Materials and Metallurgy and Reliability and Probabilistic Assessment - April 27, 2000
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
***
MATERIALS AND METALLURGY AND RELIABILITY
AND PROBABILISTIC ASSESSMENT
U.S. NRC
TWFN 2B3
11545 Rockville Pike
Rockville, MD 20852-2738
Thursday, April 27, 2000
The subcommittees met, pursuant to notice, at 1:00
p.m.
MEMBERS PRESENT:
GEORGE APOSTOLAKIS, ACRS, Chairman
WILLIAM SHACK, ACRS, Chairman
MARIO BONACA, ACRS, Member
JOHN BARTON, ACRS, Member
TOM KRESS, ACRS, Member
ROBERT SEALE, ACRS, Member
JOHN SIEBER, ACRS, Member
GRAHAM WALLIS, ACRS, Member. P R O C E E D I N G S
[1:00 p.m.]
DR. APOSTOLAKIS: The meeting will now come to
order. This is a joint meeting of the ACRS Subcommittees on
Materials and Metallurgy and on Reliability and Probablistic
Assessment. I am Dr. George Apostolakis, Chairman of the
Reliability and PRA Subcommittee. Dr. William Shack is
Chairman of the Materials and Metallurgy Subcommittee.
The other ACRS members in attendance are Mario
Bonaca, John Barton, Tom Kress, Robert Seale, John Sieber,
and Graham Wallis.
The purpose of this meeting is for the
subcommittees to review a draft Commission paper concerning
options for potential revisions to the pressurized thermal
shock rule acceptance criteria. The subcommittees will
gather information, analyze the relevant issues and facts,
and formulate proposed positions and actions as appropriate
for deliberate by the full committee.
Mr. Noel Dudley is the Cognizant ACRS Staff
Engineer for this meeting. The rules for participation for
today's meeting have been announced as part of the notice of
this meeting previously published in the Federal Register on
April 5, 2000. A transcript of this meeting is being kept,
and will be made available, as stated in the Federal
Register notes. It is requested that speakers first
identify themselves and speak with sufficient clarity and
volume so that they can be readily heard.
We have received no written comments or requests
for time to make oral statements from members of the public.
At our March 16, 2000 subcommittee meeting, the
staff introduced the different regulatory approaches it was
considering in developing the draft Commission paper. At
that meeting, the staff had not decided which of the
approaches it would recommend to the Commission. We will
now proceed with the meeting, and I call upon Mr. Mark
Cunningham, Chief of the PRA branch of the Office of Nuclear
Regulatory Research to begin.
It's a friendly crowd, Mark, here.
MR. CUNNINGHAM: I hope so. With me today is Ed
Hackett, who's the Acting Chief of the Materials Engineering
Branch and the Office of Research. Ed's branch has the
overall responsibility for orchestrating this revision to
the technical basis to the PTS rule.
Right off the bat, I'll apologize because I just
noted that we're presenting to the Materials and Metallurgy
Subcommittee and not a joint subcommittee meeting, so my
apologies, and it won't happen again.
By way of overview, we'd like to discuss this
afternoon the draft Commission paper which we provided last
week, I guess, to the committee. Talk about several things
with respect to this paper. One, the purpose of it; talk a
little bit about what's included in the paper on the PTS
screening criteria as it currently exists in the rule; talk
about three sets of information that are more recent than
when the rule was established that are relevant to possible
changes; what's going on in materials research; what's going
on in terms of the Commission guidance with respect to use
of PRA; and more information on what's our understanding of
severe accident phenomenology. The paper then has four
options for modifying the screening, potential screening
criteria in the rule, and we're going to lay out what those
options are.
Basically I think we'd like to solicit comment
from the committee or the two subcommittees on the options
that we've laid out in terms of the completeness of the set
of options, if you will. Is there an option that we hadn't
thought of that might be appropriate, or is there some
clarifications we could make on the existing options, and if
there's any opinions from the subcommittees on what option
they see is most appropriate.
We'd probably need to finish the day with some
discussion of what you'd like to hear about from us at the
full committee meeting, and I think we mentioned this at the
last subcommittee meeting, that we would ask for a letter on
this draft paper that you have.
I should note that in parallel with meeting with
the committee and seeking your opinions on this, we're
continuing to meet with the NRR staff and with the Office of
General Counsel's staff to look at these options and assure
that they're technically on target and legally on target.
The basic issue that we've got here to discuss is
how are we going to modify a rule that establishes an
adequate protection rule but has probablistic underpinnings
to it, and if we revisiting the probablistic underpinnings,
what does that mean to how we might be able to change this
rule.
DR. APOSTOLAKIS: I read it also in the -- how do
we refer to this? Policy issue. The document.
MR. CUNNINGHAM: Yes, the draft commission paper.
DR. APOSTOLAKIS: The draft commission paper.
What exactly is this article of protection thing? What does
it do to this issue that it wouldn't do -- that wouldn't be
there if it was not an article of protection issue?
MR. CUNNINGHAM: Well, one way to think about it
is that they're -- the rules that we have could be set up in
at least two bins -- adequate protection bins and cost
beneficial safety enhancements. The station black-out rule,
for example, was a safety enhancement rule.
If you're considering the cost beneficial safety
enhancement rule, the proposed rule, you look at the costs
and the benefits to be achieved from the rule. There's an
approach that you use to resolve whether or not you
implement that rule, and cost benefit is an explicit part of
that process.
An adequate protection rule does not have cost
benefit associated with it. It's determined that it is
necessary to provide adequate protection to the public
health and safety.
DR. KRESS: What criteria do you use to make that
determination, Mark?
MR. KING: Well, maybe I can help a little. This
is Tom King from the research staff. There is no
quantitative definition of adequate protection. It's a
qualitative judgment.
DR. KRESS: It's a judgment, qualitative --
MR. KING: It's a qualitative judgment, and the
issue, in addition to can you consider costs or not consider
costs, there were some numerical guidelines established that
Mark's going to talk about. Back in 1983 when the original
rule was put in place, that are not consistent with --
remember that was before there was a safety goal policy.
Before, there were the reg analysis guidelines that we have
today that are not consistent with what's on the books
today. So, that's an issue that has to be worked out in
discussing these options.
It seems to me what we want to talk about today is
technically what makes sense in terms of looking at the
options and picking the right option. As separate or
parallel with that, we'll have to deal with this adequate
protection question, but to me the first question is what
technically makes sense, given the information we have today
as to where we should go with PTS.
DR. SHACK: But you're talking about -- let me
beat that once more. I mean, if you go to one of the
options that essentially decreases the frequency, is that
you now have to do a cost benefit analysis of that --
MR. KING: That's an open --
DR. SHACK: -- or are you arguing that you didn't
have adequate protection before and now you need this to
have adequate protection?
MR. KING: That's a question I can't answer today.
DR. KRESS: But none of the options decreased the
frequency?
DR. SHACK: Now there's one that goes to one times
ten to the minus six.
DR. KRESS: Yeah, but -- oh, there is one for
that?
DR. SHACK: Yes.
DR. KRESS: I forgot. You;re right. I was
thinking the five times two to the minus six was the lowest,
but you're right.
MR. KING: Yeah, there are options that go both
ways and stay where they are today.
DR. APOSTOLAKIS: But in terms of the goals, I
don't know, are we going to talk about it later? Is it a
bad time to raise these issues now?
MR. CUNNINGHAM: It might be better when we get to
the options.
DR. APOSTOLAKIS: Okay, I'll wait.
MR. CUNNINGHAM: Okay. Just one point for what
it's worth.
DR. APOSTOLAKIS: Sure.
MR. CUNNINGHAM: When we talked to the
subcommittee the last time, we had an option on here that
it's not in the paper and it's not in the presentation, and
that was talking about a potential for a reverse backfit
analysis, and that was in effect a cost benefit analysis of
potentially relaxing the rule. The cost beneficial part of
their backfit rule today is do the costs outweigh the
benefits, or the benefits outweigh the costs of imposing a
rule.
You could think about for some rules, if you're
going to relax them, does the benefit in terms of the cost
reduction isn't justified, given the risk increase you might
get. In the context of this rule, because it's an adequate
protection rule, we took out that option because it's been
very clearly delineated that adequate protection rules and
cost benefit are two different -- are not handled together.
DR. KRESS: Two different regions in a three
region process?
MR. CUNNINGHAM: Yes, that's right. Yes, that's
right.
DR. APOSTOLAKIS: That's the implication, but I'd
like to understand it a little better, but I'll wait until
you come to that option, which happens to be B.
MR. CUNNINGHAM: Okay, so the purpose of the paper
that we provided to you is basically, just by way of
background. As you're well aware, the staff has a fairly
large effort underway to revisit and potentially revise the
technical basis for the PTS rule. This was started for a
couple of reasons. One is our experience in trying the
implement the rule and the associated reg guide in the
Yankee Rowe case a number of years ago, and also to reflect
that in times since the Yankee Rowe decision, there's been a
lot of research done on materials properties of reactor
vessels that give us a much better understanding, we think,
today of how these vessels would respond to a PTS
overcooling event.
I'll come back to some of the materials research a
little bit later, and we're not going to go into the details
of the analysis or the program today. You've heard it at
the last subcommittee meetings, and it's kind of -- we're
going to focus on one particular issue, and that is
basically the probablistic aspect of the screening criteria
that underlies the rule, if you will. So, the paper itself
is intended to provide a recommendation to revisions to one
part of the screening criteria, and it's in the rule and, in
effect, to put this before the Commission early in the PTS
revisitation program so that we'll have enough time to
respond to the Commission decisions and to modify the
program if we need to to reflect what the Commission wants
to decide on this one factor. So, that's kind of the narrow
purpose of this paper and the briefing today.
DR. APOSTOLAKIS: This criterion will be set and
without the benefit of the analyses that your staff
presented here last night?
MR. CUNNINGHAM: The way we're looking at the
criterion is kind of a top down thing. Given the basic
policies, what would that tell us about the acceptable value
as opposed to from the bottom up.
DR. APOSTOLAKIS: But sometimes it's helpful to
also know coming from the bottom what kinds of numbers
you're getting.
MR. CUNNINGHAM: Yes, that's right.
DR. APOSTOLAKIS: But you will be doing this
without the benefit of that unless there could be some
studies here and there.
MR. CUNNINGHAM: We're working first from the
general principles down.
DR. APOSTOLAKIS: Okay.
MR. CUNNINGHAM: You know, the exact figure that
we come up with in the long run may be reflected by some of
the experience in actual applications.
Go back for a few minutes to the rule itself, and
it was issued in 1983 as an adequate protection rule. As
Tom mentioned a little bit ago, this was before the safety
goal. This was before a lot of things that we had done in
PRA. It was very early application of PRA in looking at
regulations.
DR. APOSTOLAKIS: In what way was it and
application of PRA? I don't know the history.
MR. CUNNINGHAM: Okay. I'll come back a little
bit to that in a minute.
DR. APOSTOLAKIS: Okay.
MR. CUNNINGHAM: In the rule, there's an
acceptance criterion set up in terms of what they call an
embrittlement screening criterion, and that's in terms of
the material properties of the reactor vessel. If you
exceed that -- if a licensee determines that they're going
to exceed that limit, that screening criterion, at the end
of their life, the reactor vessel, then they have to do some
additional analyses, demonstrate that the plant can continue
to operate safely.
Reg Guide 1.154 was put into place to provide one
way by which you would perform that safety analysis.
DR. KRESS: Mark, I have a couple of questions
just for my education on that. One, if a plant actually
underwent -- it's got its calculations and it's got its
RTPTS value based on the calculation. Now it's going along
operating and undergoes a PTS event which supposedly would
do something to the cracks -- enlarge them, deepen them.
Does it then go back and redo its RTPTS and set a new
criteria based on the new crack properties or new crack
size? Is it a dynamic thing, or is it set one time and
that's it?
MR. MAYFIELD: This is Mike Mayfield from the
staff. You set the RTPTS based on embrittlement which
doesn't have -- it doesn't derive directly from loading on
the vessel. If you actually had a severe overcooling
transient, then history suggests that people end up doing
some inspection of the vessel just to make sure it hasn't
been harmed.
There is, in fact, an appendix in Section 11 of
the ASME code that really comes from what happens if you
exceed your heat-up, your pressure temperature limits, and
Neil Randall used to refer to this as the Friday afternoon
got you kind of rule, and so you exceed the heat-up limits.
What do you do? Well, there's some criteria that just says
well, is it really a problem and if so, what you do about
going on for some additional inspection. If you actually
had a PTS event, you wouldn't reset the RTPTS. You might
look at having to do some inspection of the vessel before
you went back to power, but that would be different than
worrying about the material property itself.
DR. KRESS: Well, what I had in mind there is that
if you went back and re-did the PTS probablistic
calculation, you would now come up with a new frequency or
through-wall crack for that vessel, it seems to me like. It
would change. Therefore, you would have a new value of
RTPTS to be equivalent to the five times ten to the minus
six.
MR. HACKETT: Let me see if I can address this a
different way. This is Ed Hackett. I think, Dr. Kress, I
think your question goes to maybe more the flaw distribution
and what might happen to the flaws from such an event.
DR. KRESS: Yeah, it changes.
MR. HACKETT: One of the things we know for sure
is that the vast majority of flaws that would be postulated
to exist in a vessel weld, for instance, would not even
participate in -- if they were hit with a PTS transient,
they're not going to react to it, the vast majority of them.
So, I think the answer to your question is yes,
there would be some adjustment, but I would wager that in a
probablistic sense, it would be a very minor impact.
DR. KRESS: I was --
MR. MAYFIELD: Well, my guess is if you did the
Section 11 inspection and you found a crack that had
initiated and grown, you'd be doing a lot of analysis.
MR. HACKETT: If you did that inspection and --
DR. KRESS: If you found one like that, you would
do so, okay. The other question I had, Mark, is are you
going to discuss sometime the actual basis for the five
times two to the minus six? They probably had some reason
that they chose that as an acceptance criteria.
MR. CUNNINGHAM: Yes. Yeah, let me come -- I'll
come back to that in a minute.
DR. KRESS: Are you going to do that later?
MR. CUNNINGHAM: Yeah, I'll do it in a minute.
DR. KRESS: Okay.
MR. CUNNINGHAM: Anyway, associated with this
embrittlement screening criteria in the RTPTS is of a value
of the through-wall crack frequency of five times ten to the
minus six. It's linked in two places in the rule and the
Reg Guide. One is the value that was chosen for RTPTS as an
acceptable value of 270 degrees for some types of welds, is
linked to the five times ten to the minus six, and I'll come
back to that in a minute.
Also in Reg Guide 1.154, if a licensee is doing
the safety analysis and does the PRA analysis that gives
them an estimate of the frequency of a through-wall crack,
the Reg Guide basically says if that value is a five times
ten to the minus six or lower, it's acceptable.
DR. APOSTOLAKIS: Now, this is on the basis of
point estimates?
MR. CUNNINGHAM: Let me jump ahead to one slide.
DR. APOSTOLAKIS: Okay.
MR. CUNNINGHAM: This slide is reproduced from --
I got it from SECY 82-465, which is a paper, obviously in
1982, which provided a lot of the technical information that
was going to support the final version of the rule. What
we've got is basically a curve -- this curve here is a
summation of all the other curves that provides information
on the frequency of different surface temperatures coming
from a risk calculation. So, you could say they've taken
information on the challenges -- the initiators that could
cause PTS -- small locas, transients and that sort of thing,
combine that with probablistic fracture mechanics
information to provide estimates on the likelihood of having
a through-wall crack if RP, RT and DT were at different
values, if you will.
DR. APOSTOLAKIS: So the vertical axis is the
frequency of through-wall cracks?
MR. CUNNINGHAM: Yes, yes, yes. That's correct,
per reactor year, as a function of RT and DT.
DR. SHACK: So, okay, you'd get enough initiators
in a year that if your RT and DT was 350, you would then get
ten to the minus something or other?
MR. CUNNINGHAM: Something like that, that's
right. The five times ten to the minus six comes in in one
way here, one particular way. If you go to 210 degrees, you
get -- I'm sorry. Associated with the frequency of five
times ten to the minus six is an RT and DT of 210 degrees.
There was an estimate made at the time about the uncertainty
in this curve, if you will, at that location. It was an
estimate that sigma, the standard deviation would be about
30 degrees.
So, the decision to make the RTPTS 270 degrees was
saying well, we have -- at five time ten to the minus six,
we have a value of about 210. We want to be confident that
it's not going to really hit that, so we're going to move
over to sigma and set the value to be 270 degrees. So,
that's what shows up on the rule.
The five times ten to the minus six, to get back
to Dr. Kress' question as I understand it, was not set from
some global standard, if you will, to say that's an
acceptable value based on high principles, if you will.
There were discussions about the safety goal
underway when this was being established, and there were
discussions about what was an acceptable frequency for core
damage from any particular type of initiator, any type of
accident, and that was generally talked about. It was about
ten to the minus five per year. The five times ten to the
minus six is what was more of an analysis of what was the
frequency. It was more of a bottom up type of calculation.
We think we can accomplish five times ten to the minus six.
We think that's a reasonable frequency, and if we get there,
it's probably okay.
DR. KRESS: That's where the five comes from.
MR. CUNNINGHAM: That's where the five comes from.
DR. KRESS: I was interested in that five.
Basically, then, you're saying that if one had an acceptance
criteria, say, on CDF, and if that value were ten to the
minus four per year, and if one looked at the set of
sequences that contributed to that -- they're being in PRA,
so there might be something like ten to a dozen sets of
them.
MR. CUNNINGHAM: Yeah.
DR. KRESS: If you say you had a principle that
you don't want to be overly influenced by any one of those,
if that were the principle, then you may divide the total by
a factor of ten or order of ten. You end up with ten to the
minus five, roughly.
MR. CUNNINGHAM: Right.
DR. KRESS: Is that the rationale you end up
getting a number like that?
All of the discussion -- in 1982, that discussion
was held with the advisory committee, among other places,
but there was a general sense, although the safety goals
weren't established, that ten to the minus five for an
individual set of sequences was probably about the right
value. We'll get to in a little bit, you'll how a little
bit later in time that showed up in the black-out rule and
the atlas rules and that sort of thing.
DR. APOSTOLAKIS: But I -- I'm sorry, go ahead.
MR. CUNNINGHAM: The difference between the five
times ten to the minus six and the ten to the minus five,
there was discussion of whether or not --
DR. KRESS: There's two really important thoughts
in that. One of them is -- number one, you have to have an
acceptance criteria.
MR. CUNNINGHAM: Yes.
DR. KRESS: All right, now we're talking here
about ten to the minus four, but in my mind, that's not
synonymous with adequate protection. That's something else.
The other important thought in there is should one
allocate among the sequences an acceptance criteria and what
rationale should one use in terms -- and how should one
factor into that allocation the uncertainties in each
sequence. Each sets of sequences have different
uncertainties associated with them, and how does that enter
into it? There's some real deep thoughts that go into that.
MR. CUNNINGHAM: Yes, definitely, and that type of
discussion was held as kind of a backdrop to this. Again,
the five time ten to the minus six per se didn't come from
that type of --
DR. KRESS: It probably come out of a judgmental
analysis.
MR. CUNNINGHAM: Out of the analysis that that was
what you would get at this --
DR. KRESS: But you could.
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: I'd like build on what Tom just
said. First of all, I really -- if this is not an article
protection rule, you can't use any of the goals we have.
DR. KRESS: That was one of my problems.
DR. APOSTOLAKIS: What's funny is you have to use
lower -- higher numbers.
DR. KRESS: Yeah, minus three probably.
DR. APOSTOLAKIS: And the second question is I'd
like to understand how this fits into the level three PRA.
I don't remember the sequence now, but is it really
appropriate to worry only about core damage frequency? I
mean, when you have this kind of failure, is the containment
going to do anything?
DR. KRESS: Well, this is another issue with PTS.
You should probably give it more thought because it's one of
those things that could lead to early containment failure.
So, yeah, this issue of is it a CDF or is it a CDF and a
LERF at the same time really is important.
DR. APOSTOLAKIS: Or is it the same all the way to
the QHO's? I don't know.
DR. KRESS: Yeah, so I think that's very
important, George.
DR. APOSTOLAKIS: But not just the core damage.
DR. KRESS: How you deal with the acceptance
criteria ought to depend on whether it's a LERF issue or CDF
issue or both.
DR. APOSTOLAKIS: That's a question then, Mark.
Why didn't you have in this write-up something about the PRA
sequences and where this fits into the picture. I think
that would have been very informative in placing everything
into perspective because you are going later on and
discussing what should I do about the LERF or what should I
do about the CDF. Then it hit me. I said well, gee, but I
really don't know. These arguments would have been much
clearer if you had had some discussion earlier on how this
phenomenon fits into the level two PRA or if necessary,
level three, which I think would be very easy to do, you
know.
DR. KRESS: Okay. I think you can almost bypass
level three and talk about LERF.
DR. APOSTOLAKIS: Yeah.
DR. KRESS: Yeah.
DR. APOSTOLAKIS: All I want is a convincing
argument. I'm not arguing, but I'm convinced, though, you
have to look at level two.
DR. KRESS: Yeah. I think you have to look at the
LERF part of level two.
DR. APOSTOLAKIS: Yeah, and some of the --
DR. KRESS: Not necessarily the vision products.
DR. APOSTOLAKIS: Yeah, yeah, that's right.
DR. KRESS: Yeah.
DR. APOSTOLAKIS: Because there are arguments in
the options you are developing that really would become much
clearer that way.
DR. KRESS: Okay.
DR. APOSTOLAKIS: The other thing, I think this
issue of article protection is going to create headaches
here because we have no numbers for article protection.
DR. KRESS: I think it's going to be a real
headache.
MR. CUNNINGHAM: That's again, this is why this is
kind of an interesting situation from a policy standpoint
that you've got an adequate protection rule then you've got
built into this indirectly and in some cases more directly
unacceptable frequency, again, set 17 years ago. That's,
again, part of the reason for getting this to the Commission
is how do we deal with that.
DR. APOSTOLAKIS: Now, the other thing I -- I have
difficulty with, you gave us an explanation why they weren't
up to 270 degrees.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: Now, it seems to me that the
uncertainties on the frequency, so it's vertical, and I
don't know, by moving to sigma -- what sigma is that, on the
RT and DT? How did they come up with that? It's a
horizontal sigma.
DR. KRESS: Horizontal, yes.
DR. APOSTOLAKIS: Now, I don't know what kind of
uncertainty it represents because there is also large
uncertainty vertically.
DR. KRESS: That is almost strictly data, that
horizontal one.
DR. APOSTOLAKIS: Okay, so the uncertainty, then,
in calculating the core damage frequency is not there at
all.
DR. KRESS: It's not in it at all, and that's one
of the things they want to improve I understand, in the new
process.
MR. CUNNINGHAM: Going back to what is my
understanding of what happened in the early 80's on this, on
that point. My understanding is it's a sigma in the
material properties.
DR. APOSTOLAKIS: Oh, okay.
MR. CUNNINGHAM: Because it was believed at the
time, with good reason, I think, that the dominant sources
of uncertainty in this calculation were the uncertainties
and the understanding of the flaws. Are they embedded flaws
or surface flaws? What's the size of the flaws? What's the
density of the flaws and that sort. It was believed that
that was the dominant uncertainty, but again, that was 17
years ago.
So, the calculation of the frequency, the
challenges, if you will, was a fairly, you know, not what
you would do today, if you will, in terms of an analysis.
It was an analysis originally based on I guess some work
done by Westinghouse on some challenges to the vessel -- you
know, perceived challenges of the vessel, and it was adopted
for more broad use by the staff. It was a quite different
type of analysis than what you would do today.
The sigma is 30 degrees. It's hard to relate it
in the context of thinking about alliatory and epistemic and
all of those things.
MR. WALLIS: This RTNDT, is the temperature
different, or am I misunderstanding something?
MR. CUNNINGHAM: It's a measure of the material
properties of the reactor vessel.
MR. WALLIS: Oh, it's a weird thing, so I can't --
MR. HACKETT: It's an index mark.
MR. WALLIS: I can't relate it to --
MR. CUNNINGHAM: Not being a materials person, I
have a hard time thinking of this as a temperature, but
people are apparently very comfortable with it.
MR. WALLIS: Yeah, so I probably won't understand
it.
MR. HACKETT: It's an index, is what it is, and as
mark mentioned, it's basically a material property that
you're relating to, you know, the degree of embrittlement.
Then what you're obviously trying to get to in this analysis
is fracture toughness, and this is a way of indexing that's
been used within ASME for many years now, indexing to the
fracture toughness curves. That's really what it boils down
to.
DR. KRESS: It actually can be looked at as a
temperature because it's where you transition into
ductility, and it's temperature related. It's related to a
lot of things about the materials, but it's a material
property.
MR. WALLIS: So, it's not a temperature
difference.
DR. KRESS: No, it's a temperature.
MR. CUNNINGHAM: It's not. Just stick on -- go
back to one point of Dr. Kress's a little bit ago. The
difference between five times ten to the minus six and ten
to the minus five, there was a discussion at that time of
should we go with something like ten to the minus five, or
should we work with what we think we can reasonable achieve,
and given our state of knowledge at the time, of five times
ten to the minus six. I believe it was Dr. Oakret on this
committee argued, given the nature of this accident, let's
keep it a little lower because in a sense, it was discussing
the uncertainty or lack of knowledge.
DR. KRESS: There's some good rationale to that.
Number one, here's a set of sequences that compared to some
of the other sequences that contribute to CDF, have
relatively large uncertainties.
MR. CUNNINGHAM: Yes.
DR. KRESS: At least perceive to have. Not only
that, it's a set of sequences that probably could be viewed
as have an impact on both CDF and LERF simultaneously.
Given those two things, they tell you well, maybe we ought
to -- in view of that kind of thing, a defense in depth
argument would say knock that down a little bit. Instead of
having, dividing the total CDF by ten, let's knock this one
down a little bit more because it deserves a little more
attention. So, there's some rationale -- I mean, some good
arguments. It's just how to put that in terms of
specificity and quantifying it. That's the problem.
DR. APOSTOLAKIS: That's why I asked for the
entries, to appreciate that.
DR. KRESS: Yeah.
DR. APOSTOLAKIS: The context.
DR. KRESS: Yes.
MR. CUNNINGHAM: Basically one way to think about
all the work that's going on to revisit the technical basis
for the rule is to go back and think about that it's going
to reformulate that line, if you will, the colored line.
We're re-looking at the frequency of the initiators. We're
looking at what we would expect to be the accident response
or the systems response to it. We're re-thinking the
thermal hydraulics given what we know today. We're
re-thinking the materials, given all the stuff that Ed will
talk about in a few minutes. So, it's going to be a
reformulation of that line.
DR. KRESS: And you're going to put vertical and
horizontal uncertainty?
MR. CUNNINGHAM: And one of the key pieces of it
is how you really assess the uncertainty in that line, if
you will, as well. I believe when Nathan Su was here the
last time we talked about trying to build a more formal
uncertainty analysis into this whole process, and both to be
able to understand what the uncertainties are and what's
contributing to those uncertainties. All of that's going
back to revisit that line. The issue for the paper is --
DR. KRESS: At the same time, you're going to
revisit the five times ten to the minus six as a separate --
MR. CUNNINGHAM: As a separate -- as another piece
of this whole process. What do you say about the right
acceptance criterion, given all these changes?
Just to stay on that for a minute and go back to
some -- we've touched on these already, but some of the key
underlying assumptions of the 1983 rule in the context of
the screening criteria was that through-wall crack frequency
of five times ten to the minus six was acceptable.
There's also that a through-wall crack was
equivalent to a large opening in the vessel. It was
equivalent to core melt. So, there was no distinction made
between starting -- once that crack may get through the
vessel, you assume that you're going to melt the core.
DR. APOSTOLAKIS: Are we going to do that
different now?
MR. CUNNINGHAM: I don't think we'll deal with
that part of it differently today. I don't think we have
the technology to really say much more about that today than
we did 17 years ago.
The last point is that the argument was made in at
the time was that if you have one of these through-wall
cracks, you would not substantially affect the containment.
So, coming back to your issue of level two and level three,
there was a fairly strong distinction made that you may melt
the core but that you're not going to fail the containment.
DR. APOSTOLAKIS: So, I can then still assume that
the condition or containment failure probability is point
one?
MR. CUNNINGHAM: The people in 1983, as I read the
Commission paper, basically said it was much less than .1.
DR. APOSTOLAKIS: And what did we say today?
MR. CUNNINGHAM: We'll come back to that. That's
a good question, and that's one of the questions we have to
face.
DR. APOSTOLAKIS: They said it was substantially
less than one? On what basis?
MR. CUNNINGHAM: The arguments that were made in
the paper were more qualitative as to why it would not be
substantially impacted.
DR. APOSTOLAKIS: So then they focused on core
damage?
MR. CUNNINGHAM: Then they focused on core damage.
So, again, this is 1983. There was -- we had never talked
about LERF's. Conditional probabilities of early
containment failure, we didn't talk in those terms back
then, but that's a key factor in where we go from here.
That's a key change in policy or practice of the agency that
we have to address.
I'll turn it over to Ed here for the next two or
three slides to talk about some of the basic materials
information that led us to start the re-visitation of the
rule.
DR. APOSTOLAKIS: Is the industry requesting that
you change this criteria?
MR. CUNNINGHAM: The industry is very interested
in working with us on this.
DR. APOSTOLAKIS: Who started this?
MR. CUNNINGHAM: The Office of Research started it
by re-examining the materials. That's what Ed will go into.
MR. HACKETT: To follow up on that, the industry
has been a full participant in what we've been doing. As a
matter of fact, we're meeting with them on an ongoing series
of meetings that I guess has been going on for about a
year-and-a-half now next week, and the industry is doing a
substantial portion of work on this project also.
MR. CUNNINGHAM: And particularly in the PRA.
They're helping us, at least for two of the plants, giving
us a lot of information where they have done PRA's for their
own plants. We've got four plants that are having things
studied here, and two of them were working from what
industry has provided us, or starting from industry has
provided us in terms of their estimates of the frequency of
these challenges. The other two, we're getting a lot of
information on frequency of initiators and that sort of
thing.
MR. HACKETT: These next few slides are pretty
much in the way of background and also probably review for
the committee because, as Mark mentioned, this has been
presented before. I think March 15, I believe, is when we
last went over this, but kind of in the order that they're
shown here on the slides, the most important driver
historically and still is the case from the materials
perspective has been the issue of flaw size, density, and
location, particularly in the reactor vessel welds, which
are usually the limiting considerations from a materials
perspective.
What had been done before, when Mark mentioned
82-465, the distribution that was used at that time was the
one attributed to Marshall in the United Kingdom. That was
based, at least in part, on as much data as they had at the
time, but it was not a whole lot. If you were to read the
Marshall report, there was a lot of extrapolation that had
to go into what they did. Since then, we've had the benefit
of a fair bit of research that's been performed out of our
branch in the Office of Research where we've actually done
detailed ultrasound examinations of welds from vessels that
never saw service, and then to confirm what was there, done
destructive examination of those welds. That's what's
described on this slide.
By and large what we've seen from that is a larger
density near the clad base metal interface of small, I guess
what I'd call indications, not necessarily flaws, and a lot
less of flaws that participate in a PTS that would be
affected, as Dr. Kress's question earlier, would be affected
by PTS transient.
So, from this we drew some hope. I think what
you'd see the theme through the materials research here is
that these are reasons to say that we thought we were
conservative previously, that maybe there's, you know, a
good technical basis now for backing off some of the
conservatism on the materials side.
The last point there is there is at least the hope
in this that we develop a generalized statistical
distribution of flaw sizes that would apply to USLWR's.
Maybe that will result, maybe not. What we hope to do is
next time we come before the committee, we'll have the
results of this study which is being done through expert
elicitation and by I think September time frame, I think, is
when we're scheduled to come back. We ought to know the
results of that study. We know pieces of it now, and I
think as a minimum, we can get down to flaw distributions
that are specific to vendors or NSSS, you know, fabricators.
So, we can at least go there, and that's a huge step over
where we were in 1982.
To move on to the next slide, another piece has
been irradiation embrittlement correlations. I think most
of the committee is probably familiar with Regulatory Guide
199, Revision 2, on which all of our embrittlement
predictions, correlations are based. We have since that
time had some ongoing research to augment and enhance those
embrittlement trend correlations. That work is now largely
completed, and just to give you an example, the database is
expanded by about a factor of four over where we were
previously, and I think it's also a lot more rigorously
defined than it used to be.
The net result is I think we have a better feel
for the rigor and statistical distributions of -- and the
uncertainties involved with the embrittlement correlations.
Again, that's an improvement. By and large what this new
embrittlement trend prediction does is taken on is own,
would tend to improve the situation for PWR's on balance, at
least the type of equation that we're looking at right now.
Not necessarily true for all plants but on
balance, there would be an improvement in the trend with the
embrittlement correlations.
Statistical distributions for material fracture
toughness, historically when we've run the probablistic
codes for fraction mechanics, the material fracture
toughness values that have been used have been lower bound
values from the ASME code. As Mark mentioned, a big part of
us going through this project is to try and not do that, and
everywhere we go making worse case assumptions and lower
bounding things. We're trying to put specific uncertainty
distributions on the individual pieces here so for the first
time, these material fracture toughness curves will be
address in a statistical fashion.
DR. APOSTOLAKIS: Now, what kind of uncertainties
are presented there? Why are you uncertain?
MR. HACKETT: This is basically -- the uncertainty
is several sources at least come to mind immediately.
There's just the uncertainty that goes with testing of that
sort in what's called the transition region for foritic
materials like reactor pressure vessel steels, that there is
an inherent uncertainty that goes with the material
variability in that region, which can be significant, and
for these materials, typically is.
There's also an uncertainty that goes with the
test data and how you got there. For instance -- well, in
this case, if you're just looking at the ASME curves, they
were based on tests of fracture toughness specimens per ASTM
standards, most of them per at the time ASTME 399 and since
that, you know, augmented by test to other standards. So,
you have the test uncertainty that goes in there, too. So,
you're at least addressing those two aspects of the
uncertainty. The overwhelming one would be just the
variation you'd see in material toughness in the transition
region.
DR. APOSTOLAKIS: So you have actual data and you
estimate variances and so on?
MR. HACKETT: That's correct.
DR. APOSTOLAKIS: Or is it the judgment of people?
MR. HACKETT: It's actual data, actually an awful
lot of data. Another thing I could add there is -- I forget
the number of data that were involved in the original bases
for the ASME curves, but there's probably about a tenfold
expansion in the amount of data that's available there now
Again, then you have to start -- you get into refinements
here because not all of that data were generated in
accordance with this original ASTM standard. There are
other standards that apply now, but I believe it's tenfold
or more expansion in the database for those types of
materials over the last 20 years. So, the answer is yes, we
are working from data.
DR. APOSTOLAKIS: So again, I have a particular
pressure vessel. This is a unique vessel. That's what I
own, and there is this uncertainty. What does this
uncertainty mean, that if I order another one from the same
manufacturer, the same specifications and so on, that
fracture toughness would be different because of some random
variations, or is it that I just don't know? I expect it to
have the same value, but I don't know that value?
MR. HACKETT: What you said is one aspect of it.
The other probably more important aspect is when you look at
these welds in specific where a lot of influence -- you're
seeing a lot of influence from the chemical composition
variability within the welds themselves, mostly related to
copper. So, what you're seeing, even within a specific
manufacturer is the type of variability you might see
through wall in that person's or in that manufacturer's weld
just due to variation of copper in the weld wire that went
into it. So, those are the kinds of things you're picking
up.
DR. APOSTOLAKIS: Just for the alliatory.
MR. HACKETT: I would agree.
DR. APOSTOLAKIS: There's an element of randomness
there.
DR. SHACK: Well, there's both.
DR. APOSTOLAKIS: Yeah, there's both, as usual.
DR. SHACK: In this particular curve, I think it's
probably alliatory.
DR. KRESS: It's 95 percent alliatory.
DR. SHACK: The other curves are mostly -- the
other uncertainties are mostly epistemic. This particular
one --
MR. HACKETT: If we were to get back to the
embrittlement correlation of curves, for instance, I
consider that largely epistemic. I mean, if we had all the
resources in the world and we could chase these things down,
we would get better and better. To some degree for material
fracture toughness, you're stuck with the inherent nature of
material variability and transition.
DR. APOSTOLAKIS: Is there any particular reason
why you're avoiding those words? I had to ask it to figure
out what's going on.
MR. HACKETT: No, other than the fact that --
DR. APOSTOLAKIS: The statistical of this division
is typically materials people use these.
MR. HACKETT: Yes.
DR. APOSTOLAKIS: But we have gone a little bit
beyond.
MR. HACKETT: See, Mark prepared these viewgraphs,
so I can blame Mark. I can say the short answer for me is
probably --
DR. APOSTOLAKIS: It would be nice to explain
those things somewhere.
MR. HACKETT: Yes.
DR. APOSTOLAKIS: Because, you know, the words
carry some meaning there.
DR. SHACK: Nathan has a chart.
MR. HACKETT: Yes.
DR. SHACK: He has all the uncertainties labeled
-- epistemic, alliatory.
DR. KRESS: That was a nice chart, by the way.
DR. APOSTOLAKIS: I can't carry that chart with
me. I read this document. I must comment on this document.
MR. CUNNINGHAM: There were a lot of things, as I
said, about the revisitation on the technical basis that we
did not get into in this paper, and we purposely stayed away
from some words like alliatory and epistemic because, at
least as I perceived the paper, they weren't necessarily
going to help us discuss the issues at hand in this paper.
DR. APOSTOLAKIS: When you get into the
uncertainties, though, it's important.
MR. HACKETT: Yes.
DR. KRESS: Nathan's document was very helpful.
That was a good document.
DR. APOSTOLAKIS: Why isn't it reflected here?
DR. KRESS: I don't know. Well, you could append
it, I guess.
DR. APOSTOLAKIS: Statistical distributions means
nothing. One of these days, we should have non-statistical
distribution.
[Laughter.]
Anyway, we're making a big deal out of it. I'm
just wondering why these words don't appear here anywhere.
DR. SHACK: It's the code for saying they have
data.
DR. APOSTOLAKIS: Well, I don't know. All three
of these -- I mean, statistical distributions for both of
these. They are both alliatory. That's what you're saying?
DR. SHACK: Yes.
MR. HACKETT: Largely alliatory, the last two.
DR. APOSTOLAKIS: Okay.
MR. HACKETT: I think the short -- Nathan's had to
put up with a lot of us crude materials types and school us.
I can at least speak for myself.
DR. APOSTOLAKIS: What's the purpose of developing
that long document that Dr. Su developed if it doesn't
influence the real ones? This is the real one, not that
one. That says an opinion.
MR. CUNNINGHAM: When we come back to the point of
coming to you and talking to you about reformulation of the
curve that I showed a little bit ago, that's where the
uncertainty analysis more comes into play.
DR. APOSTOLAKIS: Sure.
MR. CUNNINGHAM: And that's why we didn't talk
about it here.
DR. APOSTOLAKIS: Okay.
MR. HACKETT: I think we've really, just by virtue
of the discussion here, we covered the last piece, too, so
we'll probably move on to the next slide.
Another piece that we've been working for a long
time, probably longer than we thought we would be, is
revision to calculational procedures for the fluence values
per what's now draft Regulatory Guide 1053, which will
hopefully come before the Committee for approval for final
Reg Guide before the end of the year. At any rate, the
importance here is that we are doing calculations on the
fluence, basically are being updated per the three IPTS
plants that were done in the 80's, which were Robinson,
Oconee, and Calvert Cliffs. Robinson is now not part of the
project in that level of detail, but we've picked up Beaver
Valley. So, we are looking at basically the details down to
the cycle by cycle fuel loading and distributions or plant
specific fluence maps in this project, so that's another
level of refinement that was not done for the original basis
for the rule.
The last piece from a materials perspective is
just general improvements in the fracture mechanics
methodologies. Dr. Shah Malik, I think at the March meeting
presented a lot of the development and details that went
into what's called the favor code, which was itself an
expansion of the previous code, called VISA, which
originally was developed in-house by Jack Strosnider in the
development of 82-465 and then later refined it at PNNL.
So, it's been a very evolutionary treatment and
then just some of the refinements or enhancements -- I guess
I don't need to read through them all -- are listed there
that we're picking up now that were not there previously,
and that's also a major improvement and cause for optimism
in where we thought this would end up. Obviously what the
hope would be, that this criteria ends up at a higher
temperature value than where it is now, but you know, as
Mark says, the whole thing remains to be integrated and
seeing where that ends up.
I think that pretty much summarizes just a quick
overview of where we've been with the materials aspects. I
guess if there are any questions on any parts of that, this
would probably be a good time. If not, we'll move on.
MR. CUNNINGHAM: Moving on then, the last
subcommittee meeting, we talked about four areas of guidance
that the Commission had established since the 1983 rule was
established. Basically the safety goal policy statement,
station black-out and atlas rules, backfit rule, and
particularly the establishment of the reg analysis
guidelines and the tests in there for cost beneficial safety
enhancements, and Reg Guide 1.174. The paper itself
summarizes all of these things. I'm just going to kind of
quickly go through some of the key points related to these
policy documents.
With respect to the safety goals, I think the big
issue in that whole area that related to this is that the
Commission settled in on a subsidiary core damage frequency
goal of ten to the minus four. Again, back in the early
1980's, that was a thought, but it was not, you know, in
1990 basically, they said yes, ten to the minus four is an
okay subsidiary objective. They did not establish any other
subsidiary objectives for pieces of the core damage from
individual initiators. They didn't parse it out any finer
than just ten to the minus four.
DR. APOSTOLAKIS: But again, the problem here,
Mark, is what we discussed earlier. If the rule is not a
good protection rule, you cannot use any of this because
this is a goal statement.
MR. CUNNINGHAM: At this point, this is what's
changed. Yeah, when we get back, we have to sort out what's
the relevance of all of this --
DR. APOSTOLAKIS: That's right.
MR. CUNNINGHAM: -- to the PTF rule as it is.
That's what we're trying to get at. Station black-out and
atlas rules were established as cost beneficial safety
enhancements. Basically in both cases, there was a -- you
know, there were --
DR. APOSTOLAKIS: How many -- the question you
asked Tom. There are no rules how to decide whether a rule
is adequate protection?
DR. KRESS: That was my point of asking the
question.
DR. APOSTOLAKIS: That's interesting, very
interesting.
DR. KRESS: That was the whole point of asking the
question.
DR. APOSTOLAKIS: It's a brother question, of
course.
MR. CUNNINGHAM: I'm not going to say there aren't
any rules. It's just I'm not the right person.
DR. APOSTOLAKIS: I know you're cautious.
DR. KRESS: You know, George, we once wrote a
letter that called for such criteria.
DR. APOSTOLAKIS: Yeah.
DR. KRESS: It was a good letter, but it had a lot
of added comments on it.
MR. KING: No, there are no rules, you're right.
Probably the most recent discussion of this is the recent
paper that came out from NRR that talked about using risk
information and non-risk informed license submittals --
DR. APOSTOLAKIS: Yes.
MR. KING: -- where they put in all of the
qualitative things that you really don't think about when
you're thinking about adequate protection. That's probably
about the closest you're going to get to some guidance that
deals with that issue.
DR. APOSTOLAKIS: Maybe this, you know, there is
some cause to quantify what we mean adequate protection, you
know? Up, down, if you will. Maybe issues like that will
create some pressure bottom up, that there are real
decisions that have to be made, and we are making them
without guidance.
MR. KING: Yeah. In the NRR document --
DR. APOSTOLAKIS: Because we're now trying to
satisfy the Center for Strategic and International Studies,
there are real issues why we need guidance.
MR. KING: Uh-huh, and the NRR document is, in
effect, in my view, a bottoms up type document.
DR. APOSTOLAKIS: I'd like to see that, by the
way. Is that something we can see?
MR. KING: Yeah, in fact, I think this committee
reviewed it.
DR. APOSTOLAKIS: When?
MR. KING: A few months ago.
DR. APOSTOLAKIS: The context was?
MR. DUDLEY: This was a SECY paper. What we're
talking about is a SECY paper that went forward to the
Commission that tried to provide guidance on when you use
risk information --
DR. APOSTOLAKIS: Oh, yeah, yeah.
MR. DUDLEY: -- in reviewing the license
application in a deterministic arena.
DR. APOSTOLAKIS: I wasn't looking at it from that
aspect. But yeah, I know.
MR. CUNNINGHAM: Okay. Again, in the context of
the black-out and ATWS rules, they were cost beneficial
safety enhancements with different types of rules, but they
both had either fairly explicit or more implicit goals of
getting the frequency of core damage from these initiators
to be about ten to the minus five per year.
The backfit rule in the reg analysis guidelines,
what the reg analysis guidelines introduce or document is a
way of using the safety goals to screen out potential cost
beneficial safety enhancements. I think I mentioned we had
at one time an option which would somehow use that process
and invert it to look at potential burden reduction rules,
but we've taken that out because of the nature of this rule
is an adequate protection rule.
Reg Guide 1.174 brings two things, at least two
things to the table. One is it introduces a set of
principles on how you would judge the acceptability of
license amendment changes, which may be more broadly
applicable in a rule revision like this. It also
introduces, in this one context, the context of LERF, which
we'll come back to. It has important implications as to how
we might change the screening criteria.
Maybe this starts to get at the issue that you
alluded to earlier, George, about not explaining very well
of the level 2, 3 context of the PTS. In the early 1980's,
there were qualitative arguments made that there was not a
substantial challenge to the containment. Since then, we've
had a lot of work in severe accident phenomenology and that
sort of thing. The bottom part of that slide are basically
a set of the issues that I think we would have to deal with
as level 2 issues, anyway, in the context of PTS. There's
the dynamic nature of it. If this was to -- an event were
to occur and you have one of these big, through-wall cracks
or you've got the dynamic loads, what's going to happen to
the vessel? Is it going to move and that sort of thing?
The impact on the internals, what's it going to do
to the fuel itself? What's it going to do to surrounding
structures. You have the potential for pulling
penetrations, that sort of thing. You're going to have a
pressure loading if this were to occur. How severe is that
relative to other types of loading that are part of the
design basis? Again, some people have said maybe you could
just end up pulling some of the fuel or dispersing the fuel.
That has implications on source term, on the coolablity of
the fuel and that sort of thing. Another feature is what's
the availability of the containment ESF's in this context?
Going back again into the early 1980's. I think
there was a lot of credit taken for the fact that the
situation you're in here. You're breaking the vessel apart,
but you're breaking in a situation where you have a lot of
water. The argument was made, with some legitimacy, that
all that water has got to be a good thing, that you're not
melting this fuel. You've got water there. You've
presumably -- you haven't done things to compromise your
containment sprays, all of which could impact how this
accident proceeds. Maybe we need to do a better job in the
paper of explaining all of that.
MR. WALLIS: You said breaking the vessel apart?
MR. CUNNINGHAM: Creating a through --
MR. WALLIS: That's very different, though. You
can have a through-wall crack which just leaks very slowly.
MR. CUNNINGHAM: What we're --
MR. WALLIS: Not breaking this apart.
MR. CUNNINGHAM: The assumption in the analysis is
once you get a through-wall crack --
DR. APOSTOLAKIS: It unzips all the way around?
MR. CUNNINGHAM: It's going to be a large opening
in the vessel, a very large opening. The weld is going to
-- I'm sorry?
DR. APOSTOLAKIS: Conservatism built upon
conservatism. Because five ten to the minus six was in the
name of the person that I --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: Now you have another
conservatism.
MR. CUNNINGHAM: There is some conservatism in
that. How much it is --
MR. WALLIS: It is going to break the vessel apart
because a local crack isn't going to break the vessel apart.
I see it goes all the way around. It's a pretty massive
vessel.
MR. MAYFIELD: This is Mike Mayfield from the
staff. When we looked at this before and the context was
the axial cracks -- cracks running in the axial welds, both
through experiments done at Oakridge on scale model vessels,
and these are vessels that are about six inches thick and
roughly a meter in diameter. So, they're not small things,
but they're not on the same scale.
through analysis conducted at Pacific Northwest National
Laboratories, but we -- the conclusion we reached is that if
you get an axial crack that extends through the vessel wall,
it will propagate. Under these kind of pressure loadings,
it will propagate the full belt line of the vessel. That's
12 feet.
The crack opening will be measured in feet. So
these are not small, tight cracks that will just lead.
DR. WALLIS: But it still has to open, doesn't it?
MR. MAYFIELD: Yes, sir, it does.
DR. KRESS: But those are pretty spectacular.
MR. MAYFIELD: Those are spectacular, and the
flexibility of this vessel, even though it is a massive
component, the diameter to the wall thickness is such that
it will open. Even the scale model vessels tested at Oak
Ridge, which were much stiffer, did open, and they opened
rather remarkably. Tests that we have conducted on piping,
where the R/T diameter to thickness ratios were more
representative of vessels, and some of the experiments, the
pipe actually flattened. It opened so much that it looked
more like a plate than it did a pipe.
These are very high energy kind of events, and
they are spectacular when they go on, so, Dr. Apostolakis,
it's not just conservatism on conservatism. The reason that
we built in that assumption was based on engineering
analysis coupled with experimental data.
DR. APOSTOLAKIS: Good; thank you.
DR. WALLIS: So your remark about the lots of
water was that the water was in the containment. It wasn't
--
MR. HACKETT: Well, there are two factors with
respect to water. One is you're pressurizing this vessel
because of water, so there's water in the vessel as it --
DR. WALLIS: It doesn't stay in very long under
the scenario as it's --
MR. HACKETT: It doesn't stay in very long, but
there is water around the fuel when the vessel fails. The
fuel is not melted. The fuel is cool at that point.
DR. WALLIS: Yes.
MR. HACKETT: There are a lot of other questions
on what happens after that, but my second point was that
there is water in the containment also; you're right, in the
sense that you're going to have pools of water, liquid
water, standing there, and you also have the potential for
operation of the containment sprays. You're not in a
blackout situation, for example, where you cannot get
cooling, heat removal or decontamination of the containment
atmosphere, so that has some potential merit in these
accidents.
DR. APOSTOLAKIS: Okay; maybe this is a good time
to take a break. You're starting with the options now,
right?
MR. CUNNINGHAM: Yes, that is correct.
DR. APOSTOLAKIS: Yes; so we'll be back at 2:25.
[Recess.]
DR. APOSTOLAKIS: We are back in session.
MR. CUNNINGHAM: I propose that we would turn now
to some potential options for revising the probabilistic
aspect of the screening criteria. In the paper that you
have, there are four options identified; again, from what
we'd be interested in in talking to the subcommittees are
are there other options that ought to be put into the paper?
Are we clear on the options that we have? That sort of
thing. So I'd keep it as broadly or wide open as you like.
The four options in the paper, the first is just
make no change.
DR. KRESS: That's pretty clear.
MR. CUNNINGHAM: Yes; that one seemed pretty
clear, okay?
The second option is work to make the PTS rule, in
general, consistent with the blackout rule and the ATWS rule
in the context of an acceptable CDF, if you will.
DR. KRESS: When you say consistent, that bothers
me, because I think you mean make it the same.
MR. CUNNINGHAM: Okay; yes.
DR. KRESS: And consistent implies to me that
you're going to have some other thinking going into it.
MR. CUNNINGHAM: No; it's really much more narrow
than that.
DR. KRESS: Yes.
MR. CUNNINGHAM: Use the same numerical value for
the acceptable CDF, if you will.
DR. APOSTOLAKIS: But again, the issue of whether
that's legitimate is there, because these, as you state very
clearly, were cost-beneficial --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: -- safety rules.
DR. KRESS: I might want to propose an option E.
MR. CUNNINGHAM: That's what I'd like to hear,
another --
DR. APOSTOLAKIS: Do you want to wait until he
goes through A, B, C, D?
DR. KRESS: Oh, yes, yes.
DR. APOSTOLAKIS: Okay; and we'll add the comments
to your option.
MR. CUNNINGHAM: The third option is apply the
1.174, reg guide 1.174 principles and acceptance guidelines
to help us define how much we could change the acceptable
frequency of a PTS event, and basically, you'd work from
that to say how much could we afford to change this, or
should we change it?
DR. WALLIS: You mentioned defense-in-depth in
this context in your paper, in item C.
MR. CUNNINGHAM: Yes.
DR. WALLIS: I'm not quite sure how
defense-in-depth applies to a split vessel.
MR. CUNNINGHAM: This is one of the interesting
challenges to a PTS type of an event, that you could deal
with it several ways. One would be balancing the challenge
rate versus the conditional probability of vessel failure,
if you will, looking at the materials. Another way to think
about it would be vessel versus containment, that sort of
thing, but yes, this accident type introduces some unique
challenges to the issue -- or unique aspects of the
defense-in-depth concept. Basically, then, our option D is
similar to option C, except that we, in a sense, sidestep
the issue of LERF by just saying, de facto, that a
three-wall crack is a large early release, and you work from
there.
So I was going to go into --
DR. APOSTOLAKIS: So again, instead of the
equivalent, you should say the same.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: The two frequencies are the
same.
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: Because equivalent confused me a
little bit.
MR. CUNNINGHAM: Yes; we will.
DR. WALLIS: You need to say frequencies are the
same, too.
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: What I was going to go through in
the next four slides is kind of summarize what's in the
paper in terms of the --
DR. APOSTOLAKIS: Yes; you convinced us, though,
with your area presentation that this is no option.
MR. CUNNINGHAM: Okay; moving right along.
DR. APOSTOLAKIS: And why did you have that series
of view graphs telling us how much we had advanced, right?
DR. KRESS: That doesn't have anything to do with
the acceptance criteria. Advance is how you calculate it.
DR. APOSTOLAKIS: They were talking about
acceptance criteria, and for eight or nine view graphs, he
was telling us how great we are now, and these guys didn't
get much of a --
DR. KRESS: No, that was how to go about
calculating --
DR. APOSTOLAKIS: That was the acceptance
criteria.
MR. CUNNINGHAM: That's option A is that, yes, in
effect, you don't -- you just cite --
DR. APOSTOLAKIS: If you like, he could go back
and talk about it. I think you made the case.
MR. CUNNINGHAM: There's been -- there is at least
one advocate for that option, but that's not sitting at the
table here.
DR. APOSTOLAKIS: Okay.
MR. CUNNINGHAM: Okay; so, we'll move on.
DR. APOSTOLAKIS: So instead of consistent now,
what is the word? The same?
MR. CUNNINGHAM: The same, yes; utilize a core
damage frequency which is the same as that -- well --
DR. APOSTOLAKIS: Okay.
MR. CUNNINGHAM: You have to be careful, because
the ATWS rule didn't establish a goal of 10-5, from what I
can tell.
DR. KRESS: I was trying to milk you on this
consistent rule bit. What I would have thought might be the
sort of principle would be that we look at station blackout
set of sequences and the ATWS sequences; they look at how
much the contribute to the CDF, and you look at whether or
not they also impact containment, and you look at the
uncertainties in how well you can determine them. And then,
you factor those things into the acceptance criteria some
way. I don't know how yet, but some way, so that you could
-- you might actually come up with a different value for the
PTS set of sequences that would be consistent with these but
factor in these other things. That's why I asked you
whether --
MR. CUNNINGHAM: Okay.
DR. KRESS: -- you meant consistent, or did you
mean the same.
MR. CUNNINGHAM: Yes; okay.
DR. KRESS: I'm still not sure; I think you mean
the same.
MR. CUNNINGHAM: The option B was the numerical
values would be the same, in this 10-5 range. Maybe there
is another option, which is you again, factoring in all of
the other things, you develop 17 years or 15 years after the
fact, develop a consistent set of principles, if you will,
to make those three rules kind of align properly,
considering the uncertainties and that sort of thing but --
DR. KRESS: That was going to be my option E.
MR. CUNNINGHAM: Okay; well --
DR. APOSTOLAKIS: Yes, why -- what is the
contribution, what percentage of the core damage frequency
in existing plants is due to PTS?
MR. CUNNINGHAM: Typically, it's very small.
DR. APOSTOLAKIS: Right.
MR. CUNNINGHAM: Because most plants are not
anywhere near this 5x10-6 frequency. There are a few plants
that, because of the vagaries of the design of the vessel or
something like that that could approach it at the end of
life.
DR. KRESS: If it got up to that 5x10-6, it's a
code factor of 20 of the total.
DR. APOSTOLAKIS: Twenty percent?
DR. KRESS: I'm sorry; it depends on the thing.
If you had one of these reactors that's 10-3, and you got up
to 5x10-6, why, it's, you know, a factor of two orders of
magnitude. If it's 10-1 or 10-5 plants, then, it's a factor
of two, you know, two. It depends on the plant.
MR. CUNNINGHAM: Yes.
DR. KRESS: I don't know if you can do this on a
generic basis. You have to look at the plant.
MR. CUNNINGHAM: Yes; it's hard to do it on a
generic basis.
DR. SEALE: On the other hand, the station
blackout is the major -- station blackout is the major
contributor for a lot of plants.
MR. CUNNINGHAM: It is a very important
contributor.
DR. SEALE: So you're really bringing this up into
the forefront.
MR. CUNNINGHAM: And we're talking about what
would be acceptable, not what it is; it's acceptable at the
end of life and that sort of thing.
DR. KRESS: Yes.
MR. CUNNINGHAM: Which is a little different than
what it is; than what it is; that is correct. So it's a
little different, but in a sense, that's where you're going
that you would, in a sense, tolerate a larger contribution
from PTS under this option than you would today.
DR. KRESS: There's a real issue here on how you
allocate among sequences, and I think it's something that
deserves a lot of debate and thought.
DR. APOSTOLAKIS: Okay; if you remove -- the sense
I get from this document is that you believe that what we're
doing now is conservative. So if we remove that
conservatism -- no? I thought that was the idea.
DR. SHACK: Well, in the analysis, it's
conservative. I mean, the option on the acceptance criteria
is, in fact, to lower it.
DR. APOSTOLAKIS: One of the options.
DR. SHACK: One of the options, the recommended
options.
DR. APOSTOLAKIS: The recommended option says if.
They're not saying they're going to do it for sure. May
have to be used; may have.
MR. CUNNINGHAM: May have to.
DR. SHACK: I'm inserting may into --
DR. APOSTOLAKIS: But the question -- that's a
good point, in fact. Are you doing it because you feel that
better science can be applied to this, or are you doing it
because it's -- you will remove unnecessary burden?
MR. CUNNINGHAM: Let's back up. The context of
this option or in the context of --
DR. APOSTOLAKIS: The whole thing, not just this
option.
MR. CUNNINGHAM: The whole thing?
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: If they started the materials,
these folks here started the materials research long before
the issue of unnecessary burden ever made it -- this was
started because we think we could -- thought we could get a
more realistic understanding of the real risks associated
with PTS accidents, and I think that's where we're going
today. We think it also could allow some licensees to avoid
being shut down because of the vessel questions than it
might otherwise.
DR. APOSTOLAKIS: When?
MR. CUNNINGHAM: At the end of their life, at the
end of life.
DR. APOSTOLAKIS: So they might get a license
extension?
MR. CUNNINGHAM: It could impact some plants on
the -- it could impact the ability of some plants to get a
life extension, yes.
DR. APOSTOLAKIS: Okay; so that's really --
MR. HACKETT: Again, maybe just to lay out, set
the stage a little bit, right now, I think we visited this
maybe with the committee before, or NRR has. There's only
one plant right now that's predicted to reach the current
PTS screening criteria before end of license, and that's
Palisades. Every other plant is at or after the current
expiration of license. When you look at the license renewal
period, depending on -- and again, you have to get into
extrapolating some of this; what would you estimate in terms
of plants that might experience, you know, PTS difficulty
during that period, probably five to 10 or somewhere in that
range.
So, like Mark said, it's -- other than Palisades,
it's not exactly a here and now problem, but where it
becomes a problem that we hear a lot from the licensees, of
course, is that they're trying to argue now, in front of
their boards, to get approval for license renewal, and these
boards don't want to hear that there's going to be a vessel
problem or some kind of show stopper like this. So that's
where it's playing into the here and now for us.
MR. CUNNINGHAM: And they're also not interested
in hearing that the uncertainty in what the Commission is
going to do with respect to their vessel is still fairly
large, if you will.
DR. SEALE: Maybe setting them up for 80-year
lifetimes.
DR. APOSTOLAKIS: Now, coming to this option,
again, I have a problem, and since you are making very clear
on page 5 that -- you say that PTS is an adequate protection
regulation, and SBO and ATWS regulations are safety
enhancement regulations. So I don't know what it means to
make this consistent with the station blackout. I mean, if
you so clearly state that they are two different things,
it's going to be a problem, I think.
DR. BONACA: Yes; I was going to ask the question:
what was the logic for having them under these two different
criteria, I mean? You would have to go back into that issue
to understand, in fact, you know, what consistency means or
equal to, it means in the context. Also, I wouldn't see any
difference between B and A, with the exception that under A,
you would maintain it as it is today, and under B, you would
just arbitrarily, for the purpose of consistency, just
increase the number.
DR. KRESS: My option E was going to do away with
the references to things like 1.174 and station blackouts
and ATWS, because those are mixing apples and oranges --
MR. CUNNINGHAM: Yes.
DR. KRESS: -- to some extent.
MR. CUNNINGHAM: Yes.
DR. KRESS: And say derive a new criterion based
on fundamental principles, starting from some quantification
of adequate protection, which is the ringer in there, but
that was going to be my option E, and incorporate in there
some factors, something to do with defense-in-depth and
uncertainties in the termination and some principles that
you have to develop yet which don't exist on how you can
deal with allocation among sequences: do they all have to
be the same, or can there be some variation? What
principles would you use to decide?
MR. CUNNINGHAM: Yes.
DR. KRESS: And I think those are things that need
to be thought out in here. This is a good place to do that.
MR. CUNNINGHAM: Yes.
DR. KRESS: Because those are all issues, and they
all have to do with the acceptance criteria.
The other part of it and, you know, the program to
do the calculations things, that's great. I love that. I
mean, no problem at all. It's just acceptance criteria.
MR. CUNNINGHAM: Yes.
DR. KRESS: It needs some more thought.
MR. CUNNINGHAM: You're right, and maybe a better
way to characterize this is you could set up some sort of a
reliability allocation process.
DR. KRESS: I don't know what it is, but some
thought needs to be given.
DR. APOSTOLAKIS: I think it's best not to have an
allocation process per se.
DR. KRESS: But some principles to guide, yes.
DR. APOSTOLAKIS: Some principles and guidelines.
DR. KRESS: Yes.
DR. APOSTOLAKIS: As to what is good or achievable
and reasonable.
DR. KRESS: Yes.
DR. APOSTOLAKIS: And I'll give you an example:
when DOE was designing the new production reactor about 12
or 13 years ago, I was involved in that. And the same issue
came up. They wanted to use PRA and design and so on, and
balanced design -- a balanced design was considered a design
where no sequence dominated, okay?
DR. KRESS: Yes.
DR. APOSTOLAKIS: Okay? So, everyone says great,
let's do it, until somebody did the calculations for the
seismic risk, and the seismic risk was way out there,
dominating everything else. So now, the director was in
deep trouble, because, you know, he had to do something
about it, and I remember that he gave an estimate that if he
had to bring down the seismic risk to a level where it would
be comparable to the other contributions, it would cost to
the project an extra $700 million which he did not have, and
he said, you know, might as well forget it. And then, of
course, they decided to live with it, that seismic risk was
going to dominate, because overall, the risk was acceptable.
So, you have to have some flexibility, because you
never know what you're going to get.
DR. KRESS: And it may have some cost-benefit --
DR. APOSTOLAKIS: Yes.
DR. KRESS: -- considerations in there.
DR. APOSTOLAKIS: Exactly; cost-benefit.
DR. KRESS: I think you need principles.
DR. APOSTOLAKIS: Exactly, or guidelines.
DR. KRESS: Guidelines or something.
DR. APOSTOLAKIS: Yes; rather than saying, you
know, you know, Mark, that in the eighties --
DR. KRESS: You certainly don't want to say the
idle ought to be about the same.
DR. APOSTOLAKIS: Yes.
DR. KRESS: I mean, that's not the right --
DR. APOSTOLAKIS: Or you might say that would be
desirable but.
DR. KRESS: But, yes.
DR. APOSTOLAKIS: Yes.
DR. KRESS: Or it may be desirable that they not
vary by more than a factor of 10 or something like that.
DR. APOSTOLAKIS: Oh, yes, exactly.
DR. KRESS: Yes.
DR. APOSTOLAKIS: Or at least ask people to look
for --
DR. KRESS: Yes.
DR. APOSTOLAKIS: -- the reasons why there are
discrepancies and maybe give an argument why we should live
with them or --
DR. KRESS: Yes.
DR. APOSTOLAKIS: -- do something about them,
which is the same thing we do now if you are between 10-4
and, say, 10-3 core damage. I mean, people can convince you
that they have to stay there; otherwise, you have to come
down.
DR. SEALE: I have another problem with this
approach, and that is that I don't think the function or the
validity of defense-in-depth is anything like, for PTS, is
anything like as helpful as it is for station blackout in
ATWS. Remember when we were doing Shoreham, you could bring
in floating turbines and a few things like that, and you
could drive the station blackout risk down into the mud.
You're going to have fun and games for years before you're
going to drive the risk from -- assuming a PTS event -- down
into the mud.
Let's say the countermeasures are --
DR. KRESS: There's not much you can do.
DR. SEALE: That's right; the countermeasures are
in a whole different class of event. So I don't think you
want to put those on the same piece of paper.
MR. CUNNINGHAM: I guess there's an argument,
question, I guess, could you substantially affect the
frequency of challenge, frequency of a pressurized
overcooling of the vessel, and there's some arguments that
you might be able to do that through --
DR. KRESS: But work on the frequency.
MR. CUNNINGHAM: Work on the frequency.
DR. SEALE: That's not quite the same.
MR. CUNNINGHAM: No, I agree, but given that
you're in a PTS --
DR. SEALE: Yes.
MR. CUNNINGHAM: -- and you've cracked this
vessel, it's not a trivial thing to recover from that.
DR. SEALE: Yes.
MR. CUNNINGHAM: Okay.
DR. SEALE: I don't want to be the little boy
whose finger goes in that dike.
DR. APOSTOLAKIS: Okay.
MR. CUNNINGHAM: Option three is -- and perhaps
I'd already gotten, to some degree, where Dr. Kress was,
calling this the reg guide 1.174 principles is maybe not the
best way to characterize this; that there's a set of
principles that the staff has developed on how to deal with
potential changes to the license that talk about how we're
going to maintain defense-in-depth and those things. And
the question is are those principles still -- are those the
appropriate set of principles to apply to this -- to a rule
change such as this? And those principles happened to be
written in reg guide 1.174.
Some of them, I don't think, people would have
much problem with. I think the tough one is the
probabilistic aspect of it, where again, you're starting to
mix together safety goals, adequate protection and that sort
of thing.
DR. APOSTOLAKIS: And one more thing. I agree
that, you know, we are missing adequate protection and
safety goals, but also, can you talk about delta CDF,
referring to a screening criterion? That would be a very
novel use of 1.174. 1.174 refers to the actual estimated --
what, the calculated CDF. Now, you're saying I'm taking a
screening criterion, and I'm estimating the delta CDF, and
if that's less than 10-5, it's acceptable. That doesn't
make sense.
MR. CUNNINGHAM: Again, there's a principle. If
you back up to the principle, which is you might be able to
apply the principle, which is that any change either to an
actual risk or to an acceptance criterion would be small.
DR. APOSTOLAKIS: Then, you'll have a point, yes.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: In other words, but then, you
have this extra work, which is not negligible figuring out
if I change the screening criterion, what's really going to
happen for individual plants? Because ultimately, you have
to look at real CDFs.
MR. CUNNINGHAM: Yes; that's right, too;
eventually, each plant has to make some assessment against
that.
DR. APOSTOLAKIS: Or maybe you can tie it with
what you said earlier, that you don't really expect that
many plants to come close to the screening criteria; maybe a
few of them by the end of life, and maybe for those, there
will be some guidance as to what they have to do. But you
can calculate that as a change in risk.
DR. KRESS: I suspect the difference between
1x10-6 and 5x10-6 is substantial. I don't think it's linear
but, you know, it may mean five different -- five years
longer you could operate.
DR. SHACK: Well, if I look at their graph, I
would say that it would lower the screening temperature by
about 30 degrees.
DR. KRESS: It's not exactly --
DR. SHACK: Well, that would be fairly exciting, I
would suspect.
DR. KRESS: That's quite a few number years of
operation; you're right.
MR. CUNNINGHAM: I think what we're trying to
convey here is should we try to work with the principles
that were established in 1.174, the five principles, and use
those and apply those to the reconsideration of the
screening criteria.
DR. APOSTOLAKIS: As long as you make other issues
very explicit, you know, that you are -- the adequate
protection versus goals.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: And the fact that you are
dealing with the screening criterion.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: Now, there is one other comment
on the document itself, on page 5. There is something about
the tone of the second from the bottom full paragraph that I
think I find objectionable, and maybe you could change that.
MR. CUNNINGHAM: The numbering on my copy is
different than probably the --
DR. APOSTOLAKIS: It's the full paragraph under C,
just above D.
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: Do you see where D starts? It
begins this option would be most consistent.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: Okay; if you go down the
paragraph a little bit, where it says that is if it were
determined the containment performance was relatively poor,
given the PTS-initiated core melt accident, then, the
acceptable CDF may have to be reduced to ensure that the
LERF guidelines would be met. This could prevent -- could
potentially lead to a smaller value of the acceptable CDF
and potentially result in different screening criteria.
The way I read this is that the author of this
really didn't want this to happen and was apologetic, and I
don't think that's the way we should write regulatory
documents. If that's the case, that's what we're going to
do. There's something about the tone of these sentences
that I find --
MR. CUNNINGHAM: The tone -- okay.
DR. APOSTOLAKIS: -- unacceptable.
You know, if it happens that you have to lower the
screening criteria, well, what can we do? This is adequate
protection.
DR. KRESS: The way I would have read that is you
may have to -- you've got a criterion on LERF and one on
CDF, and instead of just assuming CDF and LERF equivalent,
you may have to do a lot of analysis to actually calculate
LERF on a plant-specific type basis --
MR. CUNNINGHAM: Yes.
DR. KRESS: -- and meet both criteria.
MR. CUNNINGHAM: Yes.
DR. KRESS: The one that controls would set it.
MR. CUNNINGHAM: That's right.
DR. KRESS: Which is reasonable to me. It's going
to take a lot of work to do this LERF calculation for a PTS
event like that. It's going to have a lot of uncertainty in
that problem but --
DR. APOSTOLAKIS: Yes; my point is that each of
the options, give the pros, and then, you end with the
negative, and this is the end of this; in other words, you
consider this a negative of this option, and that's what
bothers me.
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: The possibility of lowering the
screening criteria.
DR. KRESS: If it happens, it happens.
DR. APOSTOLAKIS: If it happens, it happens.
DR. KRESS: Yes.
DR. APOSTOLAKIS: Okay; good; let's move along.
Any other questions on C?
DR. BONACA: By the way, the same issue actually
was previously considered a con rather than a pro.
DR. APOSTOLAKIS: In the attachment?
DR. BONACA: The attachment at page 2-8, where the
description of these four options is discussed there.
DR. APOSTOLAKIS: That's fine.
MR. CUNNINGHAM: It's okay.
DR. BONACA: Just a note.
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: What is your deadline for
sending this to the Commissioners?
MR. CUNNINGHAM: May.
DR. APOSTOLAKIS: May?
MR. CUNNINGHAM: May. I had that same reaction.
I kind of do this doing it.
[Laughter.]
DR. APOSTOLAKIS: May what? After our meeting
here?
MR. CUNNINGHAM: On the books today, the due date
to the Commission for this is May 15. We have been talking
about whether or not that's realistic to do.
DR. APOSTOLAKIS: I think it's unrealistic as the
screening criteria we discussed.
[Laughter.]
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: You're not going to make it.
MR. CUNNINGHAM: Well --
DR. APOSTOLAKIS: Okay; option D?
MR. CUNNINGHAM: As we kind of talked about
before, this is similar to option C, except that in a sense,
it says that rather than us up front trying to analyze the
LERF of the containment performance issues in a PTS
accident, we'll just say that -- for whatever reason, we're
going to say no, we're going to assume that it's -- a
through-wall crack is equivalent to core melt is equivalent
to large early release. And that's the option. And you
adjust the acceptance criteria down, if you will, to deal
with that. That's a distinct difference between what it is
today, this option and what the rule says or the basis of
the rule back in 1983.
DR. KRESS: Is there a possible variation in that
from the --
MR. CUNNINGHAM: Yes.
DR. KRESS: It is for different containment types,
they have this assumption, whereas for others, you might
not?
MR. CUNNINGHAM: Yes; there's variations of that
possible. To some degree, it almost could be option C, the
burden is on the staff to show what the containment
performance is; on option D, you could put it on the
licensee, in a sense.
DR. WALLIS: You're going to consider
defense-in-depth, but you're going to throw away the
containment? It doesn't seem consistent.
DR. KRESS: That's consistent with
defense-in-depth.
DR. WALLIS: Throwing away the containment?
That's for defense-in-depth, isn't it?
DR. KRESS: Yes, but you're throwing away in your
acceptance criteria. That's a defense-in-depth --
DR. SHACK: But you're saying the challenge is so
large, you don't want it to --
DR. KRESS: And the uncertainties are so big that
you might as well assume it's gone. That's a
defense-in-depth concept.
MR. SIEBER: We are presupposing that the licensee
cannot or we'll never give them an opportunity to show that
its containment is robust and --
MR. CUNNINGHAM: A variation on this is to build
that in, to afford the licensee the opportunity to do that.
MR. SIEBER: I think sometime along in the future,
the technology of analysis of what happens to containment
under these conditions may advance to the point where there
is a reasonable argument that can be made. It would be a
shame to have the rule just arbitrarily preclude it.
MR. CUNNINGHAM: Yes; and that's why the paper, as
it stands today, goes more toward option C, which builds in
that flexibility, than D, which doesn't.
MR. SIEBER: It doesn't.
MR. CUNNINGHAM: But again, it's fair to say that
there's no resolution among the staff of whether C or D or A
or B is the right way to go on this at this point.
DR. BONACA: Just a question before the four
options, you know, there seems to be a significant
difference in the amount of work you have left to do to
support any one of the options, is there? And I'd like to
have a sense of what it would be.
DR. KRESS: You're right.
DR. BONACA: I mean, C seems to be much more
open-ended, too.
MR. CUNNINGHAM: I guess I hadn't thought much
about what the difference in work is.
DR. BONACA: It seems like -- I look at A; A seems
to be just --
MR. CUNNINGHAM: Yes, yes; A is --
DR. BONACA: Whatever you know deterministically
just --
MR. CUNNINGHAM: Yes.
DR. BONACA: -- plug it in, and that's it.
MR. CUNNINGHAM: That's right.
DR. BONACA: B is pretty much the same thing, I
mean, because all you're doing --
MR. CUNNINGHAM: Yes.
DR. BONACA: -- you're changing that number here.
MR. CUNNINGHAM: Yes; that's right. A and B
probably fit much better with the current reg guide and that
sort of thing. C and D would, yes, go beyond that, because
they impose some additional considerations; you're right;
that's right.
DR. KRESS: I would have a number of questions
that I would want answered before I decided on an acceptance
criterion, and they would go along this line: one, why does
a quantitative measure for CDF and LERF that's equivalent to
adequate protection or consistent with adequate protection;
number two, if I had such a measure, how -- what principles
or guidance would I use to allocate the -- that value among
a set of sequences that I have? What principles can I use
for there?
How in that determination of both the overall
acceptance criteria and the allocation among sequences, how
would I implement thinking that would be called
defense-in-depth or uncertainties in the determination of
each particular set of sequences and the overall uncertainty
in the final number? Those are the questions I think you
need to ask, and they're fundamental. They're almost
policy. Some of them are policy. It would be nice to have
a policy statement on this, but I don't know --
DR. SHACK: Just how about a purely technical one?
I mean, could you do the analysis of the containment
performance C?
DR. KRESS: Not at the moment; I don't think you
can. That's what Jack was saying: it's never been done.
It's not to say that somebody couldn't sit down and develop
the models and try to do it but, you know, I'd just say I've
never seen it done. Some people have made estimates of the
forces you get depending on the size but --
DR. SHACK: You can certainly make estimates.
DR. KRESS: Yes but --
DR. SHACK: The question is whether you would have
one that could get --
DR. KRESS: Yes, that you could defend and get
consensus on.
DR. BONACA: That's why I was asking the question
about can you do it? I mean, the effort is very, very
different and, you know, can it be done, and what would it
involve to do that?
DR. SHACK: And since the staff was recommending
option C, just what did you have in mind?
MR. CUNNINGHAM: Again, that's a draft paper, and
the staff does not have a recommendation today. Again, I
think -- you can go back several slides; we talked about
some of the key issues that you have to address in the
accident analysis, and I think you would have to go through
some sort of an event reanalysis or something like that to
deal with that. When you talk about the question of
containment performance, we're doing it in the context of
large early release. In my mind, anyway, you've got the
question of is -- can you argue that containment performance
is good enough so that you don't have a large early release?
And that's a different question than saying what is the
probability of containment failure given this, and what's
the magnitude of the release.
So to some degree, the question is a little
easier.
DR. KRESS: I don't think you want to deal with
the release at all. Just ask that question you're asking.
MR. CUNNINGHAM: You're asking; and if you can
argue, if you want to play with the numbers a little bit,
the distinction between 5x10-6 and 1x10-6, the conditional
probability of large release would have to be 20 -- the
break point, if you will, is 20 percent of 0.2. If you
can't argue that it's better than 0.2, then LERF is going to
control. If it's better -- if it's anything less than 0.2,
then the CDF would control, okay? And then, that's the
level of question you're talking about.
DR. BONACA: Well, right now, you have a lot of
conservatism that you're talking about, particularly the
flood distribution and so on and so forth. That gives
comfort that although you're not calculating a LERF with the
current rule, okay, you're really covering for it, because
probably, you're more in the 10-6 range for CDF than in the
10-5, 5x10-6, it seems to me. But now, how could you
consider A, B or even E or certainly A and B, which is
reducing conservatism, okay, in the deterministic portion
when you don't know if you are conservative on your LERF
criteria?
What I'm trying to say is that you have 5x10-6 for
CDF.
MR. CUNNINGHAM: Yes.
DR. BONACA: Okay? You have to have some
confidence that you have some margin to allow for a LERF of
1x10-6, but you are removing some of the conservatism out
there, okay, without verifying that you have, in fact, a
marginal LERF. I'm talking about independence of the CDF
that you are going to use as a criterion and the LERF; they
are not equal right now insofar as numerically.
DR. WALLIS: It seems to me you need some
technical analysis of containment failure instead of just
juggling probabilities.
MR. CUNNINGHAM: Yes, that's right, and that's
what you'd have to pursue as part of the -- again, part of
the reason of raising these issues early is to say is that
what we need to do as part of the program, or if the
Commission or somebody makes a policy decision, we're going
to go someplace else, and we may or may not have to do that
technical work.
DR. WALLIS: I think the public believes the
containment is there to contain any accident, so the belief.
MR. CUNNINGHAM: But again, the PTS --
DR. WALLIS: It provides some help with any
accident.
MR. CUNNINGHAM: Yes, but a PTS event is not a
design basis event for the containment; never was, and so,
it's a beyond design basis event for the containment, in the
sense that the containment is not specifically analyzed for
a PTS event. Again, that's -- that's where we are today.
We recognize the importance of LERF from a policy
standpoint. We have to think about the arguments of whether
or not, in a PTS event, what's the implications to LERF from
a PTS event?
MR. SIEBER: It would seem to me, though, that
there are so many variabilities in the containment analysis.
Under a vessel fracture, you know, you have pipe width, and
you have all kinds of stresses on penetrations and so forth
--
MR. CUNNINGHAM: Yes.
MR. SIEBER: -- that it would not be reasonable
for the staff to try to have a generic calculation that
would show what the relationship between CDF and LERF was.
I would leave that to the licensee to use reasonable methods
backed by good scientific and core test data that would show
that.
MR. CUNNINGHAM: In a sense, you could do that
under option C, you know.
MR. SIEBER: Yes, and, well, you'd need the data
to support that.
MR. CUNNINGHAM: Okay.
DR. WALLIS: If you get into a public meeting
which is really public, the public is going to ask you: is
the containment going to fail or not? One of the first
things that I think they want to know when you describe this
horrendous event.
MR. SIEBER: On the other hand, if you set the
criteria as in option C, you can answer that with some
surety, because you're basically saying that I'm going to
establish, under rule, the probability that it's very
unlikely that the containment will fail, and I wouldn't have
a problem answering that kind of question that way. You
know, you can't say any phenomenon in the world isn't going
to occur with certainty.
DR. KRESS: It would be nice, though, to have some
analyses that said, well, for large dry containment, maybe
not; for an ice condenser, it looks like yes, more likely to
fail; it would be nice to have some -- yes, I think you
could do some analysis that wouldn't be too costly --
MR. CUNNINGHAM: Yes.
DR. KRESS: -- that would just give you some
guidance on how to think about the LERF issue, and I think
that might ought to be part of this somewhere.
MR. SIEBER: Every containment has some kind of
analysis as part of the original licensing basis: how big
is it? How strong is it? What's the pressure increase?
How many heat absorbers are there? How much does it stress
penetration?
DR. KRESS: Yes, but we're dealing with a
different set of forces here.
MR. SIEBER: Yes; it's the penetrations that are
often unique.
DR. KRESS: Yes, that's basically the unique part,
yes.
MR. CUNNINGHAM: That's right; the dynamics of
when the vessel, if the vessel were to open up and what that
does to the penetrations --
DR. KRESS: I don't know if that's been looked at
as part of some of the seismic analysis or not; maybe you
can draw on those some way.
MR. CUNNINGHAM: That may be. That may be.
DR. KRESS: Look; you know, just some level of
analysis --
MR. CUNNINGHAM: Yes.
DR. KRESS: -- to give you guidance.
MR. CUNNINGHAM: In a sense, that's what we've
been trying to do offline, if you will, is do some of that
analysis and set it up, at least set up the problem a little
more precisely than it is here.
DR. WALLIS: If you don't do analysis, what are
you left with? Just guessing or --
MR. CUNNINGHAM: You have to go more conservative.
DR. WALLIS: -- judgment or --
MR. CUNNINGHAM: You have to go more conservative
and say go with option D rather than option C and just by
fiat say it's --
DR. KRESS: The same as the CDF.
MR. CUNNINGHAM: Yes; that's right.
MR. SIEBER: If you can't do the proper analysis,
that's where you are is in option D.
MR. CUNNINGHAM: Yes; you default to option D is
what it amounts to, yes.
DR. SHACK: You sort of have to decide how that
value coincides with an adequate protection argument.
DR. KRESS: Yes; that's one --
MR. CUNNINGHAM: We're back in that --
DR. SHACK: You're back in that ball game.
MR. CUNNINGHAM: Back in that, yes, that's right.
DR. SHACK: I mean, you can pick a CDF based on
option D, but what CDF you pick --
DR. APOSTOLAKIS: Yes.
DR. SHACK: You still have that problem.
MR. KING: I'm not convinced that's a real
problem, because I'm not convinced there's anything that's
strictly an adequate protection rule, and you never bring in
the additional step of can I add some safety enhancements
that are cost-beneficial beyond that.
DR. APOSTOLAKIS: Yes, but using the safety goal
numbers routinely in those rules I don't think is wise.
MR. KING: No but the --
DR. KRESS: It's the redefinition of adequate
protection.
DR. APOSTOLAKIS: Yes; essentially, you are
redefining adequate protection, and you're making it more
stringent.
MR. KING: No, I think what you need to do is you
say maybe the old rule was declared an adequate protection
rule, but that doesn't mean this new rule has to be declared
an adequate protection rule.
DR. APOSTOLAKIS: Then you have to do this
cost-beneficially.
MR. KING: But that, to me, is not unreasonable.
Why wouldn't we do that on any rule?
DR. APOSTOLAKIS: That might be a way out of this.
MR. KING: Then, the safety goal is really here's
what we'd like to see.
DR. APOSTOLAKIS: You just call it something else.
MR. KING: In terms of a level of safety, and
that's what we shoot for in this rule, and we have to do the
cost-benefit, and if it doesn't work out, it doesn't work,
but I think in any rule, we're obligated to do that. So I
don't really think there is anything that's strictly
adequate protection, and I don't think we ought to get hung
up on that question.
DR. WALLIS: I like what you say. I think
adequate protection doesn't exist. It's all cost-benefit,
really; with adequate protection, there's some bound,
because you don't know enough. It's really all
cost-benefit. There's no benefit, no cost at all.
DR. APOSTOLAKIS: Yes, but there are certain
benefits under the certain conditions in the country; you're
right. I mean, if there's war tomorrow, we might change
there the definition of adequate protection, but right now,
the way things have been the last 50 years, for example,
there is a certain level of --
DR. WALLIS: It's a convenient idea, because you
don't want to get into the details where it gets very fuzzy.
DR. APOSTOLAKIS: Anyway; I think Tom's idea, Tom
King's idea has some merit.
DR. KRESS: I think so, too.
DR. APOSTOLAKIS: You know, on the other hand, you
have the issue there of arbitrarily renaming things, but
well, anyway, we can't resolve that today. Do you want to
move on to 19, or you have already covered that?
MR. CUNNINGHAM: Nineteen and 20, in a sense, say
where we're going from here.
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: And if you want to do those and
then come back, it doesn't much matter to me but --
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: I did, at some point, want to
come back to Dr. Kress' option E, but if you want, I can go
ahead with 19 and 20 and just get through them and then --
DR. APOSTOLAKIS: Well, we're going to have a
round of discussions as to what should go into the letter
and what your presentation should be at the committee
meeting.
MR. CUNNINGHAM: Okay; well, let me go ahead,
then.
DR. APOSTOLAKIS: So, with this --
MR. CUNNINGHAM: This is basically where we're
going to be over the next few months. I thought about it in
the context of what it would be happening and what would be
the subjects of discussion at the September, I guess August
or September subcommittee meetings, the next set of
subcommittee meetings on PTS.
Basically, in terms of this paper right now, we've
got a May deadline. We need to talk to the full committee;
we need to continue the discussions with the rest of the
staff and with the legal staff. So that's going to proceed.
In terms of the PTS program in general, a number
of things that are going to be going on. Ed alluded to it
earlier that the development of the generalized statistical
distributions, using that term, on flaw sizes and things
ought to become available before the next meeting, so we
kind of expect that that would be a subject for the next
meeting.
DR. APOSTOLAKIS: The next meeting, you're
referring to the full committee meeting?
MR. CUNNINGHAM: No, the August-September
subcommittee meeting if you will; I'm sorry; not the May
full committee meeting.
Maybe we'll just go to slide 20 and say that this
is, in a sense, what might be an agenda for the next
subcommittee meeting. Where are we on this particular
issue? What have we learned about the flaw distributions
based on the expert elicitations that are underway now?
Other things that are going on; the materials area; what are
we doing in the uncertainty analysis, and how does that
reflect back into these other things and then maybe some
initial risk analyses for a plant or something like that.
DR. APOSTOLAKIS: So the Commission is planning to
decide on what the screening criterion soon? Or they may
choose not to do it?
MR. CUNNINGHAM: They may choose not to do it.
From our standpoint, we thought it was important to the
whole program to get these issues identified and discussed
early on in the program, because we don't want to wait until
a year from now to raise these kind of fairly fundamental
issues in front of -- before the Commission and give us no
time to react to them, depending on what the Commission
decides. So we would put these before the Commission;
again, right now, it says May, and then, the Commission may
decide -- well, the Commission will decide what it decides.
In times past, in some circumstances, they've said, well,
we're going to sit and wait and see --
DR. APOSTOLAKIS: Maybe you should give them an
option for that, like A now says make no change to the CDF
value underlying the screening criterion, and this is
permanent, right? Maybe you say make no changes now until
the staff has resolved a few issues.
MR. CUNNINGHAM: That's possible.
DR. APOSTOLAKIS: I think that's the most
reasonable --
DR. SHACK: One question; when I read through,
like, C and D, these options where you could potentially
lower the number, why is there no sort of discussion here of
a cost-benefit analysis? Wouldn't you have to do that?
MR. CUNNINGHAM: If you use Tom King's approach,
yes, you would do that. Again, the rules, as they're set up
today, is an adequate protection rule and cost-benefit are
two different things. So we don't do it that way, but it
may
be --
DR. SHACK: I mean, isn't the presumption that
you've met adequate protection --
MR. CUNNINGHAM: Yes.
DR. SHACK: -- the 5x10-6 and then, lowering it to
1x10-6 would then be judged on a cost-beneficial basis?
MR. CUNNINGHAM: That's why I say I think Tom's
idea has merit in the sense of how to tackle that.
MR. KING: And I think the 5x10-6, being a
17-year-old number, I wouldn't hold that up as some measure
of adequate protection. We don't have a measure of adequate
protection in terms of a numerical measure. I would --
maybe I'm following Dr. Kress' argument.
DR. SHACK: That number is not adequate
protection?
DR. KRESS: I think that's a risky argument, Tom,
because when you first put that number out, you said this
criterion is the -- meets adequate protection. Now, you're
saying it doesn't.
MR. KING: Well, we were giving you the historical
basis for the rule, and the word adequate protection is in
the old rule.
MR. MAYFIELD: This is Mike Mayfield from the
staff. When we modified the PTS rule in
Nineteen-Ninety-something to incorporate the latest
embrittlement trend curves, the argument against having to
do a backfit analysis was that it was, in fact, an adequate
protection rule, and what you were doing by imposing the new
embrittlement correlations, and there were some plants that
had their RTPTS value go down; others where it went up, so
it was a mixed result, but the argument was it's an adequate
protection rule, and this is redefining what you mean by
adequate protection so --
DR. SHACK: Yes, but I think that's an easy
argument. If the criterion stays the same in your analysis
of how close you get to them, I mean, that's a purely
technical question. When you're changing the criterion, I
think that's a very different kind of argument.
MR. MAYFIELD: Well, one of the other notions that
some of us have had about the various options is there is a
danger in taking a rule that was put in place to guard
against failure of the reactor pressure vessel, and you're
turning that into a fair bit of dialogue on containment
integrity, and we've had some difficulties with the level of
uncertainty in doing pressure vessel analyses; the
uncertainties and vagaries in doing the kind of containment
integrity analysis just to describe the accident
phenomenology for this kind of accident is orders of
magnitude more difficult. The uncertainties we've talked --
I think Dr. Kress mentioned large, dry containments. You're
now off into are you on a shield tank plant? Is it a
nozzle-supported plant? Is it one of these -- I guess it's
a C design that uses long columns; has to do with the amount
of movement you can get out of the vessel, which if you have
this kind of long, axial rupture, and Dr. Wallis had talked
about or mentioned the failure of the circumferential wells,
that's a different -- takes you into a whole different
scenario, but just staying with the long, axial welds; the
first thing that's going to happen is you're going to shove
the vessel up against the side of the shield structure,
whether it's a shield tank, concrete wall.
And now, how far can you drag the piping? Well,
that has to do with how much movement you can get inside
that shield. For shield tank plants, this is a completely
different scenario, so that the vagaries here would be major
to try and sort through this and do a credible analysis that
I think this committee would accept as a credible analysis
for containment failure given this scenario. I think that's
a major challenge.
So some of us have had some concern about going
down this path, because you're starting to focus on other --
things other than vessel integrity. In fact, that's how
some of the interest in getting away from essentially
mandating that analysis and looking for other ways to stay
focused on the pressure vessel.
DR. BONACA: I understand the complexity. I still
am puzzled by the fact that options C and D, it seems to me,
imply that the evaluation of the PTS rule may identify LERF
scenarios, okay, that would cause the reduction in CDF
criterion, right? So, not enough is known, but that could
be possible. Option A and B, without consideration of these
possibilities, propose to reduce margin outright, and, in
fact, in option B, they're proposing to actually increase
the CDF criterion to 1x10-5. I don't see how we can, in the
same breath, consider them on the same basis. If you really
think that there is the possibility that you're reevaluating
the rule, you may find that you're forced to really lower
the CDF criterion in some cases; how can you, then, without
evaluating LERF, go to option A or B? I just don't
understand it. I just don't see it anymore.
MR. KING: You're suggesting maybe those aren't
real options; we shouldn't even talk about them.
DR. BONACA: That's right; that's exactly right.
I'm not sure that -- do you have that option anymore if, in
fact, and I believe that you are right that there are
possibly scenarios in the evaluation of the PTS rule where
they may have LERF forcing CDF changes, and if that is the
case, you should take them out as option A and B, because
they're not options. I think you have to think about it.
MR. KING: I understand your point. I think we
were trying to cover the waterfront, not just eliminate
something because we don't think it's real at this point.
DR. BONACA: Yes.
MR. KING: But as we get feedback, that may end up
being the case.
DR. BONACA: Particularly option B; I mean, B,
it's simply for the purpose of consistency and just relaxing
the criterion, but you don't know what LERF is; really, you
don't.
MR. KING: No, that's right.
DR. BONACA: So I think you should eliminate B. I
mean, A may be a possibility. A simply states that what
you've done to date is right, okay? And all you know is
that you've been very conservative on your flow distribution
mostly, and you can give up some of the margin there. Now,
even in that case, I'm not sure you have a solid ground for
reduction of margin without looking at LERF.
MR. KING: No, I agree. I mean, my own personal
view is, you know, the Commission has put out safety goals
that express their expectations on safety. From those,
we've developed subsidiary CDF and LERF objectives, and that
ought to be the starting point, and you work backwards.
Maybe I'm supporting Dr. Kress' option E. And you work
backwards, and you say okay, how much of that do I want to
allocate to PTS? And that drives your option and your
decision, and I think there has to be some cost-benefit in
there. I wouldn't get hung up on whether it's adequate
protection or not. You're trying to meet the expectations
the Commission has put forward starting with the safety goal
policy.
DR. BONACA: Again, it may be time consuming and
complex, but it may be the only thing that you can do if you
want to revisit the rule.
MR. CUNNINGHAM: Yes.
DR. BONACA: Leave it where it is.
MR. CUNNINGHAM: But I think this issue of if you
do have to calculate LERF, can you do it? Do you have the
tools and the information and the data to do it is a real
question. Look at the direct containment heating issue. It
took us 7 years and millions of dollars to develop the
tools, the analytical tools and the data, to do away with
that issue, and it was plant specific in the sense of cavity
design and connecting compartments to the cavity.
DR. KRESS: On this business of calculating LERF,
when you do 1.174, and you specify, actually, because of the
way the safety goals are written that it's the mean value of
LERF, I think that implies that there was some known
uncertainty in the calculation of the LERF that went into
1.174 and that the mean, this mean value was acceptable.
Now, when one comes to a new LERF that has -- that wasn't
even part of 1.174, the PTS, where you have a huge
uncertainty in that LERF, and you say maybe the mean value
is no longer the right one to talk about; if we talk about
the 95 percentile of that because it's got such a huge
uncertainty in it, then, you may have a LERF that's such a
value that you just automatically say that CDF and LERF is
equivalent.
MR. KING: You mean option D.
DR. KRESS: I think in an argument like that would
support it, because the 1.174 mean value was based on some
sort of thing in mind of what the overall uncertainty in
LERF was.
MR. KING: But leave the door open for a licensee
to come in and make a plant specific case if they want to.
DR. KRESS: Or open if they want it, if they've
got the analysis tools that they can back up.
DR. APOSTOLAKIS: So we are right in the middle of
discussing options and giving advice to the staff.
DR. KRESS: Yes; I think that's what we're in.
DR. APOSTOLAKIS: And maybe we should go --
DR. KRESS: Let's get Tom's version of the
options.
DR. APOSTOLAKIS: Well, that's certainly one
option.
DR. KRESS: I still have a little problem with
that.
DR. APOSTOLAKIS: Yes, I do.
DR. KRESS: With adequate protection versus safety
goals.
DR. APOSTOLAKIS: Yes; well, would you like to go
around the table and maybe make comments, or do you want to
take a break first, 5 or 10 minutes? We'll be back at 3:30.
[Recess.]
DR. APOSTOLAKIS: So, who wants to start? Bob or
Jack?
MR. SIEBER: I can start.
DR. APOSTOLAKIS: Jack, please?
MR. SIEBER: First, I would congratulate the staff
for a good presentation. I thought it was logical; easy for
me to understand and also to congratulate the initiative to
try and use the advances in metallurgy to come up with
better ways of doing things.
As far as probabilities are concerned and what the
goals should be, I have a tendency to prefer Dr. Kress'
approach, but I think that's complex and requires a policy
adoption by the Commission. Lacking that kind of approach,
I would prefer option three, the third of the four that were
presented.
DR. KRESS: C?
MR. SIEBER: C, yes; three, C, because I think it
provides enough flexibility for the future; places the
burden on licensees and the review options on the staff and
probably, in the long run, would last longer from the
standpoint of do I need to change the rule to accommodate
some new and different situation.
From a technical basis, I don't think that option
C is as good as Dr. Kress' approach, which is applicable to
more situations than just this one and would make a greater
degree of consistency in a number of rules that are going to
come up in the risk-informing of Part 50. But otherwise,
that would be my opinion as to where we are right now.
DR. APOSTOLAKIS: Okay; thank you.
Mario?
DR. BONACA: I also think that the option C or
modified as Dr. Kress is suggesting would be the way to go.
The main point I made before and I repeat is I don't think
-- I think we should be clear whether or not we have more
than one or two options alone. I think we should try to
understand if, in fact, the only options we have are either
C or D here and then, again, a modified C as suggested by
Dr. Kress would be appropriate.
Just because I think it's important in that before
I came here, I really thought that we had four options, and
some of the simple ways to get there are attractive, A or B.
But then, because we don't understand LERF associated with
those changes for A and B, I don't think you can perform a
tradeoff of margin against an unknown LERF effect, and
that's why I'm saying that I think it would be important
that the Commissioners understand whether or not there are
more than two options. I think it's important also for the
committee to understand it. And that's pretty much my
comments.
DR. APOSTOLAKIS: Bill?
DR. SHACK: Well, I guess I'm really not thrilled
about any option that requires me to evaluate LERF, although
Mark seems to indicate that maybe it's not as -- you know,
if I had posed the question properly, it's not as horrendous
as it seems to me, and I guess, I mean, I'm always willing
to live with an option that says okay, you know if you can
present one to me, you know, it's sort of a fictitious
option, but that's okay, you know.
On the other hand, I'm not sure, you know, if I
look at even option A as one where I'm essentially assuming
a CDF equivalent to LERF, then, I still have to decide why,
you know, why do I pick five for, you know, as in D, why do
I pick one? And, you know, the arguments as to why I pick
one over the other are not clear to me, and it seems to me
that somehow that has to be made -- that case has to be made
a little bit better. But just as a practical point of view,
I guess I, you know, I kind of prefer options where I assume
that for all practical purposes, CDF is equivalent to LERF,
and yes, I'm willing to leave them in doubt.
DR. KRESS: I would have to say my feeling is
almost exactly like Bill Shack just said it, plus some
thinking along the lines of how to get those numbers. I
would not call this a 1.174 option. I would say -- I would
rephrase it to say using the principles, be sure that, you
know, and come up with a better justification, some sort of
justification for the actual numbers.
DR. BONACA: Just to understand it, so, you're
talking about option D?
DR. KRESS: Yes; yes, the equivalent.
DR. BONACA: Okay.
DR. KRESS: You know, I feel sort of like Bill
does. I leave open the possibility, but I think right now,
it's a fictitious opening.
DR. WALLIS: Well, I'm not sure that option A is
unacceptable until you know more. Things seem to be a bit
iffy, and if you had to justify C or D before a critical
Commission, you might have difficulty making a really good
case. I wonder if you can't do nothing for awhile until
you've got -- then, the question is what is it you could
learn that would help you better?
DR. SEALE: Well, I don't know that you're -- I
don't know that you really have to make a decision right
now. I understand that this is something that sort of sits
out on its own, and there's not a lot that you're doing that
helps you make this decision with any more information now
than you would later, except possibly this whole question of
risk-informing the regulatory process will be a little
further along, and so, you might have a few more insights
just by experience if you waited until everything else
catches up to this point.
But excepting that, I would say I think Tom's
approach is the one that sounds at least cosmetically the
best right now. I have a great deal of trouble, as I
indicated earlier, with writing ATWS and station blackout on
the same sheet of paper as PTS, because they're just not the
same kind of problem.
DR. APOSTOLAKIS: Thank you; anything else?
I expressed my views earlier. I, in fact, agree
with what Graham said. I don't think we can -- we know
enough right now to make a recommendation to the Commission
as to which option is best. It's probably a good idea to
offer an option for them to -- out of this and keep thinking
about these issues of adequate protection; how do you use
the principles and so on.
DR. SEALE: Yes.
DR. APOSTOLAKIS: And then come back later and
make a recommendation.
DR. SEALE: Yes.
DR. APOSTOLAKIS: That's my view at this time.
Yes; Mark, do you want to say something?
MR. CUNNINGHAM: I was just going to say one
option, if you will, is that we just use the paper that's
going up here in the near future to tell them what's going
on --
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: -- and not provide a
recommendation.
DR. APOSTOLAKIS: Yes; and some of the things
you're thinking about --
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: -- the issues that have been
raised.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: I think that would be
informative to them.
DR. KRESS: Yes; I think I would support that.
DR. WALLIS: Wouldn't you be more comfortable with
that really?
MR. CUNNINGHAM: It's a mixed --
DR. WALLIS: If you picked one of these other
options, you might find you had stepped in something.
MR. CUNNINGHAM: Yes; the down side of putting it
off, if you will, is that there is so much going on in
risk-informed regulation right now that all of these things
we talked about apply to them as well as PTS, and at some
point, we'd have to make this decision.
DR. KRESS: I think you need a policy statement on
how to risk-inform the regulations.
[Laughter.]
DR. KRESS: Yes; just ignore me.
DR. SEALE: He gets out of hand that way
occasionally. You have to pat him on the head.
MR. SIEBER: On the other hand, I think some issue
has to be first.
DR. KRESS: Yes.
MR. SIEBER: And this one is fairly clear-cut from
the standpoint of the phenomenon that is occurring, and
there is benefit even with that absent risk-informing it.
On the other hand, since somebody has to be first, why not
this one?
MR. CUNNINGHAM: This rule has advantages over
others. This is a cleaner rule than many of them.
DR. KRESS: Clean one to look at; that's for sure.
MR. CUNNINGHAM: Yes.
DR. KRESS: I like that thought.
DR. SEALE: That is a good point, yes.
MR. HACKETT: I guess the comment I would add is a
decision sooner rather than later obviously helps us from a
resource and planning perspective, because we have this
project planned to go out through 2001 now, and were we to
select one of these options versus another somewhere sooner
rather than later, we may look at very different allocation
of resources possibly. At any rate, that would be a
consideration.
MR. CUNNINGHAM: And how quickly we complete this
program has implications to decisions licensees have to make
about license renewal and things like that too so --
DR. APOSTOLAKIS: Yes; I don't think we are
suggesting that you stop it.
DR. KRESS: No.
MR. CUNNINGHAM: No I --
DR. KRESS: But particularly the other part of it.
DR. APOSTOLAKIS: The other part should go ahead.
DR. KRESS: Drive ahead with it.
DR. APOSTOLAKIS: It's just that -- you know.
DR. KRESS: And I don't think that developing the
principles for acceptance criteria would be really resource
intensive. Put one good guy on it, I don't know, Tom Keyes.
MR. CUNNINGHAM: He seems to find other things to
do most of the time.
DR. KRESS: Tom doesn't work anymore, but I don't
think that's very resource intensive.
MR. CUNNINGHAM: The issue if we want to get into
-- you can conceive of a very resource intensive program to
investigate LERF.
DR. KRESS: Oh, yes, if you had to go that route,
but, you know, I'm sort of assuming you're not going that
route.
MR. CUNNINGHAM: On personal opinion, I don't
think we have to go that way, but, so, I've been wrong about
these things before.
MR. SIEBER: On the other hand, if you went with
the option C, you certainly would have to know how to
evaluate the licensee's effort to determine what LERF is,
which I don't think is easy either.
DR. KRESS: That's right.
MR. SIEBER: There's a lot of things that go on in
containment integrity under this condition.
DR. KRESS: Yes.
MR. CUNNINGHAM: Either C or D both have
implications for some sort of a modification of the reg
guide or something --
MR. SIEBER: That's right.
MR. CUNNINGHAM: -- to lay out what we would find
at least as one acceptable way of doing it.
MR. KING: If we laid out a framework for
risk-informing Part 50, our option 3 framework which we
presented to the committee. If you -- it would seem
reasonable for whatever rule we're risk-informing, we may
want to have a similar approach, which I think to me, that
framework applies to Dr. Kress' approach more than any of
the other options we present. So maybe a realistic option
will be turn this paper into one that says this is the way
we're proceeding; we've sent the Commission this option
three framework; when you say when you apply that to the PTS
rule, it leads to an approach similar to what Dr. Kress came
up with, and we just tell them this is the way we're
heading, and if they have objections, they can speak.
If we don't hear from them, we'll assume they
don't have any big heartburn with it. I mean, that's an
intriguing thought instead of making this something where
you've got to pick and choose from four options.
DR. KRESS: I think that would be your best bet at
this point.
DR. APOSTOLAKIS: Okay; other thoughts? The
staff?
[No response.]
MR. CUNNINGHAM: Thank you for your good advice.
DR. APOSTOLAKIS: Members of the public?
DR. WALLIS: There will be a letter written on
this?
DR. APOSTOLAKIS: We will write a letter? Do you
want a letter still?
MR. CUNNINGHAM: We had requested a letter. I
think we need to talk -- the staff needs to talk about
whether or not, on our options for this paper, if you will.
DR. APOSTOLAKIS: So, you don't know whether
you're requesting a letter. That's what you're saying.
MR. CUNNINGHAM: I'm much less sure of it now than
I was at the beginning of the presentation.
MR. KING: Why don't we tell you at the full
committee? We're scheduled --
DR. APOSTOLAKIS: Yes, but then, we have to write
it on the spot.
DR. SHACK: We'll work on a draft, and then, we'll
decide on what to do with it.
MR. KING: How much time at the full committee do
we have?
MR. CUNNINGHAM: It's an hour and a half, I
believe.
DR. APOSTOLAKIS: An hour and a half? Okay; and
you will go basically over the same presentation?
MR. CUNNINGHAM: If that's what you'd like.
DR. APOSTOLAKIS: Except --
MR. KING: It may be a little different based upon
the discussion.
DR. APOSTOLAKIS: Can you add some of the
background in terms of the event trees, fault trees on how
this thing fits into the big picture of PRA?
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: You know, the comment I made
earlier? It's just a matter of pulling information from a
PRA.
DR. SEALE: And if you really do want a decision
now, if you decide, you may want to tell us more about what
all the goodies are that you can harvest if you get the
decision. You know, I mean, like the point that today, you
can put a skin on the wall and say here's our trophy from
risk-based regulation, all of that stuff.
MR. HACKETT: In that regard, I'd make one other
comment. A not inconsequential consideration in this entire
project is the industry interest in it. If the industry --
Ron Gamble is here representing the industry -- and Ron was
one of the people who identified early on that we needed to
take this on earlier rather than later; one of the issues
would be, I think, if the industry sees a significant
uncertainty on the part of the NRC or a delay, perhaps, in
decisions on this that their interest may wane accordingly I
guess is the way I might look at it, and that's something
that's just there.
DR. APOSTOLAKIS: Well, I think you should address
the issues of the benefits of the result from something like
this.
DR. SEALE: Yes; definitely.
MR. CUNNINGHAM: Yes.
DR. APOSTOLAKIS: That would go a long way.
What is the current situation? How does this
whole PTS issue fit into the risk assessments that have been
done? And then, what would the benefit be? I think that
would go a long way toward setting the stage.
DR. KRESS: We're not part of the risk assessment.
DR. APOSTOLAKIS: Oh, so we are using risk
information to do something that's not part of the risk
assessment?
DR. KRESS: It's usually screened out.
DR. APOSTOLAKIS: In all the --
DR. KRESS: It's usually screened out.
DR. APOSTOLAKIS: In all the PRAs, it's screened
out?
MR. CUNNINGHAM: Not all PRAs, but there are a few
PTS specific risk analyses around.
DR. SHACK: But again, if your embrittlement
temperature is low, it's going to screen out.
MR. CUNNINGHAM: That's right.
DR. SHACK: I mean, it's going to be zip.
MR. CUNNINGHAM: There are a large number of the
plants where it should be low because you're not anywhere
close to the embrittlement criteria.
DR. APOSTOLAKIS: Yes.
MR. CUNNINGHAM: But maybe we could provide a
better story on that, too, to the committee.
DR. APOSTOLAKIS: Yes; that's what I want to
understand better.
MR. CUNNINGHAM: Okay.
DR. APOSTOLAKIS: Because I've never paid that
much attention to it.
Anything else?
[No response.]
DR. APOSTOLAKIS: Well, we'd like to thank you
very much. It was a very good discussion, presentation.
Thanks, everyone, and the meeting is adjourned.
[Whereupon, at 3:48 p.m., the meeting was
concluded.]
Page Last Reviewed/Updated Tuesday, July 12, 2016
Page Last Reviewed/Updated Tuesday, July 12, 2016