485th Meeting - September 5, 2001
Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title: Advisory Committee on Reactor Safeguards
Docket Number: (not applicable)
Location: Rockville, Maryland
Date: Wednesday, September 5, 2001
Work Order No.: NRC-004 Pages 1-132/196-303
NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers
1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005
(202) 234-4433. UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
+ + + + +
485th MEETING
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
(ACRS)
+ + + + +
WEDNESDAY,
SEPTEMBER 5, 2001
+ + + + +
ROCKVILLE, MARYLAND
+ + + + +
The Advisory Committee met at the Nuclear
Regulatory Commission, Two White Flint North, Room
T2B3, 11545 Rockville Pike, Rockville, Maryland, at
8:30 a.m., Dr. Mario V. Bonaca, Acting Chairman,
presiding.
PRESENT:
MARIO V. BONACA, Acting Chairman
F. PETER FORD
THOMAS S. KRESS
DANA A. POWERS
STEPHEN L. ROSEN
WILLIAM J. ShACK
PRESENT (Continued):
JOHN D. SIEBER
GRAhAM B. WALLIS
ACRS STAFF PRESENT:
JOHN T. LARKINS, Executive Director
SHER BAHADUR
PAUL A. BOEHNERT
SAM DURAISWAMY
CAROL A. HARRIS
HOWARD J. LARSON
AMARJIT SINGH
. C-O-N-T-E-N-T-S
PAGE
Opening Remarks, Dr. Bonaca . . . . . . . . . . . 4
Proposed Resolution of GSI-191, Dr. Rosen . . . . 7
Michael Marshall . . . . . . . . . . . . . . 9
Art Buslik . . . . . . . . . . . . . . . . .42
EPRI Report on Resolution of Generic Letter
96-06, Waterhammer Issues, Dr. Kress . . . . . . .69
Jim Tatum . . . . . . . . . . . . . . 83, 132
Vaughn Wagoner . . . . . . . . . . . . . . 101
Reactor Oversight Process, Mr. Sieber . . . . . 214
Michael Johnson . . . . . . . . . . . . . 215
Mark Satorius . . . . . . . . . . . . . . 221
Doug Coe . . . . . . . . . . . . . . . . . 257
. P-R-O-C-E-E-D-I-N-G-S
(8:30 a.m.)
CHAIRMAN BONACA: Good morning. The
meeting will now come to order.
This is the first day of the 485th meeting
of the Advisory Committee on Reactor Safeguards.
During today's meeting the committee will consider the
following:
Proposed resolution of genetic safety
issue, GSI-191, assessment of debris accumulation on
PWR sump pump performance;
EPRI report on resolution of generic
letter 96-06, waterhammer issues;
Reconciliation of ACRS comments and
recommendations;
Reactor oversight process;
Proposed ACRS reports.
A portion of this meeting may be closed to
discuss EPRI, information applicable to EPRI report
and resolution of waterhammer issues.
This meeting is being conducted in
accordance with the provisions of the Federal Advisory
Committee Act. Dr. John Larkins is the designated
federal official for the initial portion of the
meeting.
We have received no written comments or
requests for time to make oral statements from members
of the public regarding today's 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.
I will begin with some items of current
interest. First of all, a list of topics for the
meeting with the Commissioner Merrifield tomorrow
morning has been distributed to you and also has been
E-mailed to you. The expectation is that the
subcommittee chairmen responsible for the individual
items which are in the list will take the lead in the
discussion during the meeting with the Commissioner.
A second item, I'm sorry to announce the
death of an ex-ACRS member, Mr. Jeremiah Ray. He was
an ACRS member between 1978 and 1983. He served as
Vice Chairman in 1982, and as Chairman in 1983. He
retired in 1984 due to health reasons. He passed away
on August 2001.
We will, I think, prepare a card and
circulate it for signature from individual members and
then mail it to his wife.
With regard to the items we have in front
of us, the first presentation is going to be on the
proposed resolution of GSI-191. The staff does not
have yet the proposed resolution. So the intent here
is to listen to the presentations and then make a
decision on our part whether or not we want to write
a report at this time.
Okay. So we'll decided after the meeting.
Another item, you have in front of you
items of interest. In the first page you'll see there
is a list of five Commissioners' speeches, and also
under miscellaneous items, you'll see the last item is
the announcement of the 29th Nuclear Safety Research
Conference in October 22nd-24th, 2001, and the result
of the registration form are attached.
I also believe that there is an
introduction we want to make, and for that I turn to
John.
DR. LARKINS: Yes. I'd like to introduce
our latest member to the staff, Scott Sunn, and Scott
is a senior computer science major at the University
of Maryland. He's going to be co-oping with the ACRS
ACNW staff for the next four or five months.
Hopefully he'll have an opportunity to learn
something, but if anybody needs any help in the
computer or ADP area, Scott --
(Laughter.)
DR. LARKINS: -- is more than willing and
quite capable of helping out. So I'd like to
introduce him.
Thank you.
CHAIRMAN BONACA: Welcome aboard.
With that we'll move to the first item on
the agenda is the proposed resolution of the generic
safety issue, GSI-191. Steve Rosen is responsible for
that.
DR. ROSEN: Thank you, Mario.
It's an important issue that we heard a
briefing on in July, and I understand this briefing
will follow onto that perhaps with a slightly
different slant.
Please go ahead.
MR. MAYFIELD: Mr. Rosen, if I might, I'm
Mike Mayfield from staff.
I just wanted to touch on a couple of
points before we started. Since the July meeting,
staff has been fairly busy trying to finalize the
parametric evaluation that we briefed you on in July
and completing the risk and cost benefit analyses.
And Art Buslik is with us this morning to
describe those analyses.
The other thing that we did since the July
meeting was reached a management decision to
transition this GSI from the old process under a
particular office letter to the Management Directive
6.4 process.
The committee has been briefed previously
on that process, and we felt like this was a good time
since the staff is getting ready to make 6.4 the
accepted process for handling genetic safety issues.
We're at a point in the management of GSI-191 where
the old process and the new process most closely
align. So instead of the resolution step, this is now
the technical assessment step, but it's fundamentally
the same thing, although there are some substantive
differences.
One of the things Mike is going to
describe for you today is the difference between those
two processes and the benefits, such as they are, in
making the transition at this time.
This was a management decision that we
reached in August, and we apologize for not having
gotten this to you sooner, but it was something that
we felt like this was the appropriate time to make the
transition.
Now, under Management Directive 6.4, there
isn't an explicit request for a letter from the
committee at this juncture. However, that is an
issue, as we discussed this with Mr. Thadani
yesterday. this is an issue that he feels like needs
to be revisited in the management directive. He
doesn't think that it is in the best interests of the
staff, the committee, or the public to move forward
from the technical assessment step to the -- I've
forgotten what they're called.
MR. MARShALL: The regulatory guidance.
MR. MAYFIELD: The regulatory guidance
step without having some explicit feedback from the
ACRS on whether or not you believe the proposed
approach, as this moves from research to NRR. He
feels like it is appropriate to request a letter from
the ACRS at this juncture.
So that's a step in the management
directive we are going to be revisiting in the very
near future, but it is something that we would request
a letter from the committee if you're so inclined to
write one at this juncture.
With that, I'd like to turn the
presentation over to Mike Marshall and Art Buslik.
MR. MARShALL: Good morning. My name is
Michael Marshall. I'm the project manager for Generic
Safety Issue 191, and Art Buslik and I will be making
a presentation today.
I will be talking about the change from
the old process to the new process: how does it
affect Generic Safety Issue 191? I'll describe the
proposed recommendation we'll be sending to NRR for
resolution of Generic Safety Issue 191.
And Art will build on our technical basis
for that, for our recommendation, and at the July
meeting we talked about the work that LANL did for us
with the parametric evaluation.
Here in Research, we had Sid Feld do our
cost estimates for us. Art did our benefits estimates
and the core damage frequency contribution estimates,
and he'll be covering that at the latter of the
presentation today.
And this is just to reiterate. Almost
everybody is familiar with Generic Safety Issue 191
since we are looking to see if debris accumulation on
sump screen strainers causes problems for long-term
recirculation. From our last briefing we've
concluded, yes, there's a possibility. Well, yes,
that's a credible concern.
But because of the variations, large
numbers of variations from plant to plant, we can't
say specifically if a particular plant has a problem.
So our recommendation -- I'll give a little bit of it
away -- is that plant specific analyses are required
to make that determination.
But before going on to our recommendation,
talk about the change in the generic safety issue
process. Under the old process, and the status of new
process essentially is the management directive
administration essentially are checking to make sure
it's in the right format, and so it should become
final very soon.
And under the old process, the first three
stages of both processes line up very nicely, and
after the third stage they don't line up as nicely
again, and so we thought this was a fine time to move
Generic Safety Issue 191 from the old process to the
new process for a couple of reasons.
Because Management Directive 6.4 has been
receiving a lot of circulation within our office
reviews and such, a lot of managers and staff actually
might seem a little bit more familiar with the process
that we're about to implement than the older process,
and some of the discussions we're having between the
offices we found out we would end up losing a number
of time because we're talking one process and the
other parties, assuming this management directive is
what is going to be guiding the agency's generic issue
process.
And so we found out we were talking past
each other even though we agreed on technical details
and how things should follow after that, and so that
was one reason for switching processes, was just
clarity internally.
Another reason is Generic Safety Issue
191, at this point we are not going to close it with
no new actions or no new requirements with saying that
there's no additional actions. So it's going to go on
for another couple of years possibly, and under the
old process, at this point we would have resolved
Generic Safety Issue 191 and officially on the books
it would have been closed.
In reality, we would have still been
working sump block, again, for maybe a couple of
years, where under the new process -- and this is one
of the things we think we're taking advantage of -- is
that they'll be tracking all the way through the
verification so that it will be clear that the safety
concern, the concern 191, was addressed then from
outside the stakeholder's point of view. They can
look at it and track 191 to see how it was fully
implemented.
DR. WALLIS: You mentioned the word
"closed." Now, when is the issue closed? It used to
be closed around the resolution point in the old
process.
MR. MARShALL: Right. Under the old
process, it would have been closed under the
resolution, at the resolution process. Now an issue
is closed when we determine that no further action is
required.
For instance, we went through our analysis
and determined that there's nothing here. There's no
need for backfit. There's no safety benefit with this
issue, and we'll close it with that finding.
For issues that at the end of the
technical assessment stage, where we say, "Hey,
there's something here. There's something that needs
to be addressed," we won't close it at that point
because it was truly before never closed, and then
we'll keep working the issue.
And if you're interested in the Generic
Safety Issue 191, you won't have to grope around for
finding what's the new identifier.
DR. WALLIS: Would you then close it at
the verification stage if you had to take action?
MR. MARShALL: Well, any point along -- it
will be closed any point along here if it was
discovered. For instance, let's say NEI and the
Westinghouse Owners Group, they do additional work and
decide, hey, we've found out that this isn't as big of
a concern as you thought. We don't need to do any
additional action, and they provide that to us.
And we might close it saying, "Hey, the
industry says, has proven to us that this isn't a
legitimate concern," or we begin. It goes all the way
through where there's hardware modifications, and at
that point it would be at the verification where we go
back and check either through inspections or audits of
selected utilities that it was implemented as we
expected.
CHAIRMAN BONACA: But, you know, if I
compare those two tables, I could be drawn to conclude
that before you reached a resolution without
performing a technical assessment, of course, you need
to perform a technical assessment, right? I mean, a
technical assessment was part of the resolution
process.
MR. MARShALL: Yes.
CHAIRMAN BONACA: And all you did, you
expanded. I'm still confused about what is new about
the new process, I mean.
MR. MARShALL: Well, what's new if we go
to the next page, the key differences between the new
process and the old process is not giving the
perception that something has been closed when it's
actually still being worked.
CHAIRMAN BONACA: Okay.
MR. MARShALL: That's the biggest
difference, and I believe that was probably rooted
more as a public confidence type of concern.
CHAIRMAN BONACA: Okay.
MR. MARShALL: Another one is just, again,
for ease of tracking. The generic safety issue
designation will live on with the issue all the way
through verification, where in the current process at
the end of the resolution stage, the generic safety
issue designation is no longer used as it goes through
the remaining stages of imposition, implementation,
and verification.
In the past usually that was turned into
what's termed a multi-plan action.
CHAIRMAN BONACA: Okay.
MR. MARShALL: Now, the practical impact
on us from moving from the old process to the new
process, and it boils down to two things. At the end
of the technical assessment stage we won't have a
resolution that's the agency position on this is what'
going to be done.
What happens here is Research will send a
recommendation to NRR with our proposed recommendation
for resolution, and that will be the next slide. So
instead of the consensus that we're sending to the EDO
saying, "Hey, this is how Generic Safety Issue 191
will be resolved," or sending a recommendation over to
NRR, and so instead of -- and the couple I'd already
mentioned it -- there's no longer a memo to the EDO at
the end of the stage. It's an interoffice memo.
DR. WALLIS: What is the driving force for
finishing the job? These things in the past have hung
around.
MR. MARShALL: Right now the driving
force, I would say, for finishing the job is a couple.
There's a lot of oversight for generic safety issues.
Internally there's a lot of office level attention
given to our deadlines.
Working these, there's a lot of emphasis
on finishing them in a timely manner.
DR. WALLIS: So there's some incentive for
some manager to say it's being done or there's some --
what's the --
MR. MAYFIELD: If I might, this is Mike
Mayfield from the staff.
There is a congressional oversight group.
Senator Dominici receives a monthly report on the
status of each and every generic safety issue, and
this is something that at very senior levels in the
agency has taken quite seriously.
So there is significant impetus to
continue and not lose momentum on pursuing these
issues.
MR. MARShALL: And by going to the new
process, it keep sit in that. It keeps that
visibility on this generic safety issue.
Okay. I just want to cover the last
bullet on page 5. I think we've addressed the first
two already, and so at the end of this month, by the
end of September, we plan on sending our
recommendation via memo to the office director of NRR,
and at that point, in addition to closing the
technical assessment stage, we will also be
transferring the lead for Generic Safety Issue 191
from the Office of Research to the Office of Nuclear
Reactor Regulation.
And the proposed recommendation we plan on
sending to NRR is on page 6, and there's two parts to
our recommendation.
There's two parts to our recommendation.
The first part is to conduct the plant specific
analysis, determine whether debris accumulation will
impede or prevent ECCS operation during long-term
cooling, during recirculation.
And the second part is if you discover a
vulnerability during that assessment is to implement
appropriate corrective actions.
DR. KRESS: Now, since the staff was
unable to actually do this on its own, do you think
the licensees have the capability to make this
determination?
MR. MARShALL: Well, we think they have
the capability. Yes, we do think they have the
capability.
DR. KRESS: Do you think they can actually
track, determine the source of this debris and track
its transport and end up with how much and what the
characteristics of the debris is that reaches their
sump? Do you think they have that capability?
MR. MARShALL: Yes, I do.
DR. KRESS: Is there guidance that is
given to --
MR. MARShALL: Not specifically for PWRs.
We had issued guidance for BWRs, and there's quite a
bit of overlap in the guidance, considering it's
usually done at a performance base level.
And essentially the guidance boils down to
identify the debris, estimate how much transports, and
then estimate what the head loss would be.
DR. KRESS: Yes, of course.
MR. MARShALL: And that's more or less it.
Now, the specifics of what particular debris they have
in there is something we would leave up to the
licensees to determine or whoever is conducting that
analysis would determine.
DR. KRESS: That's probably a plant
specific issue anyway.
MR. MARShALL: Right. That's true.
DR. KRESS: When each licensee makes this
look to see if they're vulnerable, what happens then?
Do they come back to you with a report or do they fix
it and you review the fix or what is the next step?
MR. MARShALL: That hasn't been decided
yet. That's where NRR will enter in the next stage of
the process. They'll map out how it's implemented.
DR. KRESS: Okay. That's up to NRR to do.
MR. MARShALL: So that's still to be done.
DR. POWERS: Let me ask more about this
debris, and some aspects of it certainly could be
plant specific, I imagine. Different types of
insulation get torn off in the blow-down process, but
I would suspect that some of it is very generic in
nature.
Do we have guidance on what that generic
component of it is?
MR. MARShALL: Let me answer your question
slightly differently. I think we would look at the
debris from the way it's created, not at a specific
material. For instance, debris would be created by
direct impact from the jet. The possibility debris
would be created by the environment in the
containment, and that will include chemical reactions
possibly.
And that's where we would direct probably
our guidance if we started assembling guidance.
That's what we would probably recommend. Then we
could say specifically what jet impact would have to
look at different types of materials.
The main one we focus on a lot because
it's a large source is thermal insulation. Then,
again, we would recommend fire barriers possibly,
especially if there's any fibrous content with that.
And then we could point out what would be
the more problematic debris types. Again, that would
be a fibroblast, your calcium silicate. So be very
careful when you're doing your assessment of sources
that you identify these types of debris because they
tend to be the worst actors.
And coupled with that would be
particulates. Again, that would be generated possibly
from the environment of the containment. During
normal operation you might have some of that material
generated and also with the jet impact.
DR. POWERS: There's been within this
general field a lot of discussion of gelatinous
material. Do you give them any guidance on that?
MR. MARShALL: Yeah. Well, specifically,
we would recommend that people look at right now --
this is Michael Marshall if I'm sitting taking notes
back from the guidance.
DR. POWERS: Okay.
MR. MARShALL: Again, point out chemical
reactions, then give examples of where this has been
seen, and then again, leave it up for licensees in
case we miss something to look for similar type of
debris generation, or whoever is doing the analysis.
DR. POWERS: Gee, I wonder how you look
for that. I mean, can you go to the Journal of
Chemical Phenomena during reactor accidents and say --
(Laughter.)
MR. MARShALL: What we did was we did our
literature search, and we started looking for just
chemistry following a LOCA, and there was a number of
things we found, such as zinc precipitates, and we
started collecting that information.
So there's some things that wasn't done
specifically for debris clogging, and again, if you
just start out with a broad literature search, you
start finding work, and we found work that the Finnish
regulators had done in this area that was very
beneficial. We shared that with industry on the 26th
and 27th of July of some of the sources that you can
look at.
And again, some of it when we went through
it, we didn't use everything we discovered during our
literature searches and our reviews, and so that's
another area where we'll probably have to do a little
more documentation than we planned to so that people
will be fully aware what we learned during this
process.
Because as we mentioned in the July
meeting, we didn't use everything we learned to prove
our case that this is a concern that we need to worry
about. So we know we might have collected a few more
bits of information that we haven't shared, and that's
one of the major comments we get from industry is,
"Please tell us what you know. Please tell us what
you know. Please tell us what you know."
And so in order to facilitate that, we've
accelerated our documentation of the work we've done,
and we right pretty much have tried by the end of
November to distribute everything we've collected.
DR. POWERS: Rain dump.
MR. MARShALL: Yes.
CHAIRMAN BONACA: Now, one thing I
remember when this issue was raised in 1995, '96, or
whatever, a number of plants did a calculation which
were plant specific, and one of the findings was that
they really had marginal NPSH and was not an uncommon
condition to have the situation, which tells me if you
have any degree of blockage, you could have no NPSH at
all.
So isn't there some sense of urgency
behind this resolution of this issue?
MR. MARShALL: Well, I speak for the
Office of Research. There's a strong urgency from my
office director down with regards to this issue. Yes,
there is a sense of urgency.
DR. WALLIS: Now, thinking back to your
presentation last time and the report that your
consultants did, there seemed to be a lot of
assumptions made about how the debris got to the sump.
I mean, you can get a sense of understanding of how
jets affect -- steam jets and so on -- affect fibrous
insulation.
But then the transport mechanism, I think
there was a lot of almost hand waving, UI mean, sort
of assumptions and so on. So there's a lot of
potential here for some licensees to hire some smart
consultants who will do some other kind of an analysis
with fancy transport equations and solving and proving
that never gets to the sump because we don't really
have a very good basis for knowing how the material is
transported to the sump.
So there's going to be a lot of debate
perhaps, and I'm wondering how that gets resolved.
MR. MARShALL: Well, after the last
presentation, I was taken aside by my colleagues and
lectured that I didn't give enough credit for the
amount of work we did with transport. There are
certain areas of transport we're pretty sure once
we've published our results, especially once the
material gets in water. There's a very strong case
that it will make it to the sump spring if it's of a
particular size.
We've also done work in trying to estimate
what that size is, and we believe we're going to get
debris of that size, and then we rely a little bit on
our work we did with BWRs on estimating how debris
transports in a dry well to the wet well, and we use
that to estimate how much would actually get into the
water.
So there's enough work we've done out
there not just on this study, but when we're working
on BWRs which demonstrates that the plausibility of
debris getting into the pool of water on the
containment floor, then transporting to the sump
spring, and in this analysis we made it even easier on
ourselves by we essentially at the very beginning
excluded debris that could transport and just focused
on the smallest debris that would accumulate uniformly
on the sump screen.
So, again, some of the stuff that would
transport sliding on the floor we didn't include in
our analysis to make it simple, but even without that
debris, with the stuff that's more transportable
because it's very fine and accumulation formally on
the screen.
So in our analysis we didn't actually
include all the different debris.
DR. WALLIS: So you don't anticipate some
real technical issues coming up where the licensees
have a different analysis. You think your technical
basis is so sound that they will essentially do the
same thing.
MR. MARShALL: I'm not going to assume
they're going to do the same thing. Some licensees,
for instance, the plant that we got some of our cost
estimates from, they did things differently because
they had different licensing constraints that they
weren't willing to change, and so they made
assumptions that whatever was destroyed got there.
And as a regulator, I don't think we would
argue with that, and the same thing with the BWRs.
There's a whole different range of ways that
individual plants handle this. I doubt there will be
a lot of uniformity as this goes forward. There might
be three, four, maybe five different approaches, and
then there will be variance on those approaches, but
for a BWR experience, everybody kind of did it based
on a little bit of what they thought was right and
what was their licensing basis and how much did they
want to deviate or try to request changes from that.
It's the only fixed increase in the screen
area?
MR. MARShALL: No, there's a combination.
One reason we picked the increase in the screen area
as a fix that's one not only with regards to the BWRs,
but through other countries, that was the favorite
solution. Other solutions were minimizing your
debris, and there's a couple ways to do that.
When we're doing debris generation testing
with the Canadians, with Ontario Power Generation, one
thing they started considering was essentially put
another sheet of jacket over top of some of their
insulations, and that significantly in our testing
reduced the amount of debris generated.
Another approach is to switch from -- and
this was an approach used, I think, by the Finns a
good bit -- was they looked at the fiberglass and the
more problematic materials, and they decided, let's
switch to the RMI.
One thing from our parametric evaluation,
the cases that were predominantly RMI, they didn't
show up as -- they weren't ones labeled very likely.
They were mostly either unlikely or at the most
possible for a large LOCA. So changing your debris,
minimizing your debris is one solution.
Other things I would expect that seem
reasonable measures to take is to reevaluate your net
positive suction head margins. I would assume people
would do that, see if they have credit for containment
over pressure, if that's allowed or if they think
that's defensible.
Another one might be operational changes.
There's a couple of things. You've got your debris,
and then you have the flow rate, and so if you were to
use flow rate, you actually would decrease the head
loss across the screen, but some people might not want
to attach that because it attaches a strong philosophy
with regards to how to respond to an accident. You
probably don't want to start off by cutting off pumps.
CHAIRMAN BONACA: Are we looking at some
scenarios that might be more likely than others? For
example, the CRDM housing breaking and debris from the
location and could happen, just understand. You know,
obviously later on in the presentation there are
evaluations of initiating event frequencies and so on
and so forth, and they would be interesting to
understand. For example, debris generation from an
event of that type, there may be something more likely
than others.
MR. MARShALL: During the study we didn't
consider the CRDMs, and I think the bulletin that went
out, they were asking for the type of materials in
that area. So at least we would have a feel for what
type of materials we would consider.
CHAIRMAN BONACA: Yeah, that's what I was
looking for. I mean the kind of debris that you would
get from the kind of break.
MR. MARShALL: Just to go back to the
presentation for a moment, our technical basis boils
down to two things: the presentation we gave you last
July, which is the parametric evaluation, and the work
that Art will be presenting today on the risk and cost
benefit considerations.
Now, we've shared all of this work, except
for the cost estimate, with the industry on July 26th
and 27th. Actually over two days we were able to get
a lot more detail, and unfortunately -- not
unfortunately -- we actually covered more detail than
we actually had published in the report we released
earlier.
That was one of the comments that we got
back from NEI, the industry in general through NEI,
and they provided several other comments we plan to
address.
But if you're interested, I could cover
the first -- just recap the parametric evaluation or
we could jump straight into the benefit and cost
estimates. I would recommend doing that.
DR. WALLIS: Well, let me ask you. Is
there agreement from the industry with your
conclusions? You made this presentation. Did they
say, "Gee, whiz, you're right," or, "no, you're
wrong," or what?
MR. MARShALL: They haven't told us we're
wrong. I think that's a fair statement.
With regards to whether we're right or
not, they would like, again -- their major comment
would be, "We know you did more than you shared with
us in writing so far. Please give us the rest of it
so we could make a better determination if we agree
with you or disagree with you."
So their position -- well, I'm going to
speak for them -- their position right now is we
probably don't have enough information to say if we
agree with you or disagree with you. We don't see
anything on the surface that seems obviously wrong,
but again, we don't have all of the information.
I think that Kurt Cozens is coming up to
answer us.
MR. COZENS: This is Kurt Cozens, from
NEI.
In all fairness to Mike, was it just
Friday that we sent you the letter with the comments?
MR. MARShALL: Yes, right.
MR. COZENS: So he's just received those
probably about the time he was wrapping up his
presentation material here, and we would be happy to
provide a copy of this letter to the staff.
Mike has properly characterized our
overall findings that we do not have enough of the
specific data to agree or disagree with the findings
that the staff has done. They have provided us a lot
more information in the meetings that we had at the
end of July that were not in the draft report that
they had put together, and you know, we are continuing
to look at that, and we'll do that once that's
publicly available.
And we would be happy to provide ACRS a
copy of that letter today.
Mike, do you have a copy that they could
have?
MR. MARShALL: I have a copy with me if
you'd like to.
MR. COZENS: Okay. So that will help you
guys, and you can see the full range.
DR. WALLIS: So that means that you folks
didn't have an assessment of your own to compare with
the NRC assessment?
MR. COZENS: We do not have the technical
details that the staff has, and we were asked to
comment on the --
DR. WALLIS: You must have some technical
evaluation from your engineers as to whether or not
this is a problem.
MR. COZENS: We are still in the process
of seeing the data. We have not seen the data yet.
So it would be inappropriate for us --
DR. WALLIS: You haven't seen anybody's
data but your own. You must have some sort of a
position as to whether or not you think it's a
problem, or has it just been something that no one has
worried about at all?
MR. COZENS: We are continuing to look at
it, and we've had questions about it, but we have not
finalized it to make a formal industry position.
DR. WALLIS: Well, that's a little
disconcerting if this is a real technical problem and
industry has no position.
MR. COZENS: There is an industry group
working on this, but until we have the technical data,
we are not able to finalize our conclusions.
MR. BUSLIK: Okay. I'll start on the --
DR. KRESS: Well, before you start,
Michael if we wanted any more information on the
parametric study since we had previously reviewed it.
I'd like to have you refresh my memory on just what
parameters were varied and why -- not the actual
ranges of those, but why -- what was the basis of
choosing the ranges of the parametric variations?
MR. MARShALL: Well, I'll go ahead and
leave that up.
In the parametric evaluation, we varied a
number of things, and usually the basis for the range
we chose was the industry survey we collected. NEI
helped us with collecting information on, let's see,
sumps, sump screen area size, height of debris curves
in containment, times that licensed plants would
expect to switch from RWST to the sump.
Sump water height was another factor we
considered, and again, that was all based on responses
to the survey.
DR. KRESS: Did you vary the -- does your
parametric variation include the source of debris?
MR. MARShALL: The only variation we had
with the source of debris was usually the amount of
debris, and we varied the combinations of debris
depending on how we -- on the varieties we saw at
different plants, but we didn't vary the debris types
beyond fiberglass, reflective metallic insulation, and
calcium silica.
And then we had a reasonable amount of
particulate debris, but the amount of those varied
from different cases, and so we had cases that were
mostly RMI, which again would show up as -- in most of
the cases showed up as not being a -- showed up as
being unlikely.
Then we had cases where the plants were
cases where 100 percent fiberblast, and again,
depending on the net positive suction and margin, size
of the sump screen area, that ranged from probably
possible to very likely.
DR. KRESS: So you took plant specific
information.
MR. MARShALL: We took plant specific
information. We coupled that --
DR. KRESS: And then coupled that with --
MR. MARShALL: We coupled that with
information we collected from two volunteer plants.
So from the volunteer plants we got the piping
configurations, and so we assumed for all 69 cases
they had one of these two piping configuration.
Now, both of those configurations were
four-loop Westinghouse units. Again, so when you look
at a two loop, as far as the capacity of the screen to
accumulate debris, we did a really, really good job
there, and that's something I would recommend industry
take because it doesn't require you to know how much
is just transported and how much is generated.
You can sit down and do a calculation of
if you have this type of material in your containment
and you assume how much of it do you need to get on
your sump screen to exceed your net positive suction
margin. That's one thing I liked about the approach
we used, is regardless of transport amount of
generation, you can always go back and look at what we
call the threshold value.
And is that threshold value 100 cubic feet
or is it just two cubic feet? And I would say those
of us that worked on the evaluation are very confident
with that point of the evaluation.
And then, of course, there's the box. I'm
assuming people remember the presentation from last
time when I referred to the box.
Then there's that box where we had the
unfavorable and favorable assumptions, and that sort
of gave a feel for how much we actually thought would
get transported to the sump screen, how much would be
generated, and then we compared that to the minimum
threshold.
DR. KRESS: Thank you. That helps.
DR. ROSEN: Mike, do I understand that in
this transition to NRR that's coming up, that NRR will
make a determination at that point or after they get
it and study the issue for some time as to whether or
not they're going to issue a bulletin? Did you say
something about an NRR bulletin that I didn't
understand?
MR. MARShALL: No, I was referring to the
bulletins on CRDMs that went out.
DR. ROSEN: Okay. So there is no bulletin
planned on this yet.
MR. MARShALL: No. Right now what -- and
I'll speak for NRR, and please correct me if I'm wrong
-- right now we're going to send over our technical
basis in this information, and NRR wants time to
consider again input from other industry groups with
regards to our work, and then they'll decide on what's
the appropriate regulatory path to take.
Is it a generic communication? If it's a
generic community, is it a bulletin generic letter?
Is the industry going to step up and propose something
which would, again, that the agency might not have to
issue a formal -- take formal regulatory action?
DR. ROSEN: Okay. I understand that.
That will be decisions made by NRR.
MR. MARShALL: Right.
DR. ROSEN: Now, let me just ask again
about the approach of not issuing detailed guidance.
I know this is a little early, but that was probed a
moment ago by some of the members, and your response
was, no, we would not issue detailed guidance on how
to do the analysis, the plant specific analyses.
MR. MARShALL: What we would avoid doing
is issuing prescriptive guidance. It would probably
be performance, and as a debate of whether how quickly
we can get guidance out there.
MR. MAYFIELD: This is Mike Mayfield.
The issue of guidance, do you issue a reg.
guide or is there some other vehicle? A reg. guide,
regulatory guidance, that specific kind of document
takes about two years to get out the door in a final
form, and there was some, I think, a question earlier
about some sense of urgency on this.
We think it's not in anybody's best
interest for the staff to take two more years to
promulgate a regulatory guide. So if we set aside a
regulatory guide is something that we're probably not
going to pursue at this stage.
What kind of guidance would the staff do
presumably if we were going to issue some sort of
generic communication? That would provide some
information , the collection of reports and analyses
that Mike and his colleagues have worked on would be
available and could be -- we could point to that as
one method that could be followed.
So it's not to just go out to the industry
with a suggestion they might go do something. We have
some -- you know, a fairly specific set of analyses
and approaches that will be published and in the
public domain and that could be used, and I think in
that body of reports, there's a lot of information and
a lot of guidance on what -- at least how we did the
analysis.
So we're not asking people to just embark
on something in a blind fashion. As the same time, we
don't see publishing a regulatory guide, at least not
in a time frame that would support the industry going
off and doing something on this issue.
DR. ROSEN: Well, there clearly is a need
for prompt action on this. I think everybody thinks
that there is some urgency here.
There is also a need for putting out
enough guidance so that you don't get apples and
oranges responses that are not into comparable.
MR. MAYFIELD: Yes, we agree. And, again,
if the staff chose to go down a path of some sort of
generic communication, a combination of information
that would be included in that document as well as
references to the reports that Mike and his colleagues
are getting ready to put out would provide the level
of guidance to provide the kind of consistency you're
talking about.
DR. WALLIS: Sorry to go back to this, but
I've just read this NEI letter which we see here which
was sent on August 31st, and all of the comments are
critical. It seems to me that we've been talking
here as if your conclusions are acceptable, but it's
not at all clear that that is the industry position.
I think you may have quite a fight on your
hands, in which case it's not clear that things are
going to be quite as smooth as has just been
discussed. You just sort of go ahead, and now I was
going to accept your conclusions, and you know, some
regulatory action will be taken. You may have quite
a debate going on in the next year or so.
That's my sense of the NEI letter.
MR. MAYFIELD: This is Mike Mayfield.
Based on some other dialogue we have had
with members of the industry and some of the staff at
NEI, we think that while there are many questions and,
indeed, the comments you see in the letter tend to the
critical or questioning side of the spectrum, we
weren't surprised by those. In fact, that's pretty
much what we would have anticipated. I think that's
what we were looking for is where they saw soft spots
or areas that they thought should be expanded.
This is an issue that will require, I
suspect, some extensive dialogue and a fair bit of
interaction. It is -- the piece of work we did is not
all that conclusive. It was a parametric evaluation.
It was a scoping evaluation to decide if there's
something there that should be pursued. We think that
the piece of work makes that case.
We will have some discussions with the NRR
staff and management as we go forward. If we were all
in complete lock step on this, then I'm not quite sure
what presentation we'd be making to the committee or
how it would differ, but the fact is there's a
process, and we've embarked on it.
To suggest to you that, like I say,
everyone is in lock step would be incorrect. At the
same time, we think there is a good case that's been
made to pursue the activity.
MR. COZENS: This is Kurt Cozens from NEI.
With regards to the letter that we
provided staff, the letter was provided in response to
a specific request that we provide them comments on a
draft research report that had been written. The
draft research report had been accelerated, and it
appeared that many of the assumptions that were taken
in it and the analyses that were performed to provide
the more conclusions and the underpinnings of that
were not provided in that particular report.
The letter that we submitted identified
specific areas where we wanted to see more detail as
to how those were arrived and the logic behind those
selections. We had the process of very thorough
evaluation and have not been able to go over those in
detail as of yet.
However, I will note that on the was it
July 26-7th meeting we had with the staff? Many of
those details were, indeed, discussed at that meeting,
but they are not in the report at this point in time,
nor are they in a format that we can actually review
them.
So, you know, I would like to compliment
staff on its efforts to coordinate its activities with
industry. We've gotten a lot of benefit out of that.
We have provided the staff with a great deal of
information to make this study possible, everything
from the basic survey of where industry is through the
effort of identifying volunteer plants to give very
explicit detail which made the study even possible.
So we have been an active participant in
this. You know, we are still evaluating the data,
however.
MR. BUSLIK: Concerning the risk and cost-
benefit analysis, the work that I did had to do with
calculating the decrease in the core damage frequency,
and doing the benefit analysis as per the reg.
analysis guidelines. Sid Feld did the costs
associated with fixing the problem, and there was an
uncertainty analysis.
An outline of the approach, I'm going to
calculate the difference in the core damage frequency
given before the fix and after the fix, and basically
you would have to look at the event sequences on an
event tree where it matters whether the sump clogs or
not.
And these basically are given as follows.
You have a LOCA. You're not able to cool down and
depressurize and use your RHR system as you would in
a normal shutdown. The sump clogs to the point where
you fail emergency coolant recirculation, and
emergency contingency action type recovery actions
fail. These are, for example, in the emergency
response guidelines of Westinghouse, ECA-1.1.
There are various size LOCAs. There are
also very small LOCAs and stuck open pressurizer
safety valves which are not considered here because,
as I'll indicate later, they don't contribute.
The initiating event frequencies I used
came from NUREG CR-5750, and the large LOCA frequency
comes from assuming that from taking the number of
leaks in large piping that have occurred and
estimating the probability of going to a rupture from
a leak.
The means and the five percent/95 percent
bounds are given there. For the reactor coolant pump
seal LOCA basically there's an error factor of three
so that the lower bound is 5.60 minus four and the
upper bound is 5.4 E minus three, according to the
table in NUREG CR-6750.
As far as the control rod drive mechanism,
whether it would be important or not would depend on
the kind of plant and how big a LOCA would be.
Also, the type of insulation may tend to
be more reflective metal, metallic insulation in most
plants. That would be the most benign, but you would
have to look at each plant.
I did look at the seismic contribution to
the initiating event frequencies for Surry using
fragilities from the old NUREG 1150 study and also
using the revised Lawrence Livermore hazard curves.
They were smaller than the initiating event frequency
listings, although there was some contribution for
large LOCA.
However, since we have arrived at the
conclusion that it's cost beneficial without seismic,
it won't make any difference if we include it.
For recirculation and nonrecovery,
basically you're going to have to go to sump
recirculation for large and medium LOCAs as I indicate
later. So these are only important for small break
LOCAs and reactor coolant pump seal LOCAs.
And it depends -- how successful you'll be
will depend on the kind of plant you have. If you
have a large, dry containment, emergency fan coolers,
and large refueling water storage tanks, then the
chances of being able to cool down and depressurize
before you've exhausted your -- you've gotten to the
point on the refueling water storage tank level where
you're forced to switch is fairly good.
For a subatmospheric plant, the RHR at
least at Surry, it's inside containment, and it's not
environmentally qualified. So there would be
questions as to whether you could actually go on
residual heat removal there.
And plants with ice condensers, the
containment spray goes on at a very low pressure, and
you would exhaust the refueling water storage tanks.
So again, there's no chance.
Some of this material in the next slide
I've already covered. For medium and large LOCAs you
have to go to sump recirculation. For very small
break LOCAs, the chances of needing to go to
recirculation was negligible.
I mean, it was pointed out to me that, for
example, if all your charging pumps failed, then you
probably would be forced to, but that's a low
probability event, and I just didn't consider it.
CHAIRMAN BONACA: And I want to let you
go. You know, you're presenting us with the cost-
benefit analysis, and I'll be very interested in
seeing this, but I'm trying to understand the whole
logic now.
The FSARs or these power plants state that
you have high pressure injection and low pressure
injection. You run through half of your RWST. Then
you switch to recirculation and you depend on that
recirculation for preventing core damage.
Now, it is a commitment of the FSAR. Now
we have doubt that the analysis provided in the FSAR
is adequate, I mean, and there is reasonable -- there
are reasons to doubt because the analysis does not
address sufficiently debris or because we find that in
some cases MPSH was very marginal, and so on and so
forth.
So there is a reasonable position that the
NRC is raising here that is basis from the analysis
done at some plants that there is a concern. I'm
trying to understand why would you need a cost-
benefit.
MR. BUSLIK: The reason is, and I can't
quote the exact document, but even for issues of
compliance, which is what you're talking about --
CHAIRMAN BONACA: Yes.
MR. BUSLIK: -- compliance with
regulations, we're supposed to do a cost-benefit
analysis.
CHAIRMAN BONACA: Okay.
MR. BUSLIK: This has been for a couple of
years now. I think there was some SECY paper where it
was mentioned, and there was an agreement with
industry, the idea being that if the issue really
doesn't have any safety significance, that you may
want to avoid -- you may want to basically have a
waiver of some sort.
CHAIRMAN BONACA: Okay. Thank you.
DR. WALLIS: It's a way of risk informing
the regulations.
MR. BUSLIK: Yes.
DR. WALLIS: Without definitely changing
them, you know; modifying them.
MR. BUSLIK: That's right.
CHAIRMAN BONACA: But in any case you
would perform a cost-benefit.
MR. BUSLIK: Yes. I don't think it has to
be as elaborate as a cost-benefits analysis for a
backfit.
Now, stuck open pressurizer safety valves
are a special case because the discharge from a safety
valve would be routed to the quench tank, and if it
got into containment, it would be through a rupture
valve there, and I am told that because of the
location of a quench tank and other things, there's
very little likelihood that that would cause a
clogging of the sump. So that was neglected.
Now, as far as the probability of some
clogging is concerned, the LANL draft report, which
you've has a presentation on, assigned -- I believe
you did -- assigned qualitative, very likely, likely,
possible, and unlikely designations for whether the
sump would clog on various size LOCAs, separate for
different size LOCAs.
After consulting with Mike Marshall and
D.V. Rao at Los Alamos, I assigned these
probabilities. More recent probabilities are possible
as .4 instead of .3. It will not make any difference,
and the direction that it would go, it's small, but
the direction that it would go would be to make it
even more cost beneficial.
I considered three aggregates of the
plant. The idea here is for any individual plant
there may be uncertainties because of lack of plant
specific information, but you consider the fact that
if you consider an aggregate of plants, these
uncertainties will somewhat cancel.
So we consider a case which at that time
had 23 plants, and according to more recent
information has 25. There are some clogs on all size
LOCAs, and there are 18 large drives and five
subatmospherics there.
The 32 plant case --
DR. WALLIS: Excuse me. That means that
they clog with any kind of a LOCA?
MR. BUSLIK: Even the reactor coolant pump
seal LOCA, yes.
DR. WALLIS: The reactor pump seal
actually --
MR. BUSLIK: I mean, they're relatively
large.
DR. WALLIS: -- actually produces jets
which remove enough material?
MR. BUSLIK: That was the question which
Westinghouse asked, and I don't really know.
PARTICIPANT: We're still collecting
marketing.
MR. BUSLIK: Yeah. I mean, it will come
out the top of the shaft, I guess, and so the 32 plant
case, there are some clogs with fuzzy certainty for
large LOCA and medium LOCA, and it can or cannot clog
with various probabilities for small break, and in the
40 plant case, it had a probability of one for large
LOCAs and either one or .6 for medium LOCAs.
Now, the change in the core damage
frequency, the mean change in the core damage
frequency associated with the 23 plant item is all
about one E minus four, and that indicates that
there's a substantial safety benefit, but we still go
on with the cost-benefit analysis.
DR. WALLIS: But it seems to me that
industry could easily come back with numbers which
instead of probability one, one, and one were
probability .2, .2, .2, and it would turn out that
nothing matters at all.
MR. BUSLIK: Yeah, I know, but if you look
at -- I mean, you need really D.V. Rao or somebody to
answer that, but if you look at some of the curves,
you have a little box which has the range of
particulates, and you have a place where if you're on
the right side there's failure and on the left side
there's not.
In some cases there's such an extreme
difference that --
DR. WALLIS: There's one or nothing?
MR. BUSLIK: Yeah, in that case for that
plant it would be the one.
DR. WALLIS: Okay.
CHAIRMAN BONACA: I mean, certainly an
argument could be that, you know, a large LOCA, it's
clear it can break and it is unlikely and so on and so
forth. So you would want to have some realistic
estimation of debris accumulation for break sizes that
are not going to be in contention. It would be
interesting to have some.
So it would probably have some sensitivity
as a function of break size.
MR. BUSLIK: Well, the --
CHAIRMAN BONACA: You have a meeting.
MR. BUSLIK: These probabilities are by
break size. That came from the report.
CHAIRMAN BONACA: Yeah, I understand.
MR. BUSLIK: So I didn't do any
sensitivity on the probability of some clogging,
except for you'll see later that it's easy to see that
it's cost beneficial even if for the ones where it
wasn't one, it was zero instead of .6 and .3.
CHAIRMAN BONACA: Okay.
MR. MAYFIELD: Art, excuse me, if I could.
Just to pursue that point, the break frequencies that
Art used came out of the NUREG 5750. One of the
points that we've talked about, without trying to
insult my colleagues that did that piece of work, I
don't think there's any question they did their sums
properly.
The problem with those frequencies is they
can only capture experience up to the point in time
when they did the analysis. It can't capture new
degradation phenomena. It doesn't capture new aging
phenomena that we haven't seen yet, and there's no way
it could.
So the frequencies that Art has used, they
reflect service data up to a point what, four or five
years ago?
He noted on the one slide that they made
an attempt to include the recent V.C. Summer
experience and just a one crack in a largish pipe made
a significant difference in that break frequency, but
there's a lot of additional analysis that goes into
that. So we wouldn't want to put forward these break
frequencies as the definitive statement the staff is
making on break frequency, but it's something to work
with for this kind of analysis, and it reflects
service experience, perhaps except the most recent
events.
MR. BUSLIK: And, of course, if we used
higher numbers like I've been using in the past in
PRAs, it would be even more cost beneficial.
DR. WALLIS: So you are not using those
PRA numbers?
MR. BUSLIK: For the initiating event
frequencies. Instead I was using these, the
initiating event frequencies from NUREG CR-5750, which
are smaller basically.
DR. WALLIS: One would expect PRAs which
are evolving to be more reliable.
MR. BUSLIK: But the initiating event
frequencies, my guess is that they're originally from
-- for LOCAs, originally came from expert judgment.
It hasn't been changed that much.
Okay. So to go into the monetized
benefits, the kinds of things you have to consider
according to our regulatory guidance are expected
averted population dose to 15 miles, monetized at
$2,000 per person-rem, expected averted off-site
financial cost, expected averted on-site cost, and
expected averted on-site occupational dose.
The largest contributor is the on-site
cost, clean-up and decontamination and replacement
power. It's about 80 percent of the benefits.
The expected averted population dose to 15
miles is about 17 percent. If you look at -- if I --
it would not be cost beneficial if this were a
backfit, now, if we only consider the expected averted
population dose, but it's not a -- I mean, that's not
what our guidance is.
And of course, in a sense, the expected
averted on-site costs should be subtracted from the
cost that the utility has to make anyway, even if you
have to consider it.
DR. WALLIS: This simply gives you dollars
per CDF, doesn't it?
MR. BUSLIK: This --
DR. WALLIS: Average plant.
Do you have to do this calculation every
time? Don't you have a sort of rule of thumb of
dollars per CDF?
MR. BUSLIK: What I did was dollars per
person-rem.
DR. WALLIS: Yeah, but eventually you're
going to relate it to CDF.
MR. BUSLIK: Oh, yes, yes. The CDF is
included there.
DR. WALLIS: So it is dollars per CDF.
MR. BUSLIK: That's right.
DR. WALLIS: What is the dollars per CDF
number, just so that I can sort of --
MR. BUSLIK: Well --
DR. WALLIS: Do you have it? If you don't
have it, it doesn't matter, but it seems that's what
eventually --
MR. BUSLIK: Yeah, I have it.
DR. WALLIS: -- it comes down to, doesn't
it?
MR. BUSLIK: Well, first of all, it would
depend, in general, whether it's a core damage
frequency, which has a large contribution, a large
early release fraction or not, but early containment
failure basically.
But for this study 23 plants gave a
benefit -- I mean, I don't have the numbers right in
front of me. I think maybe I do, as a matter of fact,
but --
DR. WALLIS: It's just very useful for the
future when we're making these assessments if we have
a rule of thumb that we can think about.
MR. BUSLIK: Okay.
DR. WALLIS: Maybe at the end of the talk
or something.
MR. BUSLIK: Yeah. I mean, I have a slide
that I could compute it from, but --
MR. MAYFIELD: Why don't we take that as
something that we can get back to you on, Professor
Wallis, if that's acceptable?
DR. WALLIS: All right, and there's no
need to do it now.
MR. BUSLIK: Yeah, okay. Because the
numbers I have depend on the number of years of
operation of the plant and things like that.
We can skip this slide, I think.
The cost analysis, the data, of course,
that's used are given on this slide, and the cost
elements consisted of three parts: up front
analytical activities; the physical modification; and
other cost elements.
The up front analytical activities, each
plant would have to do them. So it's independent of
the number of plants that have to make the fix.
Physical modifications are proportional to
the number of plants that have to make the fix, and it
was assumed that audits and inspections were also
independent of the number of plants that had to make
the fix.
So that --
DR. WALLIS: How big are the up front
activities as a fraction of the cost?
MR. BUSLIK: Okay. You'll see it on the
next --
DR. WALLIS: It will come?
MR. BUSLIK: -- the next slide.
DR. WALLIS: I was wondering if the
analysis doesn't cost more than the --
MR. BUSLIK: Well, it depends. If no
plant had to make fixes, then obviously it would, but
it's a linear function, and this is taken down to 2001
dollars. The assumption is made that the analysis is
done in two years from now and the fix is made in
three years from now, and it's discounted to the
present at a seven percent discount rate, which is the
value we're supposed to use.
And so you have six times ten to the fifth
dollars, in other words, $612,000, for making the fix
at each plant, and an up front cost of $9 million.
DR. ROSEN: That's aggregate for the whole
industry or is it per plant?
MR. BUSLIK: The aggregate for the whole
-- the nine million is an aggregate for the whole
industry, but you get an idea here. When this was
done, it was assumed that 50 percent of the plants
would go to license renewal, and there were some rough
assumptions. Really the way you should do it is you
should look at every plant, know how many more years
left, and make some decisions as to whether it is
going to go to license renewal or not, and do that.
But we did it in a rough way, which is
probably okay, but I'm told that industry may plan to
have much more than 50 percent of plants go to license
renewal. That would make it even more cost beneficial
because there would be more years with the fix in
place.
DR. ROSEN: So the hardware fixes are
about 600,000 per unit.
MR. BUSLIK: Per unit, that's right.
DR. ROSEN: And the aggregate analysis
costs for the industry are about $9 million.
MR. BUSLIK: That's right.
DR. ROSEN: And what are you expecting
that $600,000 to buy in the plants? Is there a
specific fix that that is supposed to be the cost
estimate of?
MR. MARShALL: What we assume is that the
fix would be is increasing your sump screen area, and
the costs were based on estimates of one utility that
already did that. Then estimates we got from other
vendors on how much they would charge the utility for
doing that type of work.
MR. BUSLIK: Yeah. What was the plant
that was -- Diablo Canyon?
MR. MARShALL: Yes.
MR. BUSLIK: Yeah, Diablo Canyon had
actually done such a fix.
What you get is for the 23 plants where
there was a probability of one of the LOCA on every --
for every size LOCA, the benefits were about $50
million, and the costs, 23 million. You can see that
if I considered only those 23 plants in a sense, that
has enough benefit to take care of the 32 plant case
and, in fact, the 40 plant case using mean values.
So basically even if every case where it
is possible or likely for the sump to clog, you set it
equal to zero, you would still be cost beneficial for
all of the three cases.
DR. WALLIS: But if I'm NEI, I'm going to
come back and say you've made conservative
assumptions. The benefit is really, you know, half of
that and the cost really twice that. So it's not
worth doing.
MR. BUSLIK: Well, right. And it all
hinges on the probability of the sump clogging and
whether they can argue --
DR. WALLIS: Except I wonder if it's
really -- any prediction is within a factor of two.
So it's going to be --
MR. BUSLIK: Aside from that probability
of the sump clogging, and I think probably for some
plants the probability of the sump clogging being one
is fairly robust just because of where the little box
is compared to the failure line, and that's my own
opinion, but --
DR. WALLIS: Probably nothing in nuclear
is ever one, is it?
MR. BUSLIK: No, it isn't one, but if it's
.99 it doesn't matter.
DR. WALLIS: Well, it seems to me a bit
surprising that these things have operated all this
time and engineers have looked at things and now
you're coming up with something with a probability of
one which hasn't been considered before.
MR. BUSLIK: Well --
DR. POWERS: It must have been considered
or it wouldn't have been screened.
DR. WALLIS: Well, if it's been considered
before, then we must consider the probability to be
very small. Otherwise they would have done something
about it.
DR. POWERS: Well, I think the discovery
was that that at Barseback they could produce a lot of
debris from the process itself.
MR. BUSLIK: That's right.
DR. POWERS: I mean, I think it's the
magnitude of the debris.
DR. WALLIS: So it's a new piece of
knowledge which changed this assessment from
negligible to one.
MR. MARShALL: Yes. When it was
considered before, there's a few changes.
Barseback -- well, yeah, there's a few things we knew
from Barseback that we didn't know back in 19 --
actually the agency addressed this explicitly back in
1980, 1985, that time frame.
And what Barseback showed us was that our
amount of transport, the type of debris we were
considering, not the type, but the shape and size of
it was in error.
And so when we went back from what we knew
with Barseback and applied it and a few more things we
learned along the way, such as filtering of
particulate debris, we end up with drastically
different --
DR. WALLIS: Was Barseback some event that
actually happened?
MR. MARShALL: Yes.
DR. WALLIS: When did it happen?
MR. MARShALL: A Swedish BWR in 1992.
DR. WALLIS: '92?
MR. MARShALL: yes.
DR. WALLIS: So it's going to take ten
years before anything is done?
DR. POWERS: It'll take more than that.
DR. WALLIS: Well, there are going to be
no hardware modifications.
DR. POWERS: I understand you're talking
about --
MR. MARShALL: Well, actually the NRC did
this in two steps. We addressed our BWRs first, and
all those had made modifications. The agency has
audited those modifications, have closed out,
essentially went through -- if the BWRs was handled as
a GSI, that would have been concluded probably
beginning of this year.
So we took it in two steps. We took the
BWRs first, and then we went back and looked at the
PWRs, and so we've been active sine Barseback, and
we've addressed our BWR population, and we're in the
process now of addressing our RPEs.
DR. WALLIS: Thank you.
MR. BUSLIK: I guess it was less clear
that there was a problem with BWRs, and yes, there are
some screens, but the assumption was that they would
get to be clogged only 50 percent, and in some cases
it's much more than that.
DR. POWERS: Also in fairness, Graham, the
first four years that I was on this committee, I got
to listen to just about every meeting a request from
Mr. Carroll on when was the staff going to do
something about the Barseback incident.
DR. WALLIS: So you're seeing in back in
person again.
MR. BUSLIK: Okay. These are the
uncertainties of the large and medium LOCA frequencies
here. They were on an earlier slide as well, except
for the median values, which are given there.
The values for the reactor cool pump from
sealed LOCA are not there. They were given. The
upper bound is 5.3 minus three. The lower bound is
5.6 E minus four, and I think the error factor is
three. So that the difference between the mean and
the median for reactor cool pump sealed LOCA would be
about 25 percent.
Okay. In some cases the probability of
the sump clogging may be conservative. I mean, they
use the licensing criteria for loss of net positive
suction, but in some cases it probably wouldn't make
any difference at all, but I don't know on the average
how it would affect it.
DR. WALLIS: Excuse me. There's one
screen or something? Are the screens in different
places? There are several intakes for pumps, aren't
there? There isn't just one.
MR. MARShALL: It depends on the plant.
There's one plant that has three distinctly separate
sump screen -- sumps with three separate sump screens.
More typical would be two sumps per plant, and then
that will vary between two distinctly separate sump
screens or two sumps that share a sump screen area.
I don't believe there's any -- no, there's
no single sump plant. So most of them have two, and
it's whether they have two physically separate sump
screens or --
DR. WALLIS: Doesn't it help you -- isn't
there a preference for debris from a particular
accident to be in a particular place, or is it
everywhere?
MR. MARShALL: Again, this would be one of
the plant specific things. Depending on the break, it
could be preferential in one location versus another,
and also depending on --
DR. WALLIS: Does that come into your
analysis, the LANL analysis, or do they just assume it
goes everywhere?
MR. MARShALL: We pretty much assume it
goes everywhere.
MR. RAO: My name is D.V. Rao. I work at
Los Alamos. I'm the principal investigator.
Actually sump screens changed quite --
sump screen designs are unique to each plant, I guess.
They vary quite much.
In our analysis we did take into
consideration sump screen location as relates to how
close it would be to the pipe locations where the
insulation is. In some plants it's in the remote as
packed away in some parts, and in some it could be
feet away from, literally under a recirculation line.
So we tried to take that into
consideration.
Also, another aspect that we took into
consideration is whether the sump screen is above the
floor or below the floor. In some plants, the sump
screen just looks like a storm drain of such where
it's in a pit in which the sump screen is. So the
debris actually tends to go into the pit and,
therefore, deposit, and in some plants, on the other
hand, it is a vertical screen located on the floor.
Therefore, the debris had to go up and build.
We tried to take some of these factors
into consideration and be very -- and we still have
some other experimentation and others going on on
those issues, but I do believe we tried to address
that.
DR. WALLIS: Thank you.
MR. MARShALL: I didn't go into that kind
of detail, but there's essentially no two sumps alike
between different sites.
DR. WALLIS: Which indicates that every
plant is going to have to do its own analysis and
someone is going to have to review that for technical
credibility.
MR. MARShALL: And that led us to our
recommendation of plant specific.
DR. ROSEN: I think that's absolutely
true, Graham, and my comments earlier were about every
plant has to do its own analysis, and every plant is
different. Then the need for guidance, it seems to
me, is absolutely clear in the sense that you will get
analyses that you won't -- that will look -- the
answers will be very different, and the configurations
may be the same. And then what do you do with that?
MR. RAO: Actually, may I say one other
point? It is true that every plant may have to do
separate analysis, but depending on the fix, you know,
a lot of our discussions that I've been seeing here
are going on what the status is right now.
It is, in fact, true that the sumps are
designed differently, but that doesn't necessarily
mean that the new sumps that are to be replacing the
present ones, as in the screens and others, could not
be generic or could not be more -- they share features
common to different plants, in which case it is not
necessary that you have to do analysis to that level
for each plant.
We need to think about that, that is, that
at the present time they're different from each site
or each plant, doesn't necessarily mean in the future
analysis that they have to do will have to be the same
either.
I don't know if I made my point clear.
CHAIRMAN BONACA: Just a note. We have
less than ten minutes left. So we should --
MR. MARShALL: If you don't mind, I would
like to skip to just --
CHAIRMAN BONACA: Okay.
MR. MARShALL: Just finish a couple of
slides there.
MR. BUSLIK: All right. I did an
uncertainty analysis, and using Sapphire (phonetic),
and to get some idea on the core damage frequency,
this was only for large, dry plants. And you get 6.7
E minus five per year for the mean and 1.8 E minus
four per year for the upper bound.
And if you go to -- now, this is for one
plant, one large dry. Presumably if you're
considering the average core damage frequency for a
set of, say, the 18 large dry -- this, by the way, was
for a case where the sump clogged in all size LOCAs --
presumably there the uncertainties would tend to
cancel out. The uncertainty in an average is less
than the uncertainty in an individual sample.
And that's to be indicated here. So it
looks like it's very highly likely that it's cost
effective. The only problem is, of course, if the
probability of some clogging instead of one is .2 or
something like that.
CHAIRMAN BONACA: I just want to ask you
a question about it. You know, when I look at the
cost-benefit analysis here, the benefit is all coming
from averted costs. Assume that for the case where
you have sump blockage and you give the probability of
one. That means that all the money that is going now
in supporting high pressure injections, sit tanks
(phonetic), testing, everything that a tech. spec.
requires and everything else; so much is driven by the
requirements of LOCA in the power plant.
All of these costs are totally lost, is
being wasted today because you're saying that --
MR. BUSLIK: It is all plant protection.
CHAIRMAN BONACA: Yeah. So wouldn't the
costs also have to be considered or it's just simply
simplification you don't consider that?
I mean, it seems to me that that's --
MR. BUSLIK: I don't understand what
you're saying.
CHAIRMAN BONACA: What I am saying is that
there is a lot of cost associated with running all of
the other ECCS systems in the expectations that they
will be successful. If you are telling me that when
you go to recirculation, you will not have success,
then why bother with everything else you have for
LOCA?
And I'm saying that all that is being
invested there, which is --
MR. BUSLIK: Well, in a sense, this is
included in the -- in the -- well, I'm not sure how
that's included. It's the plant which makes power,
and if you lose the plant, you lose the replacement
power. I mean you need to replace the power.
CHAIRMAN BONACA: Sure, I understand.
MR. BUSLIK: And there's decontamination.
I'm not quite sure how you --
DR. WALLIS: Well, that cost has already
been --
CHAIRMAN BONACA: I'm only saying --
DR. WALLIS: -- is gone. You've spent it
already. If you had to build the LOCA system today
and you had to figure that in, then you might well
figure out it wasn't worth doing it.
CHAIRMAN BONACA: Well, that's --
MR. BUSLIK: Well, at least for -- yes,
you might figure that for large break LOCAs you don't
need as elaborate a system or something like that.
DR. ROSEN: You know, Mario, I'm a little
troubled by the emphasis both in the analysis and in
the committee's time on the cost-benefit analysis. If
a plant has a high likelihood of sump clogging, it
would seem to me to be irrelevant whether or not, you
know, there's a two to one cost-benefit ratio or three
to one cost-benefit ratio.
They should simply verify that they do and
take appropriate measures to fix it.
CHAIRMAN BONACA: I agree with you, and I
think actually, you know, I recognize we have had
previous presentation here that was quite informative
on the generic analysis done. So but you're right.
I mean the focus today has been on cost-benefit, and
I agree with you that if there is a problem, the issue
of compliance is significant in that case.
MR. BUSLIK: Yes. I think as long as you
know there's a significant safety benefit, you don't
really -- they've figured that that's sufficient.
DR. ROSEN: Well, I take it even one step
further than you do, Mario, and you brought to the
issue of compliance, and I bring it to the issue of
responsibility for the nuclear --
CHAIRMAN BONACA: Of course.
DR. ROSEN: -- safety of the public and
the plant workers and the investment. Responsible
management faced with the finding that their plant has
a high likelihood of sump blockage, I think would take
prompt action to remedy the situation.
CHAIRMAN BONACA: Sure. That's why I
spoke before of urgency. I mean, there is some
urgency here, and --
MR. MARShALL: One reason we presented the
risk and the cost-benefit considerations is even
though this would have been very important for safety
enhancement, even if this was treated as a safety
enhancement, it still bolsters the case that this is
something that merits attention.
Even based, if this was a safety
enhancement, we would still have a case of moving
forward with it, and again, as Art pointed out, we're
required to consider or at least prepare the cost
estimate for the decision makers to look at also.
So we're presenting all of the information
we're going to be presenting to NRR as they take
action on our recommendation.
DR. ROSEN: Don't take my comments that
this work was not required, but I think we look at it
and then we get past it.
MR. MARShALL: Okay.
DR. WALLIS: Well, I like your sentiment.
It seems to me that responsible plant management ought
to figure out what to do no matter what the NRC does.
Now that there's a problem that seems to be there,
they ought to respond with the appropriate action no
matter the NRC may be doing in the meantime.
And it may be that their response will be
to show that it's not a problem, but no matter what it
is, they can't do nothing.
CHAIRMAN BONACA: And even -- I mean, I
think the lack of specific guidance should not be an
obstacle either. They know what the configuration of
the plant is. They know what the installation is, and
they have AEs that have done the original analysis.
They can be repeated with certain considerations.
And so I think that I agree with you.
MR. ELLIOTT: Can I mention something from
past experience?
My name is Rob Elliott, and I had the lead
for the Bulletin 96-03, which was issued to implement
the modifications to resolve the issue for BWRs.
At the time we issued the bulletin, there
wasn't detailed guidance out for the BWRs either. The
BWR owners group took the lead, prepared that
guidance. We reviewed and approved it after the
bulletin had been issued.
And licensees managed to implement all of
their hardware modifications within two and a half
years of the bulletin being issued.
So, I mean, if we get everybody working on
the issue, we can be working on the detailed guidance,
you know, almost immediately if there's agreement that
we need to address the issue. That's what we need to
get to.
MR. MARShALL: The important thing I think
Rob mentioned was the detail guidance was actually
prepared by the BWR's owners group. It wasn't
prepared by the NRC. We prepared, again, like a very
performance type guidance, but some people didn't feel
that was detailed enough to work from, and so they
took it upon themselves to provide their members
detailed guidance to follow, and it provided options
on A, B, C, D, on how to address debris generation.
MR. ELLIOTT: And transport.
MR. MARShALL: And they submitted that to
our office for NRC review, got an SER on it. So the
individual utilities had confidence if they followed
this and submitted it to the NRC, it would be
acceptable.
Just in closing because I think I ran out
of time a minute ago --
CHAIRMAN BONACA: That's okay.
MR. MARShALL: -- I just want to reiterate
our proposed recommendation: again, plant specific
analysis, and if a problem or vulnerability is
determined, implement an appropriate corrective
action.
And that's what we'll be sending to NRR
during this month.
DR. ROSEN: And for the committee's point
of view, what I understand from this is that you do
want an ACRS letter --
MR. MAYFIELD: That is correct.
DR. ROSEN: -- on the basis of what we've
heard today.
MR. MAYFIELD: That is correct, sir.
CHAIRMAN BONACA: What's the sense of the
membership? I think we should have one.
DR. WALLIS: Well, I'm a little concerned
because we only have one side of this. We have this
one report which does have assumptions in it. So we
don't have any kind of other view that says -- it
seems to have a vague statement that these assumptions
are conservative. We don't have a basis for knowing
what's really realistic. We just have to either
believe that LANL report or we have nothing to go on.
CHAIRMAN BONACA: Well, we received the
presentation here and read a report. It was quite
detailed and had a generic treatment of the issue.
There were representations of certain types of sumps,
one that would flush and then stepped up and different
heights of those, and they were pretty detailed
insofar as the generic representation of sumps.
I was left at the time with the sense that
all that could be done generically was done, and we
had to move into plant specific already. That's why
today I was surprised at the beginning that we were
not facing that kind of recommendation immediately.
Then I saw it coming through, but it seems
to me that we know enough to justify this
recommendation. Now, you had a different sense from
it, Graham?
DR. WALLIS: No, I just am saying I'm
anticipating that there will be another view of the
problem when eventually industry gets around to it.
It may look rather different.
CHAIRMAN BONACA: And on a plant specific
you might find that there are no problems or there are
problems, and that will be --
DR. WALLIS: And then it will come to us
again presumably. We may have to arbitrate between --
CHAIRMAN BONACA: I think we'll have to go
away from genericity and go to specificity for the
plants.
DR. ROSEN: Well, I think we clearly have
to make a choice. I think anything we write now would
have to be an interim letter. It will not be our
final word on it.
So we have to choose whether we want to
say something on the interim, on the basis of the
interim work we've heard about and seen so far or hold
off.
DR. KRESS: There's not much chance this
committee will get a chance to look at all of the
individual plant specific analyses that come in. We
need probably to make this our final letter, probably.
CHAIRMAN BONACA: Yeah, I think so, too.
I mean, do we believe that this is an issue that would
deserve, in fact, this recommendation?
DR. KRESS: That's the issue, I think.
CHAIRMAN BONACA: That's the issue, and
you know, I personally believe that. So I'm
supporting of a letter that will recommend that.
But I accept that the studies that we've
done to date may have limitations and you know.
DR. KRESS: Yeah, I think I would support
that conclusion also. I think the point of debate for
our letter might revolve around the need for guidance
and what that might take.
DR. ROSEN: Certainly that will be a point
of debate and how clear we come out on that point will
be important. But I think also, as Graham suggests,
we haven't heard the industry reaction yet, and we may
get some important input that could cause us to revise
what we might say this week.
DR. KRESS: If we get such input. We're
quite often faced with that situation though, and we
go ahead and make our judgments based on what we know,
and that's more likely to be the case here, I think.
We've got the final work probably before we write a
letter.
So I suspect we ought to resolve ourselves
to making our judgment based on what we've already
heard.
MR. MAYFIELD: If I could, this is Mike
Mayfield.
As part of the generic communication
process, there are opportunities for the committee to
be briefed on and comment on generic communications
that might issue from this.
DR. KRESS: Yes.
MR. MAYFIELD: So as the process proceeds,
there will be another look at this potentially.
CHAIRMAN BONACA: Any other comments from
members or points of view?
(No response.)
CHAIRMAN BONACA: Okay. If not, I think
we are done. So we will recess the meeting for 15
minutes and take a break until 10:20.
(Whereupon, the foregoing matter went off
the record at 10:03 a.m. and went back on
the record at 10:21 a.m.)
CHAIRMAN BONACA: Okay. Let's resume the
meeting now with the next item on the agenda. That's
the EPRI report of resolution of generic letter 96-06,
waterhammer issue. I believe Dr. Kress is the
responsible individual.
Dr. Kress.
DR. KRESS: Thank you, Mr. Chairman.
We had a Thermal Hydraulic Phenomena
Subcommittee meeting on this issue August 22nd and
23rd of this year. Not many members were there. So
we have quite a bit of time on today's agenda to try
to cover the issue.
To refresh your memory, there is a report
on the subcommittee meeting, handout 311, that you
may have already read, but to refresh your memory
anyway, this is a compliance issue for a design basis
event. A large break LOCA combined with the loss of
off-site power or a main steam line break combined
with the loss of off-site power sets up a condition in
which you're likely to get a waterhammer event in the
fan cooler units of containments.
And such an event could give you the loss
of the function for the cooling and might even set up
a bypass path from the containment.
So the generic letter in 96-06 requested
that plants evaluate their vulnerability to this
issue, and the work that was done by EPRI and industry
in a collaborative effort was to provide guidance to
licensees to do an individual plant evaluation or a
specific plant evaluation of their vulnerability to
this issue.
And the work they did was to develop a
methodology for making the determination, and this
methodology has in it a component of determining the
amount of air and steam that makes a pocket in this
event, and it's very important to know how much, what
size this pocket is, and what its constituents are
because it's a major factor in ameliorating the
intensity of the waterhammer.
So we previously had a subcommittee
meeting on this in which we looked at their
methodology, and we had basically three issues with
it.
One of them was the determination of our
release fraction that made this void region as the
event occurred. We felt the experiments that the
release fraction was based on was apparatus dependent
and might be difficult to scale to the FCUs that are
actually in the plant.
The other one is --
DR. POWERS: Can I ask -- I sent you an E-
mail asking some specific questions about the details
of the experiment on that air release fraction. Did
we ever get any clarification on that?
DR. KRESS: I was hoping we could ask that
question at this meeting and get it clarified. I've
not --
DR. POWERS: And there's a lot of problems
of nucleation and whatnot in trying to get gases out
of water in dynamic events.
DR. KRESS: Yeah.
DR. POWERS: There just didn't seem to be
enough discussion on that to me.
DR. KRESS: Yeah. I definitely think when
we get to the discussion of the termination of the
release fraction that you need to bring that up again,
Dana.
The other problem that we had previously
was to determine the amount of steam that gets
condensed and its effect on the amelioration. It was
experiments to determine an hA term for condensation
where condensation was hA delta T, and so we thought
that might also lack enough technical basis to be
scalable, and in general, scalable for the test data
to the full size was our problem.
So at the subcommittee meeting, the EPRI
group attempted to address these issues, and I think
they will also address them further in this meeting.
So with that as a preliminary, I guess
I'll turn it over to Jim Tatum of NRR.
MR. HUBBARD: This is George Hubbard,
Acting Branch Chief for Plant Systems Branch.
Before Jim gets up or Jim can go ahead and
start going forward, just Dr. Kress mentioned this
methodology. I wanted to bring in focus a couple of
things is this is not for the entire industry, as Jim
will point out in his slide. This is for about 24
plants.
I think most of the other plants have
addressed this issue, and they have satisfactorily
accepted their resolution of the issue, but for these
plants that EPRI is focusing on is they decided to go
into a group to develop this methodology.
The other thing that I'd like to point out
is that this is a low pressure system. It's probably
up to about 100 psi so that we're not dealing with the
high pressure waterhammers that we generally think of
with the, you know, 800,000 or, you know, high
pressures.
So I wanted to bring those two points out,
and then I'll turn it over to Jim, and he'll bring us
up to speed on the issues, a brief introduction, and
then EPRI.
Thank you.
MR. TATUM: Good morning. My name is Jim
Tatum. I'm from the Plant Systems Branch.
What I'd like to do, I think, just to make
sure everyone is on the same page here on this issue
is to provide a brief introduction as far as what the
issue is, and then defer to EPRI. I think they have
additional explanation that they would like to give
us, and upon completion of that, go ahead and discuss
the staff perspective on this thing.
Let's see. Now, in the way of
introductions, Generic Letter 96-06 was issued just
about five years ago.
DR. WALLIS: Excuse me. Do we have copies
of your presentation?
MR. BOEHNERT: You should have it in front
of you there.
MR. TATUM: Hopefully.
MR. BOEHNERT: It's a single page. If you
don't, I have copies for you here.
MR. TATUM: Okay. About five years ago we
issued the generic letter in response to some work
that was done at Diablo Canyon and Westinghouse in
looking at the fan cooler system and an issue that was
identified.
The specific scenario that we're talking
about has to do with a LOCA, large enough LOCA to
provide significant heat input into containment and
transfer that heat to the cooling water system.
Typical fan cooler units, this is a pretty
good schematic I borrowed from the EPRI document. I
think it came from Volume 2, but typically what
happens is you have a loss of power. You lose the
service water pumps or the cooling water pumps,
whatever the case may be, that is providing flow
through the fan cooler system, and at the same time,
the fans that are blowing air through the fan coolers
are winding down.
There is a timing difference, however.
The pumps will coast down much more rapidly than the
fans will coast down, and so what you have is a
situation in the containment where you have the heat
from the LOCA that's released rather quickly, and you
have the fans continuing to wind down, transferring
that heat into the fan coolers through the fan
coolers, which are very efficient heat exchangers.
They're designed to transfer heat, typically have
copper tubes that have fins on them, and so you get a
rather rapid, immediate heat transfer into the fan
coolers themselves.
As you get the heat transfer in there, the
concern was whether or not you would have a
significant amount of steam formation, and if that
steam formation could lead to some significant
condensation induced waterhammer event, thinking back
to the days when we were looking at the waterhammer
events associated with steam systems, steam
generators, feed rings, that sort of thing.
And not knowing a whole lot about the
response of low pressure systems and whatnot, we
thought for a level of comfort, make sure that these
systems wouldn't be compromised during the event, that
licensees really should take a look and see if their
systems were robust enough to be able to handle the
event.
DR. WALLIS: Jim, this is a very idealized
picture, and in reality, as we've said before, these
fan coolers are connected with all sorts of piping
that goes up and down. It goes into big headers, and
each plant has very specific piping.
MR. TATUM: That's true.
DR. WALLIS: Very specific connections,
very specific ups and down, and this sort of gets lost
in all this work and the connection between this
reality and some idealized view is being lost to some
extent throughout this work.
MR. TATUM: That's true, and I think EPRI
can talk a little bit about what they've done in the
way of the participating utilities. I mean, they have
surveyed and tried to get a pretty good feel for what
the specific piping arrangements are for the plants
that are involves with this particular study that's
been done.
But you're right. I mean, the plant
designs are very plant specific. There's not a
standard design. You can have the fan coolers at a
high point. You can have them at a low point.
Typically I think it's more common that
you see them at a high point in the system. You do
have headers that feed into the fan coolers, and off
of those headers then you have small tubes that form
the majority of the fan cooler itself where the heat
is transferred.
But you're right. There are different
turns, maybe different systems that are cooled by the
cooling water system in containment. It may not just
be the fan cooler. So there are some complications
that have to be considered in all of this.
MR. SIEBER: Let me ask a question. When
you have a LOCA, the containment temperature and
pressure changes pretty rapidly, but not
instantaneously. Did you take into account the
profile of containment temperature with time and
compare it to the time that that the service water is
not flowing?
MR. TATUM: Yes. Typically what the
plants have done is they have looked at their
containment profiles for the design basis LOCA, and
based on those profiles, they've maximized the heat
input typically to get the maximum steam volume that
you might be able to get from the heat that's in
containment.
That's a little bit idealized because
obviously there's difficulties in determining where in
containment the heat is being disbursed. You know,
there's going to be some complications with just
getting down to how rapidly it is going to be
transferred through the fan cooler.
So the process that utilities have
typically used is to look at worst case type
conditions, take a look at the profile, assume that
heat is there available to the fan coolers, transfer
the heat into the cooling water systems just kind of
as an approach to try to get past, well, yeah, you
have the LOCA. How is that heat being conveyed
through the containment? How long does it take to get
to the fan coolers?
I mean, there are questions that can be
asked that we really didn't go -- it wasn't the
purpose of this generic letter really.
Our feeling when we issued the generic
letter was that the bounding case, the limiting case
would be maximum steam formation with the potential
for a condensate induced waterhammer event. That's
really what our concern was going into the generic
letter, this aspect of it.
MR. SIEBER: Well, if the licensee would
respond to the generic letter by doing an analysis
that's time dependent, I presume you would accept that
kind of analysis.
MR. TATUM: If it were justified. I mean,
from the staff's perspective though, it would be
difficult because we look at design basis scenarios,
and so as design basis we look at the temperature
profile, and we go by, well, at this point in time you
have this temperature in containment, and we assume
that it's disbursed uniformly throughout containment.
So you know, we don't get, and I think it
would be very difficult to try to model exactly where
that heat would be at any point in time. So we have
to make some simplifications.
I put up another diagram here to --
MR. HUBBARD: Jim, let me add one comment
on that. This is George Hubbard.
I think part of the reason the utilities
went together is they all realized that for their
situation, that there would be this input, and they
could have the problem, and therefore, they went to
form this group to address it.
So from their own evaluation they felt
they had the problem, and they, you know, wanted to,
you know, approach it with this methodology.
MR. TATUM: I've put this slide up to
illustrate a little bit more of what Dr. Wallis was
speaking to. The header configuration that you could
expect to see for a fan cooler unit, you have the
pipes, the main pipes that bring the water into the
fan coolers, but then those pipes transition into
individual unit boxes that make up the cooler, and the
cooler itself is composed of copper tubing typically
with fins and very long lengths and winding, making
several paths through each box.
That's kind of the arrangement that we
were looking at.
DR. WALLIS: Even this figure is a bit
strange because your left-hand one shows a supply
coming in presumably on the left, going out on the
right, both at the bottom. But on the right-hand
picture the return is at the top.
Now, where is the return? Is it at the
top or the bottom in the fan cooler?
MR. TATUM: Well, typically I believe this
is -- if you look at the diagram, I think the larger
diagram over on the side there, you have a header that
comes in, and this is very plant specific. I mean,
this isn't meant to be generally applicable to all
plants, but for this particular case, I mean, it's
showing the return coming in at the bottom and going
out at the top.
I wouldn't say that that's the case --
DR. WALLIS: They both go out at the
bottom on the left, right?
MR. TATUM: Well, if you look on the left
side --
DR. WALLIS: They both go out at the
bottom.
MR. TATUM: -- it's probably hard to tell.
DR. WALLIS: They come in and go out at
the bottom, don't they, on the left?
MR. TATUM: Well, I mean it's hard to tell
from the isometric, I think, really, but it should be
showing it coming in similar to what you have here.
I mean, coming in at the bottom, going out at the top.
DR. WALLIS: And in the EPRI experiment,
they have a pipe, and then it all comes out and
bubbles up into something.
MR. TATUM: Yeah, well, they show -- and
I'll defer comments on that. I think EPRI --
DR. WALLIS: Maybe they will tell us how
their experiment is related to this sort of picture.
MR. TATUM: Right. I think they'll be
prepared to discuss the experiment and how it relates
to the actual header configuration and that sort of
thing.
But I just wanted to make sure everyone is
familiar at least generally with the system that we're
talking about.
DR. WALLIS: The headers, the big headers
that go around containment are at about the same
level. So that return if it's up has to come down
again to go into the header.
MR. TATUM: That's correct. That's
correct. If it's in a high point, typically the
piping will come back down to where the main header
is.
DR. WALLIS: It comes down. It doesn't go
up.
MR. TATUM: Right. Now, in those cases,
and I think there may be a couple where you have the
fan coolers at the low point in which case the piping
would go up to go back to the header.
DR. WALLIS: All right.
MR. TATUM: So it can be very plant
specific that way.
DR. FORD: Could I just ask another
question?
MR. TATUM: Sure.
DR. FORD: I'm assuming SS is stainless
steel. Stainless steel tubes with copper fins; does
it change from plant to plant? Do you have copper all
the time -- sorry -- stainless steel tubes all the
time, or do you have carbon steel headers?
MR. TATUM: Well, no, the piping -- the
headers themselves would typically be some sort of
carbon steel.
DR. FORD: Okay.
MR. TATUM: Typically. Service water
system, that kind of an arrangement. The tubing
itself typically, they would be what you'd find in a
heat exchanger, copper tubing, possibly fin.
This one, this particular example from the
EPRI manual is for a particular plant, and in this
case, they're talking about stainless steel, but it
varies from plant to plant.
DR. FORD: Okay. So you could have just
plain carbon steel tubes.
MR. TATUM: Well, not the tubes so much.
The header that goes into the fan cooler.
The fan coolers themselves, I think, are
typically originally commercial type units for
transferring heat. There wasn't anything special
about the design of the fan cooler itself.
DR. FORD: Okay.
DR. WALLIS: Now, while this release is
occurring, is there flow through the system or is it
stagnant pretty well?
MR. TATUM: Well, typically what we're
looking at for the Generic Letter 96-06 scenario is
that you have a stagnant cooling water system. The
pumps stop, loss of power, and you have the air
continuing the containment atmosphere continuing to
blow through the heat exchanger as the fans wind down.
DR. WALLIS: But in this part, in the
water supply here --
MR. TATUM: Right.
DR. WALLIS: -- there's no flow through
there during this event or the pumps are coasting
down. So there is a flow through here.
MR. TATUM: Well, they coast down very
rapidly. So essentially it's no flow, yeah, no flow
through on the water side. And so you may have column
separation, you know, if you have a system that's high
in the containment and, you know, would expect boiling
to occur rather rapidly, that sort of thing.
DR. KRESS: Now, what's the general source
of this water supply?
MR. TATUM: Well, it varies. I mean, the
open loop systems, you can have the source from a
reservoir. It can be from a river, a lake. You know,
the pump service water system basically, it's that
kind of a system. It would take a suction from a body
of water, whatever is available.
DR. KRESS: So it very well could be
fairly dirty water. It's not --
MR. TATUM: It could be fairly dirty
water, and in fact, we've acknowledged that and
recognized that previously by issuing Generic Letter
89-13. So there are -- you know, problems with dirty
water systems have been addressed. I don't expect
that to be a complication for this issue per se as far
as degrading the system, aging, and that sort of
thing.
DR. KRESS: Yeah. Well, I had in mind how
that might affect the higher release fraction.
MR. TATUM: The heat transfer and whatnot.
DR. KRESS: Yeah, and the heat transfer.
MR. TATUM: Right. Yeah, the quality of
water varies, and you can have silting and different
things going on there with the water supply or marine
growth, organisms, that sort of thing. But those
issues for the most part I think we've addressed with
Generic Letter 89-13.
Getting back here to just basically
introductory comments, let's see. I wanted to just
back up now with the EPRI initiative that was proposed
in August of '98. As George has already mentioned,
there were a group of utilities that were interested
in trying to come up with a less conservative
methodology than what was suggested by Generic Letter
96-06, that being NUREG CR-5220. That's a very
bounding approach that was offered in that NUREG.
Typically it goes straight from Joukowski, does not
credit air or recognize air and cushioning, that sort
of thing.
The industry felt like they could take
advantage of some of the margins and conservatisms and
maybe reduce the amount of modifications that would
have to be done to address the issue, saving the
industry money and whatnot and still providing
confidence to the staff that they had adequately
addressed the issue.
And, of course, we were very interested in
proceeding with that effort. It was really a
cooperative effort with the NRC. We observed much of
the testing that was done.
We've had discussions with them at many of
the meetings. We were involved with the development
of the PIRT analysis that was done and whatnot. So
we've provided guidance and suggestions along the way,
but the work that was done, the analysis and whatnot,
that's strictly EPRI's, and we're going to defer to
them to discuss that part of it.
DR. WALLIS: Did you ask the kind of
technical questions that we've been asking?
MR. TATUM: Yes, we have been asking those
kind of technical questions. Unfortunately the staff
has evolved. You know, this has been kind of a long-
term project, and originally we had Al Serkiz who was
working with us, and of course, he was a key player
from our side, making sure the right issues were being
addressed at least from his perspective for
waterhammer, and he was our expert at the time for
waterhammer.
Now we have Walt Jensen in Reactor Systems
Branch and Gary Hammer doing the review. So we've
transitioned in personnel, but we've tried to maintain
continuity.
We've all looked at the same documents,
and we have asked the technical questions. And I
would say that in the meetings with the subcommittee,
obviously the questions that have been asked have been
good and helped us focus also on some areas, some I
think that we were also aware of even at the time you
were asking some of the questions as well.
So we're trying to move on with this thing
at this point, but there are about 24 plants involved
with this initiative, and these are for the most part
the plants that have not really addressed the
waterhammer issue.
The other plants for the most part are
those that do not credit the fan coolers, and they're
able to take alternative measures. For example, they
can put in the procedures, restrictions on using the
fan coolers so that they don't have to worry about the
waterhammer event, and they've been able to address it
that way.
There are a few, handful of plants that
aren't involved with this initiative, a couple that
have tried to apply RELAP. We're still reviewing
those. We have not come to a conclusion on those
other plants yet.
MR. SIEBER: Just a question, I guess. A
lot of plants can't use the fan coolers after a LOCA
because the containment atmosphere density is too high
and it's too big a load on the fans. So when you get
a containment isolation, the fans usually trip and,
except for a smaller number of PWRs, they don't
restart.
So the real issue is if you have the
waterhammer and you rupture part of the piping, do you
bypass containment?
I think to answer that you have to know
what kind of a rupture you have. For example, if you
just split a seam someplace, service water pressure is
higher than containment pressure. So leakage is in
rather than out.
Has that been taken into account, any of
these factors?
MR. TATUM: Well, we have considered that.
There are many different kinds of scenarios. The
containment bypass is one, and that can be very
complicated because depending on the plant design, you
may have to have more than one rupture in the system
to get a containment bypass.
Typically service water systems are easily
isolated from outside the containment. So there are
different mitigating factors to consider here.
Also, the service water system, what you
mentioned with the load on the fan coolers and
whatnot, that's true. It's kind of plant specific
that way, but in fact, what many of the plants do is
they will operate the fan coolers in the plants
involved with this particular initiative, typically
will shift to a low speed on the fans in order to be
able to handle the load.
And so they credit those fan coolers in
some fashion. It may be just for long-term cooling of
containment. If it's a small containment, maybe it's
in clipping the peak pressure a little bit. It varies
from plant to plant, and you have to get into the
details of each specific plant in order to see to what
extent they're crediting the fan cooler.
MR. SIEBER: Well, the containment
atmospheric pressure can triple during a large LOCA,
and that really changes the load on the fans, and so
typically you don't put fan coolers on at all until
after the first hour to get the spray down and deep
pressurization. I think that makes a difference.
MR. TATUM: Well, it does, and for some
plants that the case, and for those plants, in
particular, it wouldn't be an issue, and those would
be among the group that we've already closed.
MR. SIEBER: And I guess another comment
is that a lot of things happen during the first minute
or so of a large LOCA, and even though you probably
have a radiation detector on the outlet of the service
water, I think that's pretty far down on the list of
things to look at, and so isolation is, you know,
maybe it happens; maybe it doesn't.
MR. TATUM: Well, it would be late on.
The question is, you know, if you're looking at the
severity of the event, how long do you have? And if
you're talking about a split in the seam somewhere
where it's not a major thing, you've got a lot of
time, whereas if you're talking about a major rupture
of the piping system and a direct path, maybe it's
more significant.
So there are a lot of variables that go
into this, and I think that's one of the points I
think that needs to be appreciated here, is just the
complexity and the number of variables we're talking
about, but that's pretty much all I have in the way of
introduction.
I'd like to turn it over to Vaughn
Wagoner.
DR. WALLIS: Does the staff have a
position on this work? Are you accepting it?
MR. TATUM: Yes, we do have a position,
and I'd like to defer discussion of that until we hear
from EPRI because they're going to attempt to provide
additional information, and I think for continuity of
the discussions here it would be good to have what
they intend to say here available to the other
members, and then we can go on to the staff
perspective on this.
MR. WAGONER: I guess I get the honors
now. Is there a microphone? Am I hot?
Okay. Good morning. I'm Vaughn Wagoner
with Carolina Power & Light Company, and I chair the
Utility Advisory Committee that is composed of the
members of the utilities that have been supporting
this project with EPRI, and let's see. Well, I'm here
this morning and joining with me are Tom Esselman and
Greg Zisk from Altran Corporation and Tim Brown from
Duke Energy, and Peter Griffith was going to be here
with us this morning, but unfortunately could not make
it at the 11th hour. So we'll have to try our best to
fill in if questions get to that level.
So I just want to give a brief
introduction here. Let's see, Tom. What do I do?
MR. ESSELMAN: Page down.
MR. WAGONER: Page down. Oh, that's why.
I paged up, and it wouldn't work. Okay.
Okay. Very briefly this morning, what we
want to do with you is go through an overall
description of what we've done in this thing just to
be sure everybody is on the same page, and then get
into some specifics that we've been talking with with
the Thermal Hydraulic Subcommittee, particularly in
the areas of air release and heat transfer and the
scaling issues. These seem to be continuing
questions, and we want to try to get at those and
address them for you this morning.
First, I just want to give you a little
bit of background. When we started out in this thing,
as you've heard, there were utilities or plants
generally fell into two or three groups: those that
just flat didn't have the problem because of whatever,
over pressure in their systems or whatever. They
didn't have a problem and didn't have to address it.
Others that had some facet of the program,
but could address it in terms of either operational
changes or other changes to the plant.
And then a third grouping of plants that
had -- that appear to have the issue, create the steam
voids, et cetera, but whose piping systems were very
close to being qualified in using classic systems with
the theoretical loads that you could calculate.
So then the question became is there some
mechanism or is there some activity because these are
aerated systems for the most part and there's boiling
going on; is there something going on there that we
could take advantage of?
DR. WALLIS: Well, let's ask you. You're
assuming these are aerated systems. Do you monitor
how much air is in the water in these plants or do you
just assume it's there?
MR. WAGONER: Well, there's fish that live
in the pond and they don't die. So there's got to be
some oxygen in there.
(Laughter.)
DR. WALLIS: But they aren't all like
that. Don't some of them have a storage tank and they
recirculate and so on?
Do they all bring in water from the
outside?
MR. WAGONER: The open systems that I'm
familiar with --
DR. WALLIS: Are all open?
MR. WAGONER: -- as far as I know, all
participating in this study are all --
DR. WALLIS: Are all open systems?
MR. WAGONER: They're either open or they
are closed systems, but we treat a closed system
differently with respect to the potential for gaseous
release.
DR. WALLIS: So are we talking only about
open systems here or are we talking also about --
MR. WAGONER: Yes, sir. We're talking
about both open systems and closed systems, but in the
technical basis report and the user's manual, there
are differences.
DR. WALLIS: So in the closed system we
don't have an idea of how much air is in there?
MR. WAGONER: Well, we do have an idea.
DR. WALLIS: But we don't have a
measurement or something?
MR. WAGONER: Well, you've got -- you know
there's air in the water that's put in. Then
typically there's some kind of oxygen scavaging added
to the -- because it's a closed loop to prevent rust
and stuff like that. So what you're left with then is
the other.
DR. WALLIS: So you're taking oxygen out.
MR. WAGONER: Right. So what you're left
with are things like nitrogen and what other small
constituents of things that aren't removed by oxygen
scavaging chemicals.
DR. POWERS: So when you say, "Okay. I've
got this water" -- do you say you have some idea how
much dissolved gas there is? How do you come up with
that idea?
MR. WAGONER: Typically you would -- we
don't typically take measurements of it on a routine
basis, but then again, it's large bodies of water,
surface area exposed to air. So you --
DR. POWERS: Yeah, I know, but now I still
need a number.
MR. WAGONER: Okay.
DR. POWERS: How do I get that number?
MR. WAGONER: And that number is that we
would look in a textbooks and see what the typical
dissolved gas would be.
DR. POWERS: Okay, and I look in those
textbooks and they give me the number for pure
distilled, 23 meg water. Okay? And that's a number.
Now, if I looked farther in the textbooks,
they would tell me there are section now coefficients
that will tell me how dissolved salts will reduce that
number. Do you take that into account?
MR. WAGONER: Dissolved salts?
DR. POWERS: Un-huh.
MR. WAGONER: Do you mean things that
might be dissolved in the water?
DR. POWERS: Right.
MR. WAGONER: Not necessarily. I guess
the question would be, you know, how much effect is
it. Does it take it all out or a little bit?
DR. POWERS: Well, I guess I'm asking you
what the effect is.
MR. WAGONER: And I guess I can't answer
that.
DR. POWERS: Oh,
DR. WALLIS: When these pumps pump the
stuff around, there are regions of low pressure where
maybe you get air bubbles coming out and so on. So
there are mechanisms that influence the air content of
the water. It's not as if you just take the figure
6.7, solubility of air and oxygen at one atmosphere in
distilled water and use that. I mean, there are other
things going on.
MR. WAGONER: I'll acknowledge that. I
guess I would disagree that there are pockets of low
pressure between the pump and discharge.
DR. WALLIS: Well, it's just the thing
that so surprises me is that you just sort of take
this curve and it's assumed it applies without further
discussion.
MR. WAGONER: Let me ask. Have there
been any measurements that you guys are aware of that
have actually been made in water or any other thing?
MR. ESSELMAN: Vaughn, this is Tom
Esselman.
The specific amount of gas, whether it be
nitrogen or air, that's in a plant dependent situation
depends on the plant, depends on whether you have a
bond or a cooling tower or a closed loop system with
a tank, whether it's a nitrogen blanketed tank or not,
and all of those things will enter into -- and what
the temperature is of the lake. That's clearly
different in Minnesota than Texas; what the pressure
is, whether you're taking it from the bottom of the
lake or the top of the lake.
All of those details are not dealt with by
us. What we're providing is a general recommendation
that says you need to determine how much dissolved
gas, whether it be air or nitrogen or whatever, is in
your plant at the beginning of the event.
And the kind of factors that we're talking
about are plant specific, and many of the things that
we're talking about are going to depend on the details
of the tower or the pond, and that has to be
determined, and it's clearly identified as needing to
be determined by the utility that's using this
information.
DR. KRESS: Your experiments determine the
fraction of the air that's in the water that gets
released, but they started out using clean, saturated
water.
MR. ESSELMAN: We use --
DR. KRESS: Water saturated with air. Do
you think that fraction that you determined
experimentally might have some dependence on the
initial concentration of air in the water or --
MR. ESSELMAN: We looked at the way that
air and nitrogen would come out of solution with an
increase in temperature. We saw that the behavior of
the different gases that could be in there is similar,
and that the representation or the using oxygen,
because we had a normally aerated water system; we
used tap water. We measured the oxygen and used
oxygen as an indicator of what was being released as
a percentage.
I think given the -- we will discuss this
in more detail, but given the way that we did the test
and the range of data, we believe that it applies to
a highly aerated or a moderately aerated or a highly
nitrogenated or a moderately nitrogenated system.
So the steps is, number one, the plant
needs to determine what they start with, and then they
need to determine how much water is affected, and then
they can calculate how much air would be released from
that, how much gas, noncondensable gas would be
released.
DR. WALLIS: You use oxygen as the
indicator. I'm not clear that you ever measured air.
You used oxygen.
MR. ESSELMAN: We didn't. We used oxygen
as an indicator. That's correct.
DR. WALLIS: And the assumption is that
nitrogen behaves exactly the same way.
MR. ESSELMAN: We don't presume that it
behaves exactly the same way. We know that it behaves
differently. We looked at how nitrogen and air and
oxygen behave, and their behavior is similar enough
that we were confident that using oxygen as an
indicator was representative.
But we jump ahead.
MR. WAGONER: So I guess the correct
answer to your question was that it is a plant
specific determination.
Thank you, Tom.
And that is in the user's manual.
DR. POWERS: But, I mean, I guess what's
distressing is you don't tell the user that he needs
to worry a little bit about things other than handbook
values. Pure water solubility just isn't going to cut
it for most of these. Most of these external water
sources are going to have a certain amount of
dissolved material in them. It's going to affect the
activity of oxygen strongly and nitrogen more
moderately.
MR. WAGONER: But is it not true that
within the tech. manual or within the user's manual it
does say that on a plant specific basis you need to
look at the --
DR. KRESS: If we assume the extraction of
the air during the process of the event, the boiling
event and so forth, was a stripping mechanism, which
is generally described in mass transfer texts as a
product of some sort of mass transfer coefficient and
a surface area and a driving force, the driving force
being the concentration in the difference between the
liquid and what's in the --
DR. POWERS: Activity.
DR. KRESS: -- activity. Okay. But --
DR. POWERS: Activities count in this.
DR. KRESS: Yeah, okay. But my point is
it seems to me like that activity is concentration
dependent. It depends on the concentration in there,
and you're saying --
DR. POWERS: But, I mean, the subtlety of
water is it's not dependent on the concentration of
oxygen. It's dependent on the concentration of
everything else.
DR. KRESS: Yeah, yeah.
DR. POWERS: I mean that's why water is
different than usual solutions.
MR. WAGONER: So I guess the question then
would be whether or not in your minds or our minds
that if, given the conditions that exist within the
fan coolers during this transient event, is there,
other than the total amount, is there anything that's
going to preferentially act on or not act on the
ability of oxygen, nitrogen, and whatever else is in
there to get out of the water?
And when you're taking it down to darn
near zero pounds absolute and then boiling the heck
out of it, I'm not sure that was -- I guess the
question is: is there any significant differences in
what's going to happen with the ability to --
DR. WALLIS: Well, it's not zero pound --
it's about half an atmosphere, isn't it?
MR. WAGONER: Well, it eventually gets up
to half an atmosphere, but it --
DR. WALLIS: Well, it goes through
something lower before that?
MR. WAGONER: Well, as the pumps fall away
and as the steaming starts, you have a pressure
decrease as nature is taking the water column down to
its normal 32 or what --
DR. WALLIS: So it goes down to about
zero?
MR. WAGONER: So it's headed down, and
then the steaming process starts, and then the
pressurization starts chasing the depressurization.
DR. WALLIS: And so all of your
experiments are done at half an atmosphere. Isn't
that the case?
MR. WAGONER: I believe that's correct.
DR. WALLIS: Which was chosen for some
reason?
MR. WAGONER: Somewhere between the
starting point of zero and roughly atmospheric that
some of these systems go to. So we tried to pick a
point that didn't give too much credit to just
degasification.
DR. WALLIS: But this is plant specific,
isn't it? I mean, this pressure history is plant
specific.
MR. WAGONER: Generally, yes.
DR. WALLIS: And so you're claiming that
your experiments all are operating at one half an
atmosphere are somewhat typical of all plants no
matter what the history of the pressure in that plant?
MR. WAGONER: Because of the fact that the
pressures are not -- we're not talking about hundreds
of pounds of difference. We're talking about, you
know, three to five pounds difference absolute, across
the range of plants. Because generally there's
various elevation differences, and you only go to zero
on the depressurization side, and then the
repressurization side is generally around an
atmosphere or less.
DR. WALLIS: Well, you're saying this, and
I'm not sure this is in the report. I mean, you read
the report. Someone did experiments at half an
atmosphere, and it's never really -- maybe it is. I
didn't find it -- sort of explained why this is
representative of what you're talking about here,
which is a history of pressure which can be quite
variable from plant to plant.
MR. WAGONER: I thought that we had
discussed that in the original reports. Perhaps
I'm --
MR. ESSELMAN: I would comment briefly
that we have looked at both the effect of
depressurizing a system and the effect of boiling a
system, and there are papers and references that deal
with how water behaves when it's depressurized and
agitated. The amount of gas that's given off within
this time period, which is about 30 seconds, is very
small in comparison to what we measured from the
results of boiling.
This material, including that pressure, we
will go through when we deal with the boiling test,
which is on the agenda. I guess we might defer the
details until we get --
DR. WALLIS: You're saying something which
sounds credible. If you had done the experiment at,
say, one atmosphere and a half an atmosphere and got
the same amount of air because the boiling process
dominates, that would be convincing. It would be nice
to see it. I mean, you're sort of assuming there.
MR. ESSELMAN: We ran at a half an
atmosphere because we wanted to remove the air in the
system prior to the start of the test, and we did that
by running steam through it and then closing and
allowing that to condense. So we started with an air
free system that was at a half an atmosphere.
We also researched the release that we
would have expected by pressure beforehand and
concluded that whether we ran it half an atmosphere or
one atmosphere would be immaterial, and we ran the
test on that basis and --
DR. WALLIS: This is on a theoretical
basis.
MR. ESSELMAN: Well, based upon testing
that had been performed by others, yes, not by the
testing that we had performed. But yet we looked at
that; we referenced that work in the technical basis
report.
MR. WAGONER: Okay. Let me move on
through what we're trying to accomplish in the
program, and four things that we were trying to do.
One was understand the behavior of the
system, and you heard the overviews of what went on.
And we wanted to understand in general how that
worked, what happens in terms of coast-down. Did flow
ever really quit? What happens in terms of fan coast-
down? Did fans die rapidly or did they die away
slowly such that it really was an issue?
And then where did water go? Is steam
created? Where does it go? How far does the bubble
go, and those kinds of things, and how we go about
tracking those?
We wanted to determine the safety
significance of the issue. Frankly, as you heard,
there was a lot of data around on high pressure
waterhammers. There wasn't much around on low
pressure waterhammers and what happens here.
And so we wanted to try to understand
that, and basically there's three things we had to
deal with. One is retaining cooling capability of the
fan coolers at whatever post accident requirements
that are there; maintaining containment integrity,
such that it didn't set up a bypass for containment;
and then maintaining or not flooding the containment,
not creating a flooding path for containment.
So that was the three things that we try
to deal with, and then we wanted to provide a
methodology to assure that we do maintain these
pressure boundaries and also, again, as you heard
mentioned, we want to minimize modifications that we
didn't have to make. We were willing to do anything
that we needed to do, but if we didn't have to, then
we wanted to try to pursue that.
And frankly, as we worked through that,
and that was the reason that a bunch of us utilities
got together, even though that we determined -- had
the potential for the problem; when we looked at it,
even using some Joukowski type loading, we were close.
It got down to trying to qualify the steel in the pipe
supports, and we were darn close. So we were just
looking for a little bit.
And you've heard the numbers, 20, 30
percent in load interaction with the piping support
system, and if that was possible, then we wouldn't
need to make modifications to the plant, and frankly,
the intuitive feeling is, and my experience with
waterhammers up to about 300 pounds or so, the stiffer
you make the system, the more trouble you get into.
I spent two years chasing one in the wrong
direction, and we went back and chased it in the other
direction and put rod hangers on the pipe, and it's
been banging for 15 years, and we don't have a
problem. The more steel I put in it, the more
concrete we tore out of the wall.
Okay. But moving along -- I'm sorry? Oh,
I'm sorry. I thought I heard someone.
Anyway, we put -- in order from an
industry perspective, we got Altran Corporation
together and assembled an expert panel to provide us
an independent perspective of what it was we're doing.
We wanted to get the very best in the industry that we
could, but unfortunately you're all on the ACRS. So
we had to go with --
DR. POWERS: Flattery, sir, will get you
anywhere you want to go.
(Laughter.)
MR. WAGONER: So we did assemble these
folks with a lot of experience, and I can tell you,
and I think most of you have had interaction with
them, they are independent. It didn't matter who was
paying the bill. We had some quite informative and
lively discussions on what it was we were trying to do
and acknowledged right off that we don't know
everything about the science and the details of the
interaction, but what we think we have done is provide
a reasonable approach that helps us to adjudicate the
loading, and that's what we're really working at.
And we had this utilities steering
committee. I chaired it, and we were active in it,
and our focus was to be sure that we were looking at
that stuff that would help us where the rubber meets
the road, if you will, and look at safety significance
and then look at applicability of the results to the
power plant.
Let me drop down two slides in your
handout, and I'll come back.
DR. WALLIS: Well, the one that you didn't
show us.
MR. WAGONER: Well, I was going to come
back to that one, if you'll -
DR. WALLIS: You're going to come back to
it?
MR. WAGONER: Yes, sir, I will. Okay?
Only because it's -- well, I'll get to it here.
I want to wrap up my part with just a
perspective on where I think we are in this situation.
First off, we're dealing with a very low probability
event, and the combination of LOOP-LOCA or LOOP-main
steam line break, when you sum them all up, for all
the plants that are participating, it's less than ten
to the minus six. Actually it's much less than ten to
the minus six because this ten to the minus six on
frequency is over a 24 hour period.
This thing is over in 60 seconds, and when
you do that, you take it down another couple order of
magnitude. So we're dealing with something at ten to
the minus eight, ten to the minus nine probability of
even happening, and in fact, as you know, there are
efforts underway to eliminate simultaneous LOOP-LOCA
as a design basis event. So --
DR. POWERS: I mean, I think what you're
saying is that the mean value of the probability is
very low, but if I asked my blacksmith friends if they
are very certain about that number, they say, "Well,
no." And so when I ask them about 95 percentiles,
those probabilities come up fairly dramatically, don't
they?
MR. WAGONER: Come up to -- bring them up
to -- bring them up two orders of magnitude, but then
take it down to the real time of the event, which is
60 seconds, and you add back three orders of
magnitude. So I think realistically any way you cut
it, the initiating event is pretty low probability.
DR. WALLIS: But you're not asking us to
evaluate the risk. You're asking us to evaluate a
technical report on waterhammer.
MR. WAGONER: Yes, sir, I am, but what I'm
asking you to do is look at a perspective that is at
a reasonable judgment to use to mitigate the
theoretical loading versus understanding everything
that's happening right at the interface. That's where
I'm coming from from a risk perspective.
DR. ROSEN: What you're saying is that if
you don't have a loss of off-site power, you have a
LOCA, but you don't have a loss of off-site power; you
don't have this event.
MR. WAGONER: The event never happens.
That's right.
DR. ROSEN: And I think it's generally
understood and believed that loss of coolant accidents
don't cause losses of off-site power. Generally
plants, even when they trip, as they would in a loss
of coolant accident, the grid is typically unaffected
by that. The plants continue to receive off-site
power, and in that case, this event wouldn't happen
because the fans would never coast down. They would
be starting if they weren't running, and the component
cooling water or whatever service water would never
stop.
DR. POWERS: Isn't there a lower bound on
this just given by the seismic hazard? You can never
go lower than the seismic hazard on this one?
DR. ROSEN: I think that's fair because
losses of off-site power would occur during a major
seismic event that was strong enough to cause a LOCA.
MR. WAGONER: So anyway, I think we're
starting with a low probability event. We looked at
the risk of pipe failure, again, looking at our three
safety functions, maintaining coolant capability,
bypassing containment or flooding containment. Those
last two require you to do something to the integrity
of the system.
And we think there are significant margin,
and that's why I go back to the slide that you thought
I was going to skip over, relative to the structural
integrity. If we looked at a typical tubing or
typical typing material, steel --
DR. POWERS: People never do that though,
do they?
MR. WAGONER: Huh?
DR. POWERS: I mean when we go through
ASME codes and things like that, we never look at
typical. We look at lower bound numbers, don't we?
MR. WAGONER: I've lost you. What's --
this is typical piping that's used, is carbon steel,
standard wall, .375 thickness. It might be eight,
ten, 12, 14 inch. So that's why I say this is
typical.
DR. POWERS: Well, you're going to go
through these various stresses numbers here. Are
those typical values or are they lower bound values?
MR. WAGONER: Well, these numbers are
right out of the code.
DR. POWERS: Okay.
DR. WALLIS: Well, it doesn't say use Sult.
It says use S allow, isn't it, which their number
doesn't become 3,000? It becomes less than 1,000.
MR. WAGONER: Okay, and that's true, but
you can use ultimate if you're looking at an
operability issue or looking at a real world behavior
of the pipe.
MR. BROWN: Vaughn, this is Tim Brown,
Duke Energy.
That's a faulted event. So ASME lets you
use 2.4 SH, which is very close to S-ultimate.
DR. WALLIS: It lets you use Sult?
MR. BROWN: It's very close to S-ultimate.
MR. WAGONER: Thank you, Tim.
But anyway, there's some margin. These
numbers you'd have a factor of about six. Take it
down a little bit and you've got a factor of two,
three, four, five.
DR. WALLIS: Now, this 600 -- sorry.
DR. FORD: I was about to say is B-280 as
a copper?
MR. WAGONER: That's right.
DR. FORD: Copper, copper-nickel?
MR. WAGONER: Yeah, that's typical copper-
nickel tubing, which in fact is typically what's in
the heat exchanger. Some of them have been changed to
a stainless steel.
DR. FORD: Have any of these analyses been
done on degraded piping?
MR. WAGONER: These are always -- these
are done -- well, this is a typical wall thickness.
All of these systems are monitored for degradation,
but through Section 11 of ASME code. So heat
exchangers, the tubes are monitored for degradation.
The piping systems are monitored for degradation.
DR. FORD: Is there not concern though,
Vaughn, that, for instance, B-280 -- when it goes
through that U bend, there will be erosion presumably
at that U bend. So at that point that's the thing
that's going to be hit by the waterhammer.
MR. WAGONER: Un-huh.
DR. FORD: So at that degraded U bend,
which is presumably eroded, after 20 years or
thereabouts in 8 ppm oxygenated water, what is the
safety issue then? Did not that degraded U bend be
now exposed to that waterhammer pressure? Would it
stand it?
MR. WAGONER: It could be, but again,
we're monitoring these systems. We run eddy current
(phonetic) probes through those heat exchangers to see
what the tubes look like.
DR. FORD: And that has been done?
MR. WAGONER: Yes.
DR. FORD: And there is no degradation at
that U bend?
MR. WAGONER: If there is, you have to --
you have to address it.
DR. FORD: How often is it inspected?
MR. WAGONER: Well, that depends on what
you find. If you've gone ten years and haven't seen
anything, then you -- through ASME, you're allowed to
-- through the code you're allowed certain inspection
intervals, you know, based on your findings.
DR. FORD: Presumably the -- okay, and the
same applies to the carbon steel header which is
essentially a closed tube?
MR. WAGONER: Closed with respect to the
loop that it's in, yes.
DR. FORD: And it would be a welded closed
end.
MR. WAGONER: Right, typically, yes.
DR. FORD: Okay. And that is inspected
also?
MR. WAGONER: Yes.
DR. FORD: Because that will degrade.
MR. WAGONER: Yep. And there have been
replacement programs that you heard last time. Some
folks have had to replace sections of piping due to
monitoring and indications of degradation, and that's
typical of the whole steam system.
DR. FORD: Just assume that what with the
ISI inspection periodicity you had a waterhammer
effect and it hadn't been inspected and it hadn't been
replaced. How would that affect the whole safety
evaluation?
MR. WAGONER: Well, actually --
DR. FORD: Could a degraded pipe, whether
it be the piping, the A-106 header, or the copper-
nickel tubing -- it was degraded, hadn't been replaced
-- could that withstand that water pressure?
MR. WAGONER: And, yes, it would be a
multi-degradation scenario, but in fact, from a
personal perspective, I talked with some of our
operations folks at one of the plants and said, "Okay.
What if?"
And there's a couple of things that
happen. One is our emergency operating procedures are
all symptom based. So a couple of things could
happen. You could have a containment bypass that
would be harder to detect, but it would be indirectly
indicated because you'd have to also have a loss of a
service water flow in order to get a containment
bypass.
Then the other possibility would be
containment flooding, and that's right in the EOPs
because those are all symptom based, and you would be
looking at, you know, your levels and things that are
already going on.
So the symptom based EOPs don't care where
the water is coming from. They just address it from
a flooding issue if need be.
MR. SIEBER: It seems to me that
degradation in those systems was mostly through mic.
attack, microbiologic --
MR. WAGONER: Yeah, there has been. I
think mic. has shown up in stainless steel systems on
occasion.
MR. SIEBER: It really shows up in carbon
steel piping.
MR. WAGONER: Oh, okay.
MR. SIEBER: And the ISI program uses an
ultrasonic thickness gauge, which is a spot
measurement.
MR. WAGONER: Yes.
MR. SIEBER: Those numbers there are min.
wall numbers, okay, for typically that's Schedule 80
piping, and so when you measure the thickness in the
manufacture, there's a corrosion allowance built into
it.
MR. WAGONER: Okay.
MR. SIEBER: And all of the stress
allowances are based on min. wall. Okay? So that's
how you get a service life out of it. You could
actually calculate the degradation and the bursting
pressure if you're below min. wall, but the code says
you've got to replace it when you hit min. wall or
below it.
MR. WAGONER: Okay.
MR. SIEBER: And you have to measure at
more places if you find one place that's below min.
wall.
MR. ESSELMAN: The 600, is this with air
in the lines or is that without air in the lines?
MR. WAGONER: No, that's just an
assumption at 20 feet per --
DR. WALLIS: That's just an assumption?
MR. WAGONER: Well, it's at a 20 foot per
second --
DR. WALLIS: Is this the Joukowski
pressure or is this with air?
MR. ESSELMAN: This is Tom Esselman again.
That is uncushioned. It's without air.
That's just the straight Joukowski --
DR. WALLIS: Then why do we worry?
MR. ESSELMAN: The purpose of this is to
say that a failure mechanism that we need to address
is not one that is frequent in waterhammers of much
larger pressure which causes the tube or a pipe to
burst. And in these systems, 600 psi waterhammer is
greater than any of the waterhammers we expect to see
because we have a controlled velocity of closure.
The closure velocity is determined by the
pumping characteristics. So that this is the largest
pressure that we can see from this event that we're
talking about here, again. The burst pressure which
does have to -- which has to be augmented clearly by
satisfying all of the ASME code requirements not only
for burst, but for bending, but that burst pressure
just is shown to indicate the margin that we have been
the pressure that we will see in this event and what
it takes to burst the pipe.
Now, bursting the pipe is one of the
mechanisms that have to be considered. The other is
a traveling wave that has pulled supports out of the
wall for other kinds of waterhammer, and even for this
waterhammer at those magnitudes, it has the potential
to do that.
But yet from an integrity point of view,
a piping integrity point of view, what we would like
to -- what we're trying to point out here is that
we're not concerned -- obviously we have to be
concerned, but yet this waterhammer cannot challenge
the burst pressure of the typical components. What we
are focusing on is the traveling wave, the conversion
of those waves into support forces, which is Vaughn's
second bullet, if I may, that says that we are
focusing on support failure and subsequent deformation
of the piping system that would be required to
challenge the pressure boundary integrity.
We have to evaluate for burst pressure,
but we're so far away in this case that we are
focusing much more on how to track this pressure wave
through the system and get to the point where we can
calculate support forces because that's the line of
defense.
Before pressure boundary integrity can be
challenged, you have to cause the support to fail, and
then you can subsequently challenge the pressure
boundary integrity. That's a much more difficult
failure mechanism to occur.
DR. WALLIS: This is very interesting to
me. We spent about two thirds of our time, and we
have yet to get to the EPRI report, which is the whole
focus of our meeting, isn't it?
Are you up here to take all of the shots
before we get to EPRI?
MR. WAGONER: I was going to give a brief
introduction. Let me just make one more point and
I'll quit, and that is I think to why are we worried.
Dr. Wallis, frankly, I have the same question. Why
are we worried?
Because we're really down to dealing with
a compliance issue. We're trying to make the
mathematics work in our piping analysis system.
That's where we are.
I don't believe -- we've got a low
probability event. I don't think we have a safety
significant issue, and we're down to trying to make
the mathematics work so that we can say that we have
a system that is our piping support system meets
design basis so that we're in compliance with our
design basis. I think that's all we're dealing with,
frankly, and we need a little bit to do that, 20, 30
percent, and that's what we're trying to get out of
this cushioning thing.
And with that I'll move on. Thank you.
Tom, you're up.
MR. BOEHNERT: Now, I understand we have
to go into closed session; is that correct?
MR. WAGONER: Yes, the next slides do
contain proprietary information.
(Whereupon, at 11:25 a.m., the meeting was
adjourned into closed session, to reconvene at 12:32
p.m. in open session.)
MR. TATUM: Okay. Jim Tatum again from
Plant Systems Branch.
Staff perspective on this, first of all,
we would agree with the points that were raised by the
subcommittee. Obviously when you take a look at it,
there are shortcomings in the testing apparatus. The
hA is a very difficult value to come up with. Even if
full scale testing were done, the correct analytical
approach for calculating and coming up with a value
that would be applicable to other pipe sizes would be
questionable no matter what.
So there's uncertainty, and there's going
to continue to be uncertainty from that perspective.
But I do want to acknowledge that points raised by the
subcommittee are valid. We agree, and where do we go
from there?
And basically in looking at generic letter
9606 and resolution and whatnot, there are other
factors that we need to consider, I think, from a
perspective of regulation, public health and safety
and whatnot. We really need to try to put this in
perspective in trying to determine where do you want
to go from here.
Now, in looking at the other factors, the
other factors that come to bear here, first of all, we
do recognize and appreciate that this is a complex
phenomenon. It's very difficult to model. There's
going to be uncertainty, and we need to be able to
deal with that somehow.
We believe it's important to appreciate,
I guess, the work that EPRI has done, the involvement
of the expert panel and that's gone into it. I think
by and large they've done a pretty good job with the
resources that have been available, and the effort
that they've put into it.
They're kind of at the end of the rope --
end of the road on this. We understand their --
DR. POWERS: Or the end of the rope,
either one.
MR. TATUM: Yeah.
(Laughter.)
MR. TATUM: They're as limited in
resources was we are.
DR. WALLIS: Which end of the rope are
they on?
MR. TATUM: Yeah. Maybe that was a
Freudian slip. I don't know.
(Laughter.)
MR. TATUM: Anyway, they're limited on
resources. They're having difficulty getting
additional funds from participating utilities. We can
appreciate that. We hear that on our end as well.
The NUREG CR-5220 waterhammer loads, if
you'll look at what's calculated in that approach,
which is a bounding approach, the Joukowski approach,
what EPRI is proposing in their methodology gives you
a reduction by a factor of 1.2 to possibly 1.6. If
you look at the NUREG, it talks about the fact that
the evaluation by NUREG CR-5220 could be a factor of
two to ten conservatively, depending on what's going
on, air cushioning, steam condensation, that sort of
thing.
Unfortunately it doesn't qualify how much
reduction to expect for different facets of the
waterhammer event. However, I think what EPRI is
proposing is certainly reasonable, and it's within the
expectations at least that I would have in looking at
what is said in NUREG CR-5220 and what they're
proposing. I don't think it's out of line.
LOOP events, I think in the testing and
analyses that have been done, the waterhammer group
here has shown rather convincingly that the LOOP
event, LOOP only without steam, would be bounding.
Okay. If we take a look at just the LOOP
event, that takes us back to USI A-1 basically. That
was reviewed previously, and we considered that part
of the resolution. I think it was 927, Rev. 1, talks
about resolution in there, and we acknowledge that
plants have during start-up phases experienced those
waterhammers due to LOOP, due to LOOP testing.
Any plant design weaknesses or
vulnerabilities due to LOOP have been identified
during early start-up days and whatnot, and those
problems have been corrected. So at least in my mind
the situation with steam in the piping is a step
removed really in significance from just the loop
event.
And if we were going back to resolution of
USI A-1, I'd just remind you we really didn't go out
to the plants and have them do anything to address
this issue, and I don't think it is our purpose, nor
was it our purpose, in issuing Generic Letter 96-06 to
have plants go and address this issue. It was really
the concern relative condensate induced waterhammer
that drove the waterhammer issue in Generic Letter 96-
06.
So we have sort of transitioned here in
the work that's been going on from what our concern
was to a different aspect of the concern, something I
think that is a little removed from what our real
concern was to begin with. We were thinking that
condensate induced waterhammer would be the real
severe issue that needed to be addressed, and I think
what we've learned based on the work that EPRI has
done is that, no, for low pressure systems we really
don't have to be so concerned about that. It's really
the loop event, and that brings us back to USI A-1,
and I don't think we want to try to force the industry
into doing something that we didn't ask them to do
originally and really wasn't part of the generic
letter consideration. So we do have to be a little
bit sensitive to that.
Again, I'd emphasize cooling water systems
are maintained not only for in-service testing and
ASME code or other standard requirements, but also
Generic Letter 89-13 was issued in recognition of the
problems that we were seeing, reports that were made,
LERs and whatnot with degradation and vulnerabilities
that were being identified by utilities over the years
with service water systems and cooling water systems.
So we have asked utilities, and we have
done inspections to confirm that they are implementing
programs to satisfy those concerns to make sure they
know what the vulnerabilities are, what the
degradation mechanisms are.
If it's mic, they're identifying that, and
they have established programs to address that.
Obviously those degradation mechanisms are very plant
specific. It depends on the water quality, et cetera,
but the plants are responsible to know what's going on
in their system and to implement programs to maintain
the quality of the system and the integrity of the
system.
And we're confident that they are at a
point where they're doing that. We've performed
inspections to satisfy ourselves of that.
Also, we would agree with what Vaughn
Wagoner and EPRI have said. We believe that this is
of low safety significance, primarily just looking at
the numbers for LOOP plus LOCA.
But if you go beyond that, if you had a
problem with service water in containment, we've had
other evaluations, other initiatives where you look
at, well, what is the robustness of containment, how
much can it take during, for example, maybe a hydrogen
explosion, and the containments can take more
typically than what we give them credit for, which
tells us that, well, okay; you do have some margin
there to heat up containment. If you did have a break
in the service water system, in the cooling water
system, typically those are isolatable from outside
containment. I wouldn't expect that to be a problem.
So there are actions that can be taken
should the event occur, which also helps to put this
in a different kind of a risk perspective, and early
on we were hoping to be able to address it from that
perspective.
Unfortunately, it becomes such a plant
specific evaluation that it's not something that our
staff, that the Risk Assessment staff could handle on
a generic level, and so we deferred to industry and
asked that they consider risk, and that's why, partly
why, I think, Vaughn mentioned that, was because it
was requested by the staff to see if they could handle
that or deal with that more handily than we could.
That was the reason for that.
MR. SIEBER: Let me ask a simple question.
If condensation induced waterhammer is just a small
fraction of the forces that pump driven waterhammer
has, and since start-up testing for every plant that
I know, which isn't all of them, for sure, has already
tested pump driven waterhammer and all of the
deficiencies corrected, why can't the issue be
resolved just on the basis of that logic?
MR. TATUM: Well, that's certainly a
possibility and something that could be considered.
It's not something the industry has proposed, but it
is something that I think is within the realm of
possibility.
We're still reviewing the issue and trying
to see how it fits together, but it's our expectation
that for LOOP, the plants, in fact, are able to handle
those events.
MR. SIEBER: That's right.
MR. TATUM: They have shown that during
the start-up testing and whatnot.
The complication maybe that you get into
there is the combined loads and what's required by the
FSAR design basis. Would you require plants to
combine those loads somehow?
So you get into the design basis base and
FSAR requirements and being able to address that. And
it's a possibility it's something that certainly the
industry can suggest. We have discussed it, but not
really gone into detail on that.
MR. SIEBER: Thank you.
MR. TATUM: So having considered these
other factors, I'll just put up my last slide here,
which would give you our preliminary conclusions. As
I've said, we haven't completed our review. We do
have a number of open items. One has to do with air
content.
We believe that for the work that has been
done, that the proposed amount of air is conservative.
However, we're looking at differences in plant
arrangement, for example, that maybe would explain or
argue that, well, maybe the amount of air for one
arrangement versus another may not -- maybe you
wouldn't credit that much, and we just need to think
through in our evaluation the different plant
arrangements that we would expect to see and whether
or not the proposed amount of air release would be
conservative.
At least we believe it would be
conservative for the different plant arrangements. So
we're looking really at that kind of a level or that
type of a review for air.
However, for the testing that was done and
for the limited scope testing, you know, representing
basically a stagnant tube, but without the continued
flow and whatnot, we do believe rather convincingly
that it is conservative, and it may not be the right
number, but part of what we're considering is, well,
is it a conservative approach and do we believe that
it would give us confidence that if the utility used
this approach, that they would give us an answer or
come up with a load that is conservative with respect
to the waterhammer condition.
It's not just is it the right number, but
is it a conservative number, and I think the
subcommittee has pointed out very well that it may not
be the right number, probably is not the right number,
certainly not exact,b ut we're tending to look more on
whether or not it's conservative and whether or not we
can base our evaluation on the work that was done and
use that in resolving or closing out this issue for
these plants.
DR. ROSEN: Jim, I only have one remaining
residual, remaining concern, and that is that post
LOCA-LOOP emergency operating procedures are specific
enough to assure that plant staffs will isolate
faulted fan coolers if that should happen, even though
these analyses say it probably won't.
Is that something you're thinking about?
MR. TATUM: Well, it's not something --
you know, the emergency response was touched on a
little bit by EPRI, and you do get into the symptoms
based or symptoms driven response, and to the extent
the operators are able to identify the reason for the
symptom, they can address it.
But you get into real complications with
operator response and human error and human factors
and whatnot, and we really haven't gone into that
level of detail. We have not involved emergency
response people.
I don't know. Our feeling is that it's
relatively low safety significance. We don't know
that it really warrants that level of review at this
point. That's kind of where we are on that.
CHAIRMAN BONACA: I had a question. Do
you expect us to write a report on this issue? I
mean, at the end you're telling us these are
preliminary conclusions. You told us that there are
a number of open issues, and I think you have some
judgment you're making regarding conclusions, and I am
left, you know, with a question in my mind. Are we
ready to write a report of this or should we?
MR. TATUM: Well, obviously the
conclusions I'm giving you here are the staff's views
on what we've seen, our understanding of the work
that's been done and the report as it has been
presented in our review to date.
We do have, as I say, some open issues,
but we do not think that the shortcomings of the
analytical derivations or the experimentation and the
issues that have been raised by the subcommittee, we
do not believe that those shortcoming really are show
stoppers with respect to being able to use that report
and credit it for analyzing waterhammer events at
these plants.
We think that to the extent we do identify
significant issues during our evaluation, and like I
say, the air is one. We have pulse rise time, I
think. We're considering single pulse, multiple
pulse. You know, getting back to our review of the
document itself, we may find the need to impose
certain restrictions on how the report is used.
One restriction that we know we would
impose is that the report would only be used -- we
would only accept it for resolution of 96-06
waterhammer. It would be allowed for any other
application because the testing is pretty specific to
96-06 for fan cooler systems. It would not be
applicable to RHR or other systems that typically
experience waterhammer.
So we're going to be very specific on
where we allow it. It's only this limited use
application, but we think that industry has provided
sufficient argument. It's convincing, I believe, to
provide reasonable assurance to us that if the
utilities use that methodology, they can at least come
up with a value for support loads and whatnot that's
realistic, credible, and something that we can use to
resolve the issue.
CHAIRMAN BONACA: Okay.
MR. HUBBARD: This is George Hubbard.
I'd just like to reemphasize that; that I
think really the question is: is the user manual that
they will be providing to industry -- does it provide
a reasonable method for a plant to take, do plant
specific analysis, and use this methodology to
determine their waterhammer loads? Does that provide
a reasonable method for them to use and considering,
in particular, the low safety significance of this
event?
And I think, you know, if you were to
write a letter, we would be looking for the ACRS to
tell us yes or no, that the use of this user manual is
reasonable.
DR. WALLIS: Usually what happens is the
staff takes a position and we see something written
from the staff, and then we write a letter saying we
agree with the staff or whatever.
In the absence of this final statement
from the staff, you're sort of asking us to be the
staff and to write a review of the document. It's not
really our job.
MR. HUBBARD: Okay. I guess the point is,
I think, from the management perspective we're seeing
that this with maybe some limited -- being limited to
the containment fan coolers from a management
standpoint; we're seeing that this does provide a
reasonable approach, and that any restrictions we
would be putting in our safety evaluation on how they
apply it.
But basically, considering the safety
significance of this issue, I think they've got a
reasonable approach for dealing with this.
DR. ROSEN: In fact, we do have your final
conclusions on this.
DR. WALLIS: I have another question.
This document, this EPRI report, is this going to
eventually be a public document?
MR. TATUM: Yes, it is.
DR. WALLIS: So that means that in the
presentation we kept being promised improvements to
the report, and I think that the real driving force
for that is that eventually it's going to be out there
in the public. So it's got to be a convincing
document.
MR. TATUM: Well, it will be proprietary,
and there will be a non-proprietary version. We also
have editorial comments that we've found and we will
be sharing with EPRI, corrections that need to be
made. They will prepare a final version, and also, I
think, put their own corrections and also add the
additional detail that they've promised the
subcommittee.
But once they've put that final version
together, then they will also prepare the
nonproprietary version and made that submittal.
DR. WALLIS: So I think there are sort of
two issues here. One is is this a safety issue and is
this good enough to resolve the safety issue. The
other one is is this the sort of report you want to
see out in public as typical of what the NRC accepts.
They're sort of different questions.
MR. TATUM: Yeah, and as I say, I mean,
the staff really doesn't have a problem accepting the
report for the specific limited application. We would
have a problem obviously accepting it as a way to
evaluate waterhammer in general
MR. SIEBER: Maybe I can ask one more
question. Is there a list of plants that have
resolved this issue outside of the methodology of the
EPRI process?
MR. TATUM: I do have a listing of plants.
I can't tell you off the top of my head what they are,
but there have been quite a number of more plants that
have resolved it outside this process.
MR. SIEBER: Okay, and of course, there's
a list of the members of the group who would intend to
resolve it this way. If I take those two lists, does
that include all of the plants subject to the generic
letter?
MR. TATUM: All except I'd say maybe about
half a dozen.
MR. SIEBER: What happens to them? You
know, what are they doing?
MR. TATUM: Now, the half a dozen that are
left, a couple of them have submittals in house that
we're reviewing. They have used RELAP and were not
comfortable with their use of RELAP, and so we need to
take a close look at it. So those are in process.
Others that we're looking at, I think your
concern is, well, what if they wanted to use this EPRI
methodology.
MR. SIEBER: Well, that would be one
concern, or what happens after this group has spent
maybe a million bucks or whatever to do this, and then
somebody else devises some, you know, very simplistic
approach. What is the criteria by which you would
accept all of these various methods?
MR. TATUM: Well, a particular utility is
always free to propose an approach, and we obviously
are obligated to review that. And, in fact, that's
what brought us here to begin with. Utilities were
trying to make submittals on their evaluations that we
felt were just not adequate, and we asked the
questions. We would ask the same questions that we
asked in the beginning about the evaluation. What
were the assumptions and considerations that went into
it, whether or not they followed Joukowski, if they
were proposing some other approach and what was the
justification; that's what drove this group of
utilities together to form the working group and to
develop this methodology.
It wouldn't be a trivial matter for a
single utility to come in on their own and say, "Well,
we'd like to use this other approach." We'd expect
the same kind of effort and expense, I would expect,
to justify that approach.
MR. SIEBER: Okay. Thank you.
MR. TATUM: Any other questions?
(No response.)
MR. TATUM: Okay. Well, thank you very
much.
DR. KRESS: Thank you.
And I turn the floor back to you, Mr.
Chairman.
CHAIRMAN BONACA: Yeah, I think we should
postpone any further discussion to the afternoon.
DR. KRESS: Yeah, I think that's correct.
CHAIRMAN BONACA: And with that I think
we'll recess for lunch now. Well, we do have some
discretion because the two meetings we have in the
afternoon, the first two are just internal matters.
One is reconciliation of ACRS comments.
I would propose that we do that when we
reconvene, say, at 1:45, and then after that -- and we
will do the subcommittee report at 4:00 p.m., at the
conclusion of the reactor oversight process.
Okay. With that, then the meeting is
recessed until 1:45.
(Whereupon, at 12:55 p.m., the meeting was
recessed for lunch, to reconvene at 1:45 p.m., the
same day.)
. A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N
(2:30 p.m.)
CHAIRMAN BONACA: Let's resume the meeting
now. The meeting will come to order again.
And we're going to review the reactor
oversight process. We have presentations by the NRC
staff, and I'll turn the meeting to the Chairman of
the subcommittee, Jack Sieber.
MR. SIEBER: Okay. I'll be very brief.
Actually we have had four previous meetings on this
subject where we have looked at various components of
03-05 and how it fits together, and today is a review,
which is necessary for us because we have an SRN that
we need to answer, dated April 5th, 2000.
And you'll notice on the board that it's
rated A plus, which means get it done or stay here
forever, and so what I'd like to do is we will discuss
performance indicators, initial implementation,
significance determination process, and the technical
adequacy of the significance determination process to
contribute to the reactor oversight process.
And since we are going to put out a report
at this meeting some time, I would encourage members
to ask the pertinent questions that they feel are
matters of concern to them so that we can have the
advantage of the staff's response.
And with that, Mike.
MR. JOHNSON: Thank you.
My name is Michael Johnson from the
Inspection Program Branch, and I'm joined at the table
by Mark Satorius, who is the Chief of the Performance
Assessment Section, and Doug Coe, who is the Chief of
the Inspection Program Section.
And as was indicated, we are here to talk
about the reactor oversight process. I ought to
mention that also at the side table we have Don
Hickman, who is, as you are aware, our performance
indicator lead. Chris Nolan is here representing the
Office of Enforcement, and in fact, throughout the
room are a number of folks from my branch and who
serve in various capacities, and also Steve Mays from
the Office of Research.
So we've got a pretty good spectrum of
folks in the room to listen in on and possibly
contribute on the discussion of reactor oversight
process.
As was mentioned, we have had several
briefings throughout the first year of initial
implementation for the ACRS, and those briefings have
focused on areas, I think of key importance to the
ACRS in preparing for this letter writing opportunity
that you have for the Commission.
And we focused in on the important areas,
I think, that are of interest to you. We focused in
on performance indicators, significance determination
process. We went through a fairly exhaustive
presentation, I think, and tried to demonstrate for
you the use of the SDP.
We talked about in a session, I think, in
July the action matrix and tried to respond to your
questions and provide you a good overview of what we
intended to do with respect to the action matrix and
the reactor oversight process.
At our last meeting in July, we also took
the opportunity to try to forecast for you what we
were going to -- then, at that time, we were
previewing what we were going to tell the Commission,
that we ended up telling the Commission in fact on the
20th of July about the reactor oversight process.
At that time we really used some of the
high level slides that captured the results that we
documented in the Commission paper and the fact that
we, again, did, in fact, discuss with the Commission.
Those overall results, and I'll just
repeat them briefly, right now is that based on the
input that we got from internal stakeholders and
external stakeholders, based on a very, very thorough,
I think, use of self-assessment metrics and internal
feedback through a Federal Register notice and an
internal survey, reached the conclusion that the
reactor oversight process, while not perfect, does do
what we intended it to do, in that it makes steps in
the direction of improving its ability to be more risk
informed, understandable, predictable and objective,
and in fact, goes towards meeting the agency's NRC
performance goals, maintaining safety, efficiency, and
effectiveness, those goals that you're well aware of.
Having said that, we did learn lessons
throughout the first year. We tried to characterize
those lessons for you, and in fat, we had planned
actions that we described in the Commission paper, and
we talked about those planned actions in July.
And so the point that we tried to leave
with in July, and I want to start off with perhaps
today, is, again, while we know the process isn't
perfect, we believe and have told the Commission and
I think the Commission recognizes that the ROP is a
step in the right direction. It does represent an
improvement over the previous process, and we ought to
go forward and make improvements, and that's our
mantra, the mantra that we carry for the staff, with
the staff, is that we are going to continue to improve
the ROP in this next year, in fact, the year that
we're already in, the second year of implementation of
the ROP.
I ought to also mention by way of
background that in addition to, you know, talking
about the status in that last briefing, we did
something that I thought was very useful, that is, the
ACRS subcommittee did something that was very useful
for us, and that was that we went around the table,
and each of you told us, each of the subcommittee
members told us what their primary concerns were with
respect to the ROP, and we wrote those down, and we
listened to those concerns.
And they dealt with things like confusion.
There's confusion with respect to, for example, what
is meant by a green PI and how that differs from a
green inspection finding and how we treat those
consistently through use of the action matrix.
We talked about, the ACRS subcommittee
talked about the consistency of the treatment of
issues in various cornerstones, if you will. In fact,
we talked about the ALARA cornerstone, the
occupational safety cornerstone, and the ALARA SDP and
where that gets you with respect to the significance
of issues and whether or not that's equivalent when
you look at the reactor safety SDP and where you come
out with respect to that. That was an issue.
We talked about the treatment of safety
conscious work environment and all of the cross-
cutting issues and the concern on the part of the
subcommittee members at that time with respect to
those issues in the ROP.
We talked about the plant specific
thresholds for performance indicators or the fact that
we ought to be moving in the direction of plant
specific PIs or plant specific thresholds, I should
say, associated with performance indicators.
There was a concern about rewarding the
good performance in this process, and really a
concept, I think, on at least some participants' minds
that the old process, the SALP process used to provide
something in terms of incentive for licensees to
improve their performance, and in fact, the ROP, the
existing ROP that we've gone to, does not.
There was a concern late in the meeting
about the consistency of ROP implementation, the issue
being that are we, in fact, at the threshold for
documentation level at the identification of green
issues and white issues. Are we consistent among the
very regions in terms of how we implement the ROP?
So we talked about those issues. Those
were among the issues that we raised, and, in fact,
there are other issues that we're aware that the ACRS
has continued to raise and that we've continued to
take action on.
In fact, one of the things I wanted to
tell you is that as you'll hear in a few minutes we
have taken or are taking action and moving in the
direction to address many of the concerns that you've
raised in the past, and in fact, I feel very positive
with respect to the role of the ACRS in terms of
shaping the direction of the staff with respect to
improving the implementation of the --
Has the word gotten out that we're easily
swayed by flattery?
MR. JOHNSON: The flattery is almost over.
So let me --
(Laughter.)
MR. JOHNSON: The last point I would make,
and then I'll shut up and let Mark talk, is that I do
want to tell you that we are prepared today to talk at
a very high level with respect to the ROP, and we'll
touch on all of the areas that are of interest to you,
and we'll do our best to answer your questions.
I do want to tell you though that we did
not bring the cast that I would have brought if we had
the time to do the very detailed reenactment of some
of the earlier presentations that we had for the ACRS,
for example, the SDP discussions and those kinds of
things.
So I simply tell you that to say that
welcome your questions. We'll do our best to address
your questions, although I don't think the time is
going to allow us to delve into a lot of detail on any
of the issues that we've talked about in the past.
Having said that, let me turn it over to
Mark, and Mark will start off the discussion, a very
brief presentation, I might add, on lessons learned
and actions that we're going to take on the major
areas of the ROP, and then we'll be quiet and
entertain your questions.
MR. SATORIUS: Thanks, Mike.
I'm going to talk about both performance
indicators and also assessment. But like Mike
indicated, we're here to do our very best to answer
your questions as they develop and to give you a good
briefing on where we've come thus far.
I would like to point out that unlike Mike
and Doug and the majority of the folks in the
Inspection Program Branch, I'm a relatively newcomer
and been in the branch for three months. So I don't
have, I guess, the bench strength in my memory that
some of my colleagues do. So like I said --
MR. SIEBER: Which probably won't help you
here.
(Laughter.)
MR. SATORIUS: I suspected as much.
I thought I'd start on performance
indicators with just a very brief background just to
kind of frame the performance indicator issues, and
that is we put together some guidance with NEI in a
working group that we had empaneled to develop some
reporting guidance, and that was NEI 99-02, and that
first revision was then revised again based on input
from the working group, and also our stakeholders in
the spring of 2001 after the first year of initial
implementation.
The working group primarily was put
together to provide resolution on PI issues as they
developed, insights as to where problems existed with
the PIs, and also as an avenue to develop any needed
replacement PIs should it become evident that they
were necessary.
With respect to the first bullet, that was
a replacement scram PI indicator that at the onset of
the ROP and initial implementation there was an issue
involving whether we had identified the appropriate
scram performance indicator, and that was the first
performance indicator that we took on to conduct a six
month pilot.
We performed that six month pilot in the
spring of this year, came to a conclusion that the
proposed pilot PI did not contain any advantages to
the original PI, and it was the staff's view that we
would retain the original PI for use.
Due primarily to some industry senior
executives' interest in this matter, we have drafted
a letter that would address our position on how this
PI should be retained, and that letter is at the
Commission right now for their review and consultation
prior to issuance.
Once that's issued, it would be our intent
to go ahead and inform the industry via a regulatory
information summary that would indicate that we will
retain the PI that was originally put into place.
MR. SIEBER: And I guess the difference
between the original industry position and your
current position relates to whether manual scrams are
included or not.
MR. SATORIUS: That's exactly right, and
the replacement PI proposed to do away with what's
termed unintended consequences that develop as a
result of manual scrams counting. There were some
positions that there would be unintended consequences
as a result of potentially an operator hesitating or
possibly not inserting the manual scram, and the
replacement scram we concluded to a large extent did
not remove the potential for unintended consequences.
There were unintended consequences that were developed
as a result of that new replacement PI, and that was
the conclusion.
MR. SIEBER: Well, as a former operator,
I think that when you count automatics and manual
scrams just as a scram, the operator doesn't care one
way or another.
MR. SATORIUS: We got that.
MR. SIEBER: The difference is if you
don't count manual scrams and the operator is more
likely to manually scram the plant where the automatic
set both takes it off.
So I don't know whether that's good or
bad, but that's what the original argument was.
MR. SATORIUS: Some of the feedback we got
from pros was that the operators would do the right
thing irrespective of whether they were counted or
not.
MR. SIEBER: I think that's true.
MR. SATORIUS: And we got that indication
from a lot of operations managers in direct contact
with various licensee staffs during the first year of
initial implementation.
MR. SIEBER: With regard to that indicator
though I think that one thing that I note is that the
threshold between green and white is such that it's
not particular risk significant. Okay? You know, a
plant is designed to deal with an automatic or a
manual scram so that you actually have -- before it
becomes risk significant to any appreciable extent,
you have to get into the more serious thresholds.
Another indicator that's like that is the
loss of heat sink. For example, you have to lose heat
sink to get to a red indicator three times a day every
day for three years, and boy, if your plant is in that
bad a shape, you know, I would say that indicator
doesn't tell me much
MR. JOHNSON: Yeah, and, Don, you're
welcome to jump in at this point or you can wait if
you want to a more opportune moment.
MR. HICKMAN: Okay.
MR. JOHNSON: But let me just say a couple
of words before you do, Don.
One clarification is that we're going to
count -- both indicators count manual scrams. If you
look at the primary change in the replacement PI, you
won't find the word "scram" at all. You'll find a
shutdown, and then we've gone though the effort to try
to define a shutdown that is a rapid shutdown like a
scram without saying the word "scram."
And if you look at --
MR. SATORIUS: And it introduces a 15
minute period in there, in other words, a rapid
shutdown within 15 minutes, and I think our view was
when all is said and done, the potential for
unintended consequences associated with that 15
minutes is probably more than the operator -- and like
you say, you haven't been an operator. In the heat of
battle in the control room, he's going to reach up and
do the right thing.
MR. JOHNSON: So I guess the point I was
making was that we're going to count manual scrams.
We think it's important to count manual scrams.
Now, your point is well taken with respect
to the thresholds. Typically what we find is if a
plant is going to begin to have problems with scrams,
we'll see performance problems showing up that are
reflected in other indicators, and in fact, for
example, there's a special inspection going on right
now where the plant had a scram and then had some
other complications.
And so we'll do an event follow-up type
inspection to look into that issue. So we're not --
that takes me into a good point, and that is to say
that the performance indicators are a part of the
indication that we have about the overall performance
of the plan, but it's not the sole indication.
MR. SATORIUS: I'll go now to just the
unplanned power change PI. The original PI read the
number of unplanned power changes in reactor power
greater than 20 percent within 7,000 critical hours,
and there were a number of questions within our
working group on that.
The industry and NEI had proposed a
different unplanned power change PI that they intended
to bring to the table to be piloted at some time this
summer or fall.
We had also developed one ourselves and
had entertained whether it might be useful to pilot
both of them at the same time. Through our working
group NEI has taken those, our proposed, their
proposed and, I guess, they're framing them or they're
collecting data and seeing as to where those would all
fall out, and they haven't gotten back to us with
their proposed unplanned power change PI.
We've gone ahead and developed ours and
would propose that at the next meeting that we have
with them, to go ahead and pilot that at some point in
the fall and early winter.
The last issue involves improving the
safety system unavailability PI. We've established a
separate working group to work on that specifically.
Part of the problem that we have with this one is the
fault exposure hours associated with an unknown as to
when the initiating event was.
In other words, for example, you may have
an 18 month surveillance where the previous time you
might have had an opportunity to identify that you had
a problem would have been 18 months ago, and it's --
using the standard T over two gives you nine months of
fault exposure time, and on any diesel that's going to
put you into rad.
And the consistency issue that we have
here, and we discussed this with the subcommittee
before, was a lot of times if you look at this demand
failure, in other words, during the surveillance and
you plug that into an SDP because of the chance for
operator successes, because of the chance or
opportunities for off-site power to be restored,
you'll oftentimes get a green SDP finding on a red PI
finding, and we recognize that as a consistency
problem.
Industry also has identified that, and
this safety system unavailability group is working to
develop a pilot PI that we would intend to begin
piloting. I believe it's in January, isn't it, Don?
Yes, January.
In the interim though, recognizing that
there are some challenges, especially from a
consistency standpoint, we're going to take interim
steps where for any demand failure, such as the
example I just gave, the diesel, that we would, in
fact, use the SDP to determine the actual significance
because it more closely ties it to risk significance
as opposed to the counting of hours and the use of T
over two, although T over 2 is pretty consistent from
PRA and also in the ASP analysis.
But that is an interim step that we plan
on taking until we can get -- and that interim step
would continue throughout the piloting of the PI and
until we would be able to develop a PI that would more
accurately measure this unavailability issue.
MR. SIEBER: If you continue to use T over
two in the SDP process, would you --
MR. SATORIUS: Yes.
MR. SIEBER: -- not come up with the same
result that you come up with out of the PI?
MR. SATORIUS: No, you don't, and the
reason is that the SDP takes a look at, and then Doug
probably can talk to this better than I, but the SDP
takes a look at other matters outside the simple
counting of hours. It looks at the ease or the
ability of an operator to take compensatory action and
how likely that is to be successful. It takes a look
at, for example, if you were to have a diesel that
would fail 12 hours into its full power run. If you
were to have an actual scenario with a loss of off-
site power, the chances for the recovery of off-site
power within 12 hours are relatively high.
So you take that, coupled with the
potential for operators to take -- it gives you a
better scenario and the SDP more accurately
categorizes it or addresses it from a risk
perspective.
MR. JOHNSON: This is just another one of
those advertisements that I'll try to throw in. This,
I think is one of the most substantial improvements to
the ROP in the area of performance indicators that
goes a long ways towards addressing a number of the
concerns and the recommendations of the ACRS in the
past in that I think at the end of the day what we
will have in this revised SSU is something that is
clearer, that does provide consistency in the use of
the definition of unavailability.
We've got all of the right folks in this
working group. We're talking to the PRA folks. We've
got Research participating. We've got the maintenance
rule folks participating. We've got a representative
from INPO/WANO.
And so we'll have a standard definition of
unavailability that will be used for this performance
indicator. And so when you apply this performance
indicator, again, you'll have consistency. It'll be
easier for the operators, and it will get us to the
right result.
And when you go to run through an SDP, a
finding that would reflect an unavailability for the
PI, you'll end up at the same spot. So that scratches
a lot of itches, and so we think that's a very good
change.
CHAIRMAN BONACA: Yeah, sine you have
performance indicators and you're moving to other
issues, I would like to just ask a question regarding,
again, one issue that has been brought up by this
committee many times and our Chairman who is not here
has raised this issue and I somewhat am representing
his thoughts, too.
The fact that this PI is a known plant
specific; they are generic. Okay? And you know, we
went through an exercise yesterday, just some chatting
about it, and for example, take the high pressure
injection system, which is a significant system in all
power plants because it's an element of LOCA
mitigation.
And you know, I can think of specifically
a group of early C plants out there, like St. Lucie
and Calvert Cliffs, known things, that have two high
head pumps in that system, 50 gpm each, that provide
very little floor, high head. Therefore, those plants
are vulnerable more than others to small break LOCAs
because the pressure may hang up there, and you may
not be able to add water in it.
I mean, that's a known thing technically,
and in fact, the PRAs reflect the importance of that
scenario in the risk, as well as the importance of
that system for the plant. Okay?
They also happen to be pretty limited in
auxiliary feedwater. So, therefore, you know, if you
look at the PRA, it shows a very significant
contribution, and you know, so here I have some very
specific insight on the safety aspects of that plant
tied to that system.
I also have the latest generation of
Westinghouse plants like CBER. With five I had
injection pumps that provide, I believe, 375 gpm each,
at the 2,300 psi. Two of them are charging pumps.
Two of them are self-injection pumps. One of them is
a back-up. They're interchangeable.
Tremendous capability up there, and
clearly that shows in that the fact that small break
LOCA is not a dominant sequence in those plants. You
know, these are the specifics.
Now, so if I really looked at getting
insights from PRAs and from risk regarding these two
things, I would treat the self-injection very
differently for the St. Lucie type plant than I would
call for this Westinghouse type plant. They're
telling me very different things.
I would set probably the thresholds in
different locations.
I would also even put a multiplier maybe
on the C type plant, given that the system is so
fundamental, important for the plant, and yet if I
look at the PIs, the way they are defined right now,
they don't discriminate at all in that sense.
I mean we discussed this issue to death
already, and they're not plant specific, and by the
way, when I look at the question, number one, from the
Commission that says if the PIs provide meaningful
insight into aspects of plant operation that are
important to safety, they don't provide insight at
all.
And yet the PRAs are providing that
insight right now that there is this strength for the
Westinghouse type plants, and there is this weakness.
Let me call it that way.
And I wanted to provide this example
simply because I think it's poignant in indicating how
much more one could get from existing risk information
from these plants that is not present in the current
PIs.
DR. ROSEN: let me before you answer that,
Mark, take the same point from a slightly different
angle. What we really want to measure in these
indicators is the overall risk of plant operation.
MR. SATORIUS: Yes, I would agree.
DR. ROSEN: And that's a hard job, but to
pick out a few safety systems, high pressure
injection, aux. feedwater, on-site power, et cetera,
and say those are what we're going to measure makes
them surrogates for this much more robust measure,
which is a measure of the overall risk of plant
operation. They're a stand-in for something we really
want to measure, which is the overall risk.
So Mario correctly points out that the
real thing to base this on is the PRAs because it
would get at the plant specific issues directly, and
I say to follow that on that some plants are, in fact,
doing that internally. They have to participate in
this process obviously, but some plants have risk
monitors or risk indices that are based on their
configuration risk management programs, which take in
all of that stuff.
More and more plants -- the plant I came
from had one, but more and more plants now have them
and are using them to good benefit, controlling their
configuration risk.
I suggest that long term now and in your
thinking moving towards replacing individual system
unavailability measures with a more integrated measure
based on the PRA gets to the thing we all want to
measure, which is the overall risk of plant operation.
MR. JOHNSON: Yeah, let me try to talk to
that if I can. I think what I hear and the direction
that we're headed in is synced up. I stopped short in
my discussion of what we're doing with respect to the
safety system unavailability PI to talk about the
strongest piece of that enhancement that we are
considering with the unavailability PIs, and that is
the addition of reliability indicators that are a
fallout of the risk based performance indicator
program that Research worked on.
And when you have those performance
indicators, well, what we'll do is we'll set plant
specific thresholds, plant specific thresholds, and so
what we'll look at is not a standard unavailability
percentage or a standard liability percentage, but
we'll look at a percentage that is based on a standard
delta CDF, based on the change in reliability or
change in unavailability.
And we're talking about doing that in the
near term. We're already working on the user -- we've
had a number of conversations with Research. They're
tapped into this focus group that is working on
unavailability improvement. So we're headed in that
direction in the near term, and that, I think,
scratches that itch.
With respect to this longer term use of
integrated indicator, I'll tell you right now the PIs
that we have are surrogates. They are indicative.
We've always said they would be indicative, and that's
where we were when we started this program, and that's
as far as we've been able to come.
Although if you look down the list of the
things that we're asking for and the things that are
on that risk based performance indicator task,
development task that Research has briefed you on in
the past, I know one of those things is an integrated
indicator.
And so in the longer term, I think in the
longer term there is some direction towards seeing if,
in fact, there is a capability to add something like
that.
Now, I think there's some philosophical
things that we need to get beyond before we adopt
something like that. I think right now we're more
comfortable given the limitations, given where we are
in the development. There is more comfort with this
indicative approach, this selection of a few systems
that are surrogates for the overall state of the
performance of the plan, but that's certainly on our
developmental longer range, the use of what it is
you're suggesting.
So, I mean, I think this is a good area
where we're actually moving in the direction that ACRS
would indicate is a good direction for us with respect
to the performance indicators.
DR. SHACK: Let me take a slightly
different approach that's different than my
colleagues. I mean, most of my colleagues look at
this as sort of a gigantic risk meter, that you know,
we clock in every once in a while, and I like that
approach because it sort of gives you kind of a
unifying thing.
Whereas I look at some of these
performance indicators as surrogates for ways to
measure things like safety, culture, and that, you
know, even though my Westinghouse four loop plant
could take lots of unavailability in the high pressure
injection system, it's not a good sign that you don't
keep the system up and operating.
And to my mind many of these indicators,
you know, if I base them on risk, nothing will turn
out to be safety significant. You know, everything is
unimportant until the accident happens, and there's
some measure of attitude here that is kept in by
looking at something that measures performance rather
than risk.
But that leads to sort of fundamental
problems and inconsistency because the significance
determination process is risk informed, and yet some
of the other PIs I can look at as measuring some other
kind of parameter, and that leads me to logical
inconsistencies, although I'm almost happier logically
inconsistent than I am purely risk informed at the
moment.
(Laughter.)
DR. POWERS: I don't understand that.
What's the conceptual difference between the two?
CHAIRMAN BONACA: Yeah, I really think
that, no, the fact is that could be something that you
could construe that if, in fact, you put the threshold
so close to an expected performance, that you step
over the bound because you're sloppy about it, right?
So you're measuring culture.
But you're not because you're putting the
threshold far enough that you capture only certain
cases where, you know, just you capture maybe one or
two out of 100. So the measure is --
DR. SHACK: Well, some of these plants,
and I set them on a consistent delta CDF for all
plants, some plants would have enormous tolerance, and
some plants would have much narrower ones.
CHAIRMAN BONACA: The fact you have a
question that says do they provide meaningful insights
into aspects of plant operations that are important to
safety, and you know, again, I don't think that you
get insights on the culture from the PI because, I
mean, you will see variations of that. I mean,
otherwise you would see some kind of grading
variation.
But certainly you do not get insights that
you have from existing risk assessment tools regarding
through the PIs. I mean, you don't get those because
they don't differentiate on what is important for the
plant and set certain criteria on what is important
for the plant.
In fact, I dare say that if you had a full
understanding of that through PRAs, you may have
different sets of PIs for different plants. I mean,
you could have that.
DR. ROSEN: This is the old structuralist
versus rationalist approach, and I'll come down in
between, and I'll be a rationalist with structural
tendencies.
Really having a fully integrated risk
unavailability or integrated risk monitor would be a
very good thing, and I think you should work to it,
but that's not throwing out the structural aspects,
the points that Bill was making, that Shack was
making.
CHAIRMAN BONACA: No, I'm not throwing
them out either.
DR. ROSEN: Because we have a risk
informed program here where we use risk to the extent
we can, but we have to be thinking about the fact that
the safety culture at the plant is a leading indicator
of what these things are.
I mean the safety culture goes downhill
before you ever see these numbers start to change.
CHAIRMAN BONACA: Yeah, and in fact, you
know, the inspectors have pointed out that if the
thresholds are too far, they don't count enough to, in
fact, identify trends like they should. So they
stated that actually the thresholds are allowed.
DR. SHACK: But I think risk information,
I think, will move you even further away from or at
least that's my concern. I don't know the --
DR. ROSEN: I don't think so. I think
risk basically would move you further over, but risk
informing swings you back. It brings you back to the
middle where it says we have to take into account the
safety culture.
And I suggest that it's a timing
difference, that the perfect plant has a great safety
culture and very low numbers on its indicators, but
when it begins to degrade, it degrades first in its
culture and then the indicators begin to follow it,
will begin to follow it because, in fact, the plant's
hardware starts to reflect the degraded maintenance of
whatever else.
MR. SIEBER: But the emergence of a
declining safety culture, which is a cross-cutting
issue even though it shows up as indicators, the
indicators respond, demonstrate perhaps a cross-
cutting issue is involved because you've already built
in a lot of latent defects.
And I think that that is part of Bill's
concern. You know, if you had ten -- if I had safety
injection pumps and five diesel generators for a
single unit, you would say that's pretty safe.
If you have a really lousy safety culture,
probably half of the stuff doesn't work. So I would
just assume you look at individual competent declines.
DR. POWERS: I wonder if you could speak
to those performance indicators that usually aren't
associated with any risk metrics, thinking of things
like the safeguard performance indicators and whatnot,
and in particular, I would appreciate it if you would
speak to it in the context of providing -- whether
those performance indicators provide meaningful
insight and aspects of plant operation that are
important to safety.
MR. JOHNSON: Tom, why don't I let you
start and then I'll add?
MR. HICKMAN: Did you want me to start or
you said you were going to start?
Okay. The --
DR. POWERS: This question was so easy he
asked his chauffeur to answer it.
(Laughter.)
MR. HICKMAN: Right. The indicators in
the other strategic performance areas are difficult to
associate directly with risk, as you know, and so
what --
DR. POWERS: But I'm not asking you to
associate them with risk. I'm asking you to associate
them with safety.
MR. HICKMAN: Okay. Well, I guess you
could say the same thing interchangeably there.
They're associated with performance in
those areas which have some sort of impact upon the
safety at the plant, but it's hard to tie any kind of
number with that, and that's the reason that those
indicators don't have red response -- red bands.
What we've done in those areas is to use
basically expert opinion to determine expert panel
type of approach to determine when indicator values
are to have reached a level where the NRC ought to
step in and take action.
In establishing those thresholds, as I
said, we did that with an expert panel, we confirmed
those based upon the results of the pilot program, the
six month pilot program and also the results of the
initial historical data that was provided by all
licensees prior to initial implementation.
And what we discovered was that the expert
panel process worked very well, that, in fact, we had
established levels that seemed to be very appropriate,
first of all, at the green and white level for
identifying outliers. That seemed to work very well.
As far as the higher color categories,
colored bands, as I say, we just have the yellow.
There's nor ed for those. Again, that's based upon
the expert panel opinion that those are the levels
where we need to take increased action to prevent any
further decline.
In some of those areas, of course,
licensees have to maintain those programs, and so it's
not acceptable to say, you know, the program is broken
in that regard. What we have to do is make the
program work.
So at the yellow band level, the NRC will
step in and take whatever action is necessary, whether
it requires orders or anything, whatever it takes to
make sure that the program works.
That's the process behind the development
of those thresholds.
DR. POWERS: I think what I'm really
asking you, if you could give me a thumbnail sketch of
the rationale the experts use to arrive at the
conclusion that there was some level where the NRC had
to take increased action to make the program work.
MR. HICKMAN: I'm not the expert in those
areas, but I can tell you briefly what I know about
what they did. It was based primarily on their
experience in the emergency preparedness area, for
example.
They had a lot of experience with the
number of drills that were being performed by
licensees and the amount of participation that was
involved in those.
Actually in that cornerstone or that --
yeah, that cornerstone, we achieved, we think, quite
a success because it caused licensees to do exactly
what we wanted them to do, to run more drills and put
more people in it.
And the thresholds were established based
upon their experience, and they turned out to be very
good, very close.
With regard to the other cornerstones, the
performance indicators in the public radiation safety,
for example, are not likely to be exceeded. The
industry has performed pretty well in those areas, and
it would have to be a series of serious breakdowns at
the plant for them to be exceeded, and those are what
are used in the public radiation safety area.
The safeguards area, we still have some
concerns about that, and we're still working on that,
but the security performance index has worked well and
has had some success in causing licensees to fix
system that they had not paid much attention to in the
past, although we're still working on that. There's
still a lot of concerns about the security equipment
performance index.
And there's likewise concern about the
other two indicators in that cornerstone.
DR. POWERS: Can you give us a thumbnail
sketch of what your concerns are?
MR. HICKMAN: In the safeguards?
DR. POWERS: Right.
MR. HICKMAN: One concern, I think, was
that the security equipment performance index was
probably not worded quite right. It claims to monitor
the unavailability of the security equipment, and in
fact, we don't really do that. We look at the
compensatory hours, guard postings in compensation for
degraded equipment.
And so it doesn't really do what the words
seem to imply that it does because we use a surrogate.
We posted guard hours as opposed to actual unavailable
hours for the equipment.
That was done because it's easy for
licensees to collect that data. It's more difficult
to keep track of the actual unavailable hours.
DR. POWERS: I guess one of the questions
that the licensee can legitimately ask is, "Gee, I've
discovered I've got a piece of equipment," right?
Pieces of equipment break. He discovers it Friday
afternoon. He does not have a replacement part.
He takes compensatory action for it.
Everybody agrees that it's compensatory action, and
yet he has -- he gets a degradation of this while he's
waiting for a weekend to get over, and then on Monday
he can call and get the replacement part that he
wants.
Why should that be a degraded action? It
seems to me that's a victory for him. I mean, he
should get a gold star put next to his name on that
one.
MR. HICKMAN: We've heard that type of
comment, actually maybe even a little more intrusive
into licensee performance, the case where that happens
and they have the part, but they don't want to have to
call the tech. in on the weekend and pay them extra
money to fix it when they can fix it on Monday, and
that issue has been raised by licensees a number of
times.
I guess the answer to that is that the
threshold is set high enough to accommodate some of
that type of activity. Plus, there are exemptions in
the indicator. There is a blanket exemption for
preventive maintenance.
So we're encouraging them to fix the
problems before they break and you won't count those
at all. But there is allowance. The threshold is at
eight percent. So there's a certain amount of that
kind of problem that can occur, and it still won't
cross the threshold.
MR. JOHNSON: Yeah. I mean, I think John
has given the answer that I wish I would have been
able to give right off the cuff, but that's why I rely
on Don.
Two points that Don made that are really
key. One is if you talk to NEI and ask them about
performance indicators that are working well, they'll
point to the EP performance indicators and they'll
talk in some cases about this security equipment
performance index, and it's because of what Don said,
and it is causing licensees to take actions in areas
to address performance problems that really ought to
be addressed.
With respect to EP, in fact, if you have
problems, adverse trends in your performance, if
you're not, in fact -- and you want to improve that
performance, if you want to improve your participation
and improve your drill performance, what do you do?
You run more drill sand you perform better at those
drills.
And that's what we want with respect to
performance indicators, and in fact, we've found
instances where plants were not performing as they
should have been performing with respect to EP.
Just to take you back on it, the second
point I'll make is remember the development. The
development was we said what are the cornerstones;
what's the important information that we need about
those cornerstones; and so what can we get from
performance indicators; what can we not get from
performance indicators? So we need to do baseline
inspection, and so remember performance indicators are
only a piece.
But there is a nexus. In fact, the
performance indicators, we believe, do have face
validity in that they do tie back to giving us
insights on those key attributes that we need to
measure in each of the cornerstone areas.
And so as Don points out, we need to do
more with both, with the security equipment
performance index. With 7355 rulemaking, we know
we're going to need to go back and look at those
safeguards performance indicators, to improve them, to
make them more consistent conceivably with how that
rulemaking comes out. So we know we've got some work
to do.
But those performance indicators also give
us good insights in an indicative kind of way with
respect to performance of the plant in those
cornerstones.
MR. SIEBER: Every indicator refers to one
of the seven cornerstones in the framework. I presume
unplanned power change is initiating event
cornerstone.
MR. SATORIUS: Yes, it is.
MR. SIEBER: Is an unplanned power change
risk significant at all?
We used to change power to reduce
radiation dose so we could have containment entry. Is
that a risk?
MR. HICKMAN: Do you want me to answer
that?
MR. SATORIUS: Yeah, go ahead.
MR. JOHNSON: Why don't you take that?
MR. HICKMAN: No. In fact, we say in the
guidance document, in 99-02 that unplanned power
changes in themselves are not risk significant, but
under other circumstances, they could lead to risk
significant events.
The reason that the staff is interested in
unplanned power changes is because historically we
have noted a relationship between plants that are
constantly going up and down in power and the plants
that in previous assessment process we identified as
poor performers or watch list plants or declining
trend plants.
And we've seen the plants that tend to run
steady state are also safer plants.
What we're counting, that indicators, not
just any power change, but it has to exceed 20
percent. So for smaller power changes, we don't pay
attention to those, but we're counting those that
exceed 20 percent of full power.
MR. SATORIUS: I might add to that the
scram PI falls into that same category, that it's
traditionally a PI that under previous assessment --
MR. SIEBER: It's not risk significant.
MR. SATORIUS: Right, but it matches up in
the past that plants that are scramming at lot, the
same as plants that are up and down a lot in the past
assessment process had tended to be poor performers.
MR. HICKMAN: And one other important
thing there is. The threshold is high on that
indicator. We understand that there's going to be
some of that, and we allow for that.
MR. SIEBER: Well, I worked at a plant
once that did load following, believe it or not. Are
they exempt from this PI?
MR. HICKMAN: Yes. There are a number of
exemptions, and that's one.
MR. SIEBER: Okay.
MR. JOHNSON: You'll find that discussion
in 007. We do a pretty good job of laying out why we
chose, for example, the unplanned power changes, and
it goes to what Don said.
MR. SATORIUS: I'm going to go ahead and
go to the next slide, and didn't have a lot that we
had intended to discuss at least in this presentation
on assessment. The first bullet, I think, ties into
the discussion we had just had under PIs, and that has
to do with, you know, consistent responses to PIs and
inspection issues and our endeavors to assure that the
information we're gathering through the PI process is
consistent with the system that we're using to
evaluate safety and risk significance with inspection
findings., especially the disconnect or the potential
for the disconnect where you may have, because of
fault exposure hours, had a PI that goes red and at
the same time if it had been an inspection finding and
there was an SDP associated with it, it would more
than likely be green.
So we've identified that. We're working
towards that through the safety system unavailability
working group, and as I had mentioned, in the interim
we intend on for demand failures within the PI arena
to use an SDP to analyze that risk significance and
apply a color.
The second thing we wanted to discuss real
quickly was an issue involving no color findings..
When we briefed the subcommittee, I believe it might
have been in May. We went through our rationale for
no color findings, and my recollection, there was
quite a bit of dialogue because for us to explain to
the subcommittee our bases for no color findings and
where did they fall, are they in between green and
white, are they less than green, and we've kind of
concluded based somewhat on our interaction with the
subcommittee at that time and also with some
interaction that we've had, I guess, primarily with
some other offices within the headquarters and also
with the regions that it just confuses matters.
MR. SIEBER: It certainly does.
MR. SATORIUS: You know, I've heard
anecdotally that no color to some folks made no sense,
and for the guy walking down the street, you ask him
and if I tell you I have four color and something
called no color, where would you plug that in?
So we concluded that the best approach
here would be to just call these matters green and go
on, and so that's the direction we're headed on that.
MR. SIEBER: Well, that's one element of
at least the public confusion that the color system
has, you know. You have green, white, yellow, red,
and then you have a different color, which I think is
gray.
MR. SATORIUS: It is gray. It is gray if
you go to the Web site.
MR. SIEBER: If you didn't inspect them at
all, and there's a pink or magenta color that says I
inspected it, but didn't have any findings.
And so when you look at this you need to,
as my computer has it, 256 colors to be able to figure
out what's going on.
MR. SATORIUS: And we recognize that, and
it's going to require some procedural changes because
in the past by colorizing an inspection finding, that
suggested it passed through an ADP, and these no color
findings are traditionally issues that may fall within
traditional enforcement or do not fit within an SDP,
and we need to change our guidance to reflect that.
But we think that the better view here is
just to call them green.
MR. SIEBER: Good.
MR. SATORIUS: It makes sense.
MR. SIEBER: I think another element of
potential public confusion is that people generally
associate green with good, whereas green is not good.
It's bad because now you've actually found something
that has to go into the corrective action system.
DR. WALLIS: Green one is good.
MR. SIEBER: I think that the purple,
magenta, pink is the best.
DR. POWERS: I think there's some
advantages to being color blind because the more
appropriate thing is that these no color findings are
within the licensee response band, and I mean, that's
the definition, and that's what you intend, and
everything else seems to make sense to me.
MR. JOHNSON: That's exactly right.
That's what we're doing, is we're saying those are
licensee response band findings.
I can't not react to your green is good or
whatever. You know, with respect to a performance
indicator, as Graham is pointing out --
MR. SIEBER: Green is good.
MR. JOHNSON: -- green is okay. If you
have green, if you're in the green band with respect
to scrams --
MR. SIEBER: But we're talking about
findings here.
MR. JOHNSON: -- that's a -- but if we're
talking about findings and we're talking about
everything that we find that is a green needs to go
into licensee's corrective action program, and so
there is that sort of difference in the explanation
that we've tried to be careful to make, and we
continue to have to live with based on the scheme that
we've set up.
MR. SATORIUS: Okay. Doug, you're up
next.
MR. COE: SDP is a first up. The SDP has,
I think, been acknowledged by man as one of the more
significant differences in the new program versus the
old program, and it was born of a need to address the
concerns of our stakeholders that we be more
consistent and more objective across the nation,
across the different regions and across time with our
assessments of performance.
And so given that we have seven
cornerstones, some of which are amenable to a risk
kind of evaluation and some are not, the overriding
objective for the SDP is one of objectivity and
consistency.
In the implementation in the first year of
the SDP processes, we have had some issues come up
that we know that we need to deal with, and we are
dealing with them. The first here, as indicated, is
that we need to do a better job of being more clear
about the assumptions that we are using to exercise
the SDP logic, be that in the risk informed SDPs or in
the others.
In any case, it was always our intent that
our basis for our decisions be clear, more clear than
they had been in the past, and so we do need to do a
better job of in some cases documenting the
assumptions that we use.
The other thing that has become a
significant issue for us is timeliness. A recent
audit that was performed based on the 20 issues that
have been brought to our headquarters panel between
April of 2000 and February of this year indicated that
the average time from the exit meeting to the final
panel results was about 98 days, and as you're aware,
I'm sure, the Commission has pretty much mandated that
we set a goal for ourselves of 90 days absolute.
That's not on the average. That's not a median.
That's absolute.
So we have a good deal of work to do to
improve the timeliness aspect, which certainly is our
intent because it needs to support the assessment
process, which is conducted on essentially a quarterly
basis, if not a continuous basis in some respects.
DR. POWERS: So then your objectivity
criterion for this, you've done an internal assessment
and an external assessment. You're getting yourselves
real high scores on that objectivity?
MR. COE: Well, I would say that relative
to the previous program, yes.
DR. POWERS: Yeah, relative to the
previous program, right.
MR. COE: Relative to the previous
program, I think, clearly the use of risk metrics, for
one, sa a means of achieving greater consistency from
plant to plant, from region to region, and from time
period to time period is certainly giving us a better
and more visible yardstick of measurement than when we
had in the past, which was essentially a more
subjective SALP criteria process.
And the point that was made earlier is a
valid one, that the non-risk informed cornerstones,
the ones that are not amenable to the use of risk
analysis directly, we have to make judgments regarding
the responsiveness or the level of engagement that we
would expect to have and seek to measure that or to
grade that in a way that remains consistent with the
other cornerstones, the risk informed cornerstones.
So from the standpoint of objectivity, I
think being clear about our decision logic and
employing the same decision logic from issue to issue
as we encounter across the regions and across time, I
would have to say -- and I think we said this in SECY
01-114 -- that we have achieved a greater objectivity.
We also have continuing challenges in the
risk informed arena to continue to improve the Phase
2 notebooks which are the primary implementing tool
that is in the hands of the inspectors and is intended
to provide them with the ability to improve their
understanding of the risk drivers at their plants on
a plant specific basis and to make an initial
screening kind of assessment of the potential risk
significance of the findings that they may come up
with.
We are continuing to --
DR. POWERS: Do all plants have Phase 2
notebooks?
MR. COE: We have -- all plants will have
Phase 2 notebooks issued in Rev. 0 form, we think, by
the end of September. We have the last three that
Brookhaven completed for us. We've reviewed, and it
remains for them to complete revising them in
accordance with our comments, delivering them to us so
that we can put them out via letter and then to the
Web page.
DR. POWERS: I'm not sure what phase zero
or whatever it is you called the format means.
MR. JOHNSON: You referred to a Rev. 0.
MR. COE: Revision 0 is the first official
issuance of the Phase 2 notebooks for each plant or
each plant type, and you know, we expect that there
will be further revisions. We know that there will be
because as we have issued Rev. 0 and have gone out to
do benchmarking against the plant's own internal PRA
analysis, we are finding that we need to have some
changes made in order for the notebook to better
represent that plant's design and operation.
DR. POWERS: At what point will you be
able to say all plants have these sheets that have
been benchmarked?
MR. COE: Well, we've only been able to
complete about eight benchmarking trips, I believe,
this year, fiscal year, but we are budgeted to
continue that process next year.
The short answer to your question is I
think it will take us into probably fiscal '03 to
present all plants at the current rate.
DR. POWERS: Is it a case of if you had
twice the budget you could do it twice as fast, or is
this nine women can't make a child in one month sort
of situation?
MR. COE: Certainly I've been told that
having a greater amount of money would improve -- we
could accelerate the rate at which we do these
benchmarking trips. However, you would eventually be
limited by the staffing. Okay? We have to have the
right people out there.
Typically we invite and get the senior
reactor analyst in each region to participate in
these. We think that's valuable for them as well, and
I think that's pretty much been the case for the ones
that we've done so far.
So, yes, we could accelerate it with
greater funding, but there would be a natural limit.
I'm not sure exactly what that limit would be.
MR. SIEBER: It's my understanding that
you don't have an operable SPAR model for every unit.
Is that true?
MR. COE: SPAR models are also under
development, and I don't know exactly where we stand,
but the recent, most recent development program
estimates given the budget and the funding that have
been asked for, but maybe not entirely approved yet,
would have us completing all of the SPAR models out
some time in fiscal '04, I believe.
MR. JOHNSON: And that's SPAR-3, I think.
I was actually looking for Steve Mays and
he's not around. Tom, do you have?
MR. BOYCE: Forty-three SPAR models have
been developed so far. Seventy are supposed to be
completed by the end of FY '02.
MR. JOHNSON: You've got to go to the mic.
And give you name and then --
MR. BOYCE: Tom Boyce in the Inspection
Program Branch.
I'm going to try and relate the status
that Research really should be telling you, but Steve
Mays did just depart, and the most recent data that
I've heard is that 43 SPAR models have been completed
out or 70 total. The remaining will be completed in
FY '02. They also have to go through a benchmarking
process, and only on the order of five have been
benchmarked up to this point.
They're doing them in conjunction with the
SDP Phase 2 notebooks where possible using the SRAs in
the regions.
DR. POWERS: No one has ever --
MR. BOYCE: That process takes time.
MR. SIEBER: Well, let me follow up my
thought. The last number I heard was 37, but that was
a couple of months ago. So you've made progress, but
if you lack a functional SPAR model and you don't have
a Phase 2 notebook, how do you do significance
determination? Are you relying on the licensee?
MR. COE: In many cases we will ask the
licensee for an analysis and we will review that
analysis, but I would hasten to add that, you know,
the Phase 2 notebooks are out there as high level
representations. They lack the details of the SPAR
models.
MR. SIEBER: Well, it's screening, right?
The purpose is screening and to knock out the
nonsignificant stuff at the local level.
MR. COE: It's screening, but even in the
final revision, even after we've done the
benchmarking, you know, the intent is that the
notebooks provide essentially an opening assessment,
an initial opening assessment of what we believe the
risk significance might be for a finding.
That can certainly be modified as better
information is made available to us, but in many cases
we're finding that the inputs that we make to the
licensee's models are being reflected properly in the
notebooks, in the use of the SDP Phase 2 level
process.
MR. JOHNSON: I guess I get a little
nervous about our answers that we're giving that
research ought to be more appropriately given. Keep
in mind that research does ASP analyses on any plant,
every plant based on the SPAR-2 model, and we're
talking about the SPAR-3 model, and --
MR. SIEBER: Well, that goes back to the
senior reactor analyst, the SPAR-3, right?
MR. JOHNSON: So I guess the point I want
to make is don't -- if you have continuing questions
on where we are with respect to SPAR models, and some
of the agencies' priorities are changing based on
direction from the Commission, as you're probably well
aware, with respect to that, I'd ask that you hear
from research and not my group on the final answer.
MR. SIEBER: I guess the bottom line of my
last two questions is if you don't have a Phase 2
notebook, you don't have a Phase 3 SPAR model, then
you may be in a weak position with regard to dealing
with the licensee because you're relying on the
licensee's information
MR. COE: I think one of the advantages of
what we're doing with the use of risk analysis in the
SDP though is to avoid this issue of my model is
better than your model.
MR. SIEBER: Right.
MR. COE: What we're trying to do, and
it's been my observation over the past six or seven
years that I've been engaged in the risk analysis
business that the primary impediment to furthering the
use of risk analysis in this agency, and many others
perhaps, is one of communication, and if nothing else,
the SDP process should be helping us open up the
methodologies, the analytics, the assumptions of a
risk analysis and make them more apparent and more
visible to a wider number of stakeholders, principally
those who are closer to the plant, to the physical
realities, to the physical design, to the physical
operation of a plant who can either, therefore, accept
or challenge those assumptions, that logic that goes
into this analysis, which produces a result that we
act upon.
And so I think that although we're in our
initial stages of improving our ability to communicate
with each other and with our licensees and with our
public, we are progressing in that direction. At
least at the moment, I think we are, and I do hope to
avoid the situation that you've just articulated.
MR. SIEBER: Well, one of the interesting
things is to my knowledge, there's no regulation that
requires a licensee to have a current PRA.
MR. COE: That's true.
MR. SIEBER: And so it's possible you
could run into a situation where you don't have the
information and the licensee doesn't have the
information, and the process to me becomes pretty
arbitrary.
And while you're in the process of coming
up with a decision as to what color a particular
finding is through SDP, it becomes invisible to the
public as to how you got there.
DR. ROSEN: You see, Jack, that's the
point of having a good SPAR model or good Phase 2
notebooks. For the case where the licensee is very
weak in his own PRA development, I think that's a very
useful and necessary thing for the staff to have.
On the other end of the spectrum though,
with a licensee with a very robust PRA that's highly
documented and very open, why does the staff even need
these Phase 2 notebooks and SDPs?
The right answer, it seems to me is when
a plant like that has an incident or a finding, you go
to their PRA staff, sit down, and at a clean table
discuss how the risk analysis would evaluate the
circumstances and come to some kind of joint
conclusion that both sides can support.
I've seen that process work at the place
I used to work at, and I think that's superior to your
model versus my model. There's only one model. It's
either right or wrong, and both people have access to
it.
MR. SIEBER: I think for public
confidence --
DR. POWERS: PRA is just not at that stage
yet, and there can be two, three, four dozens of
models of a plant which are equally right. PRA is
just not an exact science yet.
DR. ROSEN: I didn't say it was an exact
science. I just said that having one model that both
sides, the regulator and the licensee, can agree is
the best shot at what's right and evaluating a given
set of circumstances using that model is, it seems to
me, the way to go rather than one side having some
kind of little simplified model and the other side an
advanced model.
DR. KRESS: I think that there are
regulatory uses for these things that you wouldn't
want the staff to have to run to the licensee every
time they wanted to do some sort of risk
determination. So I think there's good reasons for
the staff to have their own models.
MR. SIEBER: I think so, too, public
confidence.
DR. POWERS: Just the capability that the
staff has when they have their own model is what's
worth the investment.
MR. JOHNSON: Yeah, we're fully supportive
of the agency's continued SPAR-3 development, and in
fact, even though I don't speak for our office with
respect to the priority and certainly not the research
in terms of the agency's priority on SPAR, we
recognize that it's the way we want to go because we
don't want to be overly reliant on licensees.
As Doug indicated, and in fact, I missed
some of the conversation, but I wanted to make one
last point, and that is, you know, there are two
opportunities for us to reconcile the significance of
findings for the SDP. One is through the SDP process
itself in our Phase 2 and Phase 3 analysis, and then
we provide that information in terms of preliminary
analysis to the licensee, and the licensee runs their
model, and we reconcile where we ought to be based on
the input that we get from the licensee.
But we have a second opportunity, and that
is through the use of the ASP program, and in fact,
research checks each of the analyses that we do where
we have a greater than -- in fact, a greater than
green finding. They'll compare what they come out
with respect to the ASP, as part of the ASP program,
of course, they do the analysis using our models, and
then they share with the licensee and they get
licensee input.
And so we reconcile those differences and
look for holes or areas with respect to the Phase 2
work sheets or the process that we have that may be
causing those holes.
So there are a couple of opportunities and
a number of exchanges with us and licensees, but I do
not want you to leave here with the perspective that
we feel like we're overly reliant on licensee models
because that's just not the case.
Having said that though, we do think that
SPAR-3 development ought t continue.
DR. WALLIS: Could you explain to me what
a Phase 2 notebook is? Is this Phase 2 notebook the
paper document with all kinds of check marks, or is it
a computer into which you can put various information
and reach conclusions based on some software?
MR. COE: No, it's the former.
DR. WALLIS: And eventually it should
hopefully be something like the latter.
MR. COE: There's thought being given to
creating a user interface to the SPAR models that look
very similar to, you know, the way that the analysis
was represented in the Phase 2 notebooks.
One of my principal concerns from the very
start has been that it's often too easy for inspectors
in the field to pass their findings off to
specialists, risk analysts, and if they don't engage
themselves in the process in some form of risk
analysis, they tend not to understand the results of
the specialists.
And so one of the distinct advantages of
a Level 2-like approach for risk analysis is that it
helps the inspectors understand both the benefits and
the limitations of a risk analysis, and it gives them
the opportunity to explore sensitivities of various
assumptions that they are in control of, and rather
than let an analyst be in control of the assumptions
and the logic that tend to drive the results, this
puts this information and the ability to manipulate
those assumptions and that logic in the hands of the
people who will then, you know, presumably have an
opportunity to accept greater ownership of the end
result.
So, I mean, in fact, one of the questions
that the committee might wish to consider in terms of
your letter would be whether or not a three phase kind
of approach for the risk informed SDP is worth our
continuing development. In other words, you know, one
of the options we had was to simply have all of our
inspection findings sent off to an army of risk
analysts.
That didn't necessarily help the inspector
better understand or guide their future inspection
activities, nor did it allow for a greater population
of individuals who were closest to the plant to
participate in achieving either acceptance or being
able to challenge the various assumptions that were
being used.
DR. POWERS: It seems to me that one of
your biggest headaches that I would worry about in the
future -- I don't know that you have it -- I would
worry about in the future is the frustration of the
inspector seeing things and not seeing anything come
about it.
I mean, right now already he's in the
position of finding things that don't even go into --
well, I guess they allow him to write on a report now,
but they don't seem to go anywhere, and you get this
problem of what good am I doing here, the thing I have
to do.
And similarly, sending things off to an
army of analysts only makes that problem worse, it
seems to me. I mean I think you've got a real morale
problem brewing among your inspectors if they continue
to get isolated as a cog in this system that you've
set up.
MR. COE: Exactly, and I feel the same
way. My emphasis has been from the start, has been to
give the inspector the tools that they could use to
find the most significant issues that might exist at
any given site.
Now, admittedly, using the risk method
that we're using for reactor safety issues, you could
arguably say that we've set the bar higher because
there is a definite objective bar that has to be met,
and the attendant basis that we have to provide to our
stakeholders to say that we've met that limit or that
threshold to carry an issue forward into a greater
than very low significance manner, apply it in that
manner.
But in addition to setting that bar
higher, we've given the inspectors the tools to help
them see how issues might get to that point, and in
the ultimate analysis, I believe that that's risk
informing our inspectors.
So, again, I think if you have thoughts on
that, you know, because there are multiple ways of
pursuing a risk based estimate.
DR. POWERS: Well, I mean, anything that
leads to the inspectors understanding that they are
essential and that, in fact, their role has been
upgraded, not downgraded, is to my mind the way to go.
MR. COE: Precisely, and I would agree.
Next I would just offer that we are
continuing development work in the areas of shutdown
SDP, which is kind of at a Phase 1 screening checklist
level at the moment, trying to develop some Phase 2
kind of sequence based tools.
Containment which has always been kind of
a place holder in our current program based on some
work that research has done for us, and we need to
carry that work forward and produce a more usable
tool, and in the fire area, of course, which we've
talked about at some length before, and we all
recognize the nature of fire analysis, risk analysis,
is probably one of the more difficult for us to
tackle.
DR. POWERS: I would like to pursue fire
just a little bit.
Go ahead, Jack.
MR. SIEBER: Well, I was just going to
comment on that. When I look at the SDP process for
fire, it is so simplified that it appears to me to be
pretty subjective, to say the least. I mean, you've
got a choice of three. It's really bad; it's not too
bad; or it pretty good.
DR. POWERS: That's the part of the SDP
that I just do not understand at all, is that we have
this rather mysterious set of numbers that I actually
think I know where they came from. I'd love to hear
somebody defend them, but be that as it may, how I
select which number to use seems to be totally up to
whether I'm a buddy with a guy that I'm inspecting or
not.
MR. COE: Well, I would certainly say that
we have acknowledged the need to be more specific
about how to characterize the various classes of the
parameters that we use as inputs to that fire
analysis. One very important one that tends to
influence it a lot, influence the outcome a lot is the
performance of the fire brigade, and we've
acknowledged that there's a need to clarify that
guidance so that it's more consistent.
And I can't explain exactly where each of
the numbers came from, but what I can tell you is at
a high level, the fire protection SDP as reflected in
Appendix F of our guidance document 06-09 is
essentially attempting to have about the same level of
detail that the reactor safety Phase 2 SDP has tried
to hit, and in fact, it's linked to the reactor safety
Phase 2 SDP.
But what we're really trying to do across
the board, across all of these risk informed SDPs is
to de-emphasize the numerics and emphasize further the
choices that historically and traditionally have been
made by risk analysts and to put the thinking, the
judgment of choosing those various assumptions more
directly into the hands of the inspector.
DR. POWERS: How do I decide that
something is low, middle degradation or high
degradation? I mean, explain to me how I pick that
number other than the fact that this guy's a good
buddy of mine. I know he's doing the right thing
versus this guy is a penny-pinching, cost cutting
dude. I'm sure that he will not do the right thing.
MR. COE: Well, first of all, I do have a
greater confidence in our inspection staff that they
wouldn't lose their objectivity in that manner, but
that doesn't mean that we can't improve that
guidance.
You're absolutely right. I mean, there is
a need to be better and more consistent, I should say,
in terms of making sure that one inspector will judge
a particular condition that they see in the same
fashion as any other inspector in another region or
across time.
DR. POWERS: If that's your objective,
that's a good one.
MR. COE: It is.
MR. SIEBER: I think there ought to be
another one, too, that whatever the outcome is,
whatever the color of the finding is ought to reflect
true risk significance potential for fire because that
is a prominent actor in reactor safety.
DR. POWERS: I mean, your priority on
fires has gone way up based on the IPEEE insights to
my mind.
Now, let's go to the numbers in the SDP.
I assume they come out of five. That's my guess.
MR. COE: And now you've just gone beyond
my level of expertise.
MR. JOHNSON: We, in fact -- Matt, I can't
remember what briefing it was, which of the briefings
it was where we talked specifically about --
DR. POWERS: The one I was not at.
MR. JOHNSON: Yeah, it was the one you
weren't at, but I guess what I would offer is if you
do have some detailed questions, Dana, that we don't
have the right folks where to deal with that. At that
earlier briefing we had the branch chief and the
section chief and we had the guy who implements the
SDP for us now, and in fact, we had the guy who
developed the fire protection SDP, and those are
really the guys who ought to be answering your
detailed questions, I think.
CHAIRMAN BONACA: I had a question.
DR. POWERS: The question is very simple,
and it explicitly addresses what the Commission has
asked. It's asked do these have any relationship to
safety, and so the question is very simple. What do
the numbers coming out of five have to do with fire
risk. Why those numbers and not some other numbers?
MR. COE: Well, I can tell you that one of
the significant issues that's being dealt with right
now is the issue of fire initiation frequency because
that does vary, and that does tend to be a significant
driver.
And from the standpoint, you know, of
what does this mean and how does it relate to safety,
you know, again, we're still using the same risk
metric, and it all boils down to whether or not the
assumptions and the logic that you're using to arrive
at your metric -- how well that comports to the actual
plant design, the deficiencies that you found, and the
way that that plant is operated.
So, again, doing a better job of defining
how to use the fire initiation frequencies and what
values are most appropriate for various situations,
how we define the levels of degradation for fire
barriers, for the fire brigade performance, and making
that more consistent from inspector to inspector is
really our intent.
And what we believe is that the closer we
get to establishing that those inputs most accurately
reflect the plant's condition gives us greater and
greater confidence over time that that risk output,
that metric is reflective on a comparative basis from
issue to issue across different plants so that we can
grade our inspection responses accordingly.
DR. POWERS: Are you thinking not
necessarily in the next three years or four years, but
maybe longer term, and I'm not going to define what
longer term is, but it's beyond 2003. I'll tell you
that -- to have the equivalent of a SPAR for fire or
other external events?
MR. COE: The current SPAR development
plan speaks of external initiating event models, but
doesn't, under the current budget forecasts, doesn't
really begin to really get started with that until I
believe it's fiscal '03 or '04.
DR. POWERS: Well, I mean, that's pretty
soon. I mean, that's more encouraging than I would
have thought.
MR. JOHNSON: Again, you're asking a
question that really is better answered by Research,
I think.
DR. POWERS: You guys are on the hook.
You can't get out of it that easy.
CHAIRMAN BONACA: I have a question on a
separate issue. It's more for information. I can't
remember.
If you have risk informed PI, say,
something that we discussed before safety injection,
and it goes from your green to, say, white or yellow,
do you perform a significance determination evaluation
of that?
MR. JOHNSON: No, we don't.
CHAIRMAN BONACA: But if you did, that
would blend the criticism we are making of not being
plant specific because what you would do, you would
then use PRA to evaluate the significance of that, and
therefore you'd absorb the blend of criticism that we
are leveling on the process.
MR. JOHNSON: I actually answered too
quickly. What I should have said was -- I think we're
rushing to correct my answer -- what I really should
have said was that in general the PI program is set
with thresholds, and crossing those thresholds alone
is enough to enter the action matrix. So if you have
a white, then you do what the action matrix would
require.
But there are a number of cases where
nothing would prohibit, for example, an inspector from
running a performance issue that happens to be also
reflected in the PI through the SDP to determine the
significance, and we've had a number of instances like
that where we have -- in fact, we're working on one
right now that is a PI reporting issue that would have
if the licensee reported it in a certain way that PI
would be red, but we know that when we run that issue
through the SDP, it's actually a white issue,
potentially a green issue, and it deals with this
issue of false exposure for demand failure that Mark
talked about.
So in fact, probably the more accurate
answer to your question is that, yes, inspectors can
run a performance issue, any performance issue,
through the SDP to determine its significance.
MR. SIEBER: Well, if you get into a
degraded performance indicator, that calls for
additional inspection. The additional inspection can
or may not result in findings. Findings are run
through SDP. So you end up having a risk input to
everything that start out as a performance issue.
CHAIRMAN BONACA: No, I'm focusing only on
the PI. What it means is that if you said, okay, I
have a PI and now it's gone from green to yellow, say,
and I'm going to run it through the significance
determination process, which essentially relies on a
plant specific PRA. Then all of the criticism we have
been leveling on the process will be eliminated
because you will have an opportunity to evaluate after
the fact, okay, whether or not it's significant, and
you would treat it like anything else that you treat
by significance.
MR. JOHNSON: I would say that the safety
system unavailability working group that we've
empaneled acknowledges that and recognizes the
problems that we have with fault exposure hours not
being plant specific, being more generic in nature,
the PI itself being generic in nature.
And we are working towards developing an
unavailability PI that I think I indicated earlier we
would want to pilot starting in January.
But in the interim, we've done, I think
exactly what you've just described, and that is for
those PIs, safety system unavailability PIs where
there's a demand failure, we would run it through the
SDP, and we would tie it more closely to actual risk,
rather than just using a generic counting of the
hours, so to speak.
CHAIRMAN BONACA: Absolutely. I mean, at
the beginning you use the reference system as you have
right now, and then you filter it through a process
where a plant specific PRA is being used to make a
judgment on the significance of that.
That would, in my judgment, you know,
address all the concerns we have raised.
MR. JOHNSON: And that's a short-term fix,
right. That's a short-term fix that we're going to
implement on -- we're hoping to implement by the first
of January. So we're pleased to hear that ACRS is
pleased with the approach we're talking.
MR. SIEBER: Well, okay. I guess that --
DR. POWERS: He's really gotten smooth
over the years.
MR. SIEBER: I guess that the ultimate
action that the staff can take is through enforcement,
and to get to the enforcement process, you have to
have inspections and findings. And it's the PIs that
generate potentially the inspection process.
So to me, you know, at least in that sense
it's tied together on more or less of a risk basis.
CHAIRMAN BONACA: It's indirectly. I
think what they're proposing here to do would make it
very direct in that, you know, from the beginning you
don't have a true risk based determination in the
calling (phonetic) of a PI, but you have a
significance determination process allows you to get
there, and so that would -- and that would not really
complicate the system.
MR. COE: No, that's right, and I'm not
sure we would want to have a system where the changing
of the color of a PI would then generate --
CHAIRMAN BONACA: I understand.
MR. COE: -- further regulatory
aggravation by having an inspection. We would want
the PI ultimately to do it all for us. That would be
plant specific enough that it would do it all for us.
It wouldn't require additional inspection because
that would be more resource on us, as well as
licensees.
CHAIRMAN BONACA: In that case then you
would consider, for example, saying, okay, it looks as
if this licensee is going from green to white. Let's
evaluate through the SDP if it is true, and then you
would have this assessment that would allow you to
keep a green, for example, if, in fact, the
significance of it was very low.
MR. COE: That's correct.
CHAIRMAN BONACA: Okay. So you were not
stepping in, and you would have the basis for keeping
it in the green, which would be based on plant
specifics.
MR. COE: That's correct.
MR. SIEBER: I would be nervous if you
attempted to, even if they were plant specific, set PI
thresholds that would skip over inspection process to
arrive at some kind of enforcement action. That's
different than what your chart that you gave us.
MR. JOHNSON: But let me -- well, I was
almost going to try to see if I could say what it was
you would be saying in terms of describing the
enforcement program and see if maybe I can clarify it
a little bit.
When we set it up, we have PIs and
inspections that are independent inputs, and each of
those are enough to get you across threshold into --
MR. SIEBER: SDP.
MR. JOHNSON: -- some assessment act --
beyond SDP, into some assessment action.
MR. SIEBER: Okay.
MR. JOHNSON: Including enforcement if
there's a violation associated with a finding, but
depending really on the action you make, you could get
an order or, you know, some other enforcement, things
that are typically considered enforcement actions.
And so as I think Mark was trying to
describe, we don't have the situation or we don't want
to set up the situation where you have a PI and then
you've got to go out and do some inspection and then
run that through the SDP and now you have what you
need to enter the action matrix.
The PIs and the inspections, each are
independent input and sufficient inputs into the
action matrix. What we're trying to deal with is this
problem that we have with unavailability PIs and the
fact that they're not, as we set them up now, risk
informed.
So in those specific cases where we have
these large blocks of exposure, that it would be
better to run those through the SDP because that risk
informs those. That takes the leap in the short term
to get us where we're trying to go.
CHAIRMAN BONACA: So rather than having
the pain of adjusting them all up front, which would
be a very big challenge, you really have a process by
which in the few cases where you have a step-down
performance potentially, you do evaluate through this
significance determination process --
MR. COE: That's true.
CHAIRMAN BONACA: -- and make the call.
MR. COE: Yes.
DR. ROSEN: In your earlier spirited
defense of the adequacy of the safeguards and
emergency preparedness indicators, you said something
like we're pleased that we've seen licensees take
actions based on these indicators to improve
performance in those areas, in a sense basically
rating the indicator by whether there was a response
by the licensee to it.
MR. COE: Backing into the answer, so to
speak.
DR. ROSEN: Yeah, backing into the answer,
and that's sort of been troubling me and gnawing at
me. I'm not quite sure what the issue is, what's
bothering me, but I think it goes back to the question
the Commission asked us, which is are these indicators
providing meaningful insights into aspects of plant
operation that are important to safety.
And we have to write to the Commission
something about that, and your answer is, well, we
don't know about that. The licensees sure are doing
something.
I can't quite connect those things.
MR. JOHNSON: Can I try to -- I think that
was my statement actually.
MR. COE: No, I think it was Don's, but go
ahead. You can defend it.
MR. JOHNSON: Don is the person who
amplified it. I probably said it in the wrong way.
What I meant to say was that with respect
to, for example, the emergency preparedness
performance indicators, we have found instances since
the ROP based on these performance indicators where,
for example licensees were, perhaps performing well
with respect to drills, but only a small percentage of
the responders were participating in the drills.
And based on these performance indicators,
they provided broader training to all of the likely
responders, and in addition, measured the performance
of those responders through this drill participation,
this drill performance indicator and the combination
of those two have resulted in improved performance in
areas that we think are important with respect to the
emergency preparedness area.
So what I said, I think, was maybe that
the licensees are improved -- if they want to improve
their performance, they run more drills, and so, in
fact, they've done that, but the point I was trying to
make was in areas where we think performance needed to
be improved based on what we believe is important with
respect to the cornerstone, we've seen licensee
performance. We've seen these performance indicators
indicate performance problems, and we've seen
licensees take action to address those performance
problems in areas that are important.
Hopefully that better clarifies what I
meant to say.
DR. ROSEN: It does, and I think what I
have to do is make the hard link between if the
licensee performs better in the safeguards area, then
that is an aspect of plant operation that's important
to safety, ergo, we are safer.
I mean, that's not something this program
can do for me. I have to have that external from your
finding. You tell me the licensee is performing
better in the safeguards area or in the emergency
preparedness area, and therefore, the plant is safer.
It's not as direct a measure as in the
mitigating systems area. It takes another piece of
information outside of the finding that comes out of
this program, if I'm expressing myself correctly.
MR. JOHNSON: I understand.
DR. ROSEN: You have to have this article
of faith first, and then you can draw that conclusion.
MR. JOHNSON: It's certainly not as easy
in the non-reactor safety cornerstones, particularly
the EP -- no, particularly the physical protection
cornerstone. It's not as easy to make that tie, if
you will.
MR. COE: But the common framework has
been that each cornerstone has been described as
having several key attributes, and the words "key
attributes" are not -- there's a definite set of
attributes as we've spelled out in SECY 007, and each
of the cornerstones has those attributes spelled out,
and each of those attributes is assessed in some
fashion, either through the performance indicator
program or through inspection findings or maintenance
rule inspections, PI&R inspections, et cetera.
And so across all cornerstones, there's
that same common basis. So your hard link is really
the adequacy with which you feel the staff has
identified the key attributes of each cornerstone and
has appropriately linked those key attributes to some
method of measurement, either PIs or inspection.
DR. KRESS: I think his problem is how to
quantify those key attributes in terms of their impact
on actual risk for safety.
MR. COE: I understand that's the problem.
DR. KRESS: Ones in one cornerstone may
have much smaller impact than ones in an attribute in
another cornerstone.
DR. ROSEN: How do you weight the
cornerstones?
DR. KRESS: And how do you weight the two,
I think, is his issue, his problem.
MR. COE: Okay. If we're ready to move
beyond SDP at this point we can go to inspections and
the challenges that we faced in the inspection. The
conduct and documentation of inspections has been one
of defining in a consistent manner what our threshold
is for documentation.
The standards are articulated in our
guidance document 0610, and we're continuing to work
on improving that in terms of how we document them and
at what threshold we document inspection findings.
We have the maintenance rule inspection
procedure, which during the first year of
implementation was felt to be -- we felt we could
improve its risk and performance focus, and so we've
engaged in pilot inspections, and we are rewriting the
inspection procedure and engaging in the pilot
inspections to test it out.
We expect that those will be ready for --
the new inspection procedure will be ready for
issuance in the next inspection cycle starting on
January.
DR. WALLIS: I'm sorry. I didn't
understand the first bullet at all. You don't mean
thresholds in the documentation. You mean
documentation of thresholds or documentation of
determinations of something?
I don't understand what you mean by
documentation.
MR. COE: The issue here is at what
threshold does the inspector document a finding. In
some part this is based on whether the finding is
deemed to be minor, in which case if it's deemed to be
minor against a set of criteria that we've tried to
provide, then the inspector does not document it at
all.
DR. WALLIS: So this word "threshold" here
has nothing to do with all the other thresholds we've
been talking about.
MR. COE: That's correct. It's a
documentation threshold. That is, at what threshold
does the inspector actually document their findings
and observations?
And because the definition of minor isn't
as precise as some of our other definitions, there's
been some variability there. We're trying to improve
that.
DR. WALLIS: So these thresholds, I mean,
you could say they're consistent. If it's white, you
have to document it, and you could relate it to the
other thresholds.
MR. COE: Yes. Well, there's no question
about findings that are green or white or yellow or
red. We document those. Okay?
The question comes in many cases as to,
you know, whether your finding -- if your finding is
minor, then you don't document it at all.
DR. WALLIS: This is sort of the no color
threshold.
MR. COE: Well, and then there's the
question of no color findings, which we've addressed
as we've indicated earlier. That was originally an
issue as well.
The no color findings were documented.
There wasn't any question about that, but how they
were documented, to what extent they were documented.
In other words, one of our objectives is to try to
reduce the bulk of the inspection report and to more
properly focus it on issues of greater significance.
So you'll see our inspection reports are
smaller in volume, and we try to be more focused and
we try to cut out a lot of the filler or not filler
necessarily, but the information that might have
historically been included in order to get to the more
significant issues.
The next point is licensee self-
assessments. We're considering that. We're starting
to think about that. I think we have to really think
carefully. We've only had a year's worth of
experience, but we certainly are beginning to think
about how to apply licensee self-assessment programs
within the ROP framework.
And finally PI&R inspection frequency went
to biennial from an annual. However, the number of
inspection hours annualized only dropped by about 25
percent because we added a few more hours in between
the biennial team inspections, which were about 250
hours now. We've allowed for about 60 hours of
inspection on specific issues.
And this was to try to reduce somewhat the
burden on the licensee by giving them a team
inspection once every two years rather than once every
year and also to allow the staff to probe, the
inspection staff to probe into areas that were
specific to PI&R concerns in between the two -- in
between the team inspections.
So that's a summary of some of the major
insights that we've gained in our first year, and at
this point I guess we'll be happy to answer any
follow-up questions.
DR. POWERS: I have a question. I'm
intrigued to know what your response is to those
plants that were, I think, SALP-1 plants in the past,
got relatively little inspection, and suddenly find
themselves being inspected quite a little bit more
under this new system and yelp about that.
What is the stock response to them?
MR. JOHNSON: I'll start, I guess. I
don't know that we have sort of a response that we've
had a lot of success with, to be honest. I mean --
DR. POWERS: I didn't say it was
successful.
MR. JOHNSON: And to be honest, there
haven't been a lot of licensees who have raised that
particular concern, although the industry in general
would say -- has, in fact, looked at where we came out
with respect to resources in general and does expect
that we continue to look for efficiencies when we go
forward.
And there are, it's true, there are plants
that were SALP-1 and, in fact, so they are getting
more inspection under the baseline.
One of the things that was interesting
with respect to the response to the Federal Register
notice from licensees, and we had generally NEI
writing in, but we had some individual licensees
writing in, and it dealt with -- it deals with the
perception of burden.
And while there are licensees who, I
think, in fact, get more inspections, there are a
whole bunch more licensees who think that the burden
is more appropriate in that they're not having to
react to the impact of inspections, that is, findings,
a lot of findings at a very low level that tend to
distract and cause licensees to expend their effort.
So I think when I talk about it, I talk
about not inspection knowledge, but I talk about the
burden of the program, and I think there's a wide
acceptance to this fact that the burden with respect
to the ROP is more right size given the significance
of the issues and what we've been able to do through
the SDP and other things.
That's sort of what I try to do to answer
that question
MR. COE: And I would only add that the
good performers get good outcomes in terms of our
assessment process still. Okay? And the extent of
inspection that they get, although it's more
normalized across all of the plants is one of the
burdens that we all share in achieving this public
confidence, one of our strategic goals.
DR. POWERS: Well, it seems to me that one
of the challenges that you face in getting public
confidence in the system is that when they look at
this system versus the old system with respect to just
inspection -- and I really liked your answer, by the
way, on look at the total thing and the burden -- but
when they look at just inspection, they say, "Yes, the
NRC has created a system. They inspect the good
performers more. That means they're inspecting the
bad performers less."
MR. SIEBER: That's right.
DR. POWERS: And I think that's a
challenge, and I really liked your answer from the
total burden is that you're putting the weight really
where it does the most good as opposed to just being
out there inspecting. I like that answer.
MR. SIEBER: Well, I'm not exactly sure
that I agree with that whole statement because no
matter whether you get a violation under the old
system where you had to write an answer back, it still
ended up in your corrective action program, and even
non-sited violations end up in the same place and
green findings end up in the same place. Everything
ends up in your corrective action system
And so the burden that the licensee has
regarding how he has to deal with all of these issues
is totally dependent on the deficiencies that are in
the plant, whether you find them or the licensee finds
them.
What does change is the licensee's
inspection fee, as a good licensee's hours went up, so
he pays more money, and a lesser performing licensee
ends up getting a fee reduction, which to me is
something the chief financial officer sees.
DR. POWERS: It would be interesting to
see the stats on that. I agree with you that the good
performers get a fee up.
MR. SIEBER: Right, and more inspection
hours.
DR. POWERS: But I'm willing to bet if
this system is working right that the bad performers
didn't see any reduction in fee.
MR. SIEBER: Well, inspections.
DR. POWERS: And fees for inspections.
MR. SIEBER: Inspection hours.
DR. POWERS: But in total, what they're
saying is it's not fair to look just at inspection
hours.
MR. SIEBER: But that's what you get
billed on, and as long as you aren't getting civil
penalties, that's the monetary --
MR. SATORIUS: But if I could add, I think
one of Mike's points also was the fact that to go
beyond just fee billing because arguably the old SALP-
1, the current program is a good performer, and the
SALP-3, the current, isn't an acceptable performer.
They're going to have more expenses with entering
things into their corrective action. They're going to
have more issues.
MR. SIEBER: That's right.
MR. SATORIUS: They're going to have more
staff hours that they're going to spend to resolve
these issues arguably than the good performer who has
a more robust corrective action system and has better
maintenance, has less issues to resolve.
MR. SIEBER: And that was my first
statement, is you're going to pay for those whether
you find them or the licensee finds them.
MR. JOHNSON: Yeah. I guess the other
point I would make is don't forget that the reason
sort of the outcry a couple of years ago, two and a
half years ago, whenever it was, that got us on this
path revising the oversight process was -- and it
didn't relate to inspection hours or fees. It related
to predictability. It related to burden. It related
to objectivity or really subjectivity being central to
the process.
And those are the things where I think
this current process offers relief that licensees --
that make them think that this is a better process.
Now, we've got challenges. The point
about -- you know, David Lochbaum still says that we
don't spend enough attention on plants with
significant performance problems. That's his
criticism of the ROP.
You know, he's looking at it from the
other perspective. When you get an IP-2 or you get a
plant that's having -- that ends up in the degraded
cornerstone column of the action matrix, he wants us
to do more than we're doing today.
So the people who fall on the other side
of the spectrum, that's the other piece of the story,
I guess.
MR. SIEBER: Well, the objectives that
were laid out by the commission, which appears in the
first couple of pages of your assessment document
which just came out, I'm pretty well convinced that
you are on the way to hitting all of them. But I
picture this process as going on for another five
years at a minimum where you can say, "Yeah, I have
all of these bases covered," and so you're just on the
doorstep of the edit (phonetic), in my view.
Would you disagree with that?
MR. JOHNSON: Not at all, not at all.
CHAIRMAN BONACA: Actually, I mean, I
think there's more even distribution of resources is
a better approach. I mean, there used to be before
the fact that they were presumed good performers that
continue to be presumed good performers because they
didn't look enough. When they looked hard, they find
they were not anymore.
So you know, that is a problem, and I
think today with a more even distribution of
resources, that's not going to happen as easily.
MR. SIEBER: Any other questions or
comments?
MR. JOHNSON: Just one last comment, if I
can. I really was serious when I suggested that we
benefit from these exchanges, and we do need the help
of the ACRS to the extent the ACRS is willing to weigh
in with respect to the SSU development work that we're
going to do, to look at the piloting in January and
going forward.
So if there is an opportunity and if the
ACRS is willing, we'd look forward to opportunities to
continue to interface and get your input.
MR. SIEBER: I think that's appropriate.
If there are no other questions, Mr.
Chairman, I'll turn the meeting to you.
CHAIRMAN BONACA: Thank you. Thank you
very much.
At this point I think we will, first of
all, go off the record. We don't need a transcriber
anymore.
(Whereupon, at 4:22 p.m., the meeting was
adjourned.)
Page Last Reviewed/Updated Monday, August 15, 2016