: I mean, that's outside of the range I was
looking for.
MR. DOROSHUK: It's outside the range; right.
DR. POWERS: I was looking for -- the answer that you
probably gave me is had you seen anything in your vessel analyses or
your containment analyses that suggested that there was a finiteness in
the plant's environment? Your answer is, well, you haven't.
MR. DOROSHUK: No, sir; that is correct.
DR. POWERS: No obvious one.
MR. DOROSHUK: And we did look from a general engineering
perspective when we did our life cycle evaluations, because as you know,
we've done a broader evaluation of the plant than just license renewal.
DR. POWERS: Sure.
MR. DOROSHUK: And even in our analysis, we, since we were
able to imagine additional large capital expenditures, we assumed a $500
million surprise on the plant, and we placed that at various points in
the time line where we would consider the operational life out of a
60-year, and we looked at the impact on our ability to compete. And so,
that's how we modeled, if you will, the unknown.
DR. POWERS: Okay.
DR. SHACK: What's your steam generator replacements going
to cost?
MR. DOROSHUK: $300 million is currently the project
estimate.
DR. UHRIG: Does that include power replacement? Or is that
in addition to power replacement?
MR. DOROSHUK: No, that would be just for the project
itself.
DR. UHRIG: The project itself.
MR. DOROSHUK: During a normally-scheduled outage; we're
looking at two 75-day outages, and it's a replacement where we're going
to cut the dome off and replace the bottom parts of the generator.
In that period of time, one would imagine if -- and Maryland
will be in a deregulated state -- that when Calvert Cliffs is shut down,
there will be no revenues. So, for Dr. Uhrig's question, in Maryland,
we will not make money after July 1 next year unless we are operating,
because we will be a deregulated plant.
The next topic that you asked me to discuss were the
one-time inspections, and I put together two slides because I wasn't
sure -- perhaps this is more of a -- we will field your questions. When
we did the evaluation of the plant, there were some areas of the plant
that we had insufficient evidence to confirm that either aging was
occurring or is minimal or doesn't represent a credible challenge to the
passive intended functions. And we were faced with do we spend a lot of
money trying to prove the negative, or do we just go in the plant and
look?
Or we might have a situation where there is little
documented evidence to confirm that where we have mitigation programs,
such as coatings, et cetera, that we don't have a history of failures.
So we were faced with either trying to engineer the problem or go look
for a problem, and that's what we decided to do, and that's the
fundamental concept of our one-time or ARDI program across the site, and
we happen to have grouped all of our one-time inspections under one
program called age-related degradation inspection program.
That was the sort of the fundamental technical basis, and
we've set our thresholds for corrective action very low. In fact, in
most cases, it's any evidence of degradation. So we would go into,
perhaps, a valve, while a valve is being torn down; it happened to be a
pressure boundary within the scope of license renewal, and while the
valve is taken apart, we would go in and look at where those low flow
stagnant areas where we might have bad chemistry, or you might expect to
see some localized degradation, and so, we would go in, and if we found
the evidence that there was either corrosion or general corrosion or
whatever the aging effect we would be looking for, we would then kick
back into a corrective action program and then look at the impact on the
operability of that particular function.
So, the program is meant to confirm, and it's also meant to
have a very low threshold so our corrective action comes into place.
What I've done on the second slide, and I'm not sure how
this is going to show up here, but this shows basically a general sample
of the type of -- on a system basis -- auxiliary feed water, component
cooling, compressed air, nuclear steam supply system which are contained
in spray CVCS sampling; emergency diesel generators; main feed water;
HVAC; RCS; reactor vessel internals; salt water; safety injection and
service water.
You can see the components of interest that we felt we
didn't have the appropriate evidence in our evaluations, and you can see
what we were going to go in and look for in these particular areas of
these systems.
DR. BONACA: I have a specific interest on this. The
question was actually -- I assume now -- and I agree with most of it; I
mean, I agree with all of them I've reviewed, and assume you go in now
and, in fact, you find that you have a problem.
MR. DOROSHUK: Right.
DR. BONACA: And so, now, you have a corrective action
there. From the perspective of NRC involvement, is there going to be an
involvement of the NRC or the process by which the NRC is able to come
in and accept the program you're going to institute because of that? I
mean, if you find you have a problem, you have the choice: you can
either fix it and say it's great for the next 20 years, and since my
FSAR commits me to one inspection only, I'm not going to do anything
anywhere, or you can be proactive and say, oh, I've found a problem; I
fix it; then, I have also an additional program to prevent recurrence
which includes a number of inspections.
Now, there are two choices there, for example; I'll give you
an example: is the NRC going to be called by some process to
participate in the decision, or what is the vehicle to doing that? That
was my question.
MR. DOROSHUK: The way the program is established is if we
do go into corrective action, we have to ask all of those questions:
number one, did we have a challenge to safety on the individual finding?
And then, when did that occur? And that's your standard operating
degraded state evaluation, but once it gets into corrective action, a
number of other things happen: we ask those -- is this a generic issue?
Or was our sample size too small? Do we have to go bigger?
But it doesn't automatically trigger a reporting requirement
into the NRC. It triggers reporting requirements only if it trips some
of the levers like tech specs or an LER, I guess would be --
DR. BONACA: Yes; I have reviewed your corrective action
program so far as the elements of that, and that's quite comprehensive.
But the point is if today, you knew that you had a problem at that
location, and you have a process or a program, the NRC would be arguing
with you on what you've got to do or not to do. It seems to me that
now, once you get into, you know, the license is granted, the NRC is not
involved in determining whether or not changes to the program that
really have to augment the programs are adequate or not. I'm not saying
it's right or wrong.
MR. DOROSHUK: Right.
DR. BONACA: I'm just trying to determine that, because
50.59 only allows you to monitor a reduction in changes, in the actual
commitments, not an increase in commitments to compensate for an
identified problem.
MR. GRIMES: Dr. Bonaca, on behalf of the staff's interest
and our evaluation basis, we concluded that the quality assurance
process and our inspection of it would operate in the same way for these
actions as they would for the rest of the licensing basis, and that is
that we inspect corrective actions; we look to ensure that they're
prompt and effective. In the event that we feel that they're not
effective, then, we can take enforcement action, and that has happened
in cases where it's, you know, either events or findings have been
challenged as adequate, and we also look to that to feed into our
generic communication process to take our experience and then determine
whether or not either information notices; generic letters; bulletins or
rulemaking is warranted based on the findings of any program.
DR. BONACA: Good.
MR. GRIMES: And we are relying on that; we're also going to
try and develop a means of getting feedback on one-time inspections in
such a way that we can build that experience to improve the review
guidance in the future.
DR. BONACA: Realize my interest was mostly for the
perspective. I mean, we are going to learn a lot about life between 40
and 60 years, and I think we have to have a means of learning and
disseminating information. And so, that was -- my intent was mostly one
of understanding, and I can accept that the existing processes, in fact,
are adequate.
Okay; thank you.
DR. POWERS: Can I understand more about the theory behind
the one-time inspection? Something goes on the list for one-time
inspection. It means you have to have a feasible degradation process
when you don't think it's a threat. Is that correct?
MR. DOROSHUK: Right, right; it's where we couldn't prove
that it was or wasn't occurring or where we have a program in place such
as a coating program that we're not sure how well that is mitigating or
a chemistry program that we're not sure how well it's mitigating,
perhaps, a particular aging effect in a particular location.
DR. POWERS: You've heard the term maybe late bloomers?
MR. DOROSHUK: Yes.
DR. POWERS: Okay; tell me what late bloomers are.
MR. DOROSHUK: Well, you're talking about latent effects
that would crop up and surprise you at some point in time, a latent
failure.
DR. POWERS: It sounds like that's what I'm talking about,
doesn't it?
MR. DOROSHUK: In other words, you don't want to inspect too
soon, and you don't want to inspect too late, and we have discussed this
with the staff, and we agree with them, and we believe you need to
inspect before year 40, and we're sequencing these inspection activities
to occur. They will be after year 30, but they will be before year 40,
and they will go into the basic scheduling process; in fact, we've
already done some of these examinations in compressed air, for example,
where we're looking for moisture effects, and in the worst location in
the compressed air system, it's been there for 20 years; it was the
discharge of a compressor that was apparently the one of the wettest
spots, and we did the destructive testing on it; cut it open, and we
found nothing. The stainless steel was still absolutely perfect; not a
drop, not even a pit.
So if you look, though, on many of the fluid systems, we're
looking for the same thing: localized corrosion or erosion. And in the
fluid area, we think that with the spread-out inspections, you're
probably going to really come up with something over that 10-year period
when there are so many outages, five outages. We're going to do all
these things. So we think that -- conceptually, we think we've got it.
DR. WALLIS: I have various questions. One is how do you
know one time is the appropriate -- it's sort of a gamble, isn't it?
You try it, and you don't see it, and then --
MR. DOROSHUK: Well, we believe that the aging effects we're
looking at are, number one, remote. They're nonaggressive. And after
30 to 40 years of operation, of giving them an opportunity to manifest
themselves, if they were going to manifest themselves in the form of a
pit or rust, then, we would see that evidence. If we didn't see the
evidence, it's hard for me to believe that there would be a rapid onset
or a late bloomer on the same aging effect.
Given that, we don't have any other experience after how
many hundreds of years not only on reactor years but on fossil stations
and other types of facilities that have discovered a new aging effect
that we're not aware of.
DR. WALLIS: Well, I am thinking about an automobile. I
mean, if you have an automobile, it doesn't show significant rusting for
about 6 years. And then, all of a sudden, it starts to rust away.
MR. DOROSHUK: Right; and if the coatings began to become --
right, exactly.
DR. WALLIS: So there's a sort of a time before things
happen. You don't quite know how long that time is.
DR. BONACA: But it's true that, for example, I'm looking at
the service water system. You do one-time inspection in many locations.
You don't do them all simultaneously. You do some locations at one
shutdown and some -- so what you're saying is that for the service water
system, really, the one-time inspection monitors different components at
different times --
MR. DOROSHUK: Right.
DR. BONACA: -- through a long period of time. So what
you're saying is that if you -- you may have one location where you have
something unique happening there. That's always going to happen. But
in general, a trend could be identified by the succession of inspections
that you perform.
MR. DOROSHUK: That is correct; but I'd like to point out
one other aspect of these. These aren't the only things that we're
doing on these particular locations, on these particular systems or on
these particular structures. We're zeroing in on this particular
activity and trying to determine is this going to be adequate? But
there are layers of other maintenance procedures, whether they be in the
license renewal space or walkdowns or activities at the plant that we do
that this is -- we have gone through all of those activities and said we
still don't feel comfortable that we have pinpointed these types of
events. So we can't discredit in our discussions the other layers of
monitoring that we do.
DR. POWERS: Let me ask you this question: if you had a
failure at, let's say, the compressed air system just for argument's
sake, would the failure of that system due to general corrosion, that's
a possibility up there, is the failure of the compressed air system due
to general corrosion any more hazardous in the license renewal period
than it would be during the normal license period?
MR. DOROSHUK: No.
DR. POWERS: You have the same protective systems available
to back up for the license renewal period?
MR. DOROSHUK: Yes, sir; all of our findings, and the staff
has agreed with us that the findings were based on maintaining the
current licensing basis under current design loading conditions that we
were licensed to. So these failures that would occur in 41 would be no
different than the failures that would occur at 32 if they were to
occur.
DR. POWERS: And you may be just as ignorant now about -- as
when you started the license renewal application.
MR. DOROSHUK: Yes, sir.
DR. POWERS: You are.
MR. DOROSHUK: Yes, sir; yes, sir.
DR. WALLIS: Now, when you do all of these inspections,
there must be a lot of questions, such as what's the threshold for us
noticing something in corrosion general term, you notice something, or
you don't notice something, and then, how much of something is
significant? What's the threshold for taking action? I go back to the
car again. If I see rust on the frame of a car, it's unlikely it's
going to rust through for 40 years. And so, who cares.
If I see rust on the body work, it implies something else.
So there must be a whole lot of thought behind this, which is what's
likely; how is it significant? If I see some evidence, how likely is it
it's going to get worse? And there are all kinds of models for what's
happening that you need to put into this as well as just looking.
MR. DOROSHUK: Yes, sir; we think that most of these aging
effects are way out on the fringes, and the likelihood --
DR. WALLIS: Do you have some sort of --
MR. DOROSHUK: -- is low.
DR. WALLIS: -- model for them so that when you do the
inspection, this is a point on some predictive capability?
MR. DOROSHUK: We're looking for any evidence.
DR. WALLIS: Any evidence at all?
MR. DOROSHUK: That's what I said earlier, is that the
threshold is on purpose set very low, and that, in fact, was something
we struggled with during the review with the staff and we ourselves
struggled with is when do you -- how do you know when you're okay? And
so, the agreement was because these are areas of the plant that you
typically don't look at that around the fringe, we'll set the threshold
down at basically zero, any indication.
DR. POWERS: Mic on your side refers to microbe-induced
corrosion?
MR. DOROSHUK: I'm sorry, sir?
DR. POWERS: Mic?
MR. DOROSHUK: Yes.
DR. POWERS: Microbe-induced?
MR. DOROSHUK: Yes.
[Pause.]
DR. BONACA: I'm sorry.
DR. POWERS: It seems to me we've got a lesson learned in
the making here. When you said you've identified a bunch of systems
here where you don't think anything is going on, but you don't have any
proof one way or the other, so you're going to go find out what, and you
set your threshold for making a decision to do that by your mission, and
I believe I've heard the staff say the same thing, that the threshold
was set very low. So is the lesson going to be that we can bring that
up?
MR. DOROSHUK: I would think that we would provide some of
the first evidence for, for example, another plant with the same type of
configuration of a compressed air system looking at is moisture over 25
years an issue at the discharge of the compressors where we took the
stainless steel piping; we brought it to the lab; we cut it open; we
documented the inspection. I would think that that would be a valuable
input to someone else deciding do they have the same effect after 25
years?
Some relative operating experience would have to come in
play, but yes, we happened to try to fit these into ongoing maintenance
activities. One of them was we were cutting out this piece of pipe for
other reasons. Another utility may not have that activity going on.
DR. POWERS: It would be interesting to see this list when
-- what? -- threshold water; in other words, which one of those systems
do you really think is really guaranteed, that you believe is absolutely
unlikely up to, well, it could or couldn't. It would be interesting to
see the list.
MR. DOROSHUK: When would you like to see that?
DR. POWERS: I'm not asking you to do anything.
MR. DOROSHUK: Oh.
[Laughter.]
MR. DOROSHUK: I thought it was an invite back maybe around
2010.
[Laughter.]
DR. BONACA: Just to complete his question, so for RCS
components which are subjected to the ISI/IST program, these are
augmentation steps --
MR. DOROSHUK: Yes.
DR. BONACA: -- that you would essentially plan
approximately when? Over the last 10 years of --
MR. DOROSHUK: At this point, that's when they would occur.
DR. BONACA: Okay.
DR. POWERS: Except for compressed air; that's just been
done.
MR. DOROSHUK: That's right. One particular area of the
compressed air system, right.
DR. BONACA: You integrate those.
MR. DOROSHUK: The procedures written onsite now actually
lay out each system; each sample is identified. And then, they go into
the quarterly maintenance scheduling process, and then, they will start
occurring as part of the normal practice, and then, we will make sure we
document that for our followup on our commitments.
DR. BONACA: Thank you.
MR. DOROSHUK: If there are no more questions on this, we
have one more presentation on regulatory commitment management, Craig?
MR. SLY: Good afternoon. My name is Craig Sly, and I work
in the nuclear regulatory matters group at Calvert Cliffs. Excuse me.
Nuclear reg matters is just a fancy name for licensing, basically.
What I'm going to cover, since I'm not exactly sure what
your interested in, but what I'm going to cover is the process that I
use for tracking and managing commitments at Calvert Cliffs, and I'm
going to do it at a very high level. First, some background
information: how did we get into the commitment tracking business at
all? And I think the answer is we had a significant emotional event
back in the 1989 time frame where we missed a commitment to implement
LTOP requirements, and it was a highly significant, from a safety
standpoint, miss. And I can tell you horror stories that I had. If
you're interested, I'll talk to you about them later, or you can ask me
the question, but prior to that time, we had no commitment tracking
system or management system, and we made a commitment to the NRC to
implement one.
We've been involved in the commitment tracking utility
group; it's called RCTG; for many years, since the 1987 time frame, and
from my standpoint, being in that group, I think it's one of the only
cases that I know of where the utility saw a problem in the industry and
addressed it without the assistance of the NRC, and actually, in the
end, I think the NRC became very aware that commitments were an issue in
the industry after Millstone, and luckily for us, we had, as a utility
or as an industry, I think, pretty much all implemented solid commitment
management programs by that time, and you guys have since gone out and
verified that. I think it was a couple of years ago.
How does my process work? It really -- the cornerstone of
any commitment process, and this is what the industry group wrestled
with for many years, was getting a definition that we really all could
work with, and if you go into the industry, you're going to find
different plants use slightly different definitions. This is our
definition, and it's what works for us and our processes. A commitment
is a docketed statement, so it has to be on the docket, to or by, so it
can be either us making it or the NRC making it. It either, one,
establishes a current licensing basis requirement, and I refer to those
as big-C commitments, capital-C commitments, or it promises a future
action that has not yet been accomplished, which is a little-c.
An example of a big-C commitment would be to commit to
implement a new program, and that program would then be required to be
incorporated into some licensing basis document like the tech specs, the
FSAR, the QA plan, et cetera. A little-c commitment example would be a
commitment to overhaul a pump or correct a minor deficiency, and we
often make those commitments in documents such as LERs and NOV
responses. The difference is fairly significant, and it's why we have
-- and I'll discuss it in a minute, but it's why we have delineated
these two categories.
Again, big-C commitments, CLB commitments, are determined by
the scope and level detail in the CLB; safety significance of
information to the public health and safety. Extent the information is
material to the decisions by the NRC; typical documents that would
contain big-C commitments are SERs and --
MR. DOROSHUK: Licenses.
MR. SLY: -- the license renewal application would contain
more than its fair share.
MR. DOROSHUK: Thank you; I was going to --
MR. SLY: Ninety-nine percent of the commitments that we
make, if I take out this big application, would not be big-C
commitments, okay? That's just the way it works out. We say a lot of
things to the NRC, and most of them just don't make their way and
shouldn't make their way in the FSAR.
This describes the way i would define current licensing
basis, and I'm sure you guys can argue back and forth whether this is
the correct one, but we've been arguing that for years. But in my
procedure, it basically says the current licensing basis commitment
would have information describing the facility and the conduct of
operations which have been evaluated by the NRC and relied upon as a
basis for meeting the applicable regulations or has been accepted by a
safety evaluation report or as part of the operating license, so that's
kind of the basic screen. We have a much more detailed screen than
this, but this is the basic information that anybody going into my
process procedure would be informed of.
This is the nuts and bolts of what we do. First of all, we
identify the commitments, okay? And the way we identify them is we
review every piece of ingoing and outgoing correspondence to and from
the NRC. Just by chance, we also do INPO and a lot of the state
regulators, because we make commitments to them, too.
And before I forget, the reason we do commitment tracking at
Calvert Cliffs is not to meet a regulatory requirement, because as far
as I know, there is no regulatory requirement to have a commitment
management program, but it's good business. When my vice-president
signs a piece of correspondence, and he hands it back to me, he knows
that those commitments will get implemented, and it's taken very
seriously at Calvert Cliffs. Not meeting a commitment is a very serious
offense. So, that's the reason we do it; good business.
So we would identify all of the commitments in the documents
that come in and go out. We then have a computerized tracking system at
Calvert Cliffs. We would assign action items to the individual owners
who are responsible for implementing those commitments.
Licensing would also tag a person to be responsible for
making sure each one of them gets implemented and reviewing the final
implementation of the commitment and also making sure that when the
commitment gets implemented that a review is done to determine whether
or not the commitment should go in the CLB or should become historical.
Closure: at closure, we would place some sort of flag or the commitment
itself in the appropriate site process for future control. That could
include the tech specs; FSAR; but more commonly, it includes a basis
captured in a procedure, okay? And then, a procedure change process
would tell you how to change those commitments in the future. I'll get
to that in a minute, right now, actually.
Managed changes: if you want to change your commitments;
for example, a commitment is bases capturing a procedure, and somebody
wants to change it; that bases capture, which is just a little
designator, usually B1 or B2, and then, in the back of the procedure, it
gives the commitment number. By procedure, when they change one of
those, they are supposed to call nuclear regulatory matters, and what we
do then is we apply the NEI guidance on changing commitments, and that
guidance would take you through a flow diagram, essentially, that starts
at -- basically, it uses a hierarchy, starting with tech spec changes,
FSAR, and moves down the hierarchy from there.
And basically, if it's something that is in the 50.59, or
it's a process for changing that commitment is defined under 50.59 in
the RTU's 50.59 process, if it's a commitment that's a tech spec, you
would use the tech spec change process. If it's a commitment that's in
a procedure, you use a little bit different process, and you would end
up normally notifying the NRC by letter or in some cases, for very
unimportant or old commitments, notifying the resident inspectors or the
project managers, something like that.
DR. SHACK: How many of these did you say there were again?
MR. SLY: It depends on what you're asking. If you're
asking how many active ones am I tracking at the moment, I would say
there are probably about 300. How many have we made in the life of the
plant? About 20,000. We, in 1994, we did a historical look at our
commitments. We hired a contractor; they read the docket cover to cover
from dumping the dirt to 1994. They identified 16,000 commitments.
We did a safety significance screen on those commitments,
and in that screen -- well, basically, the results of all of that
screening was that we identified two high -- excuse me, six high safety
significant commitments; 256 media. The rest were either low safety
significance, no safety significance or superseded by some other
commitment. That kind of gives you an idea of, you know, there are land
mines out there, but there's not very many.
DR. APOSTOLAKIS: So safety significance is the way it's
defined in the regulations, and you're using those words.
MR. SLY: We did this in 1994. I could show you the screen,
but I would venture to guess it's not exactly the way it's defined in
the regulations. We had a multipage screen that we used to figure out
what the safety significance was.
DR. APOSTOLAKIS: Well, when you say something is not safety
significant, you had better be consistent with what the regulations say.
DR. POWERS: I think it is such a specialized nomenclature
in regulatory space, it probably wouldn't be useful for the exercise you
went through.
MR. SLY: Yes, it was fun.
Commitment closure, and I think this is probably what you
guys are most interested in, you know, what happens to a commitment when
the action is complete. On my books, my goal is that all commitments
ultimately become historical and are maintained by a process that is
commensurate with their safety significance or captured in the CLB or
both. If the commitment belongs in a procedure, then, it should go to
that procedure, and we have a process for changing procedures. That
process includes making sure that if you change a commitment that you go
through the NEI change process.
If I relocate a commitment to the UFSAR, it should be
changed in accordance with the change process for the UFSAR, 50.59.
Same thing with tech specs: there is a change process for that; there
is a change process for QA documents and security plans, QA plans, those
kinds of things. And we also, when we close commitments, we do a
screen, and that screen is designed to tell us if that document or if
that commitment should go into a CLB level document, so at the end of --
when you're finished, complete with the action that you committed to
take, for example, if you committed to implement a program. We
completed the action; now, what do we do with this thing?
We have a screen in the back of my process that tells you
basically where you should put the thing: should it go in the FSAR?
Should it go in the tech specs? Should it go in the procedure? Once we
decide where it should go, and we put it there, at least from my
standpoint, I don't track it anymore, okay?
In summary, commitments at Calvert Cliffs are systematically
identified and implemented. All commitments become historical upon
completion and are located into appropriate site documents. Control of
commitments is maintained under the change process of those documents,
and we manage changes under NEI 99.04, and NEI 99.04, for your
information, if the changes to the NEI 99.04 says change in accordance
with 50.59 or change it, you know, if it's your tech specs, it says
change in accordance with tech specs. So this 99.04 covers the whole
gambit of documents at your plant or at any plant.
DR. APOSTOLAKIS: But I am again puzzled. You said that
safety significance is not what I understand by safety significance.
But if that 99.04 tells you to go and do it according to some
regulation, and that regulation uses safety significant versus not
safety significant, aren't you back to the regulatory definition? I
mean, I don't understand how you can declare something as not safety
significant when the regulations say that it is. Can you do that?
MR. SLY: I'm not declaring it -- I'm definitely not trying
to do that. Here's the -- I don't know if you can read that very well.
DR. APOSTOLAKIS: I can, yes.
MR. SLY: Can you?
DR. APOSTOLAKIS: Yes.
MR. SLY: Okay; it's pretty simple.
DR. APOSTOLAKIS: Step four, box two is not right; no, go
ahead.
[Laughter.]
MR. SLY: Okay; here is the top of the thing. If you're
proposing a change to a commitment, okay?
DR. APOSTOLAKIS: Yes.
MR. SLY: It asks you is the change process codified?
DR. APOSTOLAKIS: Okay.
MR. SLY: If it is, apply 50.59, 50.54, 50.82 as
appropriate, okay?
DR. APOSTOLAKIS: Okay.
MR. SLY: So you go to the regulations --
DR. APOSTOLAKIS: I understand that.
MR. SLY: -- and you change; is it significant to safety?
We have a figure A2 which I brought, and we'll look at that, too: do
not proceed with change; got to discuss the change with the NRC. Is the
original commitment necessary for compliance?
DR. APOSTOLAKIS: So I guess my question is in step two,
when you ask is the change significant to safety. How do you make that
decision? Okay; good.
MR. SLY: Would it impact the ability of an SSC to perform
its safety function? Yes; go to 50.92; no, document your rationale.
DR. APOSTOLAKIS: Oh, yes. It looks like the regulatory
definition to me, but I may be missing something. Everywhere there,
it's --
MR. SLY: Well, previously, you asked me did we use the
regulatory definition when we did our safety significance screening for
our historical commitments.
DR. APOSTOLAKIS: That's a different process?
MR. SLY: We would have -- we would have used a definition
that was something like this, but the screening sheets were three pages
long. We asked a series of questions that were three pages long,
including, you know, was it part of a generic letter response; you know,
was the information -- CLB-type information, there's a whole slew of
stuff that we screen for to make sure that we weren't missing something
that was important to the NRC, even on a personal level, not even a
safety level.
DR. APOSTOLAKIS: Anyway, I --
Yes, sir?
DR. SIEBER: You probably, since you manage this commitment
tracking system, have an idea of what the backlogs are in various
categories. Could you tell us what backlogs you have now, which are
items that are beyond the due date?
MR. SLY: The backlog? We don't have any beyond the due
date.
DR. SIEBER: Okay.
MR. SLY: Believe me; in 1989 -- did I say 1989? Yes. When
we had the LTOP problem --
DR. SIEBER: Right.
MR. SLY: -- the president of the company, who is now the
chairman of the board, basically said you will meet the due dates, or
you will pay the price. It came from that high up.
DR. SIEBER: Okay; now, some of these things are procedure
changes. Are there any backlogs in procedure changes, like operating
manual, maintenance manual, ISI, IST?
MR. SLY: If you're asking -- I think maybe what you're
asking me is how old is your average commitment?
DR. SIEBER: Correct.
MR. SLY: Okay; and I'm kind of guessing here. We have
commitments to pull, for example, pull test capsules out of the reactor
in 2013 and --
DR. SIEBER: Yes, I'm not interested in that.
MR. SLY: They're really old. But generally speaking, your
average commitment won't be more than 6 months old.
DR. SIEBER: Okay.
MR. SLY: And we just -- some of them might go a year if
it's something that -- like a mod; some of them may even go more than
that, if it's something like a mod that you can't implement except
during an outage, but for most procedure changes, you would reasonably
expect that they could get it done within 6 months.
DR. SIEBER: Okay.
DR. UHRIG: I have one quick question. You have an item
here on your replace electronic process and equipment. Could you define
electronic process and equipment? Is this the INC, the control room?
Or is this other?
MR. SLY: No, sir, I don't have the details of that.
DR. POWERS: That's a really good one, Bob. I'm glad you
talked about that. It would be very interesting to find out what that
one is.
MR. DOROSHUK: I don't know now what that is.
DR. UHRIG: You do not anticipate conversion from analog to
digital systems, or have you already done that?
MR. DOROSHUK: We looked at that in the early nineties, and
we went and participated with the INC upgrade efforts at EPRI and
another utility group. We decided to back off the wholesale digital
replacement and upgrades in favor of more strategic maintenance programs
that looked at establishing a better life cycle management program for
those particular items because of the cost.
DR. UHRIG: Okay; thank you.
MR. DOROSHUK: That concludes our presentations, Mr.
Chairman.
DR. BONACA: Thank you; if there are no further questions.
MR. GRIMES: Dr. Bonaca, this is Chris Grimes of the staff.
Before you let them go, since they're doing such a nice job of filling
you in on the answers to the questions, one of the other things that you
had asked at the subcommittee meeting, and I asked BG&E to be prepared
to discuss, was you had asked for information concerning the fluence at
the end of the 60-year term and questions about the -- what burnup
considerations went into those values, and if I could ask that BG&E
could respond to any specific questions on that. We also have, as you
requested at the subcommittee meeting, brought Margaret Chatterton of
the staff and Lambrose Lois, who are here to respond to the questions
you had about fluence values and burnout.
MR. DOROSHUK: Are there any questions for us regarding the
fluence calculations?
DR. POWERS: I guess the thing we wanted to know was what
the peak fluence was going to be at the vessel at the end of 60 years.
MR. GRYCZKOWSKI: Well, at the end of 60 years, for unit I,
we expect that the peak fluence will be on the order of 4.95 x 1019.
DR. POWERS: And in coming up with that, that value, what
kind of burnups were you thinking of for your fuel?
MR. GRYCZKOWSKI: Well, what we are doing now is we modeled
up to, I believe, cycle 11012 for unit I, and then, we have a standard
low fluence pattern that we're using now, and that's been running on the
order of about 620 EFPD, so we use 620 EFPD for all cycles leading to
the --
DR. POWERS: Maybe you should define the --
MR. GRYCZKOWSKI: EFPD? I'm sorry.
DR. POWERS: Effective full-power something.
MR. GRYCZKOWSKI: It's an effective full-power day, right.
DR. POWERS: Day.
MR. GRYCZKOWSKI: Yes; it's a function of the burnup. So
you take the megawatt days per metric ton of uranium, multiply them by
the amount of uranium, and divide it by your 2,700 megawatts, which is
your power output, and that gives you your effective full-power days.
There's a direct correlation between the two.
And anyway, we use that, and then, as the cycle history
changes, we update that, but it hasn't changed much recently so --
DR. POWERS: You were running an 18-month cycle?
MR. GRYCZKOWSKI: We're on a 24-month cycle.
DR. POWERS: A 24-month cycle?
MR. GRYCZKOWSKI: That is right.
DR. SHACK: What are your end-of-life fluencies for your
high fluence internals components?
MR. GRYCZKOWSKI: I don't know that right off hand, but
they're fairly high. They're much higher than that.
DR. SHACK: They're very high, yes.
MR. GRYCZKOWSKI: We're already past the 5 x 1020 value that
you used as a cutoff for almost all internals.
DR. SHACK: Yes; you're long past that; I'm sure.
MR. GRYCZKOWSKI: Yes; we hit a lot of those in cycles three
and four, I found out. I don't have that information. I might have
some of it available; hold on. I did bring some stuff.
DR. SHACK: The staff supplied me with some information for
Westinghouse reactors, and I just wondered if it was very different for
a combustion unit.
MR. GRYCZKOWSKI: We looked at several components, not all
of them. We did the calculations for what was requested, and for
extended end-of-life for unit I, for example, the core shroud, we're
talking about 1.3 x 1022; core shroud nuts, 2 x 1021.
DR. SHACK: That's not that bad.
DR. POWERS: These are pretty modest compared to what I'm
used to.
MR. GRYCZKOWSKI: Yes, and that's after 60 years, and
there's comparable for unit II. I guess it all depends on what you're
looking at. If it's right outside the core, right in the core, it's
going to be much higher, obviously, than something on the top.
DR. SHACK: Okay; now, I assume that you're running some
sort of low-leakage or low-leakage cores?
MR. GRYCZKOWSKI: We have a low fluence core. Basically, we
have twice-burned assemblies on the outside of each cycle.
DR. SHACK: When did you go to that kind of an arrangement?
Early on?
MR. GRYCZKOWSKI: Yes, we went to it -- oh, we went to it
several years ago. I have that information also. Let me tell you. For
unit I, it turns out -- that's really our core reactor; that was the one
we thought we had to shut down in 2003 originally. We went to guide
tube flux suppressors in cycle 11, okay? And on that periphery, and
then, we went to low leakage in cycle 13, which was, I guess, started in
1997.
For unit II, we were at low fluence in -- we've been at low
fluence since cycle 10, which is approximately 1994, I believe.
DR. SHACK: Now, you said low fluence and low leakage. Is
there a difference?
MR. GRYCZKOWSKI: Low fluence, no. Low fluence is the
twice-burned, and we had low fluence with guide tube flux suppressors,
which are even more -- I assume you know what flux suppressors are.
DR. UHRIG: This is just -- put around the outside?
MR. GRYCZKOWSKI: Yes, around the critical welds.
DR. APOSTOLAKIS: That's not a new law there.
MR. GRYCZKOWSKI: I'm sorry?
DR. APOSTOLAKIS: I'm familiar with e=mc2. I don't see the
square there.
MR. GRYCZKOWSKI: Oh, it's on the other side.
[Laughter.]
DR. APOSTOLAKIS: It's on the other --
MR. GRIMES: And were there any other questions that you
wanted to direct to either Ms. Chatterton or Mr. Lois?
DR. BONACA: No, not at this point.
MR. GRIMES: Thank you.
MR. DOROSHUK: Thank you, Mr. Chairman.
MR. SOLORIO: Should I proceed?
DR. POWERS: Yes.
MR. SOLORIO: Good afternoon. My name is Dave Solorio, and
I work in the Office of Nuclear Reactor Regulation in the Division of
Regulatory Improvement Programs. I'm the PM responsible for the review
of the Calvert Cliffs license renewal application submitted by Baltimore
Gas and Electric.
I'm here along with other members of the division, my
division, the division of engineering, and you just noted that we have
some safety system analysis people here and also people from the
division of regulatory improvement programs to support your review of
the Calvert Cliffs license renewal application. I want to briefly
summarize the activities that we've recently completed, and I will be
short so we can get back on track.
In fact, back on November 3, we provided the subcommittee on
plant license renewal with a draft resolution of the open and
confirmatory items that were in the March 21 Calvert Cliffs safety
evaluation report. On the 16th of November, we published this safety
evaluation report, which provided the staff's basis for concluding the
management of aging effects for the renewal term. A couple of weeks
ago, on November 18, we provided a briefing to the subcommittee
regarding the basis for the closure of the various open and confirmatory
items.
Based on the staff's presentation on November 18, the
subcommittee asked for the staff to come back today and provide some
presentations on the specific topics I have listed here as well, some of
which BG&E just got done talking to you all about. I will note that we
expect this to go a little faster in the large part because a lot of the
questions you asked related to the ARDI or some of the same things we
were going to discuss with you all.
With that and with no further ado, I'll turn the
presentation, first presentation over to Mr. Jake Zimmerman, who's going
to talk to you about the status of the license renewal generic issues if
there are no other questions for me.
MR. ZIMMERMAN: Thank you, Dave.
My name is Jake Zimmerman, and I'm a project manager in the
license renewal branch. Today, what I'd like to do is just briefly go
over where we are with the active license renewal issues. There's a
copy of the list in your handout for you to refer to that. Some of you
may recall that back on the 23rd of this year at the subcommittee
briefing, I talked about an August 25 meeting that the staff had with
NEI to discuss all of the license renewal issues, and at that time, we
agreed to come up with a new categorization scheme for these issues. If
you look at page 2 of that list, you will see at the bottom the
categories that we came up with and the titles for those particular
categories.
During a meeting that we had on August 25, NEI agreed to go
through, and based on those new categories that we developed, they
agreed to go through and take a first cut at recategorizing the 106
license renewal issues that we have open. They subsequently sent in a
letter on September 17 documenting their categorization. The staff has
reviewed that, and we met again with NEI on October 27 to discuss those
categories, and in fact, some of those categories have been changed
based on our discussions in that meeting. We did issue that complete
list on November 18 in our meeting summary. I gave an extra copy of
that to Noel today, if some of you don't still have your copy and are
interested in looking at that.
Of the 106 issues, we have deleted five. If you would refer
to the graph that I've created that's titled generic license renewal
issues, status of issues for November 1999, we've deleted five issues.
We've resolved 11. And so, there are 90 open issues in the category
two, three and four area.
During that October 27 meeting with NEI, NEI agreed to take
a look at all of the category four issues, which are easy or editorial
type comments in nature related to the SRP; come up with a markup of the
SRP and any proposed resolution and submit that information in to us by
the end of this year. So that leaves 65 remaining issues, of which 18
of them are on the active issues list that you have in front of you.
There are actually 19; one of them is a category four, which NEI is
going to address. So we're down to 47 issues that remain out of the
original 106 issues.
Those 47 issues, we are currently looking at our resources,
both NRC and NEI, looking to see if we can develop the proposed
resolutions and have them resolved before August of 2000, which is when
we plan to issue the next version of the draft SRP. We've also been
working, NRR has been working with research to resolve many of these
issues, and that's something we are going to again look at our resources
and see whether we want to go forward if we have the time and can
resolve those issues that are not on the active issues list with NEI
directly or possibly look at developing these proposed resolutions;
incorporating them into the SRP and then allowing NEI and the public to
comment on those once that's issued.
DR. POWERS: Give me a thumbnail sketch of what the issues
are with respect to IWE, AWL or jurisdiction.
MR. ZIMMERMAN: I am not familiar with the specifics of the
details as far. As where we are on status, I can address that.
MR. LEE: This is Sam Lee. I'm from the license renewal
branch. On both sides of the RWL issues, we have been clarifying what
the specific issues are with NEI, and we got those clarified, and now,
we are working the issues, and they are in the concurrence process.
DR. POWERS: The statement is that it's a jurisdictional --
MR. LEE: That turns out to be, I guess, their supports,
okay? That in the standard review plan, we said the RWL, we would
inspect those. However, the IWE/RWL, that does not include supports.
So NEI wanted to know. So there's an inconsistency, okay? So we tried
to clarify.
MR. ZIMMERMAN: Any other questions?
[No response.]
MR. ZIMMERMAN: Thank you.
Now, Ms. Stephanie Coffin will make a presentation on
one-time inspections.
MS. COFFIN: Good afternoon. My name is Stephanie Coffin,
and we already talked a lot about the processes for evaluating
degradation with Calvert, so I will try and be quick, but if you have
any more questions, I'll be glad to answer them.
DR. POWERS: Let me ask you this question: I understand the
one-time inspections; these are inspecting facilities where there is
some possibility a degradation mechanism could be active but no belief
that it's active. I mean, that's how you select them, okay? How
credible does the possibility of a degradation mechanism have to be for
it to make it onto a one-time inspection list?
MS. COFFIN: How credible does it have to be to make a
one-time inspection?
DR. POWERS: Yes.
MS. COFFIN: You have to have an environment that in a
material confirmation --
DR. POWERS: I have to have water there?
MS. COFFIN: Yes.
DR. POWERS: Okay.
MS. COFFIN: You would have to have, you know, possible
exposure to chlorides; you would have to have, like for elasmer
degradation, you would have to have it exposed to some minor radiation
or whatever mechanism you're going for, as opposed to you wouldn't
combine erosion, corrosion and stainless steel, for example. That would
not be credible. You wouldn't even have a one-time inspection for
something like that.
DR. POWERS: But I could have generalized corrosion.
MS. COFFIN: Yes.
DR. POWERS: I mean, it seems like the criterion is it's
wet; that's it.
MS. COFFIN: That's what you need for corrosion.
DR. POWERS: Yes; okay, so that means that you can impose
inspection on any component that gets wet.
MS. COFFIN: Well, not if you have a -- for example, all
there is stainless steel piping in the reactor coolant system; that's
all wet, and they don't have an ARDI for their whole reactor coolant
system, because it's mandated by their chemistry control program, but,
you know, they have dead leg portions of the RCS or valves, or your
chemistry controls may not be working as well there, because there could
be a stagnant condition, and those are the places that they're going to
be looking in.
DR. POWERS: So I'm trying to understand the threshold,
where there is some possibility it could occur, but I believe it's not.
Now, the believe it's not, I think that's in the hands of the -- of
other people. That's a religious argument.
[Laughter.]
MR. HERMANN: Bob Hermann of the staff.
DR. POWERS: The credibility argument seems like it's
something quantitative.
MR. HERMANN: Some of the items that are on the ARDI list
are things that are probably in the class three list of ASME systems.
During the life of the plant, the only inspections that those -- the
normal life of the plant, the only inspections they get are a look at
for leakage. I've got things like service water systems, for instance;
one of the items that was on there was a breakdown of coatings, okay?
There's been a lot of leaks in service water systems because you have a
finite life of some of these coatings that are in the plant, and you get
leakage. So as part of these programs, they're sort of targeted to look
at areas that people have an idea that, you know, they really haven't
had a look at before, and it's something that's looked at.
I guess along the same line, there was a question about
acceptance criteria. I think if you found something, you would even go
back to the construction code, or you would use the evaluation criteria,
especially for the things that are covered under the regulations now for
section three, using acceptance criteria that are in 11 for what it has
to satisfy.
MR. BARTON: It's all of those systems and components; when
you risk-inform them, they say they're low safety significant and don't
mean beans, and I can apply industrial and commercial practices to them;
they start rusting and leaking. It's those components.
DR. POWERS: I mean, if they were doing that, that would be
fine. What I'm trying to understand is they're doing some work.
They're saying go inspect this one time and confirm that indeed, as you
suspect, nothing is happening. Presumably, that could be an infinite
set. So it's not an infinite set; it's a pretty big set, but it's not
an infinite set.
But what I'm trying to understand is where is the cutoff in
credibility where I say that the possibility that there is some
corrosion event going on here is so incredible that I don't want you to
inspect it? Because there are things that don't get inspected, and
there are things that do. Right now, I understand things have to be
wet, and there has to be anecdotal evidence that something is happening.
DR. BONACA: Well, my understanding is really, I mean, I
began to really be much more accepting of the ARDIs, one-time
inspection, actually, after I realized how much they were complementary
or supplementary of the set of additional inspection that you normally
do; for example, you go along piping, and you inspect the whole piping
for normal external pipe inspection.
But then, you have a dead space internally, some location
that makes it unique. But you say, well, because of that, may I have
some problem? And that's where you go to have one-time inspection to
verify or not verify that.
DR. POWERS: But, you see, even when you talk about it, you
come along and say, well, there has to be a dead spot, okay?
DR. BONACA: Yes.
DR. POWERS: What I'm looking at is how much of a dead spot
for how long it has to be before I say ah, there is a dead spot.
DR. KRESS: And are there other criteria aside from that?
DR. POWERS: Yes, anything else.
DR. KRESS: Yes.
DR. POWERS: I mean --
DR. KRESS: It's a reasonable question.
DR. POWERS: I'd like to see the whole set, but I'll take
any fraction of the set that exists. Or is it it's simply a whim? I
say, well, I need to make these guys do 43 inspections, and so, I find
43 of these, I mean. That's what I'm trying to understand.
MS. COFFIN: They don't -- I mean, they have a very specific
sampling program for -- in their license renewal, they systematically go
through every system, so they're looking at every system; they're
looking at every material environment combination that they have in that
system, and then, as part of their ARDIs, they have to set out what the
staff believes is a very aggressive sampling of those susceptible areas,
and if they find degradation, they certainly have to consider expanding
their inspection scope, considering generic, you know, implications.
DR. POWERS: I think I understand what happens if you find
something. A bunch of things that don't get inspected; there are a
bunch of things here that are inspected, okay? How do we choose those?
And what criteria put them in there, or is it just a matter of we need
to have 43 --
DR. SHACK: But BG&E is the one who made the decision.
DR. POWERS: I understand, but I'm still trying to
understand by what criterion.
DR. SHACK: We should have asked them.
DR. POWERS: What kind of changes does the staff expect?
DR. BONACA: Well, I mean, the application is approved,
insofar as you have a plant assessment, it seems to me, and then, what
you do, you're also looking at the materials and the potential aging
effects that they have. So you look at the full combination of those
elements there. I particularly want to challenge even those
combinations because there may be something else that may happen there,
but I believe that is beyond the intent of the rule.
MR. GRIMES: Dr. Powers, if I may, there is a fairly
vigorous and systematic cataloging of aging effects that are potentially
applicable to combinations of materials and environments, and for each
of the license renewal applications, the applicant goes through, and
they go through all of the potential or plausible aging effects that
apply to a system and then point to the programs that are used to manage
those aging effects.
DR. POWERS: I guess the word I'm trying to get a definition
on is you said potential and plausible. I'm trying to understand what's
plausible and what's not plausible. I mean, I'm a very credulous guy.
I can believe things are plausible maybe to a far lower level of
probability than Professor Wallis who teaches classes in these subjects.
He actually knows something as opposed to me.
MR. GRIMES: We've got over a decade of nuclear plant aging
research material that was pored over in developing a catalog of aging
effects and identifying which aging effects just don't happen and can be
discarded and which do happen or are plausible, so there is a catalog
of --
DR. POWERS: You have a catalog that defines plausibility.
MR. GRIMES: That is correct.
DR. POWERS: Okay.
MR. GRIMES: In NUREG CR 64.90, we essentially pored through
all of the NPAR data, the nuclear plant aging research data, and then,
we sat down, and we wrestled with the industry for several years trying
to decide which of those things are plausible or not. Now, not
everybody uses the plausible term; some say applicable; others say, you
know, that it still boils down to if you cannot -- if you don't have
evidence that says it can be completely disregarded, because there's
just no experience at all, then, it becomes one that's potentially
applicable, and then, you've got to decide what are you going to do
about it?
DR. POWERS: But I think it's more than just a plausibility
argument here. Basically, it becomes -- is it in our catalog or not?
MR. GRIMES: Correct.
DR. POWERS: If it's in our catalog, we check it no matter
how firmly convinced you are that it couldn't happen to your system, you
check it.
DR. KRESS: Well, the catalog then must include other
things. The phenomena that cause the aging must be present, but you
must have a catalog of phenomena to look for or something.
DR. POWERS: A tree or something like that.
DR. KRESS: A tree or something, and that's what I thought
you were asking for. What does that thing consist of?
DR. POWERS: I was looking for how they got there, and now,
I'm happy. Now, I don't want to press it any further, because I may be
forced to read the tree.
[Laughter.]
DR. POWERS: But, I mean, that's good enough for me. I
mean, that's a definition of plausibility somebody sat down and did, and
you've got to have 43 one-time inspections to qualify for license
renewal. That's what I wanted to know.
MS. COFFIN: Actually, though, what I prepared to talk about
today was a process for evaluating degradation identified by one-time
inspections, and this has as its regulatory basis 10 CFR 50, appendix B
and specifically criterion 16, corrective actions. And now, BG&E has
appendix B qualified corrective action program that they described in
their license renewal application and that the staff reviewed and found
acceptable, and this is described in Section 315 of the SER.
And this corrective action program applies to all aging
management programs for license renewal and not one-time inspections,
although I'm going to focus a little bit on what's kind of unique about
one-time inspections, and I just have one other slide. What BG&E did
was take their -- the elements of their corrective action program and
reiterated them in their specific technical guidance that they developed
for ARDIs, called engineering standard ES-045. I have the title right
there. And there are essentially four elements of your corrective
action program. You have to identify the degradation and put it into
your corrective action program. Then, you have to analyze or somehow
assess the situation. From that stems your corrective action; what kind
of options you have available for corrective actions. And then,
finally, the fourth part is confirmation that your corrective actions
were effective and documentation, and this is all part of appendix B;
this is in their corrective actions program, and they reiterate it again
in their specific technical guidance for their ARDIs, for their one-time
inspections.
And the second bullet there talks about what kicks off the
corrective action program, and essentially any degradation or corrosion
kicks off the program. It has to be formally evaluated by the licensee.
The third bullet that I have on there talks about the
analysis and the assessment and the corrective action phase. And your
analysis assessment generally consists of what kind of corrosion am I
talking about? What's the extent of it? Do I have a good root cause?
And from that, you can develop a corrective action plan. And your
corrective actions can vary a lot. And that depends on many external
factors, such as you can repair; you can replace; you can accept it as
is, and things that you would consider when you decide what to do depend
on, like, do I have a qualified repair technique? Do I have a
replacement component available? Do I have a robust engineering
evaluation that would allow me to accept it as is, and we leave that
decision up to the licensee.
The fourth step I actually don't have a specific bullet for
is the confirmation and the documentation. And the confirmation
requires, either directly or indirectly, a followup to your corrective
actions to demonstrate that you -- they were effective, and
documentation means that all of the steps that you have taken, all four
of these aspects that I've talked about, are documented and are
available for review by the staff.
And so, just to sum up again, the process for evaluating
degradation is not unique to one-time inspections; it's applied across
the board to all their aging management programs, and it's their
corrective action process, their appendix B qualified corrective action
program.
DR. BONACA: And I heard before that there are means for the
staff to be informed in the event of problems; they are necessary
because of identified issues and existing commitments.
MR. GRIMES: That is correct. We expect to follow up
through our inspection program and verify that. We also will have the
licensee provide us with the results of ARDI and some kind of feedback
mechanism that we can use to monitor how all of the renewal activities
progress over -- you know, they expect this to go on for a decade or
more.
DR. BONACA: The reason I was asking also is because you
have the GALL report that is being used to support some decisions. If
you have now a finding in that inspection, that will defeat somewhat
says the GALL report is incorrect or something which wouldn't see, and
we need to update the GALL report; the staff has to understand, and its
information has to be most likely disseminated among other plants which
have renewed their licenses. That was my concern with -- I understand
they have a corrective action program, effective or not it may be,
however, but there is a corrective action program. The question is what
is the involvement of the staff, and you're telling me the staff has
adequate information right now to be informed and participate.
MR. GRIMES: That is correct, and not just for these
one-time inspections but also, we target our inspection activities to go
out and look at experience, to look at corrective action processes in
order to develop insights that will improve the whole regulatory
process. We are not just learning about aging effects. GALL can also
be affected by program changes that are stimulated by other reasons. So
we expect generic aging lessons learned and the SRP for renewal just
like generic lessons in the normal license amendment process and the
changes to the SRP for the normal license process will get a feedback
from the inspection activities.
DR. BONACA: Okay.
MR. GRIMES: The next two items on the agenda relate to
commitment management. You've heard BG&E describe their commitment
management process. As I explained to the subcommittee, for the purpose
of developing a conclusion for a renewed license, the staff has focused
on identifying those particular committed actions that it relies upon in
order to develop its reasonable assurance finding, and we have received
a proposed list from appendix -- from BG&E which we would intend to
incorporate into the safety evaluation. That is the list of things that
constitute changes to the FSAR; things that have to be screened under
50.59 for potential license amendments, and then, we would codify that
through a license condition that was described in a model license that
we sent to Baltimore Gas and Electric.
50.71(e) guidance exists to identify how BG&E can use their
commitment management process to cull through all of the correspondence
and all of the statements to make sure that the FSAR is updated to
incorporate the appropriate level of detail for all of the things on the
list that we will put on to the safety evaluation. One other aspect of
these commitments concerns timing. A number of things are described in
the safety evaluation that are going to occur over time between now and
the period of extended operation, and after looking at those things, we
concluded that the license condition associated with further actions
should only require a license amendment if BG&E elects to try to extend
them beyond the current license term rather than to try to manage
schedules for intermediate actions in the interim.
And so, we're developing a license, a second license
condition that will address the timing aspect associated with the
schedules of activities, and the other question that you raised
regarding the adequacy of the guidance for 50.59 as it would apply to
license renewal, and I'd like Dave Matthews to address that topic.
MR. MATTHEWS: Just by way of a little history that you may
or may not be familiar with, the incorporation of summary descriptions
of programs and activities for managing the effects of aging and the
evaluation of time-limited aging analyses is required by 54.21(c), the
incorporation within the FSAR and specifically the FSAR supplement as
described in that requirement.
The stated purpose for that incorporation that was contained
in the statement of consideration for the license renewal rule was that
those changes would be subject to the control of 50.59, so that, of
course, raises the question of the appropriateness of the 50.59 criteria
for controlling such changes, and we pointed out in the statement of
consideration for the revised 50.59 rule that as you know is in the
process of implementation, or at least we're at the point where we're
developing guidance such that when the guidance is completed, we can
proceed to implement that revised rule, we indicated that those criteria
were applicable to changes to the aging management programs and TLAAs.
But we also observed that, you know, judgment is needed and
must be applied when applying those eight criteria to changes to a
programmatic summary description. So we asked in our comments to NEI on
the revised guidance for 50.59 for them to address this issue in the
additional guidance, so we'll be reviewing that revised guidance. So
I'm recognizing at this point the fact that because of the judgmental
quality of those evaluations when trying to apply criteria, and for
example, I'll give you one: one of the criteria is more than a minimal
increase in likelihood of malfunction of a structure, system or
component, for instance; that's a hard evaluation to define, let alone
circumscribe, when you're dealing with the change to a possible aging
management program. So we think additional guidance is needed there,
and we're hopeful that that additional guidance will be forthcoming as a
result of the effort we're undertaking with NEI at this point in time.
DR. BONACA: All right; that was the key reason, because I
went with some of those questions.
MR. MATTHEWS: Yes.
DR. BONACA: And since I used to do it for a living, I know
that some of them were not so clear.
MR. MATTHEWS: Yes; and obviously, by your experience, you
recognize this; the suitability of the criteria for this use under the
prior rule provisions, you know, are equally judgmental. So in that
regard, we may not have improved the situation dramatically with the new
criteria in this area of evaluating programmatic changes.
DR. BONACA: Yes.
MR. MATTHEWS: So I think some additional guidance is
necessary there.
DR. BONACA: Okay; thank you.
Yes?
MR. GRIMES: Are there any questions about the commitment
management or the regulatory controls that we've proposed for a renewed
license? As Dave mentioned, we're going through a dialogue with BG&E
now in order to make sure that we both have a mutually agreed upon list
of things that the staff relied on that will be controlled under 50.59
or screened for potential license amendments until such time as the FSAR
gets updated, and then, we will, you know, do our usual monitoring of
the FSAR changes in the future, and to the extent that we can develop
lessons and experience from that that can be used to improve the
guidance on a supplement to the -- the supplement for the license
renewal applications or improvements in -- future improvements in the
guidance for 50.71(e) as it might apply to aging management programs,
we'll continue to share that experience with you.
DR. BONACA: I don't think there are other questions on
that. But there is one more item on the agenda, right?
MR. GRIMES: That is correct; the last item on the agenda is
the basis for the frequency of ASME inspections. And we've asked Mr.
Hermann to address that question.
MR. HERMANN: Good afternoon. I'm Bob Hermann of the staff.
I guess the place to start with the frequency of the inspections rather
than give you the short answer that it was based on engineering judgment
is to go back to the design a little bit and how you got to classifying
the systems and the rest of it. Basically, there was a variety of
construction for the plants for especially the balance of plant systems
and even for the primary systems; a lot of different codes, et cetera.
When the operating part of the inspection program got in
place, it tended to be a program to do inspections; to evaluate what you
find in the inspections as flaw acceptance criteria and to have a
special rules if necessary for repairs and replacements.
Given that as background, when people put together their ISI
and reclassify systems in terms of ASME designations from class one,
class two and class three, those are basically tied to the definitions
of things like ECCS systems, like the definitions like the RCS for the
safety classes, class A, B, C and D. Class D systems are systems that
are basically balance of plant systems; safety class C systems there in
the reg guide are classified as ASME 3 for inspection purposes. Two are
basically things like mostly the ECCS systems, and class one systems are
things that are the reactor cooling system basically.
The frequency of inspections were similar to what the design
considerations were. When people designed the plant, the most design
fabrication rules and inspection rules were applied to the class one
systems; the was the RCS. Risk perspective; if you look at that today,
you may not come up with the same list, but that's how they were done in
the first place, and that's how they were classified for ISI purposes.
Class one systems basically were 25 percent of the systems
getting looked at over a 10-year interval; class two by volumetric
methods, mostly ultrasonics. Class two systems, which are feed water
and steam and ECCS systems, basically a 10 percent sample, say, for
piping over 10 years; and class three systems basically were things like
service water systems, which are low energy systems in most cases. Some
of them are very important to safety, though, like aux feed water, and
some of them aren't low pressure. But the way they ended up in
inspection requirements and what the sampling was for those systems were
visual inspections, and what's getting looked at now is when you start
getting in the risk arena is looking at changes to what to look at and
how to look at it.
In general, though, most of the risk programs, though,
haven't really added any additional systems in terms of getting
inspections done. What they're doing is crediting ongoing programs for
things like erosion and corrosion to address the risk questions.
DR. BONACA: Let me ask a question on that. I understand
that at the beginning, when it was being set, we had a staff proposed
5-year inspection.
MR. HERMANN: I don't recall if it was 5. A lot of the
things, when it originally got set up, I think they originally started
breaking things down like 3 and a third years and having a program that
kind of revolved to 3 and a third years within the 10 years; and what
had ended up as 10 years; but even though it's 10 years, the
distribution of the inspections over the 10 years; like with roughly 25
percent or so in the first third and a distribution.
DR. BONACA: Let me continue my question. My question was
simply because there was a lot of judgment being used, and part of the
reason why the 10 years was set was because of the complexity in doing
these inspections and the large number of those, and so, there was a
necessity to set them. Now, for some aging mechanism, like for the
environmentally assisted fatigue, we find that there is an accelerating
effect between 40 years and 60 years, and in fact, some -- BG&E has a
program to deal with this. To some degree, ISI is still --
MR. HERMANN: Well, I'm not sure I agree that there is an
increase in environmentally assisted fatigue over the last 40 years
versus the first 40, and I'm not sure where that comes from.
DR. SHACK: Well, I think it's a cumulative damage thing, so
you notice the effect.
MR. HERMANN: We may notice the effects.
DR. SHACK: You have to build up in cycles before you
initiate anything.
DR. BONACA: But the point I want to make, none of the ISC
and ISI program, it seems to me there are so many components, and many
of them are affected by different types of aging effects, and is there
any basis for saying that if you go to 60 years, if you go to 100 years,
you're still going to have to inspect more than once every 10 years?
That was my question.
MR. HERMANN: I think things like -- certainly, I have been
through a lot of machinations on fatigue as part of license renewal and
the rest of it. I think usage factors are being used as a screen, if
you would say, to look at increased things for fatigue. I think what
you heard earlier today on corrosion; if I looked at the ARDIs that were
in here, probably half of the components that are in there are things
that are in -- not half, let's say some large number of probably class
three systems that wouldn't have got looked at before; it seems to me
what's in license renewal is a one-time look to -- if I take something
apart or if I get on the inside of it to verify what your judgment is in
terms of maybe corrosion rates and the rest of it. I don't think
there's a better answer than that.
DR. BONACA: Okay.
MR. HERMANN: Radiation damage certainly is something you
look at in the vessels, like internals for shrouds on Bs and Ps are
going to be dependent on experience today.
DR. BONACA: So a few of the programs are supplement
sufficient through the ISI IST program deal with the increased aging.
MR. HERMANN: Yes, I think so, because a lot of areas like
-- let's think of the service water system, okay? What you probably
have now are supplementary programs to look at flow blockage, and some
of them, if you get into these systems, and you see flow blockage in the
systems from MIC nodules and things like that, there is probably a
pretty good bet you're going to end up with accelerated corrosions in
those systems, and you go back and look at them. So there are already
programs for that. There are programs to look at -- say, we haven't got
the boilers yet, but when you get to boilers, there are certainly
additional programs for IGSEC and IASEC; certainly additional programs
in place for internals, both on PWRs, barrel baffle former bolts.
Any other questions?
DR. SHACK: I think it's also true to say that when you've
identified a real degradation mechanism, and you're inspecting for
cause, your interval of inspection is really set on a different basis.
I mean, on erosion-corrosion, for example, you have a different
inspection for stress corrosion cracking.
DR. BONACA: It's different, yes.
DR. SHACK: Yes; for these other ones, where you're
basically inspecting to make sure there's nothing going on, it's a
somewhat more arbitrary kind of selection for --
MR. HERMANN: I think some of it may be the nature. I think
what Chris explained earlier, kind of the nature of the rule, you go
through, and you pick out a mechanism, and you catalog it; then, you go
back and say well, I'll take a look to see if it's there or not, because
it's at least to the point of being considered credible, although you
maybe haven't really seen it in operating experience.
DR. BONACA: I understand that, but I certainly would like
to see it treated more explicitly, simply because I can understand what
you're saying, and you're saying, well, there is an augmentation we've
made because we haven't talked specifically about why this specific
interval continues to be valid, and I think it's a good question,
because, you know, I mean, I know that my HMO is supporting an
inspection on me every 2 years now, and when I was 30 years of age, I
mean, they wouldn't support any inspection on me.
[Laughter.]
DR. BONACA: So there has to be some reason, some aging
taking place in these components, and we have to question the interval
that you're looking at.
MR. ELLIOT: Barry Elliot; I'm just going to interject here.
The 10-year cycle is -- as Bob said, it's arbitrary. But the point of
the 10-year cycle is to find mechanisms and find things that we don't
expect to happen. We don't expect the vessel to fail either, so we do
things just to make sure of it.
Now, for instance, you have a 10-year ISI. There are times
that during the regular life of the plant, things happen, like steam
generator tubes fail, and steam generator tubes have problems. They
aren't on a 10-year program. We have looked at it. Our experience says
that they can't sustain a 10-year program. They have to have a
different program.
We have conditions in high pressure injection lines where we
have high thermal fatigue cycles. We figured out a 10-year program is
not acceptable. And every case has to be looked at. This is an
experience thing based on the observed mechanisms and occurrences that
happen during the life of the plant. This is an every day for the
nuclear industry and for the NRC. We go through this, and the 10-year
ISI is just a piece of the inspection. After the inspection is done,
and something is found, then, the decision is made whether we can
continue on our 10-year, and in fact, the code, the ASME code says if we
find something, instead of waiting 10 years to inspect, we inspect that
the next cycle, just to make sure that there is no problem.
DR. BONACA: I understand; I only said i would have liked to
see it treated more comprehensively, the discussion of it, that would
convince me that in fact, this integration of all the things you're
saying, it's effective to deal with these issues, and it wasn't done, so
I see this -- you know, I thought we just assumed that we will continue
our stepping of the ISI/IST at the same pace, and then, we have to build
on it and convince ourselves that all of the other things we are doing
are supplementary and sufficient.
MR. HERMANN: Maybe one comment on the -- on random or what
I'll call random, unfocused inspections. Most of the time you see
operating failures in these plants, be it fatigue, be it whatever, okay,
it's generally not that there is a problem with the design rules for the
component; it's a problem that you didn't anticipate the loading in the
location where the problem occurs, and I think what you heard before was
Barry telling you about thermal fatigue. If you're going to design for
thermal fatigue, you wouldn't have a thermal fatigue problem, okay? The
reason you have the problem is because you don't know the load.
The reason you fail something sometimes in some locations
for corrosion is because it's a part of a system might be stagnant where
you really didn't anticipate the system being stagnant, or you probably
would have put measures in place to flush the system. So when you're
getting to the risk programs later, you're still going to save 10
percent of the inspections for kind of looking at things randomly to see
if anything happens, and you pick it up and so on.
MR. GRIMES: Dr. Bonaca, I'll agree with you to the extent
that one of my expectations in generic aging lessons learned as we go
through them now and try and extend our aging effects catalog to aging
management programs is an opportunity to reflect on how does each
program, how has each program evolved, and I like your analogy about the
HMO, because I think that it is a close approximation to say that when I
turn 40, the government reminded me every 4 years that I need to get a
physical, and then, when I hit 50, they suggested that I accelerate the
pace. But in the meantime, my optometrist has suggested that I ought to
go to annual; my general practitioner has suggested that if I keep
working on license renewal at this pace, I probably ought to do ulcer
checks every 6 months.
[Laughter.]
MR. GRIMES: So there's a for cause aspect that you use to
adjust that baseline, and I think that in other countries, they still
use -- they use a, you know, 10-year periodic safety review. There are
some countries that do cycle checks, but in all of these cases,
regardless of what the arbitrary scale is that you use, you have a for
cause reason to go in and say particular things have particular concerns
that are different frequencies, and perhaps as we develop GALL, we will
be able to give you a more systematic explanation of that across all of
the aging management programs.
MR. HERMANN: But I think as part of that, Chris, that it's
part of the development of GALL, what you are doing is taking the
operating experience and putting it in GALL, and you're putting other
programs in GALL, like the BWRIP program for internals.
DR. BONACA: Okay; any other questions from the members?
[No response.]
DR. BONACA: Thank you.
With that, Mr. Chairman?
DR. POWERS: Okay; what I want to do is go ahead and recess
for about 15 minutes, and then, I want to come back and talk to the
authors of letters for this session for a little bit and take our letter
writing period; come back at the end of that letter writing period and
try to get the 10 CFR 50.55 letter done and the Calvert Cliffs letter
done and then go back to the research report.
DR. KRESS: Are we through with the --
DR. POWERS: We can terminate the transcription.
[Whereupon, at 3:02 p.m., the meeting was recessed, to
reconvene at 8:30 a.m., Friday, December 3, 1999.]
Page Last Reviewed/Updated Tuesday, July 12, 2016