484th Meeting - July 12, 2001
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
+ + + + +
484th MEETING
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
(ACRS)
+ + + + +
THURSDAY
JULY 12, 2001
+ + + + +
ROCKVILLE, MARYLAND
+ + + + +
The Committee met at the Nuclear Regulatory Commission, Two White
Flint North, Room T2B3, 11545 Rockville Pike, at 8:30 a.m., George E.
Apostolakis, Chairman, presiding.
COMMITTEE MEMBERS:
GEORGE E. APOSTOLAKIS Chairman
MARIO V. BONACA Vice Chairman
F. PETER FORD Member
THOMAS S. KRESS Member
GRAHAM M. LEITCH Member
DANA A. POWERS Member
STEPHEN ROSEN Member
WILLIAM J. SHACK Member
JOHN D. SIEBER Member
ROBERT E. UHRIG Member
GRAHAM B. WALLIS Member
. I-N-D-E-X
AGENDA PAGE
Opening Remarks by ACRS Chairman . . . . . . . . 310
Draft Individual Plant Examination of. . . . . . 311
External Events Insights Report
Proposed Resolution of Generic Safety Issues . . 374
191, "Assessment of Debris Accumulation on PWR
Sump Pump Performance
Potential Margin Reductions Associated with. . . 423
Power Uprates
. P-R-O-C-E-E-D-I-N-G-S
8:31 a.m.
DR. APOSTOLAKIS: The meeting will now come to order.
This is the second day of the 484th meeting of the Advisory Committee on
Reactor Safeguards.
During today's meeting, the Committee will consider the
following: Draft individual plant examination of external events insight support,
status of resolution of genetic safety issues, GSI-191: Assessment of debris
accumulation on PWR sump pump performance; potential margin reductions
associated with power uprates, the reactor oversight process; future ACRS
activities/report of the Planning and Procedures Subcommittee; Reconciliation
of ACRS comments and recommendations; and proposed ACRS reports.
This meeting is being conducted in accordance with the
provisions of the Federal Advisory Committee Act. Mr. Sam Duraiswamy is the
designated federal official for the initial portion of this meeting. We have
received no written comments or requests for time to make oral statements from
members of the public regarding today's sessions. A transcript of portions of
the meeting is being kept, and it is requested that the speakers use one of the
microphones, identify themselves, and speak with sufficient clarity and volume
so that they can be readily heard.
First item, draft individual plant examination of external
events. I guess I'm supposed to lead you guys through this. Well, we had the
Subcommittee meeting on the subject. We discussed primarily the seismic and
fire-initiated sequences, and the staff is here to brief the full Committee on the
subject, and who is taking the lead on this?
MR. RUBIN: Good morning. My name is Alan Rubin. I will
try to speak loudly and clearly into the mike, as the Chairman has requested.
I'm the Section Chief --
DR. APOSTOLAKIS: With sufficient clarity and volume.
MR. RUBIN: Oh, that also, okay.
DR. POWERS: It's the clarity part you want to focus on.
MR. RUBIN: Thank you. We're starting off on the right foot.
Thank you.
My name is Alan Rubin. I'm a Section Chief in the PRA
Branch in the Office of Research, and what we're going to present today, me
being myself and John Ridgely, the following feedback -- first of all, let me
mention that I think all of the ACRS Committee members have a copy of draft
NUREG 1407 in two volumes. It's right here in front of me.
As feedback that we got from the Subcommittee on June 22,
the full Committee wanted to hear some more information about it, a broad
overview of the methodological issues and needs coming out of the IPEEE
Program, which I will discuss, as well as the technical issues associated with
the resolution of generic safety issues. John Ridgely will present discussion of
picking out four representative generic issues and going into more detail and
how they were closed.
Before he does that, I'll give a brief sort of an overview and
synopsis of the generic issue closure resolution process, which I think is
important as a little bit of background information. Before, however, John
Ridgely discusses the generic issues, I'll present one slide on the generic
overview of conclusions and further actions from the IPEEE Program.
Putting things into context from the last meeting, we're not
going to repeat this in today's full Committee meeting, but we presented an
introduction to the IPEEE Program, including the background and scope of the
Program, the Program overall objectives, and the IPEEE review process. And
then for each of the major areas, the seismic fire and the high winds floods and
external events, we presented a discussion of the vulnerabilities that licensees
had discussed in their submittals, summary of plant improvements that
licensees had planned or had already implemented, perspectives on core
damage frequencies, dominant contributors to risk from the various initiating
events, as well as model perspectives, which I'll go into in more discussion this
morning.
We also discussed briefly covering the 31 IPEEE-related
generic issues and sub-issues at the Subcommittee meeting. We also
presented a discussion summary of some of the uses of the IPEEE information,
how the IPEEE Program results have been used and will be used in the future,
as well as overall conclusions and observations.
DR. APOSTOLAKIS: This use of the IPEEE results is
puzzling. I mean most licensees use the bounding techniques, screening
techniques like five for fires and the seismic margin analysis for seismic and so
on. At the same time, I think the report correctly states that the IPEEE exercise
confirmed that earthquakes and fires are among the dominant contributors to
risk even though their analyses were crude and so on. Since that's the case,
shouldn't -- I mean most of the current IPEEE studies really cannot be used for
risk management, can they, because they're bounding analysis; they're not risk
assessments.
MR. RUBIN: Well, I think they provide more than that. They
provide insights for uses, for example, in where the dominant contributors are
for a plant. So when NRR is looking at inspection findings, areas to inspect at
the plant in the fire or seismic areas, they've used the results of the IPEEE
Program. We use the results in looking at the reactor oversight process as well.
DR. APOSTOLAKIS: On a genetic basis --
MR. RUBIN: On a genetic basis --
DR. APOSTOLAKIS: -- you're right. I mean there is
information there that is very useful, but if I want to manage the risk of a
particular plant and I don't have a PRA, how can I manage it? I mean all I have
is a screening analysis with some insights. I'm not saying it's useless, but it's
not really what one would call a PRA.
And that's a little puzzling, because it seems that people think
that it's worth doing a PRA only for the reactor power and for internal events.
For shutdown modes, for external events or anything that deviates from that
screening analysis or maybe a little arm waving is good enough. And I don't
understand that, I mean especially for the external events, also for the shutdown
modes.
I mean there is strong evidence that we have contributions to
risk that are comparable to those from internal events of power. So why this
reluctance? I mean I'm just asking you now. And shouldn't we be trying to
upgrade these studies? I mean as a first step, maybe this was a successful
Program, but now that we know more, maybe we should start slowly upgrading
those so that we can, first, have a good picture of what the risks are and,
second, managing them.
MR. RUBIN: I mean I agree. On a plant-specific basis, when
I get the methodologies --
DR. APOSTOLAKIS: They're always plant-specific, right?
MR. RUBIN: Well, when I get into discuss the methodologies,
I'll talk about ongoing and planned activities to address these methodological
issues to develop standards for an improved PRA for external events. And I
agree with you. But, still, I think that there are probably more uses of the IPEEE
information that can be applied. They discuss in the report -- I don't know if we
want to discuss --
DR. APOSTOLAKIS: On a genetic basis, yes, I don't disagree
with you -- useful insights.
MR. RUBIN: Yes.
DR. APOSTOLAKIS: But I mean even that, I think, is a
compromise to have to say I have useful insights. Insights usually mean you
haven't done a good job. When a research program says, "We gained
insights," you know they produced nothing, right?
DR. POWERS: It's a fundamental theorem.
DR. APOSTOLAKIS: It's a theorem that goes back to Euclid.
(Laughter.)
MR. RUBIN: In that case, I won't disagree.
DR. POWERS: No, he was Greek, wasn't he?
DR. APOSTOLAKIS: Sorry?
DR. POWERS: He was Greek, so it wasn't respectable.
DR. APOSTOLAKIS: Who?
DR. POWERS: He was Greek; he can't be respectable.
DR. APOSTOLAKIS: The discussion is deteriorating.
MR. RUBIN: We have a short time; let me continue. Before
I get into the specifics, a general discussion of the issues, I want to make a
couple of overall comments I think relates to Professor Apostolakis' discussion
and lead-in.
In terms of methodologies, it was understood that when the
IPEEE Program began ten years ago with a generic letter, 8820, supplement
4, that there were some limitations. And the state-of-the-art of PRA external
events was not as advanced as for internal events. And it was not expected
that licensees would go beyond the state-of-the-art in their IPEEE analysis. And
even with that caveat, we felt that with these limitations the Program was able
to accomplish its objectives in terms of licensees being able to identify
vulnerabilities with the plant-specific improvements and meet the objectives of
the IPEEE Program. Can and should there be improvements? Yes, there
certainly can be improvements.
DR. APOSTOLAKIS: Now, again, this identification of
vulnerabilities, as you state in the report, and I think we discussed it last time at
the Subcommittee meeting, human error rates were not handled very well. In
fact, the report clearly states that there is no strong technical basis for the
numbers that the IPEEE showed. They either took the IPEEE numbers, which
themselves are not the most scientifically derived numbers, and multiplied them
by various factors to adjust them to external events.
Now, given that we don't really trust the human error rates,
how can we trust the results? How can we be sure that the vulnerabilities have
been identified when human performance is an integral part of these accident
scenarios, through recovery actions, right, abandoning the control room and
doing things from the outside, and trying to stop the diesels aligning the turbine-
driven pump of the auxiliary feed water system.
Let's say that the human error rates are significantly
underestimated. Are there any vulnerabilities that have been missed?
MR. RUBIN: You're leading into a couple of questions that I
was going to cover, so let me answer them now, try to, about the seismic and
the fire area. Human error rates, you're right that in many cases human error
probabilities were taken as multipliers based on judgment from the internal
events IPEEE. In some cases, in the fire for main control room abandonments
in areas, there was more detailed analysis of the human operator actions, the
human performance.
When we saw, through our reviews that these human error
rates that were indicated in the submittals, were very optimistic, in other words,
where they did not take into account the effects of fire, smoke, heat, and stress,
or in a seismic if they did not take into account the timing and the location of
operator actions, we asked plant-specific REIs for each of those. And in some
cases, licensees upped their human error probabilities, typically did not find a
significant change in their overall CDF, core damage frequency, for those
plants.
And in the seismic -- let me give an example in the seismic
area. Typically, operator actions for safely shutting down the plant would not
be required till a half hour after an earthquake. And what the licensees did in
some of their analysis for actions that were required outside the control room,
their human error probabilities, although without detailed modeling, they took
the error probability of one, but not taking credit for operator recovery actions
outside the control room.
In some cases, when actions were required, there were
different multipliers depending on the level of the earthquake. Earthquake for
a certain G level, they took a certain human error probability as a multiplier of
the internal events human error probabilities. And for higher earthquake levels,
they took higher multipliers. For actions that might have been required, say,
more than an hour after the initiating of the event earthquake, in some cases
licensees actually took the human error probability from the internal events IPE.
So there were different -- you know, although they're not a
detailed human performance modeling, we try to let them see, based on our
reviews, the timing, how the effects of the fire or earthquake were taken into the
licensee's analysis. And we felt that for determined vulnerabilities and overall
core damage frequency estimates, licensees could come up with results that
would not be too far off base.
Are there areas that look for improvements? Yes. In fact, in
the fire area, part of the Fire Risk Research Program is looking at human error
performance as a result of fire. And that information is going to be factored into
a fire risk requantification study, which is part of the Fire Risk Research
Program.
That's kind of a brief summary. I was going to get into a little
bit later, so I'll save time later on. I won't have to repeat myself. But that's sort
of the outline of the seismic and fire perspectives of how human performance
was treated in IPEEEs, why we felt there were some limitations but why we also
felt that it wasn't that bad for IPEEE purposes but could be improved. And,
again, realizing that we didn't expect the licensees to go beyond the state-of-
the-art, we really couldn't pursue it too much further for IPEEE purposes. I
know that's one of the subjects that we wanted to get into at this full Committee
meeting was the human performance and the methodologies.
DR. APOSTOLAKIS: Well, yes. The spirit of my question
was not really why didn't you do more, because you can't really do more, but I
mean the state-of-the-art is relatively weak.
MR. RUBIN: But we are planning to do more.
DR. APOSTOLAKIS: Okay. But right now, I mean there is
this doubt that all the vulnerabilities have been identified, because humans --
you know, we really don't know how to model that very well. And also the
impact of smoke, in general, not just on humans, we don't handle that in fires.
MR. RUBIN: Well, there's empirical data in the fire area. For
example, time to recover or suppress a fire manually, which does take into
account, not in the modeling, but indirectly the effect of smoke. And there are
actually some curves at times to recovery for fires in different areas of the plant.
And we looked at that in our reviews, and we saw that if an operator -- if a plant
was taking too much credit for suppressing all fires in ten minutes, that's crazy,
that's ridiculous, and we pursued that with the licensee. That was a very
optimistic assumption. So we pursued that to make sure that a vulnerability was
not missed, because there was too much credit for manual suppression, for
example.
DR. POWERS: When will you look at the results, particularly
for, I believe, at Susquehana, we see that they are remarkable in their lack of
vulnerability to fire. And I have been told, though I can't produce evidence, that
some of that lack of vulnerability they have to fire is due to extraordinarily high
levels of performance by the fire fighting capabilities. Did you not find that
overly optimistic or look that in any detail?
MR. RUBIN: We looked at it in great detail. In fact,
Susquehana initially came in with a ten to the minus ninth core damage
frequency estimate for fires. That's one of the four plants we went to and did a
site audit on to see what was going on. We knew that they had an extremely
low probability of core damage from internal events also for the similar reasons
-- high optimistic expectations on human error probabilities.
The interesting thing we found with our walkdown, even
though the Plant did not identify vulnerabilities, they made some procedural
improvements in areas where transient combustibles were, they felt, potentially
could contribute to higher risk, and they added some procedural modifications
to the Plant, even though they didn't fine, quote, "vulnerabilities."
But, yes, we pursue Susquehana. We believe the overall
whole quantitative number now is in the order of ten to the minus seventh fires.
You can question that one. It's the lowest one of all the submittals, I believe.
But there were some insights that were gained, and as a result of our review
and audit, some adjustments and some improvements.
Let me go into methodologies. I'll try to be brief. I do want to
leave some time for the generic issues portion. I have a lot to cover. So I'll
briefly touch on the issues, and if you want to hear -- if I'm going too much into
each one of them, because of time, let me know.
In the seismic margins assessment versus the seismic PRA,
both of these methodologies were acceptable approaches for identifying
vulnerabilities and meeting the intent of the IPEEE Program. Forty percent of
the submittals used a seismic PRA, which did enable them to come up with an
estimate of core damage frequency, a list of dominant contributors, as well as
a plant level fragility curve.
The seismic margin assessment, which was used by the
remainder, about 60 percent of the licensees, that was done, and the question
from the Subcommittee meeting was why did so many licensees use the
seismic margins assessment instead of a PRA, which would give more risk
insights. In part, because with the resolution of USA-46, the methodology for
a seismic PRA was very consistent with that methodology.
The IPEEE Program went beyond the A-46 in terms of scope,
but the methodology was similar, and licensees chose to integrate those two
programs together. And they came up with a list of critical components and
estimated plant and component, high confidence of low probability, of failure
capabilities, and an assessment of the margin beyond the design earthquake
but no explicit core damage frequency estimates for the seismic margins
analysis. And, therefore, there are some limitations in how you can apply risk-
informed activities as the result of the seismic margins and assessment.
And there's an ANS standard on external events that has
been issued in draft form that cover seismic events. And that standard includes
both methodologies, the SMA, seismic margin assessment, as well as the PRA,
with the acknowledgement that there are some limitations on the seismic
margins analysis for risk applications.
I think I said as much as I wanted to say on the human error
probabilities for seismic events in my introductory comments.
Surrogate elements is another area in the seismic analysis,
seismic PRAs that were used. This was sort of a short cut methodology where
groups of components are combined together into a surrogate element, which
can be screened from the analysis. And this methodology is acceptable if the
screening, one, is done in a high enough level, and obviously the surrogate
element is included in the plant logic model.
And if the surrogate element does not come out to be a
dominant risk contributor, pretty much that's what you'd like to see. But in
seven of the 27 licensee submittals that used a seismic PRA, the surrogate
element did come up as a dominant contributor. However, the total core
damage frequency for those seven plants was in the low ten to the minus fifth
range. So although the surrogate element could mask what the actual dominant
contributor is, in this case the overall core damage frequency was on the low
side. And this is an issue that is discussed and clarified more in the external
event PRA standard that the American Nuclear Society is coming out with.
The use of uniform hazard spectra and simplified fragilities
and soil evaluations, I'll just briefly touch on. There is acknowledgement that if
you take the uniform hazard spectra and don't anchor it correctly at the right
frequency ranges, you can come up with low seismic demand. We ask
questions in our IPEEE reviews; in fact, some licensees revised their seismic
analysis based on the questions. And this use of uniform hazard spectra is
discussed in the ANS standard as well as a draft reg guide that the staff is --
revised draft reg guide 1.60 that the staff is in the process of developing.
Simplified fragilities, that was a point that was brought up, I
think, by Dana Powers, at the Subcommittee meeting. There's a comment in
the report that this can mask contributions to core damage frequency. And just
to go back and reemphasize, we feel that more of the uncertainties in the
seismic PRA come from the uncertainty in the seismic hazards rather than the
fragilities. Although fragilities can be masking it, more uncertainty comes for the
core damage -- for core damage frequency, it comes from the seismic hazard
curve.
DR. POWERS: I think if you're interested in what the absolute
risk of the facility, that's the way to look at things. The seismic hazard itself is
the dominant uncertainty. But a plant manager can't do anything about the
seismicity of his area. He can do something about the components within his
facility that are most susceptible. And if you use simplified fragility curves, then
he really doesn't know which ones to do anything about if he's moved to do
anything.
MR. RUBIN: I don't disagree with that. You mentioned
something that's interesting from the last comment on soil. Soil's evaluation is
that this was typically -- it was not required for many plants that fell in the
reduced or focused scope category for IPEEE purposes.
For those plants that did perform the soil liquefaction analysis,
it was an area which if there were improvements, it would be very difficult to
show that it would be cost-effective, and an area that there is uncertainty there's
no consensus on the best approach to use for liquefaction-induced soil
displacement, but it is also an area, a topic, that's discussed in the ANS
standard.
Seismic area, I'll conclude with this slide. These are two
industry and NRC activities that are ongoing. As an external events PRA
methodology, it covers seismic events and also high winds, floods, and other
external events. It does not cover fires. There was a draft that was issued for
public comment in December of 2000. And I would mention that each of these,
both ANS standard and the revised reg guide 1.60, take lessons learned from
the IPEEE Program, try to incorporate those lessons into these standards.
The revised reg guide 1.60, there are two NUREG CR reports
that will be published shortly, one on the seismic hazard analysis and one
specifically on procedures for using uniform hazard spectra. A draft reg guide
is in the process, internal staff review process, and it will be issued for public
comment.
Now move on to the fire methodological issues. And, again,
each of the areas discussed here are pretty well covered and addressed and
following ongoing activities that I will discuss. The five versus the fire PRA, we
discussed that briefly at the Subcommittee meeting. Only about 20 percent of
the submittals used a straight five analysis. Eight percent of the licensees used
either a five or some combination of five and PRA or just PRA.
And, in general, we found that both methodologies yielded
similar results in terms of what the dominant contributors were to risk, although
five submittals seemed to result in a somewhat higher total core damage
frequency estimates. And there are some aspects of five that could use some
enhancements, such as screening multicompartment scenarios, main control
room abandonment, areas and treatment of some of the fire risk scoping study
issues.
The Fire PRA Implementation Guide was a guide that had
been issued by industry, by EPRI. Many licensees use that Guide, and in our
staff review, we sent out 16 generic requests for additional information or
questions to address IPEEE's specific related topics. Industry came in and
resolved those REIs for IPEEE purposes, and we understand that EPRI plans
to reissue a revised guidance in the future.
Human error probabilities, actions, and recovery actions I
talked about already. Discussing severity factors, severity factor methodology
is a simplified approach instead of using more detailed analysis on the
frequency and magnitude of fires. One of the issues -- the approach can be
used, but one of the concerns is if licensees tend to double count using severity
factors where they could take credit for suppression in addition to a severity
factor which can be used to reduce the magnitude or frequency of large fires.
We asked questions when this was done in our IPEEE reviews, and NRC's Fire
Risk Research Program is a task to look more at the severity factors and
analysis of treatment of large fires.
Circuit analysis is the next topic on here. This, perhaps, is
probably the number one topic in the fire risk -- NRC's Fire Risk Research
Program. There certainly were some state-of-the-art limitations in the analysis
of circuit failures in the IPEEEs -- the frequencies and likelihood of hot shorts
and multiple hot shorts. These are being addressed in ongoing NRC and
industry activities on fire-induced circuit failures. There are some tests and data
and industry activities going on in the Cooper Program, which is the cooperative
PRA Program that the NRC is participating with and sponsored with a number
of foreign countries.
The fire modeling area is the next one. Let me bring back a
point that, Dana, you also brought up at the Subcommittee meeting in terms of
multi-zone fire scenarios were not large contributors to core damage frequency.
And you referred to a quote in the draft NUREG 1742 that said, "Contribution
of these scenarios to overall fire-induced CDF range from one percent to about
30 percent." And the question was, well, 30 percent of a large number could
still be a large number. We went back and looked at the plant that had the 30
percent contribution from new scenarios, and that total core damage frequency
of that plant was around three times ten to the minus five. So the overall multi-
compartment contribution was around seven times ten to the minus six.
DR. POWERS: Which is not a trivial thing.
MR. RUBIN: Not trivial, but it's in there. But it's not something
you'd necessarily --
DR. POWERS: And the question you had is what about those
that chose to analyze their fires, assuming that all fire penetration barriers were
100 percent effective as a ground rule without putting a failure probability into
them and how much have they missed? Based on that example, you would say
they may well have missed a lot.
MR. RUBIN: Well, there are other examples, and, true, there
were some submittals that did assume 100 percent barrier failure reliability.
DR. POWERS: Most, in fact, assumed it --
MR. RUBIN: Yes.
DR. POWERS: -- 100 percent barrier reliability.
MR. RUBIN: But there were some -- when we did look at the
impact of multi-compartment of failures, it was not a dominant contributor. It
was --
DR. POWERS: I mean the question is dominant. I mean you
come back and tell me, "Well, it's seven times ten to the minus six out of three
times ten to the minus five." That strikes me as something that is something
that I would want to know.
MR. RUBIN: For most of the plants, of those plants that had
contributions from that, it was closer to ten to the minus seventh contribution.
DR. POWERS: Yes. Well, the question is who's right? And
I mean it may well have been that the one that gets the 30 percent number, they
were overly conservative. But it clearly is something that merits some attention,
it seems to me.
MR. RUBIN: It does merit some attention. It's not being
ignored. It is a task in the Fire Risk Research Program that's being looked at,
both the multi-compartment scenarios --
DR. POWERS: Well, let's hope that nobody uses this rather
optimistic view of these results as a basis for ending that research.
MR. RUBIN: I don't think they will.
DR. POWERS: I bet they will if they have a chance.
MR. RUBIN: Continue on with electrical panel fires. Just two
comments I wanted to make regarding this subject. One, the panel fire issues
are related to the heat release rates given a panel fire, and this is an area which
was under discussion with the Fire PRA Implementation Guide. When
licensees tended to use low heat release rates, we asked some questions. The
other area is some limitations on the analysis of this topic on energetic faults for
high energy electric panels that could potentially cause damage external to the
cabinet. And this is an area of topic that's also included in NRC's Fire Risk
Research Program.
Fixed detection and suppression relates to use of severity
factors and the timing in fire detection and suppression. This is also a topic
included in the Research Program. There's a draft report on this that's been
issued. It's been out for comment. It includes empirical data on reliability of
suppression. And that report is an update of some of the work that was done
earlier at UCLA.
Analysis of self-ignited cable fires, many submittals and
licensees are going along with the five methodologies screened, self-ignited
cable fires when they use IPEEE 383 qualified cables. When these cable fires
were treated into submittals, the contribution to core damage frequency was
generally small, but the topic is also included in the Research Program, the Fire
Risk Research Program.
This is a list of the activities in the fire area, both NRC and
industry activities. The Research Program that I mentioned had been
presented. A draft plan was presented to the ACRS Subcommittee on fire
protection, and the full plan was sent to the ACRS. Tasks include areas such
as the fire risk assessment methods development, fire modeling benchmark and
validation, which is an international effort that the NRC is participating in, and
the fire risk requantification study, which is a joint effort between NRC and
EPRI.
In terms of human reliability analysis, there's additional effort
to look at the quantification, the assessment of human performance following
a fire, and that effort will be factored into the fire requantification study.
I mentioned the supplemental guidance at EPRI issued on the
Fire PRA Implementation Guide. And in addition, finally, there's an ANS,
American Nuclear Society, is developing -- will be developing a standard PRA
on fire that will provide a more detailed analysis for PRA, and I guess it will
supplement the NFPA 805 standard, which does not go into the level of detail
in a fire PRA.
MR. ROSEN: Alan, when you talk about human reliability
research from fires, are you talking about operators or fire brigade personnel or
both?
MR. RUBIN: More operators. It's performance of the human
operator recovery actions in the event of a fire.
MR. ROSEN: But not of fire brigade personnel?
MR. RUBIN: I don't think so, but we could check on that.
Nathan Su is involved directly with that program. He's out on leave today. We
can get an answer to that question. But I think it's more of the recovery actions
given a fire and the actions to safely shut down the plant. It may also involve
the actions to suppress a fire. I'd have to double check.
MR. LEITCH: Do you take a look at any recent actual events
to see how operators performed to see if your conclusions are reasonable?
There's been a couple of kind of interesting fires in the past six months -- San
Onofre, there was one a couple weeks ago at Cooper.
MR. RUBIN: Yes. Typically, those have been transformers,
which are fires that have caused serious economic consequences to the plant,
long shut downs. But there has been a draft report issued on significant fires
that occurred internationally -- I'd have to check on when that's coming out --
looking at human performance, the extent of a fire, and what you can glean from
those events, in terms of analysis into a fire PRA, things that might not be
factored into the fire PRA. Large turbinal fires, for example, is one thing that --
an area we can get large fires, there's lots of combustible sources for fire. And
we've seen events -- in the IPEEEs, we've seen where those can be large
contributors to the plant risk from fires.
MR. LEITCH: Even though the balance of plant is involved,
it may give insights as to operator performance under those kind of stressful
conditions.
MR. RUBIN: Yes. I think in those recent events the plants
were shut down safely, but I don't know how much detail is being looked from
the fire perspective on that. But there is a report looking at large fires, internal
plant fires, probably ten or 12 large fires.
MR. ROSEN: Alan, you promised me an answer to my
question about operator fire brigade performance.
MR. RUBIN: Yes. John, will you take a note on that? Or if
somebody would working with me.
MR. ROSEN: But I'd like to have the answer that fire brigade
performance is also being looked at, because I think it's a very important piece
of the plant's response to a fire. And, typically, brigades are fairly effective.
MR. RUBIN: Okay. We'll get back to you on that.
MR. LEITCH: Stephen, are you referring to on-site fire
brigades or are you talking about off-site response?
MR. ROSEN: I'm talking about on-site.
MR. LEITCH: On-site.
MR. ROSEN: Yes.
MR. RUBIN: We understand the question, and we'll get back
to you. We could have a quick answer if we had Nathan Su here.
As an introduction to the next presentation by John Ridgely
on the generic issues, I think it's worthwhile just to go into a brief summary of
the processes, how the issues are resolved. There is a new process in the draft
management directive 6.4 resolving generic issues, but this is a process for the
IPEEE-related issues, as well as other generic issues.
Once an issue is identified and prioritized on a generic basis,
it goes through a resolution process. And there are various ways that an issue
can get resolved. One, an issue can be resolved because there's an
assessment that it's a low safety significance. There could be a high analysis,
cost/benefit analysis that the cost/benefit is high, or it could be transferred to
another program, and the number of issues that were addressed and
transferred to the IPEEE Program for plant-specific validation was how some
of these generic issues were resolved.
The next phase is the imposition for the resolution. In this
case, for IPEEE, this was the identification of theses issues in Supplement 4 to
generic letter 8820 where licensees were asked to address a number of generic
issues. Some issues were also included in the IPEEE Program for validation
and verification that were not specifically identified in the IPEEE Program
generic letter and guidance, and those were discussed at the Subcommittee
meeting.
Implementation is done by the licensees' plant-specific
analysis. Verification is done by the NRC's review of those IPEEE submittals
and analysis.
The issue listed under this first bullet -- and I won't go read
them all -- these are IPEEE issues that are already considered resolved. And
the IPEEE review is for verification on the plant-specific basis, and no further
generic action is required with these issues.
The fire and scoping study issues is a little different. This was
not a formal part of the Generic Issue Program, but a number of the fire and
scoping studies issues actually did become generic issues. That was GSI-57,
the effects of fire protection system actuation on safety-related equipment;
Generic Issue 147, which was fire-induced alternate shutdown, control room
panel interactions; and GSI-148, which was smoke control and manual fire
fighting effectiveness.
There is one issue near and dear to the ACRS, which is the
MS -- Multiple System Response Program, GSI-172. This issue is still
considered open in the generic issue process, although in the IPEEE reviews
and in our review we looked in 80 percent of the plants. We've verified have
adequately addressed the aspects of this issue. Typically, commonly, when a
generic issue can be resolved, there isn't a look at all 100 percent of the plants.
It can be resolved in a much more limited basis taking typical plant designs and
addressing those plant designs for resolving an issue. But this issue is one that
is going through a generic issue process. There will be a package and formal
resolution that will be submitted for ACRS review in accordance with the generic
issue procedures that are in place.
So I just wanted to comment that in terms of the generic
issues or a letter from ACRS, we're not necessarily looking for a letter to
address the generic issue aspects of the Program. The MSRP issues will be
followed up later on towards the end of this year.
DR. APOSTOLAKIS: When does the public comment period
end?
MR. RUBIN: On the --
DR. APOSTOLAKIS: You have issued the report, right?
MR. RUBIN: Yes. I'm going to get to my conclusions this
time, because at the Subcommittee meeting, John originally took too much time,
and I didn't get to my conclusions slide. So I want to get to that first before John
gets up here.
You've heard this before, and it's in the report, we feel the
Program has been successful in meeting any intent of the generic letter. The
IPEEE Program has verified resolution of a large majority of the generic issues
on a plant-specific basis. Any follow-up actions on the plant-specific issues will
be addressed separately from the IPEEE Program. Public comments are due
on the report July 31, at the end of this month. To date, we have not received
any public comments. We expect they'll come in at the end of the month, in a
couple of weeks. And our schedule is to issue a final report in October of this
year. And that concludes my presentation unless there are further comments
or questions.
DR. APOSTOLAKIS: So this Program -- I mean, again, the
connection with the GSIs.
MR. RUBIN: Okay.
DR. APOSTOLAKIS: You are not resolving any GSIs right
now. You may be using some insights from here to do it later, but --
MR. RUBIN: Formally, except for the MSRP issues, they're
considered, quote, "resolved" in the generic issue process. There's no formal
generic activities that are needed to resolve that. The IPEEEs were a
verification that on a plant basis they were addressed. So on a plant basis, if
there's some follow-up activity, that will be done separate from the IPEEE
Program. The MSRP issue, however, is going through a more formal resolution
process and a resolution package.
MR. LEITCH: Did I understand you to say that GSI issues are
typically considered resolved or closed when, in many cases, the number is less
than 80 percent of the plants have complied? Is that what you said? I wasn't
sure --
MR. RUBIN: That's what I said, yes. When a generic issue
comes up, it's not addressed on a plant-specific basis, generally, for the
resolution of generic issues. And Harold Vandermolen is here. He can correct
me if I say something incorrect. But do you want add something?
But, typically, we might look at a cost/benefit analysis on a
number of plants or plant types -- BWR, PWR -- different containment types,
different designs -- Westinghouse, B&W plants -- and then see if there is a trend
or pattern. And if the trend or pattern shows that there's low risk or high
cost/benefit, then a decision can be made that it's really not worth the limited
resources from the staff to pursue it further, and the issue can be closed. So
I think on the IPEEE Program, we're going into a lot more depth and detail in
actually verifying the resolution of an issue than is typically done.
Harold, you might want to add something on that.
MR. VANDERMOLEN: This TV set over my head is
supposed to add to clarity of -- I'm Harold Vandermolen. I run the Generic
Issues Program in the Office of Research. Just amplifying what Alan has said,
for generic issues under the original or classic process that these issues still fall
under, yes, an issue is resolved -- generic issues are resolved by taking a
regulatory action; that is, they're intended to look for whether or not we need to
change our regulations or the reg guides and other documents that go along
with them, that sort of thing, going all the way to verification of things happening
in the plant.
We don't do -- well, first of all, going almost to enforcement
actions, that's not something we do in the Office of Research, nor is it
necessary or considered necessary to do this for 100 percent of the plants for
all issues. That's more of an enforcement action that we would leave to our
compatriots in the other building. That's really how it works.
At that resolution stage, once the agency has decided what
it's going to do, then the action is handed over to another office. It doesn't mean
nothing happens, but it's no longer directly part of the Generic Issues Program.
We'll still be aware of what's going on. I hope that's answering your question.
MR. LEITCH: Yes. Thank you.
MR. RUBIN: Okay. We'll get into more specific generic
issues with John Ridgely.
MR. RIDGELY: Good morning. My name's John Ridgely. I
work in the Office of Research. I work for Alan Rubin. And once again, I see
that I have been given an abundance of time for the few slides I have, so I will
need to proceed through at a leisurely pace.
DR. APOSTOLAKIS: We have to wrap this up by 9:45.
MR. RIDGELY: Yes, I know. The generic issues were
identified -- some were identified in the generic letter that went out and in our
NUREG 1407. They specifically identified these five issues that the licensees
were supposed to address. The licensees' submittals, the NRC believed, would
be adequate if they provided a good submittal to resolve other generic issues
which were not called out. And these four issues are GSI-147, 148, 156, and
172.
The process for reviewing the submittals and evaluating them,
the staff took the position that the licensees' IPE, if it was complete with regard
to the unresolved safety issues and generic safety issues coverage, and the
licensees assessment would demonstrate an in-depth knowledge of the
external events aspects and the plant characteristics related to the issues
discussed. The licensees' assessment results are reasonable given the design,
location, features, and operating history of the plant, that this would verify that
they had indeed done all they need to do with respect to these issues.
DR. POWERS: The definition of in-depth here, is it satisfied
by a bounding and scoping analysis?
MR. RIDGELY: I think the answer is yes to that, because
usually with a bounding analysis they're using higher numbers than they would
be if they were trying to do a detailed or realistic analysis. And so the answer
would probably wind up with, for example, like a higher CDF than they would
otherwise get. And if using that bounding analysis and the higher CDF is still
low enough, then it should be adequate.
DR. APOSTOLAKIS: What's the difference between a USI
and a GSI?
MR. RIDGELY: Harold?
DR. APOSTOLAKIS: Oh, I thought it was a trivial question,
but it's not. We need an expert.
MR. VANDERMOLEN: The difference, basically, is that all
USIs are GSIs, but the reciprocal is not true. USIs are generic issues that are
considered exceptionally important, which has never been defined in terms of
anything specific. However, the existence of USIs is actually written into the
legislation that created the agency. So that's why you will always find a USI
program here.
DR. APOSTOLAKIS: I still don't know, Harold.
MR. VANDERMOLEN: Well, we have struggled with that
question ourselves. If it is exceptionally important, we would --
DR. APOSTOLAKIS: It's a USI.
MR. VANDERMOLEN: Then it is a USI as well as GSI.
DR. APOSTOLAKIS: So USIs are much more important then.
MR. VANDERMOLEN: Yes.
DR. POWERS: Harold, just tell him what the USI stands for.
MR. VANDERMOLEN: Unresolved safety issue.
DR. POWERS: As opposed to a generic safety issue.
DR. APOSTOLAKIS: See, I don't understand that.
DR. POWERS: Well, it's like beauty -- you'll know it when you
see it.
MR. RIDGELY: I may save a copy of the transcript as a
reference of that. That's a pretty good definition.
DR. APOSTOLAKIS: The Agency's striving to resolve both
USIs and GSIs.
MR. VANDERMOLEN: Yes.
DR. APOSTOLAKIS: And we are giving more importance to
the USIs? In what way?
DR. POWERS: By giving them the USI label rather than the
GSI label. That's what does it.
DR. APOSTOLAKIS: Mr. Duraiswamy, you wanted to say
something?
MR. DURAISWAMY: They are safety related. For example,
I can give a USI --
DR. APOSTOLAKIS: Both of them are safety related.
MR. DURAISWAMY: No, but they have more impact on the
safety than the -- some GSIs might not have that impact like USIs.
MR. VANDERMOLEN: Yes. The definitions --
DR. APOSTOLAKIS: Right, but how does that affect our
actions? I mean these are --
MR. VANDERMOLEN: Functionally, the difference is not as
significant as it once was. These definitions pre-date the introduction of PRA
techniques into the process. They were devised before we had measures --
they actually came about the time of WASH 1400. Now, when we use
quantitative techniques to evaluate the importance of these issues in that
quantitative sense, it's not as important to make this sort of distinction.
DR. POWERS: There's an item of history towards the -- the
USI, as Harold said, is -- it's actually in the Atomic Energy Act. It says that the
Agency will address those things, especially with its research program. The
GSI list is actually an invention of the ACRS, and it used to be that the generic
safety issues list was kept by this body. And the staff started keeping their own,
and then the decision was made to meld the two, and the GSI Program went on
from there.
DR. APOSTOLAKIS: Was there a director of some division
at one time --
MR. VANDERMOLEN: Yes, there was.
DR. APOSTOLAKIS: And he resolved safety issues?
MR. VANDERMOLEN: There was once an entire division that
worked on these.
DR. APOSTOLAKIS: You know, this Agency speaks about
risk communication all the time and makes blunders like this. I still remember
years ago David Dawkins was sending a letter here to somebody, and the
secretary typed the envelope at UCLA to the Director of Unresolved Safety
Issues, Nuclear Regulatory Commission, and the woman was panicked. She
said, "What does it mean that you have unresolved safety issues in the nuclear
business?" I mean can you imagine how a third person outside sees these
things and this terminology, "unresolved safety issues from the Nuclear
Regulatory Agency."
DR. POWERS: You need to write to your congressman, not
to this staff.
DR. APOSTOLAKIS: Well, this is a good time to say these
things.
MR. VANDERMOLEN: There's little we can do to change it.
DR. APOSTOLAKIS: I know, I know; it's not up to you. It's
really ridiculous terminology.
MR. RIDGELY: Well, the importance of this --
DR. APOSTOLAKIS: Unresolved issues, though, they have
to be safety issues.
MR. RIDGELY: Yes. Thank you. The importance of this slide
really is just to show the list of the different that was looked at and how they
relate to the generic issues and how a particular issues crosses into several
different generic issues. And the ones I want to talk about today are the four
that are highlighted there.
First one is the unresolved safety issue, A-45, shutdown
decay heat removal requirements. The objective of this issue is to determine
whether the decay heat removal function is adequate and whether cost-effective
improvements could be identified. The components that were needed for this
were identified in NUREG 1289, and the internal events aspects of this in the
PRA was performed for the IPE, and that's documented in NUREG 1560. The
IPEEE was considered with how the external events could adversely affect the
decay heat removal capability.
In the seismic findings, we found that the seismic PRAs
included this decay heat removal systems and components in their PRA. The
seismic margin analysis included the systems and components as part of their
safe shutdown equipment list. And for the equipment on this list, they
developed a high confidence of low probability of failure values, and this was
compared with what they were assigned by a plant in NUREG 1407. They did
seismic walkdowns, and this walkdown information was used as part of this
review.
Weaknesses were identified in different plants, such as
weaknesses in RHR heat exchange or anchoring. And plant improvements
were implemented to resolve these issues. No vulnerabilities were found.
DR. POWERS: That statement always -- and it shows up in
a lot of view graphs -- always kind of interests me. We don't have a definition
of what a vulnerability is --
MR. RIDGELY: Right.
DR. POWERS: -- except everybody can choose their own.
The licensees find weaknesses, but they don't find vulnerabilities.
MR. RIDGELY: Right.
DR. APOSTOLAKIS: They also find anomalies.
DR. POWERS: So I mean what significance are you
attributing to no vulnerabilities were found?
MR. RIDGELY: Well, I think the answer to that is that the
licensees have not identified a component or system or what not that plays a
large role that could adversely affect the plant's abilities, in this particular case,
to remove decay heat.
DR. POWERS: What you're saying is though weaknesses
were found, the system probably would have worked just fine.
MR. RIDGELY: Probably.
DR. POWERS: And that this is just a little extra assurance is
what you're saying.
MR. RIDGELY: Because what we're looking at in the IPEEE
is beyond the value basis.
DR. POWERS: For those institutions that found weaknesses,
did you or somebody in the staff go back and rigorously go through the system
and say, "Yes, they found them all"?
MR. RIDGELY: Well, what we did was the staff looked at and
the Senior Review Board looked at it to see if there was anything that stuck out
as something that they might have missed.
DR. POWERS: How about Project Manager?
MR. RIDGELY: The NRR Project Manager was also part of
the review process as being usually in attendance during the Senior Review
Boards.
DR. POWERS: So he may well have thought about it or at
least looked at it.
MR. RIDGELY: And so at least in a number of the meetings
that I went to contributed to the staff's understanding of the plant and how it
worked.
DR. WALLIS: You know, in a risk-informed world, I would
expect a weakness to be something with a CDF bigger than ten to the minus
seven or something. And the vulnerability would be a CDF bigger than some
number, ten to the minus six or something. There would be a measure of these
things. They would be classified on the basis of risk significance or something,
rather than by describing them.
DR. APOSTOLAKIS: It's implicit, Graham.
DR. WALLIS: Well, it's implicit, but it's not actually done?
DR. APOSTOLAKIS: It's a qualitative assessment of
probabilities.
DR. WALLIS: Well, I don't know what that means.
DR. APOSTOLAKIS: That's what it is.
DR. WALLIS: Well, that's -- I mean -- let's not get into that
one.
DR. POWERS: It's a qualitative quantification of a system,
Graham. As we talked yesterday, they qualitatively quantified it.
DR. WALLIS: Qualitative means you can argue about it
forever.
DR. APOSTOLAKIS: Yes. That's what it means.
MR. RIDGELY: I think with numbers you can argue about it
forever too, but anyway.
DR. APOSTOLAKIS: But many of these programs were
instituted ten years ago?
MR. RIDGELY: That's right.
DR. APOSTOLAKIS: When the world was not risk-informed
at the time. So now we are looking at the results of that -- quantitative risk.
DR. WALLIS: Well, is there any correlation between what one
person thinks is a vulnerability and what somebody else thinks is a
vulnerability? Is there some sort of an agreement about what's a vulnerability?
MR. RIDGELY: I think overall that there was a general
agreement, more or less, that they didn't have a plant vulnerability by virtue of
not finding anything that was deemed to be a significant contributor to a
problem. But there are also those -- there was one plant at least that I can think
of that if they found anything at all that was greater than ten to the minus six,
they termed it a vulnerability. And even though their total core damage
frequency was just some small number times ten to the minus six, they had
applied it that way. Where other plants with higher core damage frequencies
would have the same thing, and they would not identify them as vulnerabilities.
So anything they found that went above their screening criteria, they determined
it as a vulnerability, but that wasn't -- that was more like anomalies and
weaknesses that every other plant had used. So there's not a uniform
agreement.
DR. WALLIS: Well, I can see some management saying, "Go
out and make a study. Find some weaknesses, and fix them so that we don't
have any vulnerabilities."
MR. RIDGELY: With the findings in the fire area, the
licensees performed the fire PRA which would include portions of the IPE
models for decay heat removal. Licensees also looked at the fire areas by
screening criteria qualitatively. They would leave an area out if the area neither
initiated an event nor caused the loss of safe shutdown functions.
Quantitatively, they would screen it if a contribution from a fire in that area was
less than ten to the minus six.
And any areas that remained after that, they would look at on
a case-by-case basis to ensure that at least one method of safe shutdown and
decay heat removal was available. Frequently, this revolved around Appendix
R systems and functions. Fire walkdowns were used, and the information was
applied here. And, again, no vulnerabilities were found.
DR. POWERS: A great deal of significance seems to be
attached to the use of walkdowns. What is the reliability, I guess is the word I'm
looking for, of a walkdown in discovering things?
MR. RIDGELY: Well, the walkdowns were done usually by
a team, and depending upon what the function of the walkdown was for, there
would be at least one knowledgeable person on that subject.
DR. POWERS: Now, most of the people on this team walked
past the places they're walking past every day, frequently.
MR. RIDGELY: Now you're asking about the makeup of the
team that did it, and I'm not sure I can --
DR. POWERS: Well, in general, you would expect -- they're
plant personnel is my point.
MR. RIDGELY: Some of them were plant personnel, some
were contractors that they would hire, some would be consultants --
DR. POWERS: Okay.
MR. RIDGELY: -- that would be knowledgeable, like in fire if
it was for a fire walkdown, this kind of thing.
DR. POWERS: So it's more diverse than I was thinking.
MR. RIDGELY: Yes. It's more than just plant personnel.
DR. POWERS: Oh, okay. And do we -- I mean has NRC
ever tried to compile some sort of database and say, "Okay, if you -- when
people do a walkdown, they will generally successfully find these kinds of
things, and they will not find these kinds of things"?
MR. RIDGELY: I'm not aware of anything like that.
DR. POWERS: What is the significance of the walkdown?
MR. RIDGELY: Well, during the walkdowns, they have found
a lot of things that they might not have found otherwise. For example, during
walkdowns, they would find gas cylinders that were not properly secured. They
would find trash or paper combustibles that they would not necessarily
otherwise have known about being stored in a location. They would look at
anchorages of equipment and find maybe there was a bolt missing or a bolt was
loose or something. So the walkdowns, I believe, played an important role.
In the HFO area, particularly in flooding, for example, they
would look and find that roof drains were plugged or there was a hole and the
seal was missing, this kind of thing. So I think the walkdowns played an
important part in reviewing for the IPEEE.
MR. ROSEN: Are these walkdowns typically conducted in
accordance with written procedures?
MR. RUBIN: I don't think NRC procedures but I think plant
had their own checklists. There was some guidance in other documents on the
walkdowns.
MR. ROSEN: So there were checklists and guidance.
MR. RUBIN: Yes.
MR. ROSEN: Were the people trained, typically, to perform
walkdowns in accordance with the checklists and guidance?
MR. RIDGELY: Well, some of the people, at least, on these
teams would be experts in the areas that we're talking about -- experts in fire,
experts in seismic and that kind of thing. And so while they may not be -- you
know, from a plant standpoint, they may not have unique information about the
plant, but they would know what -- should know what they're looking for when
they're walking through the plant for the particular issue that they're looking.
MR. RUBIN: And, typically, there would be systems people
also making the walkdowns.
MR. RIDGELY: In the HFO area, the safety-related
equipment is protected from high winds, tornados, and tornado-generated
missiles. The external flooding induced failure was prevented by watertight
structures. If it wasn't, leakage would be limited to prevent damage or they
reviewed the equipment operability for submerged operation. If none of this
was satisfied, then they would provide an alternate means of achieving that
function.
Other external events were found to be insignificant
contributors to core damage frequency, and walkdown information was used,
and no vulnerabilities were identified.
DR. POWERS: Can you explain to me a little more about
lightening? When people do an assessment of lightening's contribution to the
core damage frequency, I can believe that they probably know something about
the number of lightening strikes per unit of time on the site. Now what do they
do? I mean what's the next step in the assessment process?
MR. RIDGELY: Well, the one that I'm most familiar with, what
they did was they looked at, as you say, the frequency of strikes, the number
of strikes they would get in a unit area. And they were particularly concerned
with the control room at this particular plant. And they --
DR. POWERS: When they get the frequency -- I mean you
know how often you have storms and you get a lightening strike, and you know
the area of the site. How do you know the lightening strikes on the part that's
vulnerable? You just divide by the area?
MR. RIDGELY: No. They looked at the surrounding buildings
in this case, because the buildings give you a cone of protection from a
lightening standpoint.
DR. POWERS: So, effectively, they're looking at the potential
function or something like that? Maybe qualitatively without calculating it.
MR. RIDGELY: Well, there is formulas and guidelines and
standards for providing lightening protection for structures and what not. And
they would use that information to review the plant. And they would look to see
then, "Okay, then what's the probability of actually hitting it?" Then they looked
at the structure of the control building, for example, and said, "Well, that's got
lots of rebar in there," so the chances of lightening hitting the control structure
and getting through the rebar and then doing damage to something important
was not a significant contributor. So that's the way at least one plant had
handled it.
MR. RUBIN: May I make just a general comment, add to what
John is saying? Typically, issues like the other external events, such as
lightening, plants screened them out as they met their standard review plant
criteria. So, generally, there was not a detailed analysis of lightening, but they
were screened on other basis that was acceptable for IPEEE purposes. And
lightening is mostly the effect on losses of off-site power and station blackout
events, sequences, so that's --
DR. POWERS: And, certainly, my approach would do -- say,
if I complied with the standards --
MR. RUBIN: Yes.
DR. POWERS: -- then I've screened it out and forget about
it.
MR. RUBIN: That was what was done for a large majority of
the plants.
MR. RIDGELY: The inclusion for A-45 is that all plants had
provided adequate information to verify they've done what they were supposed
to do. All the plants have identified at least one method of removing decay
heat, and no vulnerabilities were found.
MR. LEITCH: John, in your table, in the remarks column, it
says, "EX" opposite this --
MR. RIDGELY: That means we were looking only the
external event aspects of that generic issue.
MR. LEITCH: It's seismic, fire, and --
MR. RIDGELY: And HFO.
MR. LEITCH: -- HFO. But then this generic issue remains
open for other issues?
MR. RIDGELY: Well, the internal event aspect was verified
using the IPE Program.
MR. LEITCH: Right.
MR. RIDGELY: So I believe that this issue now would be
adequately verified.
MR. LEITCH: Okay.
MR. RUBIN: There's another report on the internal events
and the IPE Program. There's a separate report that's going to be issued, it's
not out yet, on the IPE-related generic issues. There were much fewer number
in the IPE Program than there were in the IPEEE Program.
MR. LEITCH: Okay.
MR. RUBIN: But there will be a report, and A-45 is part of the
IPE report as well.
MR. RIDGELY: I have --
DR. APOSTOLAKIS: Let me understand this, because I know
you have others, but -- so regarding the USI A-45, back to ten, you concluded
all plants have provided adequate information. They identified at least one
method of removing decay heat, no vulnerabilities. So is this then the resolution
of this issue?
MR. RIDGELY: It is the verification of the issue.
DR. APOSTOLAKIS: Verification.
MR. RIDGELY: The issue was resolved by the imposition of
what the staff decided to do, which was, for the external events, was to fold it
into the IPEEE Program.
DR. APOSTOLAKIS: So at some point, the staff decided that
this issue would be resolved within the IPEEE issue -- program?
MR. RIDGELY: Well, it would be verified through this
Program.
DR. APOSTOLAKIS: Verified. Verified means that there was
-- it was resolved already?
MR. RIDGELY: Technically, yes.
DR. APOSTOLAKIS: Let me put it a different way. After your
slide 10 -- yes, let's go back to your 10. This particular issue does not exist
anymore for the Agency since all these things are no, no, no? Is it closed now?
MR. RIDGELY: Harold is getting up to answer the question
from a generic perspective.
DR. APOSTOLAKIS: Okay.
MR. VANDERMOLEN: The answer is, yes, it is closed. We
are not --
DR. APOSTOLAKIS: It is closed because of this.
MR. VANDERMOLEN: Yes. The issue was -- terminology
is always confusing when it comes to these generic issues. The action
imposing new requirements on the plants, that's the resolution. And then that's
what we do with generic issue. It results in generic regulatory action, not
necessarily people going out to every plant. This is a verification effort. That's
all it is.
DR. APOSTOLAKIS: And this applies to all of these? We lost
the speaker.
MR. RIDGELY: Yes, it does, except for Generic Issue 172,
which you will get a package on.
DR. APOSTOLAKIS: Okay. So all the issues, the USIs and
GSIs, that you are talking about in the report, except this one that you
mentioned, are closed after this study.
MR. VANDERMOLEN: Yes.
DR. APOSTOLAKIS: Okay. That makes it clear.
MR. VANDERMOLEN: Yes.
DR. APOSTOLAKIS: Okay.
MR. VANDERMOLEN: No more plants to spend more
resources on.
DR. APOSTOLAKIS: Okay, okay, okay.
DR. WALLIS: I presume there's a feedback loop. If it's not
verified, you go back to resolution or something? Or do you always verify?
MR. RIDGELY: I believe that there's always some amount of
verification.
DR. WALLIS: But then if you're not satisfied with the
verification, you've got to go back and do something, haven't you?
MR. RIDGELY: Yes. I believe that's true.
Now, my time has --
DR. APOSTOLAKIS: Yes. There is no time for --
MR. RIDGELY: Right.
DR. APOSTOLAKIS: I mean the others you draw similar
conclusions.
MR. RIDGELY: Right. So if it's all right, then I'll skip the other
three and go to conclusions.
DR. WALLIS: The only question I have on 11, which you put
up and then took away, there have been instances where the fire suppression
system has had water hammers which have led to flushes of water flowing out
and damaging safety. It doesn't seem to be part of your list of things here.
MR. RIDGELY: Not for the IPEEE; no, it's not.
DR. WALLIS: But it has happened.
MR. RUBIN: That was an internal flooding issue.
DR. WALLIS: This is -- really, you're saying here damaging
effects caused by the fire -- activation of the fire suppression system.
MR. RUBIN: Right.
DR. WALLIS: And one of the things that can happen is the
fire suppression system itself has a water hammer.
MR. RIDGELY: And that would have been handled, I believe,
through the IPE Program. Now, what we're looking at is the activation and
effects from seismic. If you have a seismic event, will that start the fire
protection system?
Because of the time, let me jump to the conclusions. But you
do have the rest of the slides there. There were 31 IPEEE-related unresolved
safety issues and generic safety issues. Nine were explicitly discussed in the
generic letter, in the NUREG; 22 were not. We consider that this Program is a
major achievement and is verification of a large majority of these issues. Forty-
four licensees provide sufficient information to verify all 31 USI and GSIs.
Twenty-five submittals had one or more generic issues or sub-issues open or
partially verified. Saying that a little bit differently --
DR. WALLIS: You're really verifying the resolution, aren't
you?
MR. RIDGELY: That's correct. From the submittals that we
have, we have verified 100 percent of A-45 GSI-131, 156, and the other
numbers here.
DR. APOSTOLAKIS: Now, you did not, though, verify the
licensees' verification.
MR. RIDGELY: We reviewed the documentation that they
provided.
DR. APOSTOLAKIS: But you didn't send people to the plant.
MR. RIDGELY: Not for this, no. Now, there were four site
visits, but that was not for the purpose of verifying that they had done what they
said here in this.
DR. APOSTOLAKIS: Are the resident inspectors involved in
any of this? I mean would you let the resident inspector at the plant know that
this particular problem, GSI-172, claims that they have verified resolution GSI-
172 so that the inspector will do something and, say, send a message back,
"Yes, that's correct."
MR. RUBIN: Let me try to answer that. We sent the draft
report to all the resident inspectors for their information. We're not awaiting nor
did we ask for them to go back and walkdown or verify, from their perspective,
each of these issues.
DR. APOSTOLAKIS: Are they expected to verify this?
MR. RUBIN: No.
DR. APOSTOLAKIS: Is it up to them?
MR. RUBIN: Verification in this sense, when I talked about
what was in the IPEEE review process, the staff and I review, and we send our
process with the Senior Review Board. We did not validate quantitative results.
We looked at the reasonableness and completeness. So in that same vein, we
looked at the reasonableness and completeness as a licensees' process to
address these issues, to look for weaknesses and vulnerabilities in our IPEEE
review.
DR. APOSTOLAKIS: Well, that's the IPEEE's --
MR. RUBIN: Yes.
DR. APOSTOLAKIS: -- function.
MR. RUBIN: Yes.
DR. APOSTOLAKIS: But then, you know, the inspector may
have other things to do. In declaring a GSI as resolved, it's not your job, but
maybe the inspector can use your report now to actually say, "Yes, it is
resolved." But I guess that's not the case.
MR. RUBIN: It's not the intent, but they certainly have the
information in the report and can --
DR. APOSTOLAKIS: If they want to pursue it.
MR. RUBIN: Yes.
DR. APOSTOLAKIS: That's a really interesting way of closing
GSIs.
MR. MARKLEY: This is Mike Markley. Normally the
Inspection Program Office creates a temporary instruction, or through the
regional offices they do, and they go out and do a customized inspection for that
site based on the issues that are identified in the GSI translated to the
temporary instruction. And they close it out within the inspection process for
that site. This is doing it on an across the industry basis, not an individual site.
So each site had to go through that under the Inspection Program, and it's
documented in the inspection reports what the findings and follow-up issues or
things that may have still existed along those lines.
DR. APOSTOLAKIS: So each inspector then or each region
would -- well, has already received your report. They will follow up. They will
--
MR. MARKLEY: That's right.
DR. APOSTOLAKIS: -- create those temporary instructions.
MR. MARKLEY: Correct.
DR. APOSTOLAKIS: And there will be verifications.
MR. MARKLEY: That's right.
DR. APOSTOLAKIS: Okay, okay. Well, that makes --
MR. RUBIN: At least I'm not aware of that, Mike, maybe, but
it's worth checking on.
MR. MARKLEY: Well, I had to do them, so --
MR. DURAISWAMY: So did I. So we have done that.
MR. RUBIN: Okay. All right.
DR. APOSTOLAKIS: Two to one, Alan. No, they may not
come back to you at all, and that's fine.
MR. RUBIN: Yes, okay.
DR. APOSTOLAKIS: Because you have provided information
to them. But it's the region's responsibility, actually. And that makes much
more sense, that you have an unresolved safety issue, and there is a final word
that somebody inspected and said, "Yes, it's resolved." And that makes much
more sense.
MR. MARKLEY: For that site.
DR. APOSTOLAKIS: For that site, yes. That's good.
MR. RIDGELY: And so the conclusions --
DR. APOSTOLAKIS: I'm glad I'm not the only one who
doesn't seem to understand very well how the Agency works.
MR. RIDGELY: The conclusions of this is that we find that we
don't believe that a potential vulnerability was missed. Any GSI that wasn't
verified was identified as a weakness in the plant's specific staff evaluation
report. And any need for plant-specific actions would be taken up outside of the
IPE Program. And the final slide shows it for all of them. This shows, I believe,
that we've had great success in verifying the unresolved and generic safety
issues.
DR. APOSTOLAKIS: So we're going to hear about the
unverified ones sometime in the future?
MR. RIDGELY: No, this is it, as I understand it. But you will,
as I understand, get a package on GSI-172.
DR. APOSTOLAKIS: But if they are unverified, what
happens?
MR. RUBIN: Well, that's what I mentioned earlier. We're
going to take a look for those plants to see if there is some additional action that
is needed.
DR. APOSTOLAKIS: Okay.
MR. RUBIN: That is not being falled through the cracks. We'll
go to the ACRS. No plan right now except for the MSRP issue 172.
DR. APOSTOLAKIS: Is this it?
MR. RIDGELY: That's it.
DR. APOSTOLAKIS: Any comments or questions from the
members? We are five minutes behind. I don't want to make it six. Otherwise
do the members have questions? Anybody else. Thank you very much.
Appreciate you coming here.
We are now scheduled for a break until five minutes after ten.
(Whereupon, the foregoing matter went off
the record at 9:51 a.m. and went back on
the record at 10:05 a.m.)
DR. APOSTOLAKIS: Okay. We're back in session. The next
item is "Proposed Resolution of Genetic Safety Issue 191: Assessment of
Debris Accumulation on PWR Sump Pump Performance." Calling on the
member is Mr. Leitch.
MR. LEITCH: Okay. As has been stated, this issue concerns
the assessment of debris on PWR sump performance. What we're going to
hear today is a summary briefing of a contractor's report. And I understand that
in September there will be a final additional presentation, after which time we
will be expected to write a letter on this topic, but no letter is expected as a
result of today's presentation.
Mike Mayfield is here to have some opening remarks and
introduce the presenters. Mr. Mayfield?
MR. MAYFIELD: Thank you. I'm Mike Mayfield from the
Division of Engineering Technology and Research. We are here to present a
summary presentation of the results of some work done to look at the evaluation
or the technical basis for resolving generic safety issue 191. It is more a
summary briefing and a status of where we are, rather than a detailed
discussion of the proposed resolution. We do expect to come back in
September with that proposed resolution.
We are going to meet with the public and the industry on July
26 and 27 to discuss this report and see if there's any additional information that
we can glean that might influence the proposed resolution. Depending on the
outcome of that meeting, we'll be in dialogue with NRR on the proposed
resolution and bring that back to you in September.
Mike Marshall from the Engineering Research Applications
Branch has the Project Manager for this piece of work, and he's going to
provide you the technical briefing. Mike?
MR. MARSHALL: Thank you. Good morning. In the
package, my slide package, there's no background slide per se, so I'd like to
cover a little bit of background. I believe a number of the members are already
familiar with this issue and the work that was done under unresolved safety
issue A-43 and the work we recently completed in the 1990s with the regards
to the BWR strainers.
Originally, this was addressed back in 1985 under A-43. It
was resolved. No new rules or requirements were imposed on industry.
Additional guidance was provided that came out of that effort. In 1992, there
was an event at a BWR in Sweden called Barseback, and it raised a lot of
questions about the work that was done under A-43, enough questions that the
NRC actually started looking at the BWRs. And the conclusion from our look
at the BWR is that additional action was required, and they put in larger
strainers.
But based on that work, we were curious and decided to also
go look at the PWRs. There was enough difference between the Ps and Bs that
we didn't think what we did with the Bs was directly applicable to the PWRs.
And a couple years later, we started working on PWRs and that's some of the
work I'll be talking about today.
And just to let you know, what we're going to be covering in
about an hour today we have two days set aside at a public meeting at the end
of this month to go through it in greater depth. And if there's anything I cover
today that you would like us to cover in more detail in September, just please
let us know and we'll do that.
DR. UHRIG: Could you just quickly in a few couple of
sentences summarize the difference between the work that was done on the
BWRs and why it was not applicable?
MR. MARSHALL: Mainly we looked at the BWRs, and it
boiled down to the transported debris. Once you had it generated from a bray.
With the BWRs, from the dry well to the wet well, all the gases and a lot of
debris would just be forced right away down into the wet well, and then the wet
well, with chugging and condensation oscillations, it kept it well-mixed, and you
got -- and the BWRs started recirculation immediately through the strainers.
And so you got transport much more quickly in the BWR. There was enough
energy keeping things mixed up that you didn't have time for things to settle, so
you got the accumulation there.
With the PWRs, you had the water refueling storage tanks so
you have about 20 to 30 minutes before you actually switch over to
recirculation. And also once you switch to recirculation, it's not nearly as
energetic as the conditions for the BWRs. So our biggest question mark when
we started this, and our main focus when we started this work, was looking at
transport of material, because that's the biggest distinction between the BWRs
and the PWRs.
DR. UHRIG: Thank you.
MR. MARSHALL: My second slide is just a statement of the
study, and we're interested in once debris is generated it transports to the floor
of the containment and thus to the strainer and accumulates on a sump screen.
Will it cause a problem? And the problem we're looking at is creating a
resistance to flow it across the strainer -- the sump screen, and would that
resistance be great enough to take away the net positive suction margin to the
ECCS pumps, drawing suction from there?
And another focus was, again, because a lot of this work goes
back to A-43. Once we look at this time, is there something that wasn't done in
A-43 that we need to look at this time around?
DR. APOSTOLAKIS: Would you change the slide then since
you're talking about the --
MR. MARSHALL: Oh. Thank you, sir.
DR. APOSTOLAKIS: And explain the figure there a little bit.
MR. MARSHALL: The figure is just a sump.
DR. APOSTOLAKIS: I see we need the --
MR. MARSHALL: This is an ideal --
DR. APOSTOLAKIS: Wait, wait, wait, wait. You need the
microphone.
MR. MARSHALL: Oh. Excuse me, gentlemen. Okay. And
I won't touch the screen. I'll point but not touch.
(Laughter.)
DR. APOSTOLAKIS: Now you can touch.
MR. MARSHALL: Okay. This is an idealization of a sump.
Actually, from the work we did, we did a survey early on. There's no -- we
haven't quite found two sumps quite alike.
DR. POWERS: I've been there. Been there, done that.
(Laughter.)
MR. MARSHALL: But on the sump -- well, I'll start from
outside working my way in -- there's usually something called a debris curve,
and that's just a curve about, at most, 12 inches but typically closer to about
four inches. And that's to stop larger debris that just travels along the floor.
And as you approach the sump, you have a trash rack, which is usually a
course mesh about the size of floor grading. And that's to catch your larger
debris. And then you have a finer screen after that, which we call the sump
screen, and this is typically a quarter or one-eighth-inch openings in it.
Then on the other side of that, you have the suctions for the
ECCS pumps. And sometimes the plate on the solid is either a course screen,
like a floor grading, or a solid plate. And this just represents typically after you
have -- after the access and by the time you switch to recirc, this structure is
usually submerged under water, but that's not always the case, as we learned
during this study.
I'll just give you a quick overview of how we approach the
study. We started by trying to identify debris sources. What are the likely
equipment and the containment to be damaged following an accident or during
an accident and would that lead to material that's transportable that would
actually reach the sump screen? Once we identified that material, looked at
how much of it we would get generated --
DR. POWERS: When you look at something, though, as a
candidate debris source, did you look at that in terms of gas velocity over only
or were other things taken into account in deciding whether something could be
a debris source?
MR. MARSHALL: What was taken into account, essentially,
was the characteristics of the debris once we thought it would be created. For
instance, with thermal insulation, that would be broken up into different sizes,
and would it be broken up into sizes that the gases would transport? Would it
be broken up into sizes that the flow we would expect on the containment floor
would transport? And that became a legitimate debris source, and that's when
we went through further investigation.
DR. POWERS: Well, I guess what I'm fishing for is that during
the blowdown process you may have high gas velocities assuredly, but you also
have a lot of vibration and things like that that equipment vibration and what not
figure in defining something as a debris source?
MR. MARSHALL: No, it did not.
DR. POWERS: Okay. So it's strictly a velocity kind of --
MR. MARSHALL: Right. Just from the jet.
DR. POWERS: Okay.
MR. MARSHALL: Just from the jet.
DR. WALLIS: Is there a kind of conservative assumption, you
assume that if something might break loose, then the whole thing breaks loose?
MR. MARSHALL: No, we don't do that at all. That gets to the
second block of estimating amount of debris, because once we identify the
source, we had some testing and testing that was done by the BW Owners'
Group over in Sweden and Germany, and we'd see how much of that material
would be damaged into a form again that would lend itself to transport.
DR. WALLIS: So they assimilated a LOCA and then --
MR. MARSHALL: We simulated a jet. Since we have a
ruptured disk at the end of a nozzle -- at the end of a piece of pipe and blast a
piece of material at different distances and see how much debris is generated
from that.
Then after we got a good feel for how much debris would be
generated, we looked at how well it would transport once it got into a pool of
water.
DR. POWERS: This is the point that, in your draft, I guess it's
a letter report, didn't seem to go into the, at least I didn't see, a discussion of
how you did that transport calculation.
MR. MARSHALL: I will touch on that later.
DR. POWERS: Okay. Drag coefficients and things like that?
MR. MARSHALL: No. Actually, I'll go into that later, but, no.
Short answer to your question is no right now.
DR. WALLIS: Is the paint part of your debris, paint flakes?
MR. MARSHALL: We considered paint flakes as a debris
source. And we broke paint into different --
DR. WALLIS: Is the containment painted, the containment?
MR. MARSHALL: Yes, it is. It has a coating on it. And the
floor has a coating on it.
DR. WALLIS: Does that containment coating remain in tact
during a LOCA?
MR. MARSHALL: Answer to that, from the direct
impingement, we think it will come off. And then we looked at it from the
environment and the containment, whether it would come off. There's a
separate program looking at that explicitly. And there are certain conditions.
You actually have to have a rather --
DR. WALLIS: It's being looked at, but there's a lot of area that
comes off. There's a lot of stuff there that comes off.
MR. MARSHALL: Yes. That would be a very large volume.
That's one reason we had actually a separate study to look at coatings.
DR. SHACK: But in this report it's only in the zone of
influence kind of model.
MR. MARSHALL: It's only in the zone of influence that we're
looking at it in here.
Okay. And once we got to the accumulation on the screen,
we looked at the head loss, and we used the correlation that we developed
during the BWR work for that calculation.
DR. WALLIS: It's a tautology. I mean you can -- it's all where
this stuff comes off, and so you look at the zone of influence to figure where the
stuff comes off.
DR. POWERS: I guess I don't understand. Why don't you
look beyond the zone of influence?
DR. WALLIS: But by defining the zone of influence, you can
restrict it to almost anything you like, unless you have some really good criteria
for determining --
DR. POWERS: Experiments.
DR. SHACK: And then assumed it was a sphere.
DR. WALLIS: Gee whiz, spherical debris?
MR. LEITCH: Those marks on the cartoon there on the left,
are they -- is that just the depiction of debris or do those particular locations
indicate --
MR. MARSHALL: It's a cartoon just to who a depiction of
debris. That's the break, and we expect it to go in all directions, unlike the
BWRs. It's not all being driven straight down to the floor. As been estimating
that material, when it gets to the floor, where does it end up in the sump? In the
red box, there's the depiction of where the sump would be.
DR. POWERS: I'm still a little confused about the drag
coefficient business and the transport calculation. I mean how do you do that?
MR. MARSHALL: Well, I'll address it -- I'll talk about that now.
In this report here, in our parametric analysis, we didn't directly look at transport
-- address transport directly. It's rather difficult. And another is it would be
difficult to do parametrically in this structure here transport overall because of
the variability of the containment designs. It actually makes a big -- the BWRs
from this point we could all treat them similar. The PWRs, again, we can't treat
them similar. It's almost like a custom calculation each time.
DR. WALLIS: Doesn't it go up in that hot plume? Doesn't it
go up in a hot plume into the containment, because the buoyancy of the plume
lifts up the debris?
MR. MARSHALL: Some debris will go up, and then when
spray comes on there's a chance that it will also be washed back down.
DR. WALLIS: So you need a model. If you're going to do it,
you need a very fancy model for containment flows.
MR. MARSHALL: Well, let me finish addressing Dr. Powers'
question. So what we did here, since we didn't directly address transport -- and,
actually, if you don't mind, I'll probably jump ahead to a slide. We're jumping to
slide 7. And I'll remember to come back and cover some of this other material.
And, actually, on slide 7, I might end up covering some of the other material, but
we'll go back and make sure I didn't miss anything.
So the approach we took in this evaluation was -- the first step
we took was identifying the debris. The second step was identifying -- well, the
first step was identifying the case we're dealing with, then identifying the debris.
And the first step we did was to calculate this curve here. And what this curve
represents is for a given sump size, given surface area for the sump screen,
given whole size net positive suction head margin and the flow rate you have
for whether it's a small LOCA, medium LOCA, large LOCA, what's the capacity
of that sump to accumulate debris without causing degradation and net positive
suction head margin. And you could do that calculation without knowing
transport. So it just tells us what the capacity of the sump is. And usually you
get a feel for how much debris that will cause you.
Then the next thing we did is we calculated a box here. And
one part embedded in that box, again, is the type of debris, essentially the
characteristics of the debris, and how well it transports. That's what this box
represents, and also how much do you think will get to the sump screen. And
what we did here was looking at the work we've done and the work that others
have done is we looked at transport from the zone of influence, and we said
from the zone of influence to the sump screen, from all the work that we
reviewed, it looks like at the very least you'll get about ten percent of the
material. And at most you would get is about 25 percent of the material.
So this edge of the box represents ten percent of the material
from the zone of influence getting to the sump screen and 25 percent of the
material from the zone of influence getting to the sump screen. And we could
do this for all 69 cases relatively easily. It simplified the problem for us.
DR. APOSTOLAKIS: Why aren't there any numbers on the
axis?
MR. MARSHALL: I was going to use this just to explain what
we did. If you turn to Appendix B in the report we have, all those have the
numbers on them. And this is K-17 out of that report. But I just wanted the
opportunity to make sure --
DR. APOSTOLAKIS: Why does the curve go like that? Is
there something I don't understand?
MR. MARSHALL: This point here represents -- we have two
materials here: fiber material and particulate debris. In particulate debris we're
talking about something like the grains of sand when we talk about particulates.
And this point here represents if you only had the fiberglass, this is the amount
of material that the capacity of the sump screen.
This point here represents the least amount of fiber you need
and the particulates you need to exceed the net positive suction margin. And,
essentially, the line just goes straight, because there's a finite amount of
material and it doesn't get any worse if we add more particulate to it at that
point. It doesn't go to infinity, because, again, there's a finite amount of
material. Somebody asked me that question earlier.
DR. POWERS: What George may not recall is that one of the
findings in your BWR study was the synergism between fibers and particulate.
MR. MARSHALL: Right.
DR. POWERS: That's what causes the --
MR. MARSHALL: And on this study, again, one reason we
tend to -- in the report we present the results with fiber versus particulate is that
combination from the work we continue to do tends to be the most problematic
combination of debris sources.
DR. WALLIS: But it matters probably which comes first. If the
fiber comes first, then it makes a mat into which the particulates can stick.
MR. MARSHALL: Right.
DR. WALLIS: If the particulates come first, then the mat isn't
there.
MR. MARSHALL: Right. Some of the particulates will just go
through. So if you just had particulates, you have nothing to build a fiber bed,
you have no filtration that will occur. And so particulates by themselves will go
through the openings. Now the only exception to that is one insulation type,
which is called calcium silicate. And the reason is it's mostly particulate, but it
has a fiber binding, so it comes with its own fiber and particulates packaged
together. And that's used quite a bit in the PWR population.
DR. WALLIS: It looks like your debris is way more than you
need to clog the screen. Is that --
MR. MARSHALL: Yes. That's what this shows.
DR. WALLIS: What does very likely mean?
MR. MARSHALL: Actually, let me go start my presentation
--
DR. WALLIS: Sure.
MR. MARSHALL: -- back over again, and we'll get to that.
(Laughter.)
DR. APOSTOLAKIS: You're coming to 11.
MR. MARSHALL: Let's go back to the slides.
DR. APOSTOLAKIS: Don't start over too many times.
MR. MARSHALL: Oh, no. I'm going to go back to slide 4.
DR. POWERS: He's just managing his reviewers, that's all.
And doing a good job of it.
(Laughter.)
MR. MARSHALL: Go back to four. I just want to make -- we
use the shorthand of something we call sump failure. When we say sump
failure, we don't mean the sump collapses or it breaks. Essentially what we
mean is the resistance across the screen exceeds the net positive suction head
margin. And for a few cases, for a few plants -- well, cases, the screen's not
going to be fully submerged, and so you're just looking at the height of water
needed to overcome that resistance. And that's the criteria we used for saying
the sump failed.
Just to go on a little bit more about how we went about doing
the study, we started out with individual plants, and we had a survey, and we
collected information on their sump screens, what's the size, what's the area,
what flow rates you might have, what's the net positive suction head we got
from responses to generic letters to the NRC? And so we had a good feel for
the variety of the -- the range of variables in industry. And instead of starting
with a base case then varying a particular variable, what we did is we created
cases based on each individual unit. And we did that for, again, the large
LOCA, medium LOCA, and a small LOCA.
And we wanted to get a feel for generically if this is something
that's -- I guess the way we looked at it originally, is this a credible concern for
PWRs?
DR. APOSTOLAKIS: Now, in the first line, you mean the
formal analysis that will demonstrate generically whether debris accumulation
will --
MR. MARSHALL: Yes, yes, whether.
DR. SHACK: I noticed you had a typo. I was really
impressed --
MR. MARSHALL: Oh, where is it?
DR. SHACK: -- by a draft report that said you did 1,350 break
locations for each three-foot pipe segment.
(Laughter.)
MR. MARSHALL: Oh, not for each three, no. That was the
total after going three feet each.
DR. SHACK: But now I find that was -- the amount of work
you did just went down by orders of magnitude.
(Laughter.)
MR. MARSHALL: Yes, it did.
DR. SHACK: In your large break LOCA, I noticed you had
this very wide range in distribution. Is that because you did pipe sizes from six
inches -- you did break sizes all the way from six inches on up to the largest
pipe?
MR. MARSHALL: Yes. What we built -- we have two
reference plants, which gave us more information than we collected from the
survey. And we started out building AutoCAD model evolver piping in the
containment. And then we put that into a MATLAB model. And, essentially,
what we do is tell MATLAB to go every three feet, then calculate for that zone
of influence what material is in that volume.
So if you have a large pipe with very few targets around it, the
only material you'll get is on that pipe. But if you have a large pipe with a lot of
-- when I say targets, other piping, other pieces of equipment -- so if you have
a large pipe that's surrounded by a lot of piping, that's when you get the larger
volumes of material.
DR. WALLIS: If you have a large pipe which is covered with
boron stalactites, do the stalactites become the debris?
MR. MARSHALL: We didn't look at boron stalactites.
DR. POWERS: But the boric acid is water soluble.
DR. WALLIS: Instantly?
DR. POWERS: Oh, very quickly.
MR. ROSEN: Especially in hot water.
DR. POWERS: Especially in high water.
MR. MARSHALL: I believe I touched the other points on that
slide already. But, no, we didn't look at boron.
DR. WALLIS: So there's no precipitate from chemical
reactions that becomes part of the debris?
MR. MARSHALL: We considered that, actually. But one
thing we could show from the work we've done here is if you just -- again, we're
looking at mainly the thermal insulation, and then we took what's called other
particulate sources, and we used the minimums that the BWR Owners' Group
recommended. But just looking at the thermal insulation and not considering
some of the other debris sources, we still get a large number of what we call
very likely for sump failure cases. And so once we got said -- and, again, we're
looking is this something credible to pursue later? If we get it without counting
all the debris sources, we think that makes our case.
DR. POWERS: If you harken back to TMI and the sumps
there, the sump's kind of interesting in that it's kind of cold water at the top, but
as you go deeper and deeper, it becomes progressively murkier until in the
lower foot of it, it's a sludge and it's fibrous kinds of materials and things like
that.
There is a precipitate for the boric acid. It does react with
calcium to form a gelatinous precipitate. I don't think it changes your study.
MR. MARSHALL: Yes. Instead of calling it gelatinous, we
started using the word "flocculent," which I'm still not.
DR. POWERS: Yes.
MR. MARSHALL: But we did look at those other debris
sources. And, again, I think I already the covered the cases when I went to the
other slide, but let's see if there's anything I missed. One thing, again, each
case was based on an individual PWR unit. Most of the information was from
the survey. One important part is our two reference plants, those were both
four-loop Westinghouse units. And so even if we were dealing with a two-loop
unit or a C unit or a Babcock & Wilcox unit, that's the model -- that's the
information we used to estimate debris. That's what we call a limitation in the
study later on.
DR. POWERS: We harp a lot on that in the draft report. How
big of an issue do you think it is?
MR. MARSHALL: Actually -- and, DV, correct me if I'm wrong
-- but started doing -- we treated both reference plants like we did all the other
plant -- cases, sorry, cases in here. And we started going to look at the
individual reference plant from where we know where the material is instead of
the way we assumed the material would be distributed here.
And we found one interesting thing when we went to the small
LOCA for one of the reference plants. And that is here, in this calculation, we
have fiber content in our small LOCA volumes. When we look at that plant by
itself, we don't get the fiberglass in the small LOCA. So, actually, knowing the
location of the debris -- it's not debris at that point -- of your debris sources is
very important.
And because we're limited to our two reference plants here,
and that's one reason we started calling them 69 cases instead of 69 units,
because that can make a big difference. We're unable to address generically
how the location of the material will affect it, but we are able to see from the
sump design, if you look at the sump, we varied all those conditions and got a
good feel for how this would work with that.
And another point is -- if you look at the next slide, another
important thing we've learned from here, and we call it one of the insights on the
last slide, is by looking at the results this way, we're able to identify what's the
minimum amount of material that's necessary to cause sump failure? And if you
go through the report, you'll see some of these values are down to a couple
cubic feet. We're not talking very large volumes of material necessary for some
of the cases, for some of them to reach the point of what we would call sump
failure.
Now, to explain the likely, possible, and very likely.
Essentially, it's quite -- it boils down to if you're on the right of this curve, we
labeled the results "very likely."
DR. APOSTOLAKIS: Which curve?
MR. MARSHALL: This curve here; sorry, sir. This curve
here.
DR. APOSTOLAKIS: That means very likely clogging will
result?
MR. MARSHALL: Some failure. Very likely you'll have some
failure.
DR. APOSTOLAKIS: It doesn't mean a likely event. It means
likely clogging.
MR. MARSHALL: No. Likely -- yes. Unlikely, if the box was
completely on this side, and we have a number of those cases, we said, "You're
unlikely to have sump failure." Then we looked at the cases that straddle the
line, where you would have the favorable conditions on one side and the
unfavorable conditions on the other side. And we took a harder look at each of
those. And if the case had a large RMI -- RMI is fiberglass and particulates, the
most problematic. If you're looking at the least problematic, you're talking about
your metallic insulations.
And so if the survey information said there was 75 percent or
90 percent reflected metallic insulation and let's say fiber was only five percent,
we would label that possible, because the material would dictate that it's less
likely to occur. But if it turned out that that was 100 percent fiberglass for that
case, we would have called it likely, because, again, there's a chance that
enough material would get there. You're closer probably to the unfavorable
conditions than you are to the favorable conditions.
DR. WALLIS: Now that box of range of expected debris is ten
percent to 20 percent of all the debris that's generated in the zone of influence?
MR. MARSHALL: Ten to 25 percent.
DR. WALLIS: All right. So there's a point there which is all
the debris somewhere off the top there, which would be where all the debris is
simply ten times the ten percent.
MR. MARSHALL: Yes. We didn't think all the debris would
go, but --
DR. WALLIS: But there is a point there.
MR. MARSHALL: Yes. What I gather from this is that the
particulates are the problem, that you have to reduce the particulates by a factor
of about 100 in order to get down into the unlikely region. You have to slide that
box down by a factor of 100.
MR. MARSHALL: Well, even if you only had fiberglass, you'll
be along this line.
DR. WALLIS: That's right. But I mean the particulate's the
biggest one.
MR. MARSHALL: Yes, exactly.
DR. WALLIS: You pull it all the way down.
DR. APOSTOLAKIS: So one-thousandths of the particulate
is enough to give you a major problem.
MR. MARSHALL: It boils down to a filtration problem, and so
once you have the filter there and you start filling up the pores, and that's what
the particulate does, it starts filling up the pores.
DR. WALLIS: So maybe you need an active screen shoveling
device that keeps it clean or something.
MR. MARSHALL: Well, that's been considered. The BWR
had decided that wasn't the best approach. What they wanted was --
DR. WALLIS: Rotating screens or something.
MR. MARSHALL: -- rotating screens. Actually, it was a
problem actually getting them to work right, so that idea was abandoned.
DR. KRESS: Would you repeat again? I'm afraid I wasn't
listening closely on where the ten percent and the 25 percent came from.
Those were estimates of what would get transported?
MR. MARSHALL: Yes, from the zone of influence.
DR. KRESS: Yes. And how did you make those numbers?
Those come out of what kind of an analysis?
MR. MARSHALL: It comes out of, essentially, a review of the
work that's been done and calculations factored into that.
DR. KRESS: This had stuff in it like the flow rate and how the
particulates would move with the flow and how far they had to go and stuff like
that.
MR. MARSHALL: Well, the part to work with the transport
and water, that's true for that.
DR. KRESS: Oh, okay.
MR. MARSHALL: Because that was -- we have that
information, and we had not only other programs, in this program we spent a lot
of work.
DR. KRESS: But there was a lot of judgment involved in this
particular --
MR. MARSHALL: But from the transport and the volume
itself, there's a lot more judgment in there and looking at some testing that was
done over in Sweden to sort of get a better feel for that.
DR. KRESS: Okay. Thank you.
MR. LEITCH: Looking at the drawings here, the charts
related to the various plants --
MR. MARSHALL: What page?
MR. LEITCH: For example, page 132 of the report. It doesn't
seem to have any boxes on it. Versus page 133, I can clearly see the boxes
you're talking about.
MR. MARSHALL: Oh, my. The short answer to that is it's
probably hidden under the label, and they didn't move the label.
MR. LEITCH: Okay.
MR. MARSHALL: Sorry about that. That's a problem. We
missed that.
MR. LEITCH: Okay. Thanks.
MR. MARSHALL: Okay.
DR. WALLIS: Sorry, your code with the likely and unlikely and
the one you just showed us, when people designed this sump, they presumably
did some sort of similar calculation. I guess that Westinghouse or someone
would have a box which is in a very different location from your box, otherwise
they would never would have designed it the way they did.
MR. MARSHALL: We don't think they did. The guidance that
the NRC originally had for the sumps was essentially make it twice as big as
you think you actually need it. What the assumption was in the original reg
guide, rev 0 of the reg guide, was if you had 50 percent of the sump blocked,
would it cause a problem? Would you lose net positive suction head margin?
So, essentially, all you would have to do is take your sump,
cut out a piece of plyboard, paste it up there, cover 50 percent, and see if you
would still have adequate flow and you wouldn't have high resistance, excessive
resistance. And some result was that was just make it twice as big as we think
we actually need it to be. I don't think explicitly the debris was taken into
account. I'm pretty sure it wasn't taken into account from our reviews.
The tally of all the cases is on slide 8 here. And, again, for the
small LOCA, medium LOCA, and the large LOCA, and the report defines how
we defined those and how we thought the progress of the event, the accident
would occur. With the large LOCA, because of the volume of debris you
generate, 57 cases came out being very likely. The cases down here with the
large LOCA that are unlikely, most of those cases had large RMI content. And
that is actually another difference nobody -- between back in A-43 and now is
there's a lot more RMI use back in A-43, and that would form part of the
rationale for how they resolved that issue the way they did.
DR. POWERS: It seems to me we had a gentleman before
us that had looked at some high pressure flows on the reflective metal insulation
and saw some fragmentation of that.
MR. MARSHALL: Yes.
DR. POWERS: And I never heard what the resolution on that
issue was. RMI breaks up and goes to the sump? It doesn't do anything?
MR. MARSHALL: Oh. With RMI, especially in the cases with
the PWRs, is it doesn't transport as well. And it depends on the -- there's two
different types of RMI, major types of RMI.
DR. WALLIS: What is the design?
MR. MARSHALL: Oh, sorry, reflective metallic insulation. It's
a thermal insulation made completely out of aluminum or stainless steel. And
there's one type that has a bunch of small components inside of it, so that lends
itself to a lot of fragmentation. A lot of others is just sheets, and when you rip
the sheets off and they get into a pool of water, they don't travel so well.
And once you -- another thing we use in our -- from looking
at transport and what gets to the screen, is things that tumble or a sliding
transport or tumbling transport that's along the floor. The curbs are pretty
effective in removing that type of debris from concern. So if you have a debris
source that tends to travel on the floor and not suspended in a liquid, the curbs
are very good to have. So you eliminate a lot of the heavier RMI types of debris
that way. So that's another reason why it's less problematic for the PWRs.
DR. WALLIS: But if you do get a sheet of that stuff, it's pretty
dramatic, isn't it? It goes up, it could flatten against the screen.
MR. MARSHALL: If it gets that far, but chances are it wouldn't
get that far. Okay.
The only major point we wanted to make here, and Dr.
Powers mentioned we sort of make the point quite often in the report, is we
think this work gives us a good feel for whether this is a credible concern for
PWRs by looking at it with all the variables we varied and all the different cases
we produced. But because of the limitation and one of the biggest ones is
where is the debris, we can't go and look at each case and say, "This case, we
know, was built on this unit. And since it's very likely, it's going to be very likely
there." We only have two units that we have that much detail on where we
know where the location of the materials is. But based on that limitation overall,
this, we think, the work says this is a credible concern.
DR. UHRIG: Where do you draw the line there?
MR. MARSHALL: Of?
DR. UHRIG: Very likely or likely?
MR. MARSHALL: What do you mean by "draw the line"?
DR. UHRIG: Well, you've got four categories there, and at
some point, let's just take for a large LOCA there, if you draw the line between
possible and likely, you've got 61 of the 69 plants that you have a potential
problem with. Do you draw it between likely and very likely?
DR. WALLIS: And does it matter?
MR. MARSHALL: Actually --
MR. MAYFIELD: This is Mike Mayfield. I can -- in response
to your question, we have tried to not draw that line yet. We're looking at this.
My personal view is that, by and large, once you start migrating into the likely
category, you start wanting to ask more questions, and that's, by and large,
where we are as we've moved into an area that we, as Mike characterized it, is
enough that we think it's a credible concern. And there are enough plants in --
or enough cases in those two categories for all three LOCA sizes that motivates
us to continue to be interested in the issue.
DR. UHRIG: The largest group are within the large LOCA,
which is less probable than the others.
MR. MAYFIELD: Yes. And we have some ongoing work now
looking at risk significance, same kind of problem with that analysis that Mike
and his colleagues have faced here. You need a lot of plant-specific information
to try and put risk numbers to it, but we're seeking some insights as to what this
might or might not mean and the likelihood of the various break sizes gets to be
a factor.
DR. WALLIS: What sort of risk significance do you come up
with to do that?
MR. MARSHALL: We haven't finished that work yet.
DR. WALLIS: But you must have done something. I mean
you can look at some sort of typical to make an estimate of --
MR. MAYFIELD: That work -- our PRA folks are continuing
that work literally as we speak, and I just don't have an answer for you today.
DR. WALLIS: What does this do? This makes high pressure
injection not function after a certain time or something?
MR. MAYFIELD: Yes.
DR. WALLIS: Oh. So you have to go and look at the whole
--
MR. MAYFIELD: In recirculation.
DR. WALLIS: You have to deal with all the calculations to
estimate the effect of that.
MR. MAYFIELD: In the recirculation mode. So that this is a
recirculation issue.
MR. MARSHALL: The next slide --
DR. WALLIS: Well, at least it means that some function,
which is specified in the plant specification is no longer functional, isn't it?
MR. MAYFIELD: That's what it would imply; yes, sir.
MR. MARSHALL: The next slide covers some of the
limitations. And as the first bullet says, most of it's due to lack of plant-specific
information. And another root for some of the limitation was with the small
LOCA. Some of the small LOCAs, conceivably you don't need to go to
recirculation. Then others we're not sure how uniform individual utilities would
respond to a small LOCA if they would all behave similarly. We defined how we
thought the progression would go, but we're not clear of the deviation from unit
to unit. With the medium LOCA and the large LOCA, there's much more
uniformity on how we expect plants to respond to those.
And I already mentioned the first sub-bullet there, which was
the location of the different debris sources. One reason transport wasn't directly
addressed in this analysis was there's such a large variability in the containment
designs we just didn't address it. We tried to address it indirectly with the curve
and the location of the curve to the expected box.
There's another point that's also -- the net positive suction
head margins that we use we drew from licensee responses to a generic letter.
And we're pretty sure most of those responses were, what, licensing basis and
not necessarily what the plant expected to have in the net positive suction head
margin. So some of the cases might not accurately reflect the net positive
suction margin and the different units.
And the last bullet essentially repeats what I mentioned earlier
is that we mixed actual debris-expected cases and some licensing information.
We took the information that was available to -- readily -- best available to us
and what we could collect during the first year of the study. And there are
certain gaps that we just couldn't fill as far as making it expected conditions.
And that was one of the criteria placed on me for the study was it had to be as
close to what we expected in the plant. It couldn't be a bounding or
conservative analysis, and this is the driving force.
And the last slide just covers, again, I already mentioned this,
but I'll touch on it again, for sumps with very large -- for the cases, rather, where
we had very small net positive suction head margins and very small screen
areas, and that usually translates to very high approach philosophies, it took
very little fiber and particulate combination to reach sump failure. Most of the
large LOCA cases, again, because of the volume material generated, turned out
it's very likely you'll have sump failure. And just some of the small LOCA,
because of the much smaller amount of material that's generated, likely to get
sump failure.
DR. SHACK: Wasn't the original argument that the PWRs, by
and large, had rather large sump screen areas compared to --
MR. MARSHALL: The first year of the study we did a survey,
and we found out the PWRs -- there's some that do have very large areas, but
there's also some that fall under what the BWRs had. Actually, in the report on
page 61 -- page 64, we have a listing of the sump areas, and they range from
11 square feet to I think the high was 400 and something.
MR. LEITCH: Yes, I was surprised at the variability in the
design.
MR. MARSHALL: But many of them, a good number of them,
was under 100 square feet. And at one time, we thought all the PWRs were
above that. That was one surprising thing we learned from the survey. The
other surprising thing was the variability amongst the plants, we didn't think it
would be that great. Our original approach was to essentially look at the --
collect the survey information, pick three, maybe five at most, plants that would
represent the entire population and do calculations just for those five instead of
parametrically like we've done here. And we found out we couldn't bin them.
We couldn't bin them by AE, we couldn't bin them -- you could, but it wouldn't
be very defensible by grouping them by containment, reactor type, AE or sump
design. There was not enough of them alike that we could get a small enough
group where we could do representative calculations.
DR. BONACA: What's going to be next step? I mean you
have enough information to know that at least for some plants recirculation may
not be effective.
MR. MAYFIELD: Again, as Mike's characterized it, there are
a number of, we believe, limitation sites. It gets to be problematic to say, "Plant
X has or doesn't have an issue." But I guess it's easier to say when they don't
have than when they do. So the next step is this meeting with the public on the
26th and 27th to discuss in some detail the report, the basis -- I should point out
that the industry has been very actively involved up to the point that Mike and
his colleagues started doing the calculations. At that point, we took a step back,
did the work, and we're now prepared to present the results and discuss them
in detail.
If there is additional information that can -- if there's something
we've missed, we wanted to have that dialogue, discuss the results, see if there
is additional information, insights that should be brought to the table for the staff
to consider in developing a generic resolution -- or a resolution to the generic
issue. That's the dialogue that we'll have over the next month or so, and come
back to the Committee in September with a proposed resolution.
And I got to tell you, today I am not prepared to even
speculate as to where we might go with this. It's something where we -- there's
enough uncertainty in the plant-specific information that you would need to do
this analysis that we just don't have access to. But if additional information can
be brought to the table to help us sort this out, it could have significant impact
on the response the staff might propose.
DR. BONACA: In the report, it seems there are three pieces
that ought to be done, and two you could possibly do a generic evaluation. The
third one, you want it very clear that you couldn't possibly have a generic
evaluation or resolution of it. You have to have a plant-specific evaluation. And
for how long are you going to be taking the lead in gathering information and
attempting to, you know -- it seems to me, at some point, it has to get on the
other side and get the licensees involved in addressing the specifics that you
cannot do by yourself.
MR. MAYFIELD: That's correct. And that's -- we have
colleagues from NRR behind me and beside me. We have been in discussions
with them on this issue. The meeting on the 26th and 27th is actually a meeting
that they organized and are hosting. But the subject of the meeting is to present
and discuss the report results. And they -- I think I can safely say they are very
interested in what responses and additional information we may get from that
meeting on the 26th and 27th. And we'll be in discussion with them about next
steps from there.
DR. KRESS: For those plants that plant-specific analysis
show they may have a problem, it looks to me like the simplest fix is a bigger
surface area of screen.
MR. MAYFIELD: Sure.
DR. KRESS: Is that possible in all cases or is there a problem
with that?
MR. MAYFIELD: I guess I would be hesitant to speak for the
licensees, but you would think so. But I don't know that for a fact.
DR. POWERS: If I owned a plant and was concerned about
this, can I call up Mike and get the calculation and procedure to evaluation my
plant, because I know all the specific information?
MR. MARSHALL: You mean could they?
DR. POWERS: Yes, yes. I mean is there something you
could hand me and I could off and follow this script, and I can come up with a
number like you have, except I would have specifically the information you don't
have.
DR. KRESS: You could put a dot in first for your plant.
MR. MARSHALL: Yes. Yes, we could provide that
information.
DR. POWERS: It's a fairly straightforward calculation and
what not. I don't need to run a cray or anything like that?
MR. MARSHALL: No, you don't need to run a cray.
DR. POWERS: Can I run a spreadsheet?
MR. MARSHALL: Actually, yes.
DR. POWERS: Okay.
MR. MARSHALL: The way we did it here.
DR. UHRIG: What experience, incidents where pumps have
been put into operation, there must have been, over the years, a number of
cases where water has been pumped out. I know specifically Crystal River had
400,000 gallons at one point in the basement. Has there been any other
incidents that you know of where you could determine how much debris might
have been picked up there, just as a validation point?
MR. MAYFIELD: The answer is I don't know. I don't know if
Mike has any --
DR. BONACA: I don't think there was the impingement that
you would have from --
DR. UHRIG: There was no impingement on this, but --
MR. MAYFIELD: I think the impingement is the thing that
drives it.
DR. UHRIG: Drives it, okay. So this would not be a valid
piece.
MR. MAYFIELD: I don't think you can go back to that and
infer what you'd get if you had a LOCA -- a pipe break and you started pealing
insulation off.
DR. UHRIG: Okay.
MR. ROSEN: I'm interested in a little dialogue you had a
moment ago with Dana about the chemical reactions that occur in the formation
of this flocculent type of gel, I think, you talked about. What is actually going on
there, and what is the effect of that on the flow rate? Does it make it worse or
better or what?
MR. MARSHALL: I'm not sure. I'm trying to make sure I
understand your question. When you say make it worse you mean once it
accumulates on the screen, the flow rate?
MR. ROSEN: This is gel formation of some kind you
mentioned.
MR. MARSHALL: Right.
MR. ROSEN: I'd like to know what that is. But just for the
moment we'll just assume that this gel forms on the screen. Does that make
some of these cases that are okay not okay or is this --
MR. MARSHALL: We didn't look at that. We didn't include
the flocculence in here. This is only the thermal insulation -- the particulates,
the cal sil, RMI, and fibrous material in insulation.
DR. WALLIS: Can we talk about chemistry a bit more? When
we talk with Westinghouse about AP 600, we had a discussion what would
happen to these sort of metal pieces which were supposed to sink to the bottom
of the pool. If you take something like foil and put it in a pool, which is acidic,
and it produces gas evolution, it doesn't take many bubbles to make a piece of
foil buoyant. And I think we made some calculations about that. Is not gases
evolved by chemical reactions in the sun?
MR. MARSHALL: We didn't look at gases produced from the
chemical reaction.
DR. WALLIS: Because they make particulates buoyant, and
then they wander around.
MR. MARSHALL: And one thing we did notice with the
buoyant materials, things that trapped air or trapped gases, they tended to
travel on the surface, and when they got to the screen they didn't readily go
under the surface to accumulate.
DR. WALLIS: It depends what the level is, doesn't it?
MR. MARSHALL: Well, yes, if it was one of the partially
submerged. But, Dr. Rosen, may I get back with you with regards to your
questions on the chemical formations and the flocculence?
MR. ROSEN: Sure. You've got a perfect chemical reaction
machine if you have a LOCA in one of these containments. It's got high
temperature, lots of stirring, and if you've got chemical reactions going on
between the materials that are available that can make the situation more
adverse than you have portrayed it, I think we all need to know that.
MR. MARSHALL: Okay.
DR. POWERS: I can tell you what I know or what I think I
know. Any of your aluminum is going to react, form gibbsite. It's aluminum
hydroxide, fairly gelatinous material. Boric acid's going to precipitate as
mineral, which I think is cordorite, which is a calcium borate. Those are the
ones that most quickly happen and seem to be the ones that we see in the TMI
sump.
MR. ROSEN: Does that make the situation worse or better,
Dana?
DR. POWERS: Well, I think these gelatinous materials, when
coupled with the fiber they do the same thing as the particles do. When you
don't have the fiber, then they can actually be sucked through most of the
screens.
MR. ROSEN: Just pumped like -- if you have the fiber, you
have what we -- being a chemical engineer, I suffer from some knowledge of
how --
DR. POWERS: Well, you're just looking at a filter aid.
MR. ROSEN: -- a chemical engineering unit operations
proceed, one of which is -- one of those unit operations is filtration, and one of
the kinds of filtration is pre-filtration with laying down a bed of fibrous material
before you try to filter out a particulate material.
DR. POWERS: Yes. Any kind of filter aid is just that, it's a
fibrous material. That's what they found in their BWR work. Fibers and
particles are much worse than fibers alone or particles alone.
DR. WALLIS: I haven't read your report; I'm just looking at it
now. It seems to me that what might be the weakest is to split the leads to the
ten percent box up there. If I were a licensee, I might try to hire some experts
in multiphase flow to try to prove by CFD or something that the stuff really
wouldn't get into the sump area the way that you say it does. And is this based
on this casanova? What's the source of the --
MR. MARSHALL: Casanova --
DR. WALLIS: Is this casanova --
MR. MARSHALL: -- concerned debris generation model.
DR. WALLIS: -- a theoretical thing or is it a recent
experiment?
MR. MARSHALL: It's empirical. The way casanova works
are debris generation models. Essentially AutoCAD converted over into a
MATLAB model. We have the entire piping and containment structure digital.
And based on the testing, debris generation testing, we go and say, "For this
size pipe, so many L over Ds, this amount of material will be damaged." And
we tell the casanova model at MATLAB that, "For this material, that
corresponds to ten L over Ds."
DR. WALLIS: So you get the amount of debris generated, but
the transport is the bit which is --
MR. MARSHALL: Correct. The transport is the weakest part.
DR. WALLIS: -- the weakest part, right.
MR. MARSHALL: We're very good, I think, with debris
generation. We're very good with the head loss and accumulation. I think we're
very --
DR. WALLIS: And the transport is based on some integral
experiment or something somewhere?
MR. MARSHALL: Different experiments.
DR. WALLIS: Because I just see sort of five percent, ten
percent stuck in a table, but where does that come from?
MR. MARSHALL: We didn't discuss that in this report. One
other thing we have to do is a lot more documentation of our work. We haven't
published our transport testing and our ongoing debris generation testing. The
main purpose of this was just to tell people what we did. And in the public
meeting, again, since we have so much more time, we will go into the details of
where some of these values came from.
MR. LEITCH: Are there any more questions?
MR. ROSEN: One more point I'd like to make, because I'm
a little suspicious of easy answers to hard questions. And one of the easy
answers we've got is, "Well, we'll just double the sump screen size, area," when
we're talking about potential fixes. If you have significant chemical reactions
going on which could exacerbate the sump blockage phenomena, that might not
be enough. And so we need to hear -- I need to hear a whole lot more about
how effective these things are as filters in this situation, including the chemical
reaction part of it so that we have some sense of what fixes might be like given
the true nature of the blockage.
MR. MAYFIELD: We didn't mean to suggest that we would
propose simply doubling the size. That was earlier regulatory guidance. Today,
I think if somebody was going to increase the screen size, there would have to
be a lot more information addressed. So we agree with you completely. And
that's why the staff went away from the earlier regulatory guidance.
DR. KRESS: With respect to the chemical reactions, if it were
me, I would first try identify what those chemical reactions are and try to judge
their rate, the kinetics, and compare with the kinetics with the kinetics of the
accident itself. You may be able to exclude the significant contribution by that
process without -- if you could exclude it, why then you've got an easier life. If
you can't, then you've got to do something else.
MR. MARSHALL: One thing I'd like to mention is in this work
here we didn't cover all the debris sources. And one reason why was because
from just the thermal insulation we thought we could make a good case that this
is something that needs to be or doesn't need to be addressed. If somebody
was going to redesign their sump screen or try to -- let's say somebody was
required to fix this, they would have to address a larger variety of debris than we
did.
It wouldn't be a calculation that's as simple. It would be
complicated by looking at the larger variety of debris. Then also once you have
this different variety of debris, you have the sump screen, what actually is the
head loss with that material there? And that's going to be different than a
combination of fiberglass and particulates. The work we did here was to more
or less, again, show if this was a credible concern. It wasn't to devise a fix.
MR. LEITCH: Any other questions? Did you have any
concluding remarks? Mike?
MR. MAYFIELD: No, sir. Just we wanted to thank the
Committee for letting us have this opportunity to come brief you.
MR. LEITCH: Then, Mike and Mike, thanks for the
presentation. Mr. Chairman, back to you.
DR. APOSTOLAKIS: Thank you, Graham. We'll recess until
11:15.
(Whereupon, the foregoing matter went off
the record at 11:06 a.m. and went back on
the record at 11:15 a.m.)
DR. APOSTOLAKIS: Back in session. Okay. Next topic,
"Potential Margin Reductions Associated with Power Uprates." Professor
Wallis, this is your topic. Graham?
DR. WALLIS: We have an hour, Mr. Chairman?
DR. APOSTOLAKIS: We have one hour.
DR. WALLIS: The way I propose to proceed is to present to
you a summary of Subcommittee meeting, and then we'll hear from the NRC
staff about the question of whether there's a need for MSRP, which is one of the
questions raised in our Subcommittee meeting. Do we have a quorum?
DR. APOSTOLAKIS: Yes, we do. What, we lost it again?
DR. WALLIS: We don't have a quorum.
DR. APOSTOLAKIS: We don't have a quorum. Okay, now
we do.
DR. POWERS: You're a pivotal member; you can demand
all kinds of things from him.
DR. APOSTOLAKIS: He just did.
DR. WALLIS: I think it's interesting that we should have a
quorum, because one of the questions is going to be do we write a letter, and
we felt that --
DR. APOSTOLAKIS: Among other things.
DR. WALLIS: -- one of the reasons the Subcommittee feels
that we should not write a letter is that this was a Subcommittee study; it's not
been a full Committee study, and full Committees write letters. The other
reason we don't think a letter is appropriate is we don't feel that enough of the
story has been put together for it to be the basis for a letter at this time.
The Subcommittee on Thermal Hydraulics met with General
Electric, the NRC staff, and fellow Gus Korenberg, who made a presentation on
June 12. And we learned that GE has devised a method for increasing BWR
power by the order of 20 percent. How much is actually plant-specific, the plant
has to make its own calculations to justify a plant upgrade -- power upgrade.
This is achieved by flattening the neutron flux profile and
increasing the steam and feed water flow and otherwise changing as little as
possible. So GE's contention is that very little has been changed. There's no
increase in burn-up, most of the LOCA ATWS analysis remain the same, except
perhaps for time of operator response. And, therefore, there is no change in
safety margins, and they stay within established regulations.
The written material we got from GE was a topical report on
constant pressure power upgrade, and we found this to be composed mostly
of words, no analysis, and mostly concerned with classifying what is generic
and what is plant-specific.
Now, we heard from the NRC staff. We heard from Ralph
Caruso. And from that point of view, approved methods are continuing to be
used for safety analysis. As long as the results fit in within the existing
regulations, they can approve a power upgrade. They have few concerns. One
is that result of the power upgrade increases the instability region in the flow
versus power map. And, therefore, the operators have to be more aware of this
and rely on the Solomon-GE method for maintaining stability, and they have to
work around the instability region.
Another concern of the staff is that the use of a code could be
tracked to generate a database in order to establish a correlation for boiling
transition for GE fuel bundles, GE-XL14. Whether it's appropriate for a code to
generate data is the question.
We also heard from NRR about risk. Now, this submittal is
not risk-informed, but the staff follows the guidance of RG 1.174 in order to
ensure that there are no significant risk changes. They concluded that there is
an increase in CDF of about nine percent and in LERF of about 16 percent.
This is almost entirely due to the decreased operator response time and ATWS
events.
Now, we had put together -- ACRS has put together six
questions, six specific questions. You may remember those, my colleagues,
and I propose to very briefly give the staff's answers to those questions. We
asked about power uprates leading to increased frequency release of small
amounts of radioactivity, which are not reflected increases in CDF or LERF. Do
we need to have additional risk-informed acceptance criteria? And the
conclusion of the staff that we need no additional risk metrics to assure that the
risk impact is consistent with the Commission's safety policy statement.
The second question was, the power uprates can be
accommodated by reducing margins between calculated values and limits. Can
the licensee use up all these margins? The conclusion of the staff, I think it's
also the conclusion of the Subcommittee, is that there are rules, such as staying
below 2,200 degrees F, Appendix K. As long as the licensees stay within this
box defined by these limits, they may use up the margin to those limits. That's
the margin we're talking about. I think our conclusion and the staff's conclusion
is that that particular margin belongs to the licensee.
We had a question about power uprates leading to significant
increases in burn-up, and do we reflect this in our PRA evaluations related to
core melt behavior for efficient product release? The answer here is that delta
LERF is used, and the consequences have changed. There is increased
sufficient product release from increased core power, but this is not a concern.
Will power uprates change sufficient product source
associated with gap release and iodine spiking? The conclusion of the staff was
that power uprate has no direct impact on gap fraction and that the 500 spike
multiplier will continue to compensate for uncertainty in the iodine spike rate.
Now we had a concern about the times required and times
available after power uprates for operator actions. Yes, indeed, there is a
shorter response time available, particularly in the ATWS situation, and this
does increase the --
DR. APOSTOLAKIS: How much more is it?
DR. WALLIS: Gee whiz, I have to look it up. It's significant.
MR. BOEHNERT: I think for Duane Arnold it was two
minutes. It went from six to four or something like that. That's what I recall.
DR. WALLIS: Yes, something like that. I remember it was
about a 30 percent change.
MR. BOEHNERT: Yes.
DR. APOSTOLAKIS: Six to four. That's really the important
thing.
DR. WALLIS: So there are larger AGPs, larger human error
probabilities. And this is reflected in delta CDF.
DR. APOSTOLAKIS: Well, I'm very curious how they did that.
DR. WALLIS: And we had a question about the need to
evaluate the live start risk study. And the conclusion there was, yes, indeed,
there was a two percent increase in live start from these increased AGPs due
to operator response time. The increase, because of the increased burn-up,
would increase by 25 to 25 percent. But the overall risk remains small and
within the Commission's objectives. And, therefore, things were okay. So the
picture we got from the staff was that things are okay.
Now, we also heard from RES, and Jack Rosenthal presented
a very extensive research program to assess the effects of operation and
synergistic effects in particular, though in fact they proposed to look at all kinds
of effects of upgrades. This work would take two FTEs and 850k in FY '02. It's
a very broad program, and it's sort of interesting that everything is fine. The
operates are proceeding in September. We have to make a decision on Duane
Arnold, and now we're starting a research program to see if there might be
problems.
What was the response of the Subcommittee? Well, we
heard from Gus. He made a presentation, which this Committee has heard.
Gus' conclusion was that there was a need for the NRC to make independent
calculations to check the validation of the licensee contentions. He felt there
was a need for an SRP. We're going to hear about that from the staff, including
acceptance criteria, that the staff was proceeding, though they hadn't sort of
spelled out what they were specifically looking for or specifically with regard to
operates and were there any criteria that needed to be focused on. And Gus
also felt there was a need to evaluate the potential for diminished margins. He
did make some estimates. The margins, in most cases, had not changed very
much.
What were the members' concerns? Members had some
concern about this SRP issue. They felt that the staff should have laid out more
of the plan for how they were going to evaluate uprates. The members felt that
many of the presentations were very general. There was very little analysis or
data presented. There was a sort of qualitative feeling that everything was
okay, but we didn't see much in the way of hard evidence.
And members had some concern that I've already, I think,
mentioned, that the ACRS is coming in at a very late date to get answers.
We're supposed to sign off on Duane Arnold in September, and we have a
feeling there might be some issues, but we're not quite sure what they are yet.
We believe the operate does indeed decrease safety margin,
it almost has to, and we're concerned -- some members are concerned about
allowing the licensees to use up all the margins when the GDCs call for
sufficient margin. And yet I think we also feel that as long as they stay within
regulations, that particular type of margin can be used up. And one of our
members had a concern about whether the stability monitor was adequate for
dealing with the new power distribution within the corps.
So there we are. It looks, from NRR's standpoint, GE's
standpoint, everything is straightforward. Some of us have an uneasy feeling
that there might be some technical questions to be addressed, materials
questions, for instance, but we don't have a very good knowledge of what those
questions might be.
I think some of the members who are here now who are not
members of the Subcommittee had some concerns about our uprates, which
led us to ask some of these questions earlier, and that may have stimulated the
RES research program. Maybe it's time now to hear from NRR.
MR. ZWOLINSKI: Good morning. I'm John Zwolinski. I'm the
Division Director for our Division of Licensing Project Management in NRR. I
have not had the opportunity to meet with you in some time. My counterparts
are Gary Hollohan and Jack Strosnider. We work directly for Dr. Sharon. I'm
interested in some of the comments that have been made, because as we do
go forward with our technical reviews, it would strike me that those would
certainly be areas that we would want to ensure our staff is focused on,
although I have high confidence that the Systems Division, especially, which
Ralph Caruso and others are working, feel very confident that they have an
approach that is very defensible. And I think we'll have an opportunity to get
into that, as we bring that before this body.
With me, I have the senior executive responsible for power
uprates, Mr. Singh Bajwa, and Claudia Craig is the Section Chief responsible
for this area. And Mohammed Shuaibi is our Lead Project Manager for the
effort, and he coordinates and ensures uniformity and consistency across our
review activities as a Lead Project Manager and ensures that activities with the
project's organization are undertaken consistently.
We're here at your request to discuss the staff's views on the
need for an SRP section for power uprate reviews. However, before we get into
that, I'd like to take this opportunity to highlight some recent correspondence
related to power uprates and some staff activities in this area.
As you may already know, in a staff requirements
memorandum of May 24th, 2001, the Commission directed the staff to make
power uprates a high priority, noting the situation throughout the country as far
as the need for power. I surmised that was part of the genesis of directing the
staff to assure we conducted business in a timely and effective and efficient
manner. The Commission also directed the staff to work with stakeholders to
identify potential areas for improvement in the current power uprate review
process to assure that these processes do not impose needless requirements.
This comes back to one of our key pillars of unnecessary regulatory burden.
Earlier this week, Dr. Travers signed a Commission paper
documenting the staff's response to the staff requirements memorandum. I
believe that was late Monday. The Commission paper provided a status of
ongoing activities in the area of power uprates, including plant-specific reviews,
generic topical reviews, and ongoing and future work related to improving power
uprate processes; that is the actual review itself.
In summary, we told the Commission that we considered
power uprate applications to be among the most significant actions being
conducted and reviewed by the staff, and we will, to the extent practical,
accelerate review schedules, assigning additional staff where appropriate in an
attempt to ensure that we fulfill the Commission's expectation of maintaining the
focus on these being a high priority.
We also told the Commission that we are evaluating the
review process for the measurement uncertainty to capture uprates to improve
the effectiveness and efficiency of these processes. This is the small, typically
on the order of about 1.5 percent, uprate. Some of us struggle, is that really a
power uprate or is that a flow uncertainty elimination, and thus we're more
accurate in measuring flow.
DR. WALLIS: Yes, we've already considered that one. I think
that the Committee is in agreement that this is a good idea.
MR. ZWOLINSKI: Okay, very good.
DR. WALLIS: We did that, I think, about a year ago that one.
MR. ZWOLINSKI: We did have topical reports in from a
number of vendors on this particular issue, and that particular measurement
device has been in use for some time.
We will be conducting a similar evaluation of the review
processes for extended power uprates following the completion of the first few
that are currently under review. And this was alluded to earlier. That's Duane
Arnold, Quad Cities, and Dresden. In other words, once we've got these
technical reviews completed, we will go back and ask ourselves, are there
efficiencies that can be identified or garnered or areas that we may or may not
need to look as deep into? So scope and depth of review and the areas that we
review will be at least challenged by us.
DR. WALLIS: So that's the time that if you -- I think you're
proposing if you did devise an SRP, that you would do it then after you've
learned from these three reviews.
MR. ZWOLINSKI: We really do want to have the lessons
learned of these reviews. And as we get into some of the presentation, I think
you'll see the flow of thought.
The staff, as you said, is scheduled to present these reviews
of these applications in the September and October time frame.
Contained in the paper, which is, as I said, a status paper to
the Commission, are enclosures, and it's interesting to note that we did survey
the industry. This is a very interesting concept, especially for the boiling water
reactors. But we are doing some reviews of power uprates, albeit much smaller
than -- something as much as 20 percent for the PWRs.
With the economics and the deregulation and various factors
that are coming into play, it's not surprising that we found a considerable
number of licensees contemplating some amount of power uprate, whether it
was simply the uncertainty power uprate or something a little bit more than that.
A fair number of licensees, on the order of 40 to 50 percent of the industry, is
looking at this issue very hard.
By the way, there have been a lot of power uprates that we
have reviewed and approved over the last 20, 25 years, and so there's learning
from those that led to a fairly extensive study that was performed following the
issues that arose at Maine Yankee. And out of the issues of small break LOCA
and things of that sort, the staff did do a rather rigorous analysis of prior power
uprates, and thus that led to the process and procedures and review
methodology that we're currently using today.
Mohammed's going to get into this. I'm stealing a little bit of
his presentation, but I felt it was important to say that we had baselined our
technical work, and I'm confident that was part of the basis that led people such
as Mr. Caruso and others to speak. So with that, unless there's questions --
DR. WALLIS: While you're giving the overview, power
uprates are also part of the industry interest in risk informing some of the
regulations, I think. They're hoping that risk-informed regulations will give an
opportunity to justify some power uprates.
MR. ZWOLINSKI: Yes, sir.
DR. WALLIS: So you folks are also following that line, I
imagine.
MR. ZWOLINSKI: Yes, yes, very much so.
MR. SHUAIBI: Thank you, John. Again, we're here today to
talk about the staff's views on the need for the SRP section for power uprates.
I would like to just add one more thing to what John said about the comments
earlier. When we came here, we did say that the reviews for the Duane Arnold,
Dresden, and Quad Cities were still ongoing. We didn't have a lot of detailed
information at the time, because we're still conducting our reviews. Those are
not done. And we have received some questions from ACRS, and we're
looking at those now, going back to see that we have all bases covered. And
if there were additional questions to the ones that we had identified, we're going
to be asking the plants to address those. Those questions were specific to
Duane Arnold, but we did send them out to the other PMs as well, project
managers, to make sure that in their review of the Quad Cities and at Dresden
they also look at those areas as well.
With that, I'd give you an overview of the presentation today.
What I'm going to do first is I'll give you a brief background on how the idea for
an SRP section came about. Following that, I'll talk about what use today for
current guidance for power uprate reviews. I'll get into ongoing activities that
have the potential to change the power uprate review process. And some of
that came out of the Commission direction in the SRM. And I'll conclude with
the staff's views on the need for a power uprate and also the timing -- I'm sorry,
a power uprate SRP and the timing for putting one together if we should decide
to do that.
As far as background goes, in December of 1995, we
received an allegation that the Maine Yankee licensee used inadequate
analysis for their power uprate. Following that, in January of '96, we ordered
the Plant to be limited in power level to the original power level that was
licensed that was prior to the uprate in question. In April of '96, the NRC formed
a Maine Yankee Lessons Learned Task Force, and the purpose of the Task
Force was assess performance in three key areas: One was code reviews, the
other was the power uprate review process, and the third was really staff
interfaces related to issue closures and document generation.
The focus of this presentation is on the second item, which is
the power uprate review process. But let me say that on the code reviews, we
have actually drafted and issued a draft SRP section and a reg guide, which I
believe the ACRS was involved in. We received comments from the public, and
we're in the process of evaluating those comments.
DR. WALLIS: Since you mentioned those, we have been
waiting a long time for those to be finalized. I think ever since I've been on the
ACRS this has been coming, and it seems -- we just don't want to get it stuck
somewhere and forgotten about. We'd like to see it finished.
MR. SHUAIBI: Well, I think we have actually issued them for
public.
DR. WALLIS: They've been out for public comment --
MR. SHUAIBI: Yes.
DR. WALLIS: -- for quite a while.
MR. SHUAIBI: Yes. We're in the process of evaluating public
comments, but I'll take that back as you'd like to see that accelerated.
DR. WALLIS: Yes. Don't divert the effort to somewhere else
and then forget about those.
MR. ZWOLINSKI: As a memo, we should be able to provide
the Committee with current schedule and bring some closure to that particular
question, and we'll do that.
MR. SHUAIBI: We'll take that back. On the last item on the
-- as far as staff interfaces and document generation, we have gone back since
Maine Yankee and reviewed some of the TMI action items related to the Maine
Yankee experience. And we also have added guidance to the reviewers of
updates to the SRPs -- I'm sorry, updates to the FSAR sections and plant
documentations to make sure that power uprate-related material gets reflected
in the updates to the FSARs.
After the Main Yankee Lessons Learned Task Force in July
of 1996, and independent safety assessment was conducted at Maine Yankee.
In October of '96, the Independent Safety Assessment report was issued, which
recommended that we go back and review our process for the power uprate
reviews. In November of 1996, we were directed by the EDO to address the
recommendations in the Independent Safety Assessment report. In December
of 1996, the Main Yankee Lessons Learned Task Force issued their report, and
that also recommended that we go back and look at the review processes for
power uprates. And, finally, in April of 1997, NRR committed to develop a
standard review procedure for power uprates. So that's how this came about,
the idea for --
DR. WALLIS: So you don't want another situation like this
where as a result of, let's say, a future uprate, you get a similar situation where
--
MR. SHUAIBI: No, we certainly don't want that.
DR. WALLIS: -- someone says, "You should have had an
SRP. That's why you got into this mess."
MR. SHUAIBI: That's right. No, I'm going to talk about what
we do today, which has a lot of --
DR. WALLIS: So you had a talk in '97 to do something, and
it hasn't yet appeared, is that the case?
MR. SHUAIBI: Well, that's not totally true. I mean we have
a process that incorporated the lessons learned from Maine Yankee. I'm going
to get into that, and I'll actually go through an example of how that works with
you. We're pretty confident that what we're doing today is --
DR. WALLIS: So you have a sort of pseudo-SRP, is that what
you call it?
MR. SHUAIBI: Yes.
MR. ZWOLINSKI: Before Mohammed goes on, I was deeply
involved in a lot of this activity, either leading or co-leading much of these
efforts. One of the sensitivities had to do with our indexing and maintaining a
currency as far as our knowledge of all the codes that we rely on in the Office
of Nuclear Reactor Regulation. There are a lot of codes that go beyond thermal
hydraulics.
For example, over in the Engineering, in Jack Stresnider's
Division, we have a wide variety of codes that were part of this overall comment
that came out of this Lessons Learned Task Force. In so many words, does the
Agency have a good means to understand which codes have been
reviewed/approved and which codes are being used by the industry? And we
found that we were -- we needed to improve in the area of thermal hydraulics.
We concluded that many of the codes were used more in the Engineering area
had much better documentation.
So this was a very broad effort that narrowed into the thermal
hydraulic area. So while we share with the Committee kind of a one-liner, this
had a lot of horse power and a lot of effort behind, and it was quite a broad
activity.
DR. BONACA: These are the codes that you used to
evaluate the performance of the plant, right?
MR. ZWOLINSKI: Yes, sir.
DR. BONACA: And the condition it's on.
MR. ZWOLINSKI: Yes.
DR. BONACA: Now, at some point, since I was not on the
Subcommittee, one point of I've expressed is typically evaluation consists of
analyzing outcomes of events and accidents and so on and comparing to set
the limits. And what you do for a new plant you compare against the design
limits of components for that particular application. What I mean is that if you're
evaluating the blowdown forces or stress on a component, you compare that
with the capability of the component for that stress level.
When you do a power uprate, you do an evaluation using the
same hydraulic codes, but it seems to me there is an implicit assumption that
those design limits you're comparing to don't deserve any evaluation to address
issues, such as, for example, aging. And that's one of the concerns we have
expressed, I think, in more specific examples. I just would like to know from
you, at some point in the presentation, whether in fact you are asking a licensee
to give you, for example, a summary of the operating experience of the plant?
You know, we have seen for BWR we looked at some
components we have compared things for had to be replaced, because they
were found to be cracked or possibly cracked. Certainly, those components
were not capable of their original design capability. So that raises the question
of whether or not when you increase the power plant by 25 or 30 percent, there
should be also an assessment of the design capability of certain components,
which now are much close or somewhat closer to the performance of the plant
in case of accidents.
MR. ZWOLINSKI: I think I understand your question. Can I
ask that Mohammed continue with his presentation --
DR. BONACA: Oh, yes, please.
MR. ZWOLINSKI: -- and if we don't answer, I'll be more than
happy to revisit that.
DR. BONACA: Yes. If there is a way, at some point, where
we can talk about that.
MR. ZWOLINSKI: Yes. Okay. Very good.
MR. SHUAIBI: As far as current guidance, what we use today
we have four approved GE topical reports. We have two for stretch operates,
one from 1991, one from 1992, and two for extended power uprates. One was
approved in '96 and the other in '98. These are only for BWRs, but they do
provide guidance for the staff in terms of what areas need to looked at for a
power uprate and the depth of review that has to be done. And I believe ACRS
was involved in the reviews of those topical reports.
DR. WALLIS: I guess our criticism, or the criticism I've voiced
on behalf of this Committee, was that many of these reports are full of words.
Do you go thoroughly into sort of the technical basis for these words when you
review the reports?
MR. SHUAIBI: Let me -- can I just continue, and after I'm
done with this slide, I may addressed your question. But we can come back to
it right after I'm done with this slide.
In addition to those topical reports, we have implemented the
use of template safety evaluations. We have a safety evaluation for Monticello
and one for Farley, which were both done in 1998. Both of those incorporated
all the lessons learned from Maine Yankee, so we felt that those were very good
safety evaluations, and we use those as templates for reviews of any of our
power uprates today.
In addition to that, when our reviewer gets into a section,
whether it's because a GE topical report included it or whether it's included in
the safety evaluation for these plants, and the reviewer needs further guidance
on the area they're reviewing, they would go to the applicable sections of the
current SRPs.
Now, to address your comment in terms of how this gets
done, a reviewer could pick up the safety evaluation and see the scope of
review that has to be done and what has to be done. They also see from the
write-up the level of detail that they have to address or how deep they have to
go. And they could either use a safety evaluation for one of these plants or if
they need more guidance, they can go to the SRP, and the SRP would have
that guidance for them.
DR. WALLIS: So the assumption is that you already know
enough to do this. There isn't something about these extended power uprates,
which are fairly large, that take you beyond some region of knowledge or
judgment that you used in the past.
MR. SHUAIBI: Well, we recognize the recent submittals at
Duane Arnold, Quad Cities, and Dresden and Clinton now as first-of-a-kind
power uprates. And we will be learning from those. There may be areas where
we may need to focus more effort.
DR. WALLIS: Maybe after they've done the power uprate and
the correct experience with --
MR. SHUAIBI: Well, I think as we're going through them,
hopefully not after the fact. As we're going through them.
DR. FORD: I guess the frustration here is that in the topical
reports, for instance, from General Electric, there's numerous references to the
increased flow rate will not have an effect on carbon steel, for instance.
Whereas we know phenomenologically that it will have an effect. But there are
no data to support the contention in the reports. That is the frustrating part, and
we are being asked to sign on that, "Hey, everything is fine." We've got no data
to --
MR. SHUAIBI: Are we talking about the new topical report
that's been submitted?
DR. FORD: Yes.
MR. SHUAIBI: I think what I'd like to do is go back to the
topical reports that have been approved a little bit. In there, I think what GE
tries to do are two things: They try to propose a way or a format for submittal
for the plant to come in, and then they propose a second topical report which
gets into the technical details of what is to be submitted. So in the first case,
yes, GE would say, "This is the kind of information that you have to provide."
The second report would get into the technical discussions of what that
information is.
On this new CCPU topical report that you're talking about,
actually that one has been put on hold. We've given them a lot of comments
back on that report, and they've asked us to stop work on it. So I don't know
that we're at the point that we want to discuss what that report's going to look
like at the end. We don't know what's going to happen with that topical report.
DR. WALLIS: I noticed quite a bit in that topical it leaves quite
a bit up to the licensees. So maybe some of the sort of question that my
colleague asked here is not to be answered by GE; it's got to be answered by
Duane Arnold or whoever it is.
MR. SHUAIBI: That's correct. Again, we will identify those
areas. In most cases, we will go back to the -- we do go back to the safety
evaluations, for example, for Monticello, and then the SRP and use as guidance
for how we would do those reviews. But, yet, there are plant-specific areas that
GE isn't going to analyze or put in their topical reports.
MR. ZWOLINSKI: This most recent issue that arose where
we asked Duane Arnold a set of questions and then we passed also along to
the other licensees, we don't believe those individual licensees are going to be
able to answer those questions. I think it's targeted to General Electric. So we
get to General Electric through the licensee so that we place the burden on the
licensee to --
DR. WALLIS: And they go right back to GE, because GE has
the expertise to answer the question.
MR. ZWOLINSKI: Well, the vendor needs to undertake work,
at least in the set of questions that were just generated. We're aware that they
have done a lot of work. And to make a long story short, we believe that they
can answer the questions. We'd like to see what the answers are.
DR. WALLIS: I think some concern is that maybe -- whether
the old questions you've been asking are adequate for the new situation. That's
the bit of concern we have is that you're extrapolating your judgment and
knowledge from lower power and so perhaps assuming or judging that there's
nothing new that you have to worry about.
MR. SHUAIBI: Well, I think, to go back to an older point that
was made earlier, when we do these power uprate reviews, we don't just review
the, for example, Chapter 15 analyses. We do component reviews to make
sure that those components can still operate in the same fashion that they're
described in the submittal. So I think all that gets addressed.
Now, are there any new areas? I think research, like you
said, is to undertake a study to see if there are any new areas. I think some of
the synergies that were mentioned during the last meeting had to do with things
that aren't being done in parallel here for these power uprates. But we do look
at not just the Chapter 15 analysis; we look at components to make sure they
can work in the uprated condition.
DR. WALLIS: These research results may come in after all
the GE plants have been uprated.
MR. ZWOLINSKI: I believe that was a two-year program --
MR. SHUAIBI: Yes.
MR. ZWOLINSKI: -- and the number that we'll be processing
over the next two years is not a large number.
DR. WALLIS: It's not a large number?
MR. ZWOLINSKI: No. And by the way, I would also suggest
that we've encouraged Research to ensure that they've given a healthy look at
PRA methodology to look across these facilities as to changes that have been
made over time. And now this touches on a myriad of different areas within a
plant and to what extent does it affect the PRA? So I think this -- I view the
research initiative to be one of complementing the staff's activity and
independent validation of our review process.
DR. WALLIS: It's very striking if you look at some of the PRA
results. It seems that nothing has changed except for this operator response
to ATWS. And it seems remarkable. Maybe it's just a function of the excellent
design of the plant, but that's the only thing that changes.
MR. SHUAIBI: In some areas, they are changing the way that
these plants are being operated. I know one example is a plant that's using
more feed water pumps for full power operation than what we use today. And
our PRA folks are looking at that to make sure that that --
DR. WALLIS: Where balance of power changes.
MR. SHUAIBI: That's right. That's right. And that would be
reflected in their risk assessment.
DR. WALLIS: And that's plant-specific, isn't it?
MR. SHUAIBI: Yes.
MR. ZWOLINSKI: And that's why I think the PRA can be of
great use.
MR. SHUAIBI: I think I've done this already, but I'll walk
through it anyway. What I have here is an example of how this process would
work. Again, I randomly pick containment system response. I actually opened
the Monticello safety evaluation to a page and said, "I'll just use that one." You
could do it with, I believe, anything in there. For a boiler, I take the Monticello
safety evaluation -- a reviewer would take that in his area or her area. They
would be looking for containment system response. They'd look in there and
see that, yes, that has to be addressed, and it would have a discussion to give
you an idea of the depth of review that has to be performed. For BWRs and
BWRs only, we have the GE topical reports as well. So if a reviewer also looks
at the GE topical report, he'll see that they need to review that area.
DR. WALLIS: So, I'm sorry, what you're saying is that
someone who's reviewing Duane Arnold, instead of turning to an SRP to say
what shall they do, turns to the precedent of Monticello and then goes with that.
MR. SHUAIBI: Well, if we were today to write an SRP section
for power uprates, what we would have to do is identify the areas that have to
be reviewed for the power uprate and then provide the guidance that goes
along with that area. Now, we wouldn't go back and rewrite the SRPs. What
we would actually do is identify the areas that have to be reviewed and then
point the reviewers to the right sections in the current SRPs unless those need
to be modified, which I don't think they would.
So what the Monticello SE effectively does is that. I mean
you've got an SE that outlines the areas that have to be reviewed, and then the
reviewer can go to the SRP for further guidance.
DR. WALLIS: So you're saying that there already exists an
SRP and that we would have to --
MR. SHUAIBI: That's why I agreed when you said pseudo-
SRP, yes.
DR. WALLIS: All right. And you don't need add substantially
to it or maybe you don't need to add to it at all. You just need to point to
sections of it.
MR. SHUAIBI: Right.
MR. ZWOLINSKI: And then the question is if we've got
pointers to a myriad of different sections of our standard review plan, do we
need to construct an overarching power uprate SRP that has these pointers and
any other -- and for lack, I think, of a better term -- synergistic effects or effects
that would not be considered individually but more as they accumulate. Those
types of things, I would think, would be the part that would actually be in the
SRP section if we were to develop it, along with all these pointers. I don't want
the Committee to think that we're not using the existing standard review plan
extensively. As you go through the safety evaluation, you will find the pointers
into our SRP. We call this particular type of review the template review.
DR. BONACA: Could I ask you for the percent power
increase that you had for Monticello and Farley?
MR. SHUAIBI: I'm sorry?
DR. BONACA: What was the power uprate --
MR. BOEHNERT: Monticello was about 6.6, and Farley I
don't remember. That was below five, wasn't it?
MR. SHUAIBI: Yes, I believe that was five.
MR. BOEHNERT: Yes, five, which we didn't look at it.
DR. BONACA: Does the extent of the power uprate have any
bearing on the guidance that you need to provide?
MR. SHUAIBI: The areas you would look at would be the
same. I mean the SRPs and what we would be reviewing are the same. Now,
the extent of the power uprate is obviously going to have an effect on the
analyses and the results of the analyses. And possibly even, when you change
the operation of the plant, like the example I gave earlier, where you have more
feed water pumps required to run the plant, yes, we would be looking at that.
But that's part of the review today. When we look at the risk impact, we look at
how the plant will be operated and how that is reflected. And when we look at
the analysis, we do look at what effect does this increase in power have on it?
Now, whether it's a one or a 20, we would still be looking at that effect.
Again, this process was implemented following the Maine
Yankee experience to address the Maine Yankee's lessons learned. So we see
this as a pseudo-SRP, as was said earlier.
What I'd like to talk about now are potential changes, ongoing
work that has the potential to change the current review process for SRPs. I
think John already talked about the staff requirements memorandum of May 24
this year, where the Commission directed the staff to make power uprate
reviews a high priority, assure that our review processes don't have any
needless impediments, and to not unnecessarily delay licensees' plans for
implementing power uprates. We responded to that SRM, and John already
talked about that, so I'll just move on to the different power uprates and what
we're doing in those areas.
DR. WALLIS: So you have an incentive to proceed quickly.
MR. SHUAIBI: Yes.
DR. WALLIS: And when one proceeds quickly, there's always
a risk that one forgot something. But I agree that this is appropriate.
MR. ZWOLINSKI: What we said in the paper, and I think it will
be released in the near-term, is that the Agency has performance goals for
licensing actions. And we see this as the type of action that should be
completed within about a year's time. Direction to our staff is to ensure it's
carried as a high priority item. So staff will be made available to conduct the
reviews to meet that type of review schedule, which is not abnormal.
I will say some of the previous power uprate reviews possibly
took on the order of 18 months and even a little bit longer. But had we
dedicated our staff, they would have been much shorter reviews. We're
assuring our staff is targeted to complete the review in about a year's time.
MR. SHUAIBI: And to address your point about moving
quickly and what happens when you move quickly, we recognize and stated in
the paper that the Duane Arnold, Quad Cities, Dresden, and the big power
uprates are really first of a kind. And we said that improvements to that
process, as far as gaining efficiencies, we will look at that following our review
of those applications.
My bullet on this slide, the measurement uncertainty power
uprates, we're currently reviewing the process for those. Again, those are the
ones on the order of one and a half percent. We're looking to see where we
could gain efficiencies in that. And we also have a GE topical report that
addresses the small power uprates.
With respect to extended power uprates, we're currently
reviewing, again, first of a kind. We recognize the first-of-a-kind submittals from
Duane Arnold, Quad Cities, and Dresden. We've also received an application
from Clinton, which I believe you're aware of, for a 20 percent power uprate. All
of these are for uprates greater than 15 percent, and will be presented to the
ACRS soon. I think the Duane Arnold will be presented in September, Quad
Cities and Dresden are coming in October, and Clinton is probably next year.
It's not -- it just came in.
DR. UHRIG: Which plants are the BWR-3, 4, 5?
MR. ZWOLINSKI: I think the Clinton is BWR-6.
DR. UHRIG: Six. What are the others?
MR. SHUAIBI: Threes? Threes.
DR. UHRIG: They're all threes, okay.
MR. ZWOLINSKI: Yes, these are all threes.
DR. UHRIG: All threes. Okay. Thank you.
MR. SHUAIBI: Well, following reviews of the Duane Arnold,
Quad Cities, and Dresden, we are planning to have a lessons learned workshop
with industry to share our lessons learned from those reviews, possibly areas
where we will need to get into, areas where we could gain efficiencies, that kind
of information, and looking for feedback from them on how we could improve.
We will also conduct a review of the process following the
reviews of the Quad Cities, Duane Arnold, and Dresden to see what we can do
in terms of gaining efficiencies. And we did have the GE topical report on
CPPU power uprates, like I said earlier, that we were asked to stop work on that
because of comments that we gave back to GE, but we'll be meeting with them
soon to discuss their plans on that.
MR. ROSEN: What's CPPU?
MR. SHUAIBI: Constant pressure power uprates, I'm sorry.
DR. FORD: Given the time content that we've got here that
we're going to be asked to decide upon Duane Arnold or give comment on
Duane Arnold, is there any way at all that we can see what the questions are
that you are giving on, for instance, the CPPU report, so at least we can give
our advice on the adequacy of those questions?
MR. SHUAIBI: We haven't issued any REIs on the CPPU
topical report. That was a separate effort from the Duane Arnold activity. As
far as Duane Arnold, Quad Cities, Dresden, and now Clinton, I think we've
provided ACRS with all the incoming and REIs that we have to date. If you
would like the comments that we shared with GE on the CPPU topical report,
we could look into getting that for you. But those are separate activities.
DR. WALLIS: But these are uprates that use the CPPU idea,
constant pressure.
MR. SHUAIBI: Well, they don't use the topical report. Now,
remember, the topical --
DR. WALLIS: The issues would be the same. The method
of achieving the power uprate is as described in that report.
MR. SHUAIBI: True, but I think what you're getting into on the
CPPU topical report is, again, the GE -- usually, again, GE comes in with two
types of reports. The first one lays out the format of a submittal. The second
one talks about the technical discussion -- you know, provides a technical
discussion of the uprate. The CPPU topical report was more on formats. A lot
of the comments were on format. As a matter of fact, many of the comments
was on their approach of a checklist.
DR. WALLIS: Yes, it's in regulatory space; that's not in
technical space.
MR. SHUAIBI: That's right. Many of the comments are in that
area. I don't know if they would help. That's why these are separate efforts.
I think Ralph Caruso wants to make a comment.
MR. CARUSO: This is Ralph Caruso. I just want to let you
know with regard to the Duane Arnold, Quad Cities, and Dresden, I believe that
those were done in accordance with one of the earlier topical reports. The two
that are used most often are ELTR-1 and ELTR extended. Something licensing
-- Extended Licensing Topical Report 1 and 2. Both of those reports have been
presented to ACRS, and we've discussed them with you. And these three
uprates, and I believe also Clinton, are in accordance with those two topical
reports.
MR. SHUAIBI: I mean they may be constant pressure, but the
CPPU topical report was mostly on what the licensee presents in terms of
content or in terms of format. The CPPU --
DR. WALLIS: This is a problem that I think ACRS, or I always
have with these things, is the scent of a paper trail. In order to find out if there's
any technical issue we ought to worry about, you have to find out which report
refers to which report, which refers to which report, who's really relying on what.
And that can be a lot of work.
MR. SHUAIBI: For the Duane Arnold, Quad Cities, Dresden,
and Clinton, those should point to the 1996 approved topical report and the
1998 approved topical reports, which are called ELTR-1 and ELTR-2, not the
CPPU topical report.
DR. WALLIS: We saw those before. I remember them, yes.
MR. SHUAIBI: Okay. Yes.
DR. UHRIG: One other question. Going back to the
uncertainty recapture power uprates, these are all ultrasonic measurement
devices or are they the correlation?
MR. SHUAIBI: Ultrasonic.
DR. UHRIG: Ultrasonic. Do you know if this is a four-
channel, single-pass system or the X?
MR. SHUAIBI: What we have approved to date is the single-
pass system, and there has not been a submittal for the --
DR. UHRIG: The new one.
MR. SHUAIBI: -- for the new one.
DR. UHRIG: Okay.
MR. SHUAIBI: We understand that some of the plants would
like to use the new one, but we have not approved that yet. We would have to
review that submittal first.
DR. UHRIG: Okay. Thank you.
MR. ROSEN: Do you have any corresponding activity such
as this on the pressurized border reactor side for extended power uprates?
MR. SHUAIBI: As far as the template reviews, we do use the
Farley, which is a PWR. Reviewing the process, we're reviewing the process
for all, not just for BWRs. As far as topical reports, we don't have an approved
topical reports for a PWR. If they want to submit a topical report, we would
review it, but we don't have one approved today.
MR. ROSEN: And there are no licensees who have come in
-- PWR licensees.
MR. SHUAIBI: Not for 15 or 20 power uprates, no. No,
they're not at that --
MR. ZWOLINSKI: It's our understanding that PWR
community is looking at the feasibility of trying to provide the staff a topical
report, but I think if you stop and assess each of the PWRs, there are -- the
likenesses that you find in the boiling water reactor device -- the pressurized
water reactor, they're going to have some difficulties in making that a broad,
generic-type report. Whether we talk to various venages or we get into various
manufacturers or two or three flow, it's just a lot of different issues that will
probably make that effort very difficult.
DR. UHRIG: It's also not likely to be talking 20 percent either.
MR. ZWOLINSKI: I agree with that.
DR. UHRIG: Maybe five.
MR. ZWOLINSKI: Many of the power uprates that we've
approved over the past several years have been for very small power uprates,
typically less than five percent for the PWRs. And those were part of the review
when we did the Maine Yankee lessons learned. Did we do a quality review --
and I'm going back in history now -- when we approved any power uprate prior
to 1996? And that led to the template that we're talking about that we used for
Monticello and Farley.
DR. UHRIG: Well, there were many plants that were -- the
analysis was done at one level, and then their initial license was a lower level.
And then they went up to the higher level. We went through this in St. Lucy.
MR. ZWOLINSKI: Yes, you're correct. Their analysis is
essentially at a higher power level, and for a variety of reasons they operated
at some number less. And then they ask to be able to go back to the higher
number. Those types of reviews were not very extensive. They're predicated
primarily on --
DR. UHRIG: Because the original review was done on that
basis.
MR. ZWOLINSKI: Exactly.
MR. BOEHNERT: Bob, ANO has a review in-house for a 7.5
percent uprate. Yes. And Point Beach announced that they're going to be
seeking a 10.1 percent uprate.
DR. POWERS: If the staff were asked to approve the 1994
decay heat standard, what kind of power uprates do you think would be feasible
for PWRs?
MR. SHUAIBI: I don't know. I would have to get back to you
on that. I don't know if Ralph has an answer, but I don't.
MR. CARUSO: This is entirely speculation on my part, but I
don't think decay heat is what's limiting the PWRs. I think it's steam generators,
and that's why the power uprates, the ten percent power uprates are the plants
that are replacing steam generators. They can't get the heat out.
DR. BONACA: And flow. They're flow-limited too.
MR. CARUSO: Right.
DR. BONACA: I mean, simply, you just have -- there in PT
envelope in DMB in a good portion of the operating range.
MR. ROSEN: They're RCS flow-limited.
DR. BONACA: Sure, RCS flow-limited, yes.
MR. ZWOLINSKI: In fact, I'm reminded on the Arkansas
docket they did just replace their steam generators, and in fact we are finding
as licensees replace their steam generators, many are not exactly like for like.
They are going to larger steam generators. And Ralph accurately characterized
that's the probably unequal component. Very expensive uprate.
MR. SHUAIBI: I guess the last bullet --
DR. BONACA: I had -- okay, I'm sorry.
MR. SHUAIBI: No, I'm sorry, go ahead.
DR. BONACA: I asked the question before, but I didn't get an
answer yet. So before we get to conclusion, I would like to hear an answer.
MR. SHUAIBI: Okay.
MR. ZWOLINSKI: If I could go back to the review processes
that we intend to undertake, I don't want to predict what the results of that will
be. It could lead to additional review in selected areas or it may modify the
scope and depth of the review in some areas. It would be premature to
speculate the outcome. We are looking for efficiencies in our processes.
DR. BONACA: I understand that. The question I asked was
purely to do with the fact that since the plants are aging, wouldn't it be -- and
you're pushing the plant closer to these operating limits. Wouldn't it be prudent
to look at past history, try to understand if in fact all those components which
are being compared are still capable of what they were capable when they were
designed.
MR. SHUAIBI: Power uprate reviews involve a lot of different
branches within NRR, and you have the Chapter 15 analysis, which reactor
system DSSA does, and you have the component analysis, which DE does.
And when we review power uprates, we do look at component's ability to
function at the operated power level. So we do look at that. We do rely on ISI,
IST programs, as well, to feed back into the process, make sure that, you know,
as far as history goes, IST, ISI will provide that kind of information.
I don't know if that fully answers your question or not, but we
do --
MR. ZWOLINSKI: Just as I believe there is an integrated
effect when you make changes to your plant, the Agency's appropriate
response would be there's an integrated regulatory effect, including our
inspection program, but also the fact that through our oversight program there's
a number of areas we continuously probe, and during the license review there's
a variety of areas that we are probing.
Some of these we have come to the conclusion the best
vehicle or device that we have to assess more macroscopically might be use
of the PRA. But that will not necessarily get me to is my equipment qualified to
perform over the next 40 years, which could be a question. We're faced with
those kinds of questions, for example, in license renewal. We're essentially
asking ourselves, "Do we need to ask those kinds of questions during this
review?" And as appropriate, we do.
DR. BONACA: Okay. I think I got an answer, and the answer
includes an answer satisfactory, in part, because my main concern was that this
would not become, when you get to this kind of power uprates, just simply
checking certain components against a checklist.
MR. SHUAIBI: Oh, no. No. And maybe when we come and
talk about some of the applications that we're reviewing now, maybe we could
talk -- maybe pick an example and go with that. Maybe that will give you a
better feel for what we do in those areas. But we do review the components
that we rely on for the analysis to make sure that they will operate at the higher
power level and provide the functions that they're relied upon.
DR. WALLIS: Are we ready to move to conclusions?
MR. SHUAIBI: My last bullet, though, I'd like to say that we
will issue guidance based on these reviews or whatever changes we make. We
have a risk process where we could issue guidance through a risk. We may
also put things on the NRC's external web site to make sure everybody's aware
of the results of these activities.
In conclusion, I guess using the template safety evaluations,
approved topical reports for the BWRs, and the current SRPs, we believe that
sufficient guidance exists for the reviewers to conduct power uprate reviews.
On the bigger power uprates, of course we'll be looking at that as we're doing
the reviews of those applications to see if there's anything that needs to be
added or areas where we can improve.
We're considering explicitly identifying the Monticello and
Farley safety evaluations -- those are the ones that we use as templates -- in
the project managers' handbook. What that will do is it will identify to a project
manager for a plant that when he gets a power uprate review to go back to the
Monticello or Farley, and maybe even include it on the work request that goes
to the technical branches. Everybody's aware of those, but we could formalize
that by putting it in the project managers' handbook.
Again, the processes may still change. They are changing.
We expect the measurement uncertainty power uprate process to change.
There may be changes as a result of the first-of-a-kind applications for the
major power uprates. So the process is still dynamic. If we were to write an
SRP section, I don't know if this is the right time to do that.
Resources right now are really needed for reviewing the
current applications that we have. If we were to write an SRP section, we would
have to take the experts working on those applications to develop the SRP or
make sure that it's done correctly. And we need those reviewers on the
applications that we have. And we can always reevaluate. We will reevaluate
the need for an SRP section in the future, especially when we do the review of
the process.
DR. WALLIS: Is there any other questions?
MR. LEITCH: Your arguments are compelling, but John and
I were both deeply involved with the Main Yankee situation, and then there was,
as you indicate, a Maine Yankee lessons learned. And one of the outcomes
from that lessons learned was that a standard review procedure would be
developed. And I guess what I'm hearing is that that has not been done.
And I guess one of the benefits of any procedure is to
document and formalize your institutional learning so that you have a structured
way to go through a particular process and benefit from the experience and so
forth. And I just hope we're not digging ourselves down in that same hole that
we were in at Maine Yankee again. And I mean this conclusion was definitely
coming out of that report, and yet we're saying now that, well, we're not going
to do that now.
MR. SHUAIBI: Yes. I don't think that we're quite saying that
we don't need it. What we're saying is the concerns that the Maine Yankee
lessons learned recommendations were after was a scope and depth of review
when it came to the review process. What is it that you want to review and how
deep do you need to go when you're reviewing it?
Now, following that, we have implemented this template
review. You know, we didn't have this template review process before Maine
Yankee. So we see it as, yes, we have taken some action. We haven't actually
written an SRP section, but to address the scope of review, in terms what you
look at, and the depth of review, in terms of how deep do you go when you
review it, I think that those template safety evaluations probably capture that
pretty good.
Now, again, if I were to go and write an SRP section for
review of power uprates, what would I want to do to address scope and depth?
Probably the same thing that these templates do. It would be in the SRP. That
would be the only difference. It may not include as much of the text in the safety
evaluations, because they're plant-specific reviews, but, essentially, you'll get
the same thing. I mean that's the argument.
DR. WALLIS: The material is there already, you're saying.
The material that you need is --
MR. SHUAIBI: Yes.
MR. ZWOLINSKI: If I can help bridge the gap just a little bit.
We looked at all the power uprates that we had done, and we looked at what
technical branches contributed to those power uprates. So if this power uprate
had four branches and this one had seven branches, and this one had three or
nine, we asked ourselves, "What is the appropriate set of review sections in the
standard review plan?" And that ultimately was codified and used in the
Monticello docket and on the Farley docket. So we took the areas that were
reviewed and not reviewed, the plants of yesterday, and going forward laid out
a list that was much more detailed and very extensive that we asked our staff
to look at each one of those individual areas. So that's a fairly complex and
lengthy review list.
For the uncertainty reviews, we don't believe it's necessary
to review all those various areas. We've actually gone back to our technical
staff and requested that we don't believe it's necessary, and we think there's
justification that can be provided to not review every section that we are asking
to review, for example, on a more complicated review. But that's formed the
basis, and I think we've captured the thought that we have all the sections. It
just hasn't been codified into a particular single section called "Power Uprate."
DR. WALLIS: Are we ready to wind up this discussion? Then
I'd like to thank our presenters for giving us a far more in-depth discussion of
this issue than we were able to have with the Subcommittee. It's been very
helpful. Thank you.
MR. SHUAIBI: Thank you for having us.
DR. APOSTOLAKIS: Thank you. Okay. We will recess until
1:30.
(Whereupon, the foregoing matter went off
the record at 12:23 p.m.)
Page Last Reviewed/Updated Monday, August 15, 2016