Protecting People and the EnvironmentUNITED STATES NUCLEAR REGULATORY COMMISSION
Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title: Advisory Committee on Reactor Safeguards
Reactor Fuels Subcommittee
Docket Number: (not applicable)
Location: Rockville, Maryland
Date: Wednesday, April 4, 2001
Work Order No.: NRC-146 Pages 1-242
NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers
1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005
(202) 234-4433 UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
+ + + + +
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
REACTOR FUELS SUBCOMMITTEE
+ + + + +
MEETING
+ + + + +
WEDNESDAY,
APRIL 4, 2001
+ + + + +
ROCKVILLE, MARYLAND
+ + + + +
The Subcommittee met at 8:30 a.m., at the
Nuclear Regulatory Commission, Room T2B3, Two White
Fling North, 11545 Rockville Pike, Rockville,
Maryland, Dana A. Powers, Chairman, presiding.
PRESENT:
DANA A. POWERS, Chairman
GEORGE E. APOSTOLAKIS, Member
MARIO V. BONACA, Member
THOMAS S. KRESS, Member
WILLIAM J. SHACK, Member
ROBERT E. UHRIG, Member
PRESENT (Continued):
AUGUST W. CRONENBERG, ACRS Fellow
ACRS STAFF PRESENT:
MEDHAT EL-ZEFTAWY
ALSO PRESENT:
MICHAEL ALDRICH
SUDHAMAY BASU
EDWARD BURNS
RYAN T. COLES
MARGARET CHATTERTON
SKIP COPP
RALPH CARUSO
DAVID DIAMOND
GARRY GARNER
RICH JANATI
STEVE LA VIE
RICHARD LEE
EDWIN LYMAN
BOB MARTIN
LARRY OTT
JACK ROSENTHAL
HAROLD SCOTT
UNDINE SHOOP
C-O-N-T-E-N-T-S
AGENDA ITEM PAGE
Introduction . . . . . . . . . . . . . . . . . . . 4
Research Activities on High Burnup PIRT . . . . . 5
Framatome Testing and Assessment of LOCA
Ductility of M5 Cladding . . . . . . . . . .77
Westinghouse Testing and Assessment of LOCA
Ductility of ZIRLO Cladding . . . . . . . 116
Summary of OECD Topical Meeting on LOCA
Fuel Safety Criteria . . . . . . . . . . . 137
Recent Operational Issues and Experience with
High Burnup Fuel . . . . . . . . . . . . . 158
Research Activities on MOX Fuel . . . . . . . . 191
Presentation by Dr. Edwin S. Lyman . . . . . . . 209
P-R-O-C-E-E-D-I-N-G-S
(8:31 a.m.)
CHAIRMAN POWERS: The meeting will come to
order.
This is a meeting of the ACRS Subcommittee
on Reactor Fuels.
I'm Dana Powers, Chairman of the
Subcommittee.
ACRS members in attendance are George
Apostolakis, Thomas Kress, William Shack, Mario
Bonaca, Robert Uhrig. We also have the ACRS Fellow,
Dr. Gus Cronenberg, attending this meeting.
The purpose of the meeting is to discuss
the safety issues associated with the use of high
burn-up and mixed oxide fuels. The Subcommittee will
gather information, analyze relevant issues and facts,
and formulate proposed positions and actions as
appropriate for deliberation by the full Committee.
Medhat El-Zeftawy is the cognizant ACRS
staff engineer for this meeting.
The rules for participation in today's
meeting have been announced as part of the notice of
this meeting previously published in the Federal
Register on March 14th, 2001.
A transcript of the meeting is being kept,
and it will be made available as stated in the Federal
Register notice.
It is requested that speakers first
identify themselves and speak with sufficient clarity
and volume so they can be readily heard.
We have receive done request for time to
make oral statement from a representative of the
Nuclear Control Institute regarding today's meetings.
Do members have any other comments they'd
like to make before we enter into today's rather
interesting discussions?
(No response.)
CHAIRMAN POWERS: Seeing none of those,
then I think we'll just proceed directly ahead, and
I'll call upon Dr. Ralph Meyer to begin us in this
discussion of some of the most interesting research
going on in the Agency.
PARTICIPANT: This is when we figure out
if Ralph is a theoretician or an experimentalist.
(Laughter.)
CHAIRMAN POWERS: Know the answer? I know
the answer. Dr. Meyer can organize both research and
analysis to produce useful outcomes for the regulatory
process, right, Ralph?
DR. MEYER: Couldn't have said it better.
(Laughter.)
DR. MEYER: Okay. I have a lot more
material later for the second presentation, which is
a summary of a meeting that was held recently on the
subject of the embrittlement criteria. So I'm going
to try and stick with the time period that's been
provided here, and that means that I have a couple of
slides that I just want to throw on for background so
that they'll be in your package, and I didn't mean to
dwell on every single slide in the package.
I'm going to spend most of the time
talking about the PIRTs and trying to say what we
learned from them and what we're going to do about it.
If there's a little time left over, then I can talk
about the status of some of the various research
programs.
CHAIRMAN POWERS: Okay. That would be
useful. Because this is a Subcommittee meeting, I'm
pretty liberal with the time allotments because
there's no other opportunity we have to discuss
things.
DR. MEYER: Okay.
CHAIRMAN POWERS: So I'll hold the
schedule roughly correct, but if you have things that
you think we ought to hear, feel free to tell us.
DR. MEYER: Okay. The first slide is just
some background information on the alloys that we'll
be talking about. Zircaloy has ten in it. I think
everybody knows that. Low tin Zirc is tin with a
concentration in the range of 1.2 to 1.4 percent.
ZIRLO is like low tin Zircaloy. Add a percent niobium
and M5 is zirconium and one percent niobium, more or
less.
We will be referring to these off and on
throughout the day.
Also I want to point out some of the
criteria that we are looking at. We are looking at
criteria for postulated accidents. These are the
things that were identified in the agency program plan
a couple of years ago, and just in general, we have
criteria on fuel damage to make sure that the damage
is limited and that we don't get uncoolable core
geometry.
Specifically for over power events, we
have a criterion of 280 calories per gram fuel
enthalpy as a limit for a rod ejection accident, which
is the big over power event in the PWR that's
analyzed.
We have embrittlement criteria in the
regulations for the loss of coolant accident. We'll
talk about those a lot today.
There are similar limits on fuel damage
during dry storage, and these are related to creep
deformation and also to peak temperature during the
early stages of dry storage.
This work on the fuel damage limits for
dry storage is also going on in our program. I don't
plan to talk about that today unless I get questions.
CHAIRMAN POWERS: I think I would try to
keep the two separate, but it doesn't hurt to
parenthetically note if one result relates to the
storage issues.
DR. MEYER: Okay.
CHAIRMAN POWERS: Yeah, parenthetically
noting where there's overlap is fine, but I don't
think I want to go into the storage stuff in great
detail right now.
DR. MEYER: Okay.
CHAIRMAN POWERS: Stay with the real
stuff.
DR. MEYER: Okay. The safety criteria
that we used for all developed for fresh or low burnup
Zircaloy clad fuel rods. We believe for many years
that low burnup also provided the limiting conditions,
but with the movement to the higher burnup fuels and
the large concentrations of burnable poisons, you can
now see peak powers occurring later, not at the
beginning of life, but as late as end of second cycle.
So we have to take a look at the criteria
at higher burnups, and of course, we started doing
this in a general way some time ago.
The criteria that apply to these
situations were also developed for Zircaloy cladding,
and in the beginning at least there was an assumption
which seemed like a good assumption, that if the
advanced alloys improve the performance during normal
operation, that it would do so during the accidents as
well, and in some cases that may be true. In some
cases it might not be true.
But in any event, we are now looking at
high burnups and other cladding alloys to try and
confirm these assumptions or find other results if
that's what happens.
DR. CRONENBERG: Ralph, why did you think
that early ripe (phonetic) conditions were more
limiting? You didn't have much fission product
buildup. You didn't have much corrosion,
embrittlement. So what was the original thoughts on
that?
DR. MEYER: Yeah. Usually the big actor
is the power, is the linear heat rating of the fuel
rod. In a loss of coolant accident, both the stored
energy and the decay heat from short-lived species is
proportional to the power, and that often dominates
other things.
We've been looking for a very long time at
things like rod pressure and gap opening at high
burnup during normal operation because they also have
a fairly significant impact on the conditions during
a loss of coolant accident. The gap conductants is a
big player.
DR. CRONENBERG: I'm surprised at that
view because water side corrosion was an early -- you
know, a phenomenon identified early with Zircaloy,
when new corrosion was a problem.
DR. MEYER: It's just a historical fact.
DR. CRONENBERG: Okay.
DR. MEYER: I mean, we're not being
governed by this point of view at the present time.
DR. CRONENBERG: Yeah.
DR. MEYER: But this is sort of how we got
here.
Now, the status of where we are right now
is that we have burnups approved to 62 gigawatt days
per ton. This is average for the peak rod in the
core. In Europe they tend to report their license
limits in terms of average for the peak assembly,
which is a lower number by about ten percent. So you
have to keep that in mind. We are not that far ahead
of the rest of the work in our burnup approvals.
And this applies to the three major alloys
that are in use at the present time. Specific
questions now have been raised about these criteria
for postulated accidents. A long time ago we learned
from both the Cabri program in France and the NSRR
program in Japan that the 280 calorie per gram number
that we're using for the reactivity accidents is
probably not valid at high burnups.
Oh, four or five years ago we raised
question about the effect of corrosion during normal
operation, the oxide buildup during normal operation,
and how that should be added into the corrosion during
a high temperature transient in LOCA in order to
compare with the 17 percent criteria and whether there
would be some other effects.
And so we've recognized some -- and more
recently, the questions that will be addressed heavily
in this meeting by the later presenters and then in
the summary of the meeting that I'll be describing
about the possible effects of niobium on the
embrittlement criteria for loss of coolant accident.
So there are now some -- we had general
questions about whether we should be looking at the
validity of these criteria for high burnups and other
alloys. Now we have some specific questions, and
we're just continuing a broad approach to this whole
thing.
We have in our agency program plan of a
couple years ago agreed that we would not ask the
industry to do the confirmatory work for the currently
approved burnup range, that we would do that
ourselves. And that's the big mission in the research
program.
So we are specifically addressing all of
these criteria, effects of burnup and alloys for
burnups up to 62 gigawatt days per ton.
The industry has been told that they will
have to do all of those things for the burn-up
extensions above that.
In order to try and improve our progress
on the work with the NRC's confirmatory obligation, we
organized these PIRT panel meetings which I'm going to
talk about here.
PIRT is a phenomenon identification and
ranking table. You build tables of phenomena that
occur during the events that you're studying, and you
try and learn something about that by discussing the
importance in each of those, of each of those
phenomena.
DR. KRESS: Ralph, when you talk about a
burnup limit, like the 62 --
DR. MEYER: Yeah.
DR. KRESS: -- that's for the limiting
high power assembly?
DR. MEYER: That's average burnup for the
peak rod.
DR. KRESS: For the peak rod?
DR. MEYER: That's a peak rod, yeah.
DR. KRESS: Now, what does that translate
into for the average burnup of the whole core?
DR. MEYER: Well, it's a lot lower.
(Laughter.)
DR. KRESS: Yeah, I would assume.
DR. MEYER: You know, you go from the peak
rod --
DR. KRESS: The peak rod is like 1.4?
DR. MEYER: Mitch Nissley from
Westinghouse probably has an answer right on the tip
of his tongue.
MR. NISSLEY: These are very approximate.
Mitch Nissley from Westinghouse.
I would say at the beginning of the cycle
a reasonable core average burnup would be in the order
of 20,000 gigawatts or 20 gigawatt days per metric
ton, and that by the end of the cycle they're probably
in the low 30s.
DR. KRESS: Okay. That's --
MR. NISSLEY: And that's for a fairly
aggressive core design.
DR. KRESS: Okay. Thank you.
DR. MEYER: Okay. The dry storage issue,
the dry storage situation is a little different. The
task had been proved for fuel burned up to 45 gigawatt
days per ton, and we are able in our reactor oriented
programs to look at the dry storage conditions. So
these are folded into one of the big programs that
we're doing.
So we did three different PIRTs, which we
refer to together as the high burnup PIRT. One was on
the rod ejection accident. For a PIRT activity,
you're supposed to assume a very specific sequence,
and so in this case, we assumed that the rod ejection
accident occurred in TMI-1 with high burnup fuel at
hot zero power.
TMI-1 was chosen because it had been used
for an international standard problem. There were
input decks. We have done extensive analysis in our
cooperative work with IPSN in France and Kurchatov in
Russia.
So we had a lot of analysis on TMI-1 rod
ejection accident, and we chose that as the base case
for that PIRT.
For the BWR power oscillations, we chose
Lasalle-2, which had some oscillations and a lot of
analysis. So, again, there was an analytical base
that we could build on, and again, we assumed high
burnup fuel in that core.
When we came to the loss of coolant
accident, however, we did not pick a specific plant.
We didn't even specify whether we were talking about
a BWR or a PWR or a small break LOCA or a large break
LOCA.
We did, however, have discussions on each
of those. We had major presentations given to the
PIRT panel members prior to their ranking activity on
small break, large break and BWR and PWR. So all of
that information was given to the panel members, and
in the end, we decided to just go with a generic loss
of cooling accident with Zircaloy clad fuel at 62
gigawatt days per ton.
Now, I think last year at this
Subcommittee meeting we were already into the PIRTs,
and so we had talked about them. I don't go into a
lot of detail. We had about 25 fuel experts from all
over the place. The approximate sign is not because
we can't count to 25, but because it varied from time
to time, and we would tend to have a slightly
different mix of people for the BWR events and the PWR
events.
We held eight meetings, a total of 25 days
of meetings. This is really quite a large commitment
of resources to this activity. We prepared three
NUREG reports, and I think most of you, if you have
not seen the reports, you at least have had access to
them. They're quite large. They are on the Web, and
they're nearly finished. We have final draft
versions, which are out electronically to the PIRT
panel members for final comment, and our hope is to
publish them at the end of this month.
We also have a staff report which I wrote
that tries to give our interpretations of what we
learned and some suggestions about how we can move
forward with that. That is also written up as a draft
report. It's not on the Web in its final form. We're
trying to decide how to publish that at this time.
However, the three main components of that
report are on the Web. They were developed as we went
through the PIRT process as little white papers, and
they're on the Web, along with the PIRT reports.
CHAIRMAN POWERS: Well, you've been
through the PIRT exercise. Would you do it again if
you had a similar problem?
DR. MEYER: Probably. It's an imperfect
process when you apply it to a mixed situation like
this. I think the PIRT process probably works best
when you apply it to development of a computer code,
like one of the large thermal hydraulics codes, and I
believe that was the environment in which the
technique was developed.
When you apply it to a more general
subject, the we found that we had to be a little bit
fast and loose with some of the concepts and a little
bit creative in the way that we tried to put it
together.
In fact, at these eight meetings that we
held, the first three-day meeting was basically
written off as one where we just floundered around and
tried to figure out how to go forward, and we started
over again with the rod ejection accident in the
second meeting.
So there's a high cost to this because we
had people from the industry, from overseas, from all
over the place coming in largely at their own expense,
and I don't know how many times you can generate
enough interest and enthusiasm to do that.
We are trying again with the source term.
It's an important subject, and probably we'll be able
to generate the same kind of interest in the source
term.
I'm not sure that we could do this every
four or five years as a routine matter.
Also, I would say since we're on the
subject of opinions, the result of a PIRT ranking by
and large are boring. I mean, you list a lot of
phenomena and you rank each one as high, medium and
low importance with regard to some outcome, and you
usually get what you knew at the beginning.
So we got a lot of tabulated results that
just summarized what we already knew. The thing about
it was that there were for some of us in any event,
there were some surprises and some light bulbs that
went off, and this just would not have happened
without the broad discussion with all of these people
in the room.
And I think that's what made it
worthwhile. It also makes it risk because if a light
bulb doesn't go off, then maybe you've spent a lot of
money and didn't get anywhere.
So let me now try and go through these
three PIRTs very quickly. Just for calibration
purposes, the rod ejection accident occurs when you
postulate to the control rod drive mechanism, brakes,
and is ejected from the vessel by the pressure
differential.
You get a prompt critical power pulse. In
a power reactor the width of the pulse at half maximum
is about 30 milliseconds. You get the cladding
temperature rise that lags this a little bit. You get
a strong negative Doppler feedback due to the power
pulse, which basically shuts it down.
DR. KRESS: Now, is this local or --
DR. MEYER: It is local. It's localized
to several neighbors around the ejected rod, and so it
is not a core-wide event.
DR. KRESS: Not core-wide event.
DR. MEYER: Right.
DR. APOSTOLAKIS: Which one is regulatory
guide to 177?
DR. MEYER: One, seven, seven is for the
rod ejection accident. It's -- I don't know the exact
title, but it's the methods and assumptions for
analyzing a PWR rod ejection accident, specifically
for that event.
And it has the assumption of 280 calories
per gram in that --
DR. APOSTOLAKIS: I'm confused. Don't we
have a risk informed guidance on 177 as well?
PARTICIPANT: Seventy-four.
DR. APOSTOLAKIS: Five, six, seven?
PARTICIPANT: Those are the same.
PARTICIPANT: It's 117.
DR. APOSTOLAKIS: Oh, 11?
PARTICIPANT: This is an oldie.
DR. MEYER: Oh, it's very old. I think
this was safety guide 77 in the prehistoric time.
DR. APOSTOLAKIS: It's one. But you said
interesting things about PIRT, and for years now I've
been hearing people talk about PIRT in awe. What's so
big deal about it? Why are people so impressed by
PIRT? Was K used before?
DR. MEYER: That's a fair question. I
think to some extent, I think there is a little over
expectation. I've felt this from the beginning and
have tried to make the best of it, and I think we have
come out pretty well on this one because we learned a
lot.
It is not much more than a little bit of
organization in a big discussion of a lot of experts.
So it's a way of getting experts around to get their
opinions in a more or less organized way. That's what
it turned out to be for us.
CHAIRMAN POWERS: George, I would say that
in this context and having attended one day of one of
the PIRT discussions --
DR. MEYER: I hope it wasn't the first
one.
CHAIRMAN POWERS: No, in fact, it was the
second one, but I think this floundering that you
encountered on the first one is typical even among the
thermal hydruaulicists when they undertake a PIRT.
The first round is always a bunch of floundering
because you're asking everybody to get on the same
page at the same time, and that's difficult because
they come in with different imperatives in which their
expectations are.
But it seems to me that when you're
struggling to understand how to approach a problem
that is calling into question things that are as old
as 1.77, and it's not a question of is it 280 calories
or 220 calories or 100 calories. Is the whole concept
any good or not?
When you're struggling with that, you want
to get the best people to look at it and say, yeah,
you've thought about all of the things that are likely
to be important.
Now, they can be flat wrong because they
don't have a great deal of experience working in this
regime, but you're confident that you've tapped into
as much knowledge as you're likely to have in setting
up and planning something.
Now, the idea is that you go on and you do
some research and some experiments and things like
that, and you're going to learn more about it, but at
least you start off knowing what you ought to be
looking for.
DR. APOSTOLAKIS: And there is consensus
at the end? You said that there is a ranking of high,
medium and low, and so on and so forth. Are we at the
end of this phenomenon?
DR. MEYER: We had a very large panel,
atypically large, and we're told by our panel
organizer, Brent Boyack, who's done a lot of these,
that typically with the panels on the order of six to
eight people, that they do, indeed, reach consensus
just naturally on these.
We did not, and we did not attempt to
reach a consensus. Instead we voted, and we recorded
the votes and the rationales, and so you tend to get
a distribution of answer, high, medium, and low, and
often there's a sizable majority, and you can go and
look and see what the reason was for that and why some
other people didn't quite agree with it.
DR. APOSTOLAKIS: That assumes, of course,
that everybody's vote is equally important.
DR. MEYER: Well, you know, we even
addressed that. We asked the PIRT panel members to
vote only when they felt that they had a good basis
for voting and that we didn't expect them to vote on
every item because we had a range of subjects from
analytical to experimental, and so there was some
restraint on that.
DR. KRESS: If you had a split vote, 16 --
DR. MEYER: If you had a what?
DR. KRESS: Sixteen of your members voted
high and the rest of them voted it low.
DR. MEYER: Yeah.
DR. KRESS: Would that automatically make
it high? Is that the way you would have ranked it?
DR. MEYER: What we did in the end was we
agreed on some -- I forget what we call them -- but
some scoring criteria, and we went back and had a
little formula for deciding two things about a
particular phenomenon. If it was important and if it
was well known because we would address both of those
at the same -- you know, in the same discussion, and
what you're really looking for are things that are
believed to be important and not well known, and those
are the items that you ought to focus on.
And the tables are so large that we
developed a little formula and put the numerical score
in the table. So you could run down the table and
pick these out.
And that's exactly what I did in
developing this implications report that I prepared,
was I went down the tables, and I skimmed off the
items that were of high importance and not well known.
You also sometimes find something from the
inverse of that. You look for a subject that is not
thought to be very important that you might have felt
was important, and I have one of those on this list.
DR. KRESS: The final product is you're
looking for where you need more research or finally
decide --
DR. MEYER: Well, some people would use it
that way. What I was looking for was insights on how
I could plan a way to resolve the issue, and it
involved doing additional work, but it also involves
a method to get there. So that was -- I mean, you
could do a lot with the PIRT, and the information is
all recorded. So you can do other things as well, but
that's what I tried to do with it.
So let me try and move through this now,
and you'll look at some of these items here and see
that they're perfectly expected results, but not all
of us knew all of these things at the outset.
The first one, for example. I have to
confess that I saw this as a little bit of a surprise.
I always thought that, you know, the energy deposition
was just a function of something that could never be
changed, and if you went over 280 or 220 or 100,
whatever it was, you were just out of luck.
But core designers know that that's not
the case. You can design the core. You can put high
burnup rods near or far from high worth control rods
and do other things.
Another thing where a real light bulb went
off had to do with that discussion and with the
calculations that David Diamond was doing for us on
the rod ejection accident.
We have believed for some time now that
the 280 calorie per gram number should come down in
the neighborhood of 100 or 80 calories per gram for
high burnup fuel, and so we asked David Diamond to do
calculations of a rod ejection accident where he gets
100 calories per gram deposited in the fuel rod.
And so he makes the presentation to the
PIRT group members, and somebody asked him what
control rod worth did you assume, and he says, "Two
dollars."
And you hear a chorus of utility people
and others say, "There's no way you can have a control
rod worth two dollars and 50 calories per gram,
$1.20." Well, maybe.
And so the idea comes up that perhaps for
screening a large number of operating reactors, the
current ones up to the current burnup limit, that
maybe we can do some generic calculations based on
some enthalpy limit in the range of 80 to 100 calories
per gram, discover something about the core design
that you would have to have in order to achieve that
energy deposition, and then use those to screen the
reactor population.
And if, for example, you have to have two
dollar control rod worth, and NRR knows for sure that
we don't have two dollar control rod worth out there,
then you're done.
DR. BONACA: I have just a question. Did
the group discuss the high level objectives that set
the --
DR. MEYER: Yes.
DR. BONACA: -- pure enthalpy limit?
DR. MEYER: Yes.
DR. BONACA: I seem to remember in ancient
times as you said one of the concerns was challenge to
the vessel.
DR. MEYER: Yes, we did, and this is where
that first meeting went, and so I probably shouldn't
characterize it as a waste of time, but we started out
considering the general design criteria.
There are two general design criteria that
govern these two event, 23 and 27 or something. I
forget the numbers, but one on the LOCA and one on the
rod ejection and rod drop accident. And they talk in
terms of maintaining coolable core geometries, of
pressure pulses that don't damage the vessel more than
just a little bit of yielding or something like that.
And for the first couple of days we
decided how we could adopt those directly as the high
level criteria for the ranking exercise, and a
conclusion from that discussion was that was going to
be really difficult because neither the codes that we
were looking at, nor the experiments we were
considering would take you all of that distance.
We were not looking at codes that
calculated the coolability of a debris bed, and we
were not looking at experiments that would get
pressure pulses large enough to threaten a pressure
vessel.
And so as a practical matter, we backed
down to another level, which seemed to be
conservative, but workable, and probably not
penalizing in any significant way, and we ended up
using a concept of fuel damage with significant fuel
dispersal.
So we know that there's going to be some
fuel damage, and that's not a problem, but it's the
fuel dispersal that's the problem, whether you're in
a loss of coolant accident where you fragment the
cladding and you lose the structural geometry of the
core, you get fuel spilling out or in a very high
energy rod ejection accident you actually expel fuel
through the cracks.
And so those were things that could be
addressed with the codes and the experiments that we
were talking about until we settled down to that
level, and we used that throughout.
DR. CRONENBERG: I think you might want to
respond. Didn't you have a tutorial? Even though
these were experts, there were some tutorials on -- by
like Phil MacDonald -- on experience, fuel behavior
experience for the various accidents; is that correct?
DR. MEYER: Yes, that's right.
DR. CRONENBERG: For each one of these?
DR. MEYER: We tried to do this with each
of the PIRTs. We would start out the PIRT discussion
with two or three tutorials. Phil MacDonald gave one
of them on the reactivity accidents.
David Diamond back here in the audience
gave one on the same subject.
Larry Hochreiter gave a couple on PWR loss
of coolant accidents.
Jens Andersen from GE talked about LOCAs
and also about the power oscillations.
So we had a lot of tutorials. We, in
fact, used a court recorder for most of the sessions.
We captured the tutorials on transcript, and we took
the transcripts and edited the transcripts, send them
back to the authors, the presenters for editing, and
included a select number of those presentations as
appendices in these PIRT reports.
So those tutorials, some of them, are in
the PIRTs.
DR. KRESS: Ralph, how many calories per
gram does it take to go from normal operating
temperature up to fuel melt temperature?
DR. MEYER: It takes -- fuel melting is
about 267 calories per gram and normal operating fuel
enthalpy is -- it's in the range of 15 or 30. So it
takes a lot, 230 or 240 to get to melting.
And you know the technical background
here. Originally with fresh materials we thought that
you had to start melting something to get some real
action, and with high burnup cladding, you see a
completely different mechanism come in where the
expansion of the pellet against the cladding, which
has lost a lot of its ductility results in splits, and
you also then have the gassy microstructure of the
pellet, which can blow particles out through these
splits.
So that's the kind of thing we've see.
Well, okay. Some other results of the PIRT was the
majority thought that you needed to run tests in the
burnup range that you were really looking for because
part of the action is in the cladding, but part of the
action is in the pellet, and even if the properties of
the cladding are dominated by oxidation or hydride
distribution, the loading is going to be determined by
the pellet, which is affected by burnup.
We talked a fair amount about testing the
MOX rods because of plutonium enriched agglomerates.
This was a subject where there really wasn't any big
change in views because we all knew this going in, and
we knew it coming out.
Testing in the right coolant environment,
we talked about that before, and that came out highly
ranked.
This one is a little bit of a surprise for
the reactivity accidents, the PIRT panel members
didn't think that the alloy was such a big deal, but
this was in the context of did you have to run an
integral test like in the Cabri reactor or the NSR
reactor. Did you have to run those tests for all
different alloys?
And their thought was, no, probably not.
As long as you knew the relative mechanical
properties, you could extrapolate from some base case,
and so, in fact, the cladding alloy was not ranked
high, although you might have expected it.
Also, near the end of the discussion of
the rod ejection accident, we realized that there may
be some of the newer alloys which have so much
ductility even at high burnup that they don't fail by
this pellet cladding-mechanical interaction, and in
those cases, then you would be able to go on up to
higher energy depositions before you failed, and that
the phenomena that would come into play would be more
like the high temperature transient effects in a loss
of coolant accident.
And we have some experience with the
Russian cladding that showed that. The E110 Russian
cladding that was tested in IGR reactor and later with
short pulses in the BIGR reactor always shows
ballooning type deformation and gas pressure rupture
rather than a PCMI, even at 55 or 60 gigawatt days per
ton. The stuff is very ductile.
DR. BONACA: I am still surprised that you
did all this work and there was no linkage to some
high level objectives as discussed before. Two,
eighty used to be, if I remember, was a true
threshold. If you demonstrated that you were below
that, you didn't have to consider effects on the
vessel. For example, the pressure pulse that may
cause a challenge to the vessel were all issues of
coolability, too.
I understand what you're doing. You're
trying to say, well, you know, pragmatically let's go
to a lower value. To accomplish what? I mean, it's
not clear yet that you have linked a value, whatever
value you're searching for, to a high level objective
such as coolability or pressure pulse.
And without that, you could always have
the industry coming back and saying, "Well, I want to
go to 70,000 or 80,000 megawatts per metric ton," and
there is no basis for 100 calories per gram.
DR. MEYER: Yeah, yeah. Well, we talked
about that, and we decided as a practical matter to
tie it to fuel dispersal. If you don't have fuel
dispersal, you're not going to have pressure pulses
because you won't have a fuel-coolant interaction.
DR. BONACA: Okay.
DR. MEYER: And you won't lose coolable
geometry. So we tied it to fuel dispersal, and I
think there was a general belief that if you work with
an enthalpy level that corresponds to fuel dispersal,
that you will always be able to get under that
comfortably and won't be penalized.
DR. BONACA: Oh, okay. So you have a
linkage to that. I mean --
DR. MEYER: There is. Yes, there
definitely is.
DR. BONACA: Because I haven't heard the
NUREG so I don't know, but all right.
DR. MEYER: Okay. Now, I can't remember
whether I discussed this last year or not. So I'll
just go through it very, very quickly, but the idea
now to bring some resolution to the reactivity
accident is, first of all, to improve an empirical
correlation that we have, and you've seen it before.
I've stuck it in as the next slide. This is what we
call our paint brush slide. It's not really a
correlation yet. It's just sort of a failure map of
the tests that have been done.
But it's that kind of a plot that we would
look at and try and draw some boundary between
survival and failure, looking at enthalpy increase as
a function of either oxide thickness or some
fractional oxide cladding thickness to accommodate
different cladding diameters.
DR. KRESS: What do you do with those
black dots that are below the line?
DR. MEYER: Yeah. Well, this is kind of
reminiscent of NUREG 0630 and the ballooning and
rupture data from before. You have to know the
personality of these data points to realize that these
things ought to be moved up on the plot.
Those were tests in NSRR. They were
tested at room temperature. The accident isn't at
room temperature. It's at hot zero power, which is
pretty hot. It's about 280 or 300 degrees Centigrade.
So there's a big ductility.
DR. KRESS: It just tells you you've got
the wrong parameters plotting.
DR. MEYER: Well, in the past in the NSRR
reactor, they've only been able to test at room
temperature because they didn't have a high
temperature capsule, but now they're building a high
temperature capsule.
And one of the things that we want to wait
for are some data from the high temperature capsule
because if they can quantify how much too low their
room temperature test was, then we have a basis for
bringing these up.
Here's another one. This is REP Na-1.
This is the very first test done in the Cabri reactor.
Then intense discussions going on still to this day.
DR. KRESS: That's the one that got
everybody excited.
DR. MEYER: Got everybody excited, and it
probably is an anomalous result. I think we
understand this one now. The understanding that we
believe we have is not universally accepted, but it
looks like that the precondition of that fuel rod was
at such a high temperature that it caused hydride
redistribution that affected the ductility.
We've been looking at that at Argonne
National Laboratory and have been discussing it as
recently as two weeks ago, a full day meeting, and
it's very controversial because this was a pitfall
that was recognized.
When they prepared this rod, they realized
that they shouldn't take it up too high in temperature
before the test and thought they had kept the
temperature low enough, and the only thing we can
conclude is either their temperature measurement
wasn't real good or we just didn't quite understand
where this boundary was because it seemed inescapable
when you look at the microstructures before and after
the test, that the hydrides were redistributed before
the test.
DR. KRESS: That's why I thought maybe you
had the wrong parameter. Oxide thickness must be a
surrogate for --
DR. MEYER: Oxide thickness -- well, it's
largely the hydrides that affect the ductility in this
temperature range, and the --
DR. KRESS: -- and ductility of the
remaining material in the clad or something.
DR. MEYER: You've got a little bit of
LOCA thinking coming into that question about the
remaining metal thickness. It's --
DR. KRESS: Well, those are only microns,
aren't they? Yeah.
DR. MEYER: Yeah. This is the --
DR. KRESS: Pretty much.
DR. MEYER: This is the corrosion. This
is the amount that was accumulated during normal
operation, and approximately 15 percent of the
hydrogen that is released during the dissociation of
steam that results in the oxidation. So about 15
percent of the hydrogen that's formed is also
absorbed.
DR. KRESS: So it's a surrogate for the
amount of hydrogen --
DR. MEYER: That's exactly right.
DR. KRESS: Okay.
DR. MEYER: That's exactly right.
DR. KRESS: Thank you.
DR. MEYER: It's easy to measure the oxide
thickness. It's hard to measure the hydrogen
concentration. It's a surrogate for hydrogen.
CHAIRMAN POWERS: I guess what puzzles me
a little bit about the discussion of REP Na-1 is,
okay, these guys tried very hard not to redistribute
the hydrogen, but despite their best intentions, they
did.
DR. MEYER: Yeah.
CHAIRMAN POWERS: Okay. Does that mean
that hydrogen can never be redistributed in a real
core?
DR. MEYER: Well, it is distributed in the
real core in a very characteristic way because you
have a temperature gradient across the cladding and
the hydrogen congregates to the cooler outer shell,
and this tends to embrittle the rim of the cladding,
but leave a lot of ductile material underneath, and
when you look at the fracture surfaces, this is
exactly what you see.
You see a blunt cracked tip through the
hydrided rim, and then a 45 degree shear through the
ductile part of the cladding, and what you saw in REP
Na-1 was a blunt cracked tip throughout the specimen.
It's the only one that looked like that. It's the
only specimen that they took the temperature up to 390
degrees Centigrade during preconditioning. All of the
rest were kept at much lower temperature.
I don't know if there are any conditions
in the reactor that could do that. What we are asking
ourselves though is if there are conditions during
vacuum drying for storage which could cause this to
happen because this redistribution happens when you
don't have the normal pellet expanding putting stress
on the cladding, and in the storage casks when they
dry them, you get -- I don't know the exact numbers,
but I've heard them talk about numbers in excess of
400 degrees Centigrade sometimes.
And so I think one of the things that we
have fed back from this experience into the dry
storage work that we're doing is to look specifically
at the ductility of this material after it's gone
through a range of vacuum drying conditions, in
addition to just looking at the creep rupture, which
is what is currently used to get the limits for dry
storage.
CHAIRMAN POWERS: The redistribution of
hydrogen that you're talking about, it's really an
equilibrium phenomenon. It's driving itself from
being dispersed hydrides along the grain boundaries
into a more coherent hydride to reduce surface area of
hydrides.
So, I mean, the hydride redistribution
that you want, I mean, it wants to do this, and it's
just a question of whether you have enough temperature
and time for that to accomplish.
DR. MEYER: That's right.
CHAIRMAN POWERS: So there's a time-
temperature tradeoff here.
DR. MEYER: Right, right.
CHAIRMAN POWERS: And it's not clear to me
that you don't have time even though you might have
modest temperatures --
DR. MEYER: Yeah.
CHAIRMAN POWERS: -- to accomplish that in
a real reactor.
DR. MEYER: Yeah.
CHAIRMAN POWERS: In which case it would
not be an anomalous point. It would be characteristic
of a point where there had been redistribution of the
hydrogen.
DR. MEYER: Well, the only thing I can say
is there have been a lot of rods looked at out of the
reactor, and they have this characteristic high
hydrogen concentration near the OD. They do not look
like this one did.
DR. KRESS: The higher burnup implies
they're going to stay in there longer.
DR. MEYER: Implies that?
DR. KRESS: Those high burnup rods will
stay in there longer and will have more time to
potentially redistribute the hydrogen.
DR. MEYER: Yeah, well --
MR. SCOTT: Ralph, also the orientation.
I mean there's always hydrogen, but sometimes the
orientation of what the hydrides look like --
DR. MEYER: Yeah.
MR. SCOTT: Is that part of it?
DR. MEYER: It certainly can be part of
it, but in this case, Hee Chung (phonetic), who is
examining this issue, has not made the reorientation
a big issue. The orientation of the hydrides is
affected by the street that you apply to the cladding
when it's hot enough for the hydrides to be mobile,
and he's not arguing that they reoriented from
circumferentially aligned stringers to radially
aligned stringers, which right away will really ruin
your ductility.
There just seems to be a redistribution,
a sort of homogenization of the hydrides. They are no
longer all packed up on the OD, and there are a few
radial ones, but it's not predominantly radial.
It just looks like you annealed it and
gave it a chance to relax the highly organized
distribution into a more random distribution.
DR. KRESS: Those are predominantly axial.
You said circumferential.
DR. MEYER: When you look at them in
cross-section, they are stringers around the
circumference.
DR. KRESS: They are circumferential?
DR. MEYER: Yeah. So to try and wrap this
one up, what we want to do is improve the correlation,
to get mechanical properties for all three of these
because the correlation is predominantly Zircaloy, and
so we have to have the relative mechanical properties
of all of these, use our FRAPTRAN code to try and make
the adjustment for the mechanical properties
differences, and then use the three dimensional
neutron kinetics code to do the plant analysis and
hopefully relate some enthalpy limit to control rod
worth or some other parameters that could be easily
used to screen the core.
DR. KRESS: Well, does FRAPTRAN deal with
the hydrization of the plant?
DR. MEYER: That's going to be just
imbedded in the mechanical properties. The mechanical
properties are being measured under the conditions --
DR. KRESS: You'll input mechanical
properties.
DR. MEYER: That's right, and the
mechanical properties for the reactivity accident,
which compared with the LOCA these are low
temperature, high strain rate, whereas the LOCA are
going to be high temperature, low strain rate.
The mechanical properties for ZIRLO and M5
are going to come from the Cabri program. We have a
commitment from ENUSA in Spain to provide a ZIRLO rod
for testing in Cabri and a commitment from Framatome
in France to provide an M5 rod, along with the
permission to do mechanical properties testing on
these and provide all of that to the participants in
the Cabri program.
And these, there will be one test of each
of these in 2002. That's next year, in the sodium
loop.
CHAIRMAN POWERS: And one test, and the
uncertainty in the outcome is?
DR. MEYER: I'm sorry?
CHAIRMAN POWERS: What's your uncertainty
in your outcome when you have one test?
DR. MEYER: Large, but we have --
hopefully we'll have ample mechanical properties
measurements, and we'll have other tests. We have all
of these other tests with Zircaloy.
I know it's not going to completely
satisfy you in terms of the quality of this
correlation, but what my proposal is to my office,
which is trying to resolve this issue, is that we go
ahead in 2003 and try and go through the exercise and
see if we get an answer that's favorable.
I think the answer is going to be
favorable. This is one where we now have enough
information to have a "seat of the pants" idea of
where it's going, and hopefully the margin will be
enough that we can discuss the uncertainties and see
where we are.
The reason for pressing to do this in 2003
is that there's going to be a three-year delay before
the water loop starts, and I think it's better for us
to go ahead and try and go through the resolution with
what we have from the socium (phonetic) loop and from
NSRR and hopefully a few tests and a high temperature
capsule from NSRR.
We're going to be on a plateau of
understanding for at least three years, and so we
might as well go ahead and try and go through the
exercise, see if we can finish it off, and then when
we get to the water loop if we see any surprises, then
we'll go back and make an adjustment.
DR. CRONENBERG: So what is it, 2003 you
go to the standard review plan and say for 50,000
megawatt days per ton, the enthalpy will be 100
calories per gram and anything less it remains 280 as
in the original review plan or what?
DR. MEYER: I can't say that that's what
we would do. What I'm saying is that in 2003 that the
Office of Research will try and write a paper of some
sort that says we have assessed the operating reactors
with the current fuel up to the current burnup limit,
and we have this database. We think the enthalpy
limit -- a reasonable enthalpy limit to use for this
is such-and-such. We've done the neutron kinetics
calculations. Everything is honky-dory. We have some
big uncertainties. There will be some additional work
in the future to look for mistakes. Case closed,
and --
DR. CRONENBERG: But case closed means we
remain with 280 calories per gram?
DR. MEYER: That would depend on how I
think NRR wants to handle this, and we haven't had any
discussion on that. How you implement this into the
regulatory framework is another step. At the moment
I'm just talking about establishing the technical
basis to do it.
I would expect during the same time period
that the NRR will address the regulatory guidance and
maybe even the Office of Research might be asked to do
that. I just don't know.
DR. CRONENBERG: There's things on the
docket now that are kind of pressing, like the power
upgrade for I don't know if it's Commonwealth Edison
anymore, but the Dresden, Quad Cities. They're going
for 17, 20 percent power upgrades with extended fuel
burnup. I think with the new GE design to above 50 or
55, maybe even 62. So where does research come into
play with NRR that NRR has to review these
applications?
DR. MEYER: Ralph Caruso from NRR wants to
answer your question.
MR. CARUSO: I just wanted to make the
comment about the power up rates. The power up rates
for the BWRs do not involve raising any of the burnup
limits above 62,000. They do not involve changing any
of the burnup rates for any of the fuel.
DR. CRONENBERG: Okay. I guess it's more
on the power oscillations when we get to the BWRs, not
this rod ejection, but still I'm sort of seeing how
the research falls into near term licensing, licensing
amendments.
MR. CARUSO: Well, right now what we're
doing is we're following the work that's being done by
the Office of Research, and we take it into account as
we make our licensing decisions.
But right now none of the power up rates
involve any changes to any fuel licensing limits.
We've not changed any fuel licensing limits to
accommodate the power up rates.
DR. CRONENBERG: So you look at the
standard review plan as it is written right now, and
that's what you base your review on, the 280 calories
per gram. If PWR comes in, what is it? Two, thirty
or BWR? It's all based upon the old standard review
plan.
MR. CARUSO: The vendors have approved
methodologies for their existing fuel designs, and
they are going to continue to use those approved
methodologies to analyze the behavior of the plants at
the higher power levels, and as long as they continue
to meet the standards that have been already approved
at those higher power levels, we'll find them
acceptable.
MR. ROSENTHAL: Yeah, Jack Rosenthal,
Research.
You have to do this very piecemeal. Okay?
For the ejected rod, if you say that the limiting
ejected rod action is at hot zero power because at hot
full power you have far less rods in the core, then
the fact that when you are at full power you're going
to be running at a higher power doesn't enter into
that hot zero power calculation.
Like I said, you just have to piecemeal it
through, you know, think it through event by event and
what's limiting with.
CHAIRMAN POWERS: It's what I'm still
wrestling with a little bit, Ralph, is how one selects
the fuel and clad combination that one would test.
Grant you you cannot test all conceivable clads, all
conceivable fuels, all conceivable degradations of
that clad, and you get around that by saying, well,
I've got these computer codes that are going to allow
me to extrapolate and interpolate within the data set
I've got, but the question comes up: which one do I
test?
Do you test a representative piece of a
rod, or do you test the worst piece of a rod?
DR. MEYER: We have done both, but we're
generally focusing now on the worst piece of the rod.
The worst piece of the rod is -- well, the one that we
select is the uppermost span between grids where the
power is still level. So we don't take the end where
you have a big power gradient, but we take the next
one. It's from the hottest elevation in the core. It
has the highest oxidation on it of the other grids,
and those are the ones that we almost always select
now.
We had some interesting -- we had three
pairs of tests. If you go back and look at both the
NSRR and the Cabri test, you can find three pairs of
tests were -- Span 5 and Span 3 were tested, and each
of those three pairs, the Span 5 failed, and the Span
3 didn't fail. They had exactly the same burnup
level, but their oxide thicknesses were quite
different.
CHAIRMAN POWERS: Okay.
DR. MEYER: Okay. Can I go on to the --
CHAIRMAN POWERS: Please.
DR. MEYER: -- the next one? I'm a little
anxious about the time here.
CHAIRMAN POWERS: Well --
DR. MEYER: But I'll go on.
So the next PIRT that we did was for
boiling water reactor and for power oscillations that
were not stopped by a SCRAM, and this is an accident
for which we do not have clear regulatory guidance,
but for which GE has in the past done some analysis
and have used the same 280 calorie per gram limit to
show adequacy in this analysis.
And that limit probably -- it either
suffers from the same problems that it does for the
PWR or maybe it's not appropriate at all for this
event, and so we just worked our way through this
event with some interesting understanding of an event
that hasn't been understood very well before, at least
from the point of view of fuel behavior.
Just a few basics. The accident that we
considered started at about 85 percent power, and the
recirculation pumps tripped, and then you got some
oscillations and you didn't get a SCRAM. So the
oscillations build.
Now, the oscillations come at about three
second intervals, and this three second interval, two
to four seconds what's seen in all of the analyses
that have been done.
It takes about eight seconds for a fuel
rod to transfer its heat out. So this is less than
the time constant of the fuel rod. So if you look at
this part, it looks like the rod ejection accident on
a small scale. These little pulses have about 15,
one, five, calories per gram in them instead of 50 or
100 calories per gram.
And so they cause the cladding temperature
to start warming up, and it starts to cool down, and
it warms up again, and pretty soon it gets to a high
temperature, and the experts expected that you would
get to a point where you would dry out and you would
not rewet.
And now you had a transient that looks
something like a LOCA transient.
So the opinions and insights that we got
from discussing this accident are highlighted here,
was nearly a unanimous feeling among the experts that
you would not get failure by this mechanical
interaction of the expanding pellet pushing on an
embrittled cladding because the energy was just too
small in that pulse, and by the time you get to the
second pulse the cladding is now heated up and it's
more ductile, and so forth.
They did expect that you would eventually
get a high temperature transient during which you
would have oxidation, high temperature oxidation,
something like you have in a LOCA, and you might even
have ballooning and rupture depending on the pressure
in the rod.
The BWRs I don't think tend to run quite
as high a differential pressure as the PWRs, but I
believe they do use the same liftoff criterion. So
there can be a positive pressure differential, but it
might be a negative pressure differential.
If you get this kind of high temperature
excursion with oxidation, you would get classing
embrittlement just like you do in the LOCA. There was
a fairly lengthy discussion about what bad things do
we have to worry about. Do we have to worry about
embrittlement of the cladding? Do we have to worry
about melting of the cladding? Do we have to worry
about melting of the fuel pellets?
It was decided that we don't have to worry
about melting of the cladding or melting of the fuel
pellets because you're going to embrittle the cladding
at a far lower temperature than those two events, and
so what we really have to look at it embrittlement of
the cladding.
I did not expect runaway oxidation. We
had a number of discussions on that. There doesn't
seem to be any magic temperature at which you get some
autocatalytic reaction that runs away. It's simply a
matter of heat balances, how much heat from the
chemical process and how much can you pull away?
And it was not thought that that would be
a problem, particularly since we're going to run into
our problem at a fairly low temperature. Well, fairly
lower temperature means around 1,000, 1,200 degrees
Centigrade.
And it was further thought that LOCA-like
criteria may be even the LOCA criteria, might just
apply to this transient.
DR. BONACA: I assume that this event is
bounding with respect to a drop for BWR?
DR. MEYER: We decided to focus on the
power oscillations a couple of years ago when we did
our little agency program plan Commission paper, and
we focused on this as a result of our perception of
the risk.
We looked at the probability of occurrence
and the risk, and what we know is the power
oscillations without SCRAM are a -- I don't want to
overstate it, but they're a significant risk
contributor in BWR PRAs, whereas the rod drop is not.
The rod drop is of very low frequency.
So we just focused on this one. I think
that, in fact, a lot of what we learn for the PWR rod
ejection accident in terms of fuel behavior and damage
limits can be transferred, but not all of it because
the Japanese continue to study BWR power pulse events
and have recently looked at some high burnup BWR
cladding in their NSRR reactor and find unusual
behavior that hasn't been seen before that seemed to
be related to the bonding between the pellets and the
classing, which in the BWR cladding that they were
looking at has this soft zirconium liner on the ID.
So, you know, working is going on on
things that aren't at the center of focus for some
regulatory agency, and we're plugged into it.
DR. BONACA: The reason why I asked it,
yeah, was that maybe embrittlement is not the issue if
you have that kind of transient.
DR. MEYER: Well, I guess it might not be,
but the group of experts thought that that was going
to be the issue, and so following that --
DR. BONACA: Even for rod drop? Okay. I
just --
DR. MEYER: Well, for this -- well, look.
For the rod ejection accident, embrittlement is a
different -- it's embrittlement from a different
temperature range from a different cause, but it's
still embrittlement.
Anyway, now --
DR. SHACK: But you're not proposing to
use LOCA type embrittlement criteria for a BWR rod
drop. I mean --
DR. MEYER: Not for BWR rod drop.
DR. SHACK: You got rid of that on your
frequency argument.
DR. MEYER: Right, right. I think what we
tend to do is if BWR rod drops continue to be
analyzed, you'd probably use the criteria that emerge
from the PWR
DR. BONACA: Okay. Because, I mean, right
now still in the FSAR if you were licensing a plant
today, you would still have to analyze rod drop.
DR. MEYER: Right.
DR. BONACA: Not necessarily power
selection. That's why I was leaving that --
DR. MEYER: Again, this is some decision
that NRR would make and that --
DR. BONACA: So you would have to infer an
equivalent temperature or enthalpy, the position from
the PWRs, and I was intrigued by that process, how you
would go from one to the other.
DR. MEYER: I think it would make sense to
use the criteria that are developed for the PWR for
the BWR rod drop.
DR. BONACA: Okay.
DR. MEYER: Although there may be some
differences because of the cladding.
Now, for the power oscillations, we are
still lagging behind on attacking this issue. This is
the one that we know the least about and that we're
doing the least on, but it looks like that resolution
of the power oscillation question is going to depend
largely on analysis. We're going to have to calculate
our way through a high temperature transient and look
at dry-out and rewet and cladding oxidation.
We have talked to JAERI, the Japan Atomic
Energy Research Institute, about doing some repeated
pulse test just to confirm that the pulse part of this
isn't playing a role, and hopefully they'll be able to
schedule a few tests like that in the next two or
three years.
DR. UHRIG: That would be in that three-
year reactor that they have?
DR. MEYER: Yes, yeah. We talked at
length about the test, and they don't have to do them
every three seconds. They might do them every three
days. They just do one, leave it in there, raise the
temperature up a little bit and do another one, and if
you do two or three of these, you can probably see
what is going to happen or not going to happen, and so
that's the kind of repeated pulse testing that's being
talked about for NSRR.
Halden has done a number of dry-out tests,
and are interested in doing a test specifically
planned for this BWR event. We're trying to help plan
that test. I wouldn't say that we're very far along,
but the capability is there. The interest in the
project, in doing this kind of testing is there, and
if we can get our act together and define a good test,
I think they will do the test as part of the joint
program.
CHAIRMAN POWERS: When you think about
these ATWS and the embrittlements that occur, do you
think about the ATWS processes?
DR. MEYER: I'm sorry. I didn't
understand you.
CHAIRMAN POWERS: The ATWS recovery
processes, you know, where you drop the core down and
then try to promote mixing by raising the coolant
level back up.
DR. MEYER: I'm afraid the only thing that
we considered was that some time the process, the
oscillations would stop, but we did not look at the
process of stopping in any detail.
DR. KRESS: You probably don't do much
more oxidizing of the clad.
CHAIRMAN POWERS: It's not the oxidizing
that I'm worried about. You know, bring the core down
and then bringing it back up to prolonged mixing where
you must be putting some sort of forces on the clad.
DR. KRESS: Yeah, looking at forces on it,
okay.
DR. MEYER: Yeah, but see, these are
exactly the considerations that we're talking about
now for LOCA. What are the forces on the rods and how
do you cover? And we'll get to that in just a few
minutes.
DR. KRESS: It looks to me like, Ralph,
with the frequency of these oscillations for BWRs
being what they are the only difference between that
and the single pulse is just the integrated energy
that you put in, other than how you deal with it
otherwise, put different forces on it.
DR. MEYER: Well, the second thing is only
the first one is going to take place with cold
cladding.
DR. KRESS: Yeah, and then you're heating
up.
DR. MEYER: And then you're heating up,
and you're less vulnerable to the brittle failure.
DR. KRESS: Yeah. So I think this would
be amenable to calculation rather than --
DR. MEYER: Yeah. Well, that's what we
hoped, and the code, the code that we're hoping will
solve this is a combination of our FRAPTRAN code and
a code you might not have heard of before called
GENFLO, which is a Finnish, sort of a utility
thermohydraulics code that has been coupled in Finland
with FRAPTRAN more or less specifically to do this
calculation.
Keijo Valtonen, who is known by a number
of people here at NRC as the principal person at STUK
in Finland who is doing this with support from their
laboratory at VTT, and just a couple of weeks ago I
was given two reports on the progress of this, and I
want to say to you that this is a completely voluntary
effort on the part of the Finns. We don't even have
a formal agreement with them on this, but we have been
working cooperatively with them on a voluntary basis
for four or five years.
They're doing actually more work on this
than we're doing, and so, you know, if you have any
interaction with people from Finland, tell them the
research people certainly appreciate this.
MR. ROSENTHAL: Let me just make the point
that you know, you use the systems code like RELAP or
TRACK to drive a hot channel code, to drive a fuel
code in an integrated, you know, sequence and
calculation. When we're all done, we're still going
to have to sit back and say what do we really know,
and we're planning that.
And so that, you know, I mean, it's still
a piece of work to do, and we shouldn't be dismissive
of it. I mean, we'll do the work, but it's --
DR. CRONENBERG: Can you run fuel codes or
do you use still contractors to do most of your
FRAPTRAN or can you do it in house now?
DR. MEYER: We do it in house.
DR. CRONENBERG: Okay.
DR. MEYER: I don't want to oversell
either the capability of the code or our in house work
at this time, but we do run both of the codes, FRAPCON
and FRAPTRAN. We are running LOCA scenarios and ATWS
scenarios in house and at the lab.
DR. CRONENBERG: And at PARCS is Purdue
still doing that or you guys can run that yourself?
DR. MEYER: Gee, I don't know whether
anybody on the staff can run it, but David Diamond at
Brookhaven is doing the rod ejection calculations for
us. PARCS is a Purdue University developed code, but
it's run other places, and David runs it at
Brookhaven.
MR. ROSENTHAL: And as we speak, we're
moving PARCS into TRACK M as an integrated product.
DR. CRONENBERG: So you'll be able to run
that in house.
MR. ROSENTHAL: Yes, sir.
DR. MEYER: Okay. I'm well behind now.
So let me move on and talk about the loss of coolant
accident where we have both embrittlement criteria and
evaluation models. EM stands for evaluation models.
PCT is peak cladding temperature, and ECR is
equivalent cladding reactant. That's the jargon of
the LOCA trade.
The PIRT tables for the loss of coolant
accident were extremely long, and I only skimmed off
a couple of things of interest here. One was it
surprised me that these fuel experts who had also some
experience with the large system codes -- at least
some of them did -- they identified a lot of thermal
hydraulic models that were of high importance and not
well understood, and these are the traditional thermal
hydraulic models that are in our LOCA code.
CHAIRMAN POWERS: You don't even need to
understand the momentum equation.
(Laughter.)
DR. MEYER: So I just had to point that
out.
CHAIRMAN POWERS: We're desperate.
DR. MEYER: They also found that for the
loss of coolant accident that the cladding type was
very important, but the most interesting result of the
discussions on the loss of coolant accident was the
second bullet where George Hache from IPSN in France
got up and gave us a summary of our own U.S. history
of the development of the ECCS criteria and reminded
us that the embrittlement criteria, these numbers
2,200 degrees Fahrenheit and 17 percent oxidation
were, in fact, based on ring compression tests made by
Hobson in the early '70s, and that the quench tests
were only confirmatory because there had been a lot of
discussion about whether the quench tests could
reasonably represent the axial forces or other
constraints that might be on a fuel rod during the
quench.
I should say that another way. The
discussion was that the external forces on the fuel
rod, whether they come from the quenching process or
from some other source, including things like
earthquake, could be adequately represented in these
quench tests. It was felt that they could not, and so
the quench tests were used only as confirmatory tests,
and the criteria themselves were derived from these
ring compression tests.
Well, I didn't know that, and I think most
of the people who knew about the details of the
development of these criteria during the ECCS hearings
are retired, and we had not planned such a test in our
program at Argonne National Laboratory. So the very
first thing, you know, as soon as this presentation
was made, we knew that we had to modify our program at
Argonne where we had only planned quench tests to
include some measure of post quench ductility from a
test, either a ring compression test or something
better than a ring compression test.
So that was the immediate result. There
was a sort of delayed reaction to this when in France
and in my office we discovered some Eastern European
papers from the early and mid-'90s reporting on ring
compression tests with the Russian alloy, E110, which
is zirconium, one percent niobium, which is very
similar in composition to M5.
So all of a sudden light bulbs are going
off. Here is some information on a similar alloy that
shows a marked reduction in the amount of oxidation
that can be tolerated during a loss of coolant
accident.
And so this then led to meetings with
Framatome and Westinghouse. It led to modification of
a conference that had already been planned under the
OECD framework, and you'll hear directly from
Framatome and Westinghouse on this subject, and then
I'll come back and give you a summary of the
conference which focused on that subject.
So just quickly to go over some steps in
trying to resolve this, we do have a test program at
Argonne National Laboratory with what we think of as
an integral test or a LOCA criterion test where we
take a piece of a high burnup fuel rod with the fuel
inside, pressurize it, run it through a LOCA type
transient, ballooning rupture, oxidation, cool down,
quenching, everything present, and try and look at the
results.
We also have a number of separate effect
tests in the same laboratory where we're looking in
separate measurements of oxidation kinetics and
mechanical properties, including now the post quench
mechanical properties.
The work started with real specimens last
summer when we received the BWR rods from the Limerick
plant, and it's slow going. We have done a number of
the oxidation kinetics measurements, and I can just
give you a qualitative result of that.
Oxidation kinetics seem somewhat faster
for high burnup fuel than for fresh fuel. So we get
oxidation rates that are higher than Cathcart-Pawel
correlation, for example, whereas when we measure for
fresh tubing, we can reproduce the Cathcart-Pawel
correlation.
CHAIRMAN POWERS: And do you exceed Baker-
Just?
DR. MEYER: I'm sorry?
CHAIRMAN POWERS: Do you exceed Baker-
Just?
DR. MEYER: I don't think so.
CHAIRMAN POWERS: That's harder to do.
DR. MEYER: Yeah, it would be harder.
CHAIRMAN POWERS: But which in a
regulatory world, that's the one that counts.
DR. MEYER: The Halden reactor is also
planning to do what we would call an integral test.
Take a piece of a fuel rod and run it through a
transient.
The principal interest in the Halden
program is to look at the possibility of relocation of
fragmented fuel into the balloon section, but it,
again, will allow you to look at a lot of things,
including oxidation, ballooning rupture.
There are a lot of related studies going
on in Japan and in Russia, and our FRAPTRAN code will
be used in performing the work, but not in a major way
in terms of coming to some resolution of this, unlike
resolving the BWR power oscillations, where it looks
like our job is going to be to analyze our way through
the transient.
In this case, analyzing your way through
the transient will be done with the large LOCA codes,
and our job is limited to just looking at what the
embrittlement criteria and the modeling for oxidation
and ballooning and rupture are.
We also are interested in doing the same
kind of testing for ZIRLO and M5 cladding, and in
fact, in the meetings that were held at the end of
February with Framatome and Westinghouse, we asked
them if they would cooperate with us on this work and
provide the materials, and we'd do the work right at
Argonne and involve EPRI in the program at the same
time, and so we're kind of waiting for a response on
that.
I think that's all I want to say at this
time. There are two other slides in your handout.
This one is a list of the work that we're relying on.
I put NRC in quotation marks here because we don't
fund or direct all of these programs. There's a
range.
For example, the JAERI program, we neither
fund it nor direct the work in it, but we have full
cooperation with JAERI on this, and they do provide us
with all the information.
Some of these programs we participate in
as paying members. The Russian work, we provide a
portion of their funding and a lot of the direction of
that work, but this is pretty much a list of the
research programs on which we will be depending for
information on fuel behavior.
CHAIRMAN POWERS: One of the questions
that came up in a previous discussion of the Argonne
out of pile test was the question of what temperature
scenario you put them through to simulate the LOCA.
DR. MEYER: Yeah.
CHAIRMAN POWERS: And do you track some
sort of average temperature history or do you try to
find the temperature history of a particular rod in
those experiments?
DR. MEYER: I'm not sure that the
temperatures have been set for this, but our current
thinking is to run these integral tests at 1,204
degrees Centigrade. So we would run them up.
We have a linear -- Harold, what is the
run-up? Five degrees per second heat-up or is it
higher than that?
MR. SCOTT: It's about that.
DR. MEYER: Sud knows the numbers for
that.
MR. ROSENTHAL: Let me just offer that we
need to be thinking this thing through because the
heat-up rate of the evaluation model, large break
LOCA, is going to be different from a small LOCA, is
going to be different from the best estimate LOCA, and
so we need to think it through, and we don't have all
of the answers yes.
DR. MEYER: I know that Dana is concerned
about some stressed that might be applied on the way
up. We have, in fact, focused more on the way down
and have given more attention to the cool down part of
this because this is when the oxygen and hydrogen and
distributing themselves in the alpha phase and in the
prior beta phase, and we believe that the cool down
conditions are going to ultimately determine what the
ductility is, and then you do the test when you're
down at a relatively cold temperature. You run it
through the oxidation transient, come down, and then
the ultimate challenge is near the end down at the low
temperature.
So I know that you have for some time
asked us to look carefully at the heat-up. We've
brought this question up. We haven't found much there
to accommodate. You know, if there's something more
specific that you can help us with, these conditions
have not been set in concrete yet.
CHAIRMAN POWERS: Yeah, my concern is that
when we look at an individual rod in one of these
scenarios, nearly always -- I can't say always, but
frequently -- what you see is the rod heats up, then
it cools down, and then it heats on up and hits the
plateau, whatever it is.
On the average, if you plotted the core
average, it looks like you ramp up to a plateau, sits
in a plateau, and then it cools down, but by looking
at the individual rod, it's actually going through a
fairly complicated scenario, and it does have this
cool down period, and it is, indeed, that cooling off
that you become most concerned about.
DR. MEYER: I think if it had a cool down
period prior to the ultimate cool down and quench
following a long period at a very high temperature,
then you might have some interesting effects. It was
my impression that the ups and downs occurred at a
relatively low temperature as you're approaching this
high temperature period, and I don't think those would
have a very big effect because you still have ductile
cladding and a very small amount of the oxidation
taking place.
We can continue to --
CHAIRMAN POWERS: Well, I mean, when you
do the tests, you're going to have to have some
justification --
DR. MEYER: Yeah.
CHAIRMAN POWERS: -- for that, I mean, and
what you outlined is probably an appropriate
justification, but it would have to be substantiated
with something quantitative, the analysis.
DR. MEYER: Okay.
CHAIRMAN POWERS: The heat-up that shows
that all of these things are taking place at
relatively low temperatures, and they don't go up, sit
in a plateau, oxidize for a while, then cool down,
then heat back up again.
I think you'll find though --
DR. MEYER: Sud Basu is the project
manager for this program, and I'll at least say that
we will go back to the project and tell them that
we've been reminded of this again, and to make sure
that we have either a justification for what we do or
we change our says.
DR. SHACK: I mean, I think if you look at
it, you know, you're pumping all of this hydrogen in
during this oxidation. Then the tricky thing about
this thing is as Ralph said. You know, you don't
really get the big thermal shock until you've cooled
the thing down, in which case, you know, while this
thing is hot, it's ductile as hell. It's after you
cool it down again that it re-embrittles, and then you
hit it with the big thermal shock.
But the embrittlement that you get because
you've pumped all of the hydrogen in because of all
the oxidation that's occurred at the high temperature
and the huge thermal shock that you finally get when
this thing re-wets, you know, that really does seem to
be the limiting material and stress condition that
you're ever going to see. You know, one of these
cycles before you haven't pumped all of the hydrogen
in. You certainly haven't got a stress that's
anything like the re-wet stress.
DR. MEYER: Well, I mean, I understand the
argument.
DR. SHACK: You're at the worst --
DR. MEYER: But that's all I ever get is
this argument, and I get other people showing me
calculations and individual fuel rods that don't seem
to be consistent with the argument, and nobody ever
coming back to me and saying, "Look. All right.
Here's the calculation we've done with our code that
we're happy with, and here's how the fuel rods behave,
and indeed, the limiting stress conditions are always
calculated to be in the quenching."
I mean, you can wave your hands make
those --
DR. SHACK: Well, the stress and the
limiting --
DR. MEYER: You can make those arguments
as long as you want to until you come back and
quantitatively show me that that's, indeed, what you
expect to be.
The problem with the old scenarios is when
we were worried about just oxidation, then sitting at
the high temperature plateau was the conservative
case. It's not clear now that we're worried about
fuel embrittlement that sitting at the high
temperature condition is the limiting case.
And how you get there suddenly become
important, and making a qualitative argument all the
time, I've heard it. I agree it. Now show me
quantitatively that that's the case.
MR. NISSLEY: Mitch Nissley, Westinghouse.
We've done a number of calculations with
both evaluation model and realistic codes, and I would
support the general conclusions that Ralph has
offered, and we'd be more than willing to share that
information with the staff to help resolve this issue.
I'd also say that some of the higher
stress in the cladding during re-wet are really very
early in re-wet at the bottom of the core where you've
not had much oxidation. It's the higher levels in the
core generally we will have a slower cool down and a
less severe quench load because there's a lot of
precursor cooling as the reflood front progresses up
through the core.
But we would be willing to provide
quantitative information to the staff to help address
this concern.
DR. KRESS: Ralph, the research and the
PIRTs deal with three design basis type of accidents.
Are you planning an additional PIRT to look at severe
accidents and effects on the core melt behavior and
fission product release?
DR. MEYER: There's a PIRT that's been
organized to look at source term, which is kind of
serve accidents.
DR. KRESS: Yeah.
DR. MEYER: And I won't be running that
directly, but Charlie Tinkler and Jason Shaperow, who
have been involved with the severe accident program,
will be conducting that PIRT.
DR. KRESS: So questions about effects on
high burnup on core melt and source terms will be
addressed later.
DR. MEYER: Yes.
DR. KRESS: So it's not part of this.
DR. MEYER: Yes.
DR. KRESS: The other question I have is
has anybody raised an issue of the potential effects
of high burnup on the iodine spike and steam generator
II rupture accidents? Has that ever been brought up
as a potential issue?
DR. MEYER: I can't answer that question.
Jack, can you?
MR. ROSENTHAL: Yeah, in response to the
ACRS report, et cetera, we're just now planning out
how to take on the iodine spiking issue. So actually
it's a very timely comment, and in my own mind you
make so much iodine per fission, and it's a question
of where is that iodine before the hypothesized event
occurs. Is it in the fuel or the gap, or is it
already outside in the --
DR. KRESS: I think that is very relevant.
MR. ROSENTHAL: It is probably more
dominant than the fact that at higher burnups you'll
end up ultimately with some sort of equilibrium iodine
concentration. That is the time we have to take it
on.
DR. KRESS: Yeah.
MR. ROSENTHAL: A different project.
DR. KRESS: And also the spike is a rate
at which things get out of clad, and that's not just
a function of where the iodine is. It's a function of
what has happened to the clad.
So, you know, it could affect both of
those things, but anyway, it's something I think ought
to be thought about.
MR. ROSENTHAL: Right.
DR. MEYER: And finally, I just want to
mention EPRI's cooperation in the big program at
Argonne National Laboratory and to say to you that we
finally have the H.B. Robinson fuel rods in a hot
cell.
So we have Odelli Ojer at EPRI to thank
for a lot of hard work on that, and also John Siphers
at CPNL in the end stepped in and was a big help.
So that's all I have right now. I don't
know if Med is -- we're really going to be pressed for
time. We have a 17 minute video on the Cabri program
that Med might show at lunchtime.
CHAIRMAN POWERS: Yeah, I think we're
planning on doing that at lunchtime.
DR. MEYER: Or some other time.
CHAIRMAN POWERS: What I want to do now
because I don't want to break up the next presentation
is go ahead and take a 15 minute break now and we'll
come back and listen to the presentation on the
assessment of LOCA ductility of M5 cladding, and we
can understand better the difference between quench
and ring compression test.
DR. APOSTOLAKIS: When are you going to
show this video at lunch because I had other -- at the
beginning, 12 o'clock or 12:30?
CHAIRMAN POWERS: When I get around to it.
(Whereupon, the foregoing matter went off
the record at 10:18 a.m. and went back on
the record at 10:33 a.m.)
CHAIRMAN POWERS: Let's come back into
session.
Ralph, I have TBD on my speaker for the
Framatome testing assessment of LOCA ductility.
DR. MEYER: Garry Garner will give the
presentation.
CHAIRMAN POWERS: Okay. So it's actually
Garry Garner is TBD. Strange initials.
MR. GARNER: If you like what you hear,
it's Garry Garner. If not --
CHAIRMAN POWERS: It's that other guy,
right? Good. Good strategy.
MR. GARNER: Well, good morning, gentlemen
and ladies. My name is Garry Garner. I am a
metallurgical engineer, materials engineer at
Framatome ANP in Lynchburg, Virginia, and I will be
speaking this morning of the LOCA ductility with M5
clad testing results.
At the end of February, latter part of
February, this presentation was given to the NRC
staff. We took about three hours and we had about 100
slides.
I've pared that down a little bit for this
morning. We had our in-house LOCA man give part of
these results, and I gave primarily the mechanical
test results at the end. It's just me this morning.
I'll, of course, try to answer all of your questions.
If I don't know the answer, I'll defer, and we'll get
it for you.
I want to stress at the beginning that our
primary mission in life is not pure research. Our
goal with getting alloy M5 developed and licensed and
in reactors is to do those tests that are required by
the codes and the criteria and compare the results to
Zirc-4.
And you'll see, I hope, this morning that
those results compare favorably or are the same in
some cases.
The way I would like to proceed through
this subject material is to start off with just a very
brief review of a couple of things about in-reactor
operating experience, not LOCA, but just normal in-
reactor.
I want to talk a little bit about the
alloy composition and fabrication parameters, and then
I want to show you that it is a low oxidizing alloy
and that it has a low hydrogen pick-up, and so I'll
show you the oxidation curve and the hydrogen curve.
And then that is just as a way to set the
table for the LOCA, post LOCA discussion that will
follow, and we'll talk about the oxidation tests that
we did, the quench tests, and the post quench
mechanical testing, and then we'll follow with a brief
conclusion and a summary.
So for the in-reactor performance, M5 is
a binary alloy primarily of zirconium and niobium.
Tin is an impurity in this alloy.
Three things that might differentiate this
particular Zirc-1 niobium alloy from an E110 or from
someone else's zirc, one percent niobium are we do
target iron in this 250 to 500 ppm range for improved
corrosion. Oxygen is targeted rather high. The spec
limit is 11 to 17. We target it right in here for
improved creep performance.
And sulfur. Sulfur is an impurity. It's
not called out even in the spec as anything other than
an impurity, but what we found -- and if you've kept
up with the work of Mr. Sharke (phonetic) and others
from Framatome -- we found that a very small change in
an impurity element has a fairly dramatic change on
macro properties like creep.
So when people talk about M5 being of a
similar nature, similar chemistry to other alloys,
yeah, on the surface, but on the other hand, very
small changes can have very drastic effects.
Thermal mechanical processing also plays
a role. There's more to the alloy than just its
chemistry. This particular alloy is fully
recrystallized, and in the tube making process and in
the strip making process also, we do all of the
intermediate temperature anneals below the transition,
below the 610 transition.
DR. SHACK: While we're at though, I mean,
if the sulfur has such a big effect, why isn't it spec
then rather than just left to float as an impurity?
MR. GARNER: We found out sulfur, when we
were developing the alloy during the creep tests, we
were noticing that the thermal creep properties were
all over the place with each ingot, and it turned out
that some of the raw zircon coming from some of the
beaches had an unnaturally higher sulfur content than
the others, and some of them were low.
So we did the research. We found out
where the knee in the curve was, and now we specify
ten to 35 ppm sulfur in our spec.
By the way, we also found that same effect
for Zirc-4 to a lesser degree, but I think all of the
zirconium alloys are sensitive to that.
So just to make sure that we always get
the right creep properties, the good creep properties,
the best thermal creep properties that we can get, we
do specify it now between ten and 35. But you won't
find that in the ASTM zirc specs.
Again, my point on the mechanical
processing, we do the anneals below the transition.
We found with this alloy that that makes a marked and
significant difference in the microstructure, the
appearance of the microstructure, and the stability of
the microstructure of the alloy.
If we can go into a LOCA and a post LOCA
with the stablest microstructure possible, that's what
we want. So it's not only a stable microstructure and
a good chemistry. It's not only important in the
normal operation. It's important in an accident
condition as well.
The two properties that I would highlight
this morning are the corrosion, and you've seen these
kind of curves before. This curve -- and I apologize.
It's hard to read because it is just so small on this
viewgraph -- but it's the maximum oxide thickness
versus fuel rod average burnup, and you can see that
all of the colored dots are M5 data points. They come
from a wide variety of reactors, from 14-14 to 17 by
17, and the colors are just differentiating those.
And there is a linear behavior up to a
burnup so far of 63 gigawatt days. This is sort of
the line through the middle of the Zirc-4 data.
The points that I would make here is that
we're getting more and more additional data in the 50
to 60 gigawatt area. We're seeing no increase in the
oxidation rate at the higher burnups. The highest
oxidation so far has been about 40 microns at 60, 63.
So it is a low oxidizing reactor, and
that's important when we start talking about what's
the condition of the alloy, when you go into an
accident condition.
CHAIRMAN POWERS: If I look at the data
points from the 16 by 16 --
MR. GARNER: Yeah, the red ones.
CHAIRMAN POWERS: It looks to me like you
could probably convinced yourself as you went out
toward 60 you would get the same kind of upturn that
you see for Zircaloy-4 based on those data points.
MR. GARNER: I don't really think so.
These reactors are different duties, granted. There
does seem to be a little bit higher effect in the 16
by 16s. I think the behavior still though is rather
linear. I don't see any kind of a two slope upturn
like you do see with the Zirc-4 type alloys.
Yeah, when we get more data out here for
16 by 16s, we'll see what that's doing, but so far I
would point out that the max oxide there is 41
microns, and only in that point.
So if it does turn up, it's going to turn
up at a significantly different rate than these guys
are turning up. Hopefully.
Similarly, the hydrogen plot for these
alloys, I had hoped to have because the results are
going to be given to us in April, some burnups in the
mid-50s to almost 60 or right in here, to show you
that this linear trend with M5 continues, but you have
the hydrogen content MPPM versus fuel rod average
burnup here, and there's the Zirc-4.
As you would expect, the source of most of
the hydrogen for these alloys to pick up is the metal
water reaction that's going on. So you would expect
a similar kind of behavior. This alloy has a
significantly lower pickup fraction than does Zirc-4,
and so we get a flat behavior.
Again, this is going to get important when
we talk about how much hydrogen is in the alloy in the
event of the LOCA, either at the beginning, middle or
end of life.
As you can see on this curve, it's going
to be less than 100 if this trend continues out here
as we expect it, of course, to do beyond 60 gigawatt
days.
So just a summary for just that brief
portion of this presentation. It is a low oxidizing
alloy. We don't see any increase in the oxidation
rate at the highest burnups that we've achieved, which
are 63 gigawatt days.
If the alloy is lower in sensitivity, to
temperature and rod power, we've seen that it has
less, dramatically less response to those kind of duty
factors, temperature and power, than do the Zirc-4
alloy.
The low oxidation rate and the low
hydrogen absorption, the low hydrogen pickup fraction
for this alloy end up with a low hydrogen content at
high burnups, end of life burnups.
DR. CRONENBERG: When did M5 go into use?
'95?
MR. GARNER: Yeah, it went into just rod
by rod demonstration rods in the early '90s. It went
into our first batch deliveries were in '98, full
batch reloads, and now we're well on the way of
delivering those full batches now.
DR. CRONENBERG: And that's all in France?
MR. GARNER: No, no, no, no. We have full
batches at North Anna and Oconee at this point and
some more being delivered later this year.
Our North Anna reactor burnup is after --
we just finished our second cycle, and we're on our
lead assemblies there, and our burnup was 40 to 46
gigawatt.
MR. ALDRICH: Mike Aldrich in Framatome.
I think right around 46 peak rod.
MR. GARNER: I think it was 46, 300 peak
rod of gigawatt days. So, yeah, we do have it in the
-- the alloys in TMI, North Anna, Oconee.
MR. ALDRICH: Yeah, the full batches that
we have are at Davis-Besse, Oconee Unit 1. We're
supplying Oconee Unit 2 right now, and at TMI will
also be getting a batch this fall.
DR. CRONENBERG: And then for the hydrogen
pickup, you take them back to Lynchburg and do your
constructive testing there or --
MR. GARNER: I didn't mean to mislead you
on that. We haven't done a hot cell within the U.S.
M5 yet. Those are planned.
These hydrogen analysis were done from the
European exposures, yeah.
Okay. Now, I would talk about the results
in the high temperature testing, the oxidation quench
test and post mechanical quench test. It was called
the CINOG. That was the facility in Grenoble where
the work was done, and beyond that I don't even know
what CINOG means.
The test matrix for the high temperature
oxidation tests were we tested both M5 and Zirc-4. It
was a double sided oxidation experiment. Length of
the samples, about 20 millimeters.
We tested as manufactured, unradiated
cladding, just as received from the cladding from the
tube vendor, at temperatures between 700 and 1,400 C.
At 1,200 C. we tested some pre-hydrided
cladding, which was pre-hydrided at 200 ppm for the M5
alloy and 200 and 450 ppm for Zirc-4. The reason that
we didn't go to the 450 for M5 is for the obvious
reason that we're not even going to get 200 possible
in normal behavior, plus the oxidation. We're going
to show you that in a few minutes. We're not going to
get so -- 200 was felt to be very bounding for M5.
We did three oxidation times at each test
temperature. To try to get these, you know, you time
it, and you try to get 50, 100, and 200 microns per
side, and for three samples for each test conditions
we're done.
The results of the oxidation testing are
presented on this plot. It's a little bit busy, but
really the results aren't as busy as it might seem.
On the left is oxide in terms of weight
gain, milligrams per centimeter squared, versus the
oxidation times square root of seconds, and you'll see
a series of lines here.
For instance, in like I say it was 700 to
1,400. At 1,400, Zirc-4 and M5 oxidation kinetics are
right on top of each other. If you had the time and
inclination to go through this legend, you'll see that
at that temperature they're the same. At 1,250
they're the same. At 1,150 they're they same. At
1,100 they're the same.
At 1,050 the Zirc-4 and the M5 are parting
company rather dramatically with the M5 having a much
lower oxidation kinetic than the Zirc-4.
Now, I didn't draw lines through the data,
but the NFI did some independent research on our
alloy, on M5, and got the same results, and that's
what you see right here. The open triangles are Zirc-
4, and the closed triangles are M5. So, again, you're
seeing that behavior.
Mr. Lebourhis at the OECD meeting two
weeks ago in France presented the results on this
curve of another French test at 1,000 degrees saying
the same thing.
And then down here at 900 again, the
alloys are having the same oxidation kinetic again,
900, 800, and 700. So this area here between 1,100
and 1,050, lower than 1,100 and greater than 900, the
M5 alloy is clearly oxidizing at a lower rate. It's
the only place in that spectrum that that's happen.
I put the 17 percent for folks that want
to think about weight gain in terms of the ECR, the
equivalent clad reacted. It's right in there, about
24, 25 milligrams per centimeter squared. So that's
about 17 percent ECR.
You can see that we behave better or
similar to Zirc-4 at these temperatures. The values
are consistent with the literature, and they were
verified by independent folks, NFI in this case.
CHAIRMAN POWERS: Do you know why you're
slow in the oxidations in the 1,050 to 1,100 degree
range?
MR. GARNER: I don't, no.
CHAIRMAN POWERS: There's a phase
transition in there someplace, isn't there?
MR. GARNER: Yes, yes. You know, and the
alloy -- we know that the chemistry of the alloy has
to do with what temperature that phase transition goes
in and like that. That's certainly the speculation,
but I'm not an expert on that. I don't know exactly
why that is, but it's very well documented, and it is
confirmed.
DR. CRONENBERG: Do you have the
diffusivity measurements at these temperatures, too,
that for the two different alloys, oxygen diffusivity
measurements?
MR. GARNER: We did not make diffusivity
measurements, no.
DR. CRONENBERG: Is there in the
literature that show that, yeah, this is all in sync,
that there's a phase change, there's a diffusivity
change, therefore, there's an oxidation rate change?
MR. GARNER: Right.
DR. CRONENBERG: I mean, is that all --
MR. GARNER: It's all consistent.
DR. CRONENBERG: It's all consistent?
MR. GARNER: Yes, sir, yeah.
Okay. In those results compared with
literature results, compared with the correlations,
this is the weight gain function again, and in this
case one over the reciprocal of temperature. So
temperature is going down as you go this way.
We've plotted the Baker-Just correlation
with the solid line. The dotted line is the Leistikov
correlation, and the points here are the M5 and Zirc-
4. The open squares are Zirc-4. The solid, the
diamonds are M5, and you can see at the higher
temperatures that the data are consistent with each
other, and also shows that Leistikov does a fair job
of predicting actual data, whereas were conservative
to Baker-Just.
At this lower temperature, and this
corresponds to about 1,300 degrees C., you see that
difference again where M5 and Zirc-4 are behaving
differently, and with the lower oxidation kinetic
associated with M5.
So we are bounded by Baker-Just in all the
encountered configurations, and I think we were
surprised that Leistikov does a fairly good job of
predicting the real data.
DR. CRONENBERG: Prater-Cartwright was
used during -- developed during severe accident
program here for Zirc-4. Have you benchmarked
anything against Prater-Cartwright for severe accident
conditions with the M5 class?
MR. GARNER: We did have a slide in the
presentation at the end of February where we showed
consistency with the Prater-Cartwright data, yes, and
I can --
DR. CRONENBERG: It's less?
MR. GARNER: Yes.
DR. CRONENBERG: Your data is less than
what would be predicted by Prater-Cartwright?
MR. GARNER: Yes.
DR. CRONENBERG: Okay.
MR. GARNER: Now, in terms of what we
saw --
CHAIRMAN POWERS: Radiation has no impact
on these?
MR. GARNER: Excuse me?
CHAIRMAN POWERS: Radiation has no impact
on these oxidation rates?
MR. GARNER: I think radiation can be
expected to have a small impact on them, yes.
When we looked at the oxide coming from
these oxidation tests, at the high time, 1,000
degrees, these were two sided tests, and so in this
picture you see the oxide, the base metal to both the
alpha and the prior beta, and then the inner layer
oxide. This is the mounting, the medium here.
And you see for Zirc-4 that you do have
this layer, this flakiness, this layering. It's a
trace amount of it, but it is present, and we saw that
on both the inner layer and the outer layer of the
Zirc-4 samples.
When we looked at the M5 alloy, same
magnification, you see the less oxide here in this
case. This is the mounting material. This is the
base metal, and where all of the etching in these
photographs to see the oxide and any flaking.
You see that it is a less, but the
important thing is that there is a homogenous barrier
there. There are no cracks through it. There are no
-- none of these delaminations through it that we saw
a slight bit of in Zirc-4 that you're going to see a
whole lot more of in the E110 in a few moments. So we
didn't see that.
Now, just to put some numbers to these
pictures, I thought it might be interesting if on the
two sided test you have the external zirconium layer,
the external oxide, the internal oxide, and then you
have the oxygen stabilized alphas next to both of
those, and then in the middle the beta layer.
And for Zirc-4 you can see the expected
difference in the thickness of the oxides, both on the
inner and the outer, but you can see that the alphas
and the beta phase are about the same.
This is interesting because in other
results for Zirc-niobium alloys, and specifically the
Bohmert paper, he explains in there that he had a hard
time differentiating the alpha and the beta, and he
couldn't find it.
In a picture that I'll show you in a
little bit you can sort of see what he's talking about
there.
In this alloy and you'll see it in some
other pictures in a little while, those layers are
very discernable, and you'll see that in a minute. So
those numbers sort of just go with those pictures.
That's the magnitude of the thicknesses involved
there.
Now, the quench test, the quench test
matrix, again, comparing M5 and Zirc-4, double sided
oxidation test. Failure was defined as if you put a
slight after the quench, if you put a slight over
pressure in that and you see some bubbles coming out;
that's failure. It's a fairly conservative definition
for failure because just a pin hole is a failure under
this criteria.
The temperatures tested at were 1,000
through 1,300 degrees C. in 100 degree increments.
Again, as manufactured tubing.
At 1,200 degrees C., again, the pre-
hydrided samples, 200 ppm for M5 and the higher ppm
added for Zirc-4, and generally you did five or more
tests to establish where that failure occurs. You
test until you get that failure, and so generally that
took five or more times.
And then there was post test metallography
and hydrogen analysis, which I can show you. The
results, just in a nutshell, on this plot you can see
that the two alloys in this column, that the
temperatures 11, 12 through 13 and the time to
failure, and you can see at these higher temperatures
they're fairly consistent, the two alloys.
At the lower temperature, the 1,000
degrees, the M5, it took twice as long to fail,a nd
you'll see this again on the curve in a moment.
For events of equal duration, alloy M5
seems to be superior to the Zirc-4.
Plotting that up as a function of ECR, we
have ECR on the left and temperature on the bottom
here. This is the Baker-Just correlation points. I
hope nobody asks me why that dips because I sure don't
know. It surprises us.
This is the Leistokov correlation points,
lower understandably, and uniformly. And then this is
our data. We're plotting failure points up here, and
as you can see, at the 1,300 degrees temperature, the
red line is the 17 percent linking the criterion, and
so that's a failure point at 1,000 degrees, and it
took four and a half hours to get there.
And this is the last unfailed point, and
it took three and three quarters hours. So somewhere
between three and three quarters hours and four and a
half you fail this alloy, and it looks like it's
pretty close to the 17 percent criterion.
It's really for this kind of reason that
we think that 17 percent criterion is a decent
criterion for this alloy, because it's of no concern
until you get to times of failure that are just so
ridiculously large that it's no longer interesting.
We measured the hydrogen content for the
two alloys. Zirc-4, at these oxidation temperatures,
this was, again, the durations of these tests, and you
can see that the hydrogen content here -- these are
the results of three different measurements, and you
can see that they're in the 20s, and they're fairly
consistent. M5 might be just a tad lower. It's not
significant.
The significance of this chart to me is in
some of the Eastern European papers, specifically
Bohmert again, at 1,100 degrees where we're showing 18
to 20 ppm of hydrogen in our oxidation and quench
test, that study produced over 400 ppm of hydrogen.
Don't know why.
So the results, just a summary of the
results. The oxidation and the quench. It's clear
that M5 is performing equivalent or superior to Zirc-
4. The hydrogen uptake is low. That's clear.
The M5 accident survival is definitely
superior to Zirc-4. At temperatures greater than
1,100 they're about -- they're the same. At
temperatures less than 1,100, it's surviving up to two
times longer than Zirc-4. That's consistent with
those oxidation curves and that small band of
temperatures where M5 has the greater oxidation
resistance.
The oxide itself in the quench and in the
oxidation, it's not delaminating. It's not showing
any signs of breaking down. It's not cracked or
delaminated.
If you use Baker-Just to establish the
criteria, of course, M5 always meets it. We do
successive oxidation times to achieve -- if you want
to get down to 17 percent criterion, it takes a long,
long time to get there with a low oxidizing alloy, and
again, we agree with the criterion.
Now, in our efforts to license this with
the utilities and the power authorities in Germany,
they are very aware of the Bohmert paper, and they
wanted to see how we did in post quench mechanical
testing similar to what he did, and so a year and a
half, two years ago, Framatome undertook to do some of
those tests.
This was the test matrix. We tested at
1,100 degrees C. We did it for times that would give
ECRs from three to 17 percent. This series of tests
was a single face oxidation, and again, we used as
fabricated M5 and compared it to Zirc-4 cladding.
After oxidation it was water quenched, at
which point we did mechanical tests. We did a three
point bend test, an impact test, and split ring
compression test.
That begs the question. That matrix begs
the question: why did you test at 1,100 degrees?
And, again, we go back to this chart. We wanted to
test in an area where the alloys of M5 and Zirc-4 are
oxidizing at a similar rate.
It's not very interesting down here to
test M5 because it takes so long to get anywhere close
to 17 percent. It's really out of the realm of what
we're interested in. So we picked 1,100 degrees,
where the two alloys are oxidizing at a fairly steep
rate, and they're oxidizing the same. A test in that
region might learn something was the thought.
This just briefly was the test rig that we
used for that series of tests. It's just a four zone
heater with the sample hanging here. This is the
little quench tank. The reason I wanted to show this
slide is mainly for that little piece of white cotton
that's sitting in there. That collects the oxide that
falls off of the sample upon quenching, and we wanted
to show you the results of that.
So each sample, that oxide was collected
and weighed and compared to the weight gain that that
sample achieved in its oxidation phase.
DR. CRONENBERG: Were you measuring any
hydrogen off-gassing besides hydrogen pickup?
MR. GARNER: No.
DR. CRONENBERG: No?
MR. GARNER: No.
And here are the results of those tests.
At 1,100 degrees Centigrade, again, for the longest
exposure times, the Zirc-4, these were the weight
gains observed, and that was the oxide spalled in
grams, and this is expressed as a percentage of the
weight gain.
And you can see that with the Zirc-4
because of that slight delamination that we saw, that
slight flakiness in that alloy, it's losing a lot on
quenching. It's losing between 65 and 80-some odd
percent of its oxide, whereas the M5 oxide seems to be
very tenacious. It's losing only between two and four
percent.
That confirms quantitatively what those
pictures were attempting to show qualitatively about
the difference in the character of the oxide in M5 and
Zirc-4.
Now, the pictures, again, also support
those results. This is the Zirc-4 at the high time.
On this sample you can see clearly the oxide layer,
the alpha layer, the oxygen stabilized alpha layer,
and the prior beta, and you see the large greens.
In this picture, and you can see it a
little bit here, but more in this picture that was
etched specifically to bring this feature out, the
oxide is up here and you can't really see it, but this
is this alpha area here, and you can see these cracks.
That oxide is cracking, and it's breaking down, and
that explains the results that we just saw.
Now, in contrast to that, the M5 oxide
looks like this. Again, it's the same kind of
picture. There's the oxide, and then there's the
alpha, and then the beta below that, and again, over
here you can't see the oxide, but you can see this
alpha area.
And I guess you have to take my word for
it a little bit. Those are not cracks. They're
shadows. Most of what they are is this linear
distribution of niobium particles.
At these temperatures, what we noticed,
and you can see it here, within the matrix of the
grains, you see the particles lining up in a linear
fashion. That's a microstructure that we specifically
prohibit in the alloy for a normal operation, but in
a LOCA event, that's what happens.
When you go above that oxygen or alpha-
beta transition, you tend to get that, and that's
what's going on these, these agglomerations of beta
Zirc or beta niobium sitting there.
Again, no cracks, and again, you get that
linear distribution.
Now, to compare that with what people have
observed in some of E110 alloys, this is a picture
that was not in Mr. Bohmert's paper. It is in a
Russian report, and I can give you the reference of
that if you need that, and in a second, I'm going to
show you a quote from Mr. Bohmert's paper where he
describes in words what he's seeing here, and other
folks have seen this, too.
Again, the stratified oxide, in this case
highly stratified. In the Zirc-4 that we looked at a
little while ago, it typically had, you know, one of
those going through there. This alloy is full of
them.
Mr. Bohmert also makes the point that he
can't find what's going on in the base metal between
alpha and beta. This picture, although probably not
optimally etched for that, tends to support that.
The point here is that it's a very
stratified and cracked oxide layer, and it has a
completely different morphology than M5. In words,
Mr. Bohmert said that not at a late stage -- that
photograph that I just showed you was taken after like
9,000 seconds -- but Mr. Bohmert and his work said
that at an early stage he found the same thing in
multi-layer oxide scales formed which tend to flake.
We saw that flakiness in the Zirc-4.
And, again, we just didn't see that. We
don't see it in M5. We've never seen that kind of a
morphology, and in the quench test, you can see that
when we weighed the amount of oxide that's falling
off, falling off, flaking off, it's not there.
DR. CRONENBERG: I think 110 has higher
niobium and higher tin or --
MR. GARNER: I'm going to say that I don't
know. Nominally it's the same niobium. Nominally
it's a Zirc one percent.
DR. CRONENBERG: I thought it was like two
percent.
MR. GARNER: No.
DR. CRONENBERG: No?
MR. GARNER: There are alloys that are
two, two and a half, and even Framatome has fooled
with those from time to time. E110 is nominal one
percent, but as far as their tin, their impurities,
their other things, I don't know, and I specifically
don't know with respect to the version of E110 that
Mr. Bohmert tested back in the early '90s. It could
be vastly different from the E110 that's in reactors
now for all we know.
DR. CRONENBERG: Did he put in his paper
what the --
MR. GARNER: He put the chemistry in
there. Yeah, and like I say, it's a nominal one
percent.
DR. CRONENBERG: Do they use the one
percent now or is it two percent?
MR. GARNER: They use the one percent.
Post quench mechanical tests, the three
point bend test was the first one that was done. This
is just a picture of the test rig showing the two
mandrels with about a nine millimeter rod, tube going
through there and pushing down on the center of it.
The maximum deflection that they got on
all of these was about seven and a half millimeter
displacement off of that line. That's the rig. Did
it for M5 and Zirc-4, and that's the results.
And you can see that the Zirc-4 and the M5
in this case are right on top of each other in terms
of the displacement versus weight gain. They are
behaving similarly in three point bend tests.
The next test was an impact test. I don't
have a picture of the test rig for that, but it was
like any impact test. It was a tube made with a notch
and a hammer coming down, and you're measuring the
energy that's absorbed in the material here called
resilience joules per square centimeter, again, versus
weight gain, and you can see again the two alloys, M5
and Zirc-4 behaving very similarly.
When you look at the fracture surface like
you like to do it with impact tests, you notice that
the Zirc-4 was a ductile ruptured in the ex-alpha-beta
phase and brittle in the oxygen alpha. M5 was
essentially the same, just a tad more ductility.
Maybe that explains that in the alpha phase.
DR. SHACK: Now, if you did a sort of
typical LOCA transient, what would your expected
weight gain be?
MR. GARNER: A LOCA transient.
DR. SHACK: Just to calibrate myself on
this curve.
MR. GARNER: Yeah. Well --
DR. SHACK: It would be less than 17
percent.
MR. GARNER: Yeah. What we saw in one of
these curves back here a minute ago, the weight gain
for 17 percent is about 24 milligrams per square
centimeter. So on that curve you could see where we
would be relative to that.
DR. SHACK: Okay.
MR. GARNER: Yeah.
DR. CRONENBERG: Well, then what's going
on between the E110 and the M5 if it's not
composition? Was it --
MR. GARNER: I didn't say it wasn't
composition.
DR. CRONENBERG: Okay.
MR. GARNER: In fact, I tried to imply
just the opposite of that. The compositions are
nominally the same, but what we found out in the
development of M5 was very small changes can have very
large effects. So it might be something like that.
There might be a compositional --
DR. CRONENBERG: And it's not in the
annealing process. So --
MR. GARNER: It could very well be. If
you don't anneal below the alpha-beta transition, you
will not get a stable microstructure. One of our
developmental precursors to M5 was we called it 5R,
and we even put it in test rods in reactors, and it
didn't do as well as M5 does, and that's when we made
the change.
What we were doing with 5R was we liked to
anneal above that transition because we got better
creep properties. What we found out was that that had
detrimental effects on some of the local oxidation,
specifically oxidations under spacer grids and like
that.
DR. CRONENBERG: But it's also time and
temperature for annealing and so it's not sorted out
then. You said you think it's probably chemistry and
trace.
MR. GARNER: All we can speculate is it
has to do with the stability of the microstructure.
Beyond that I wouldn't care to speculate because,
number one, I don't know much about E110. It's not
our position in life to compare our alloy to E110.
We're trying to compare it to Zirc-4.
Sure, we're as interested and curious as
anybody as to why these differences might be, but
we've not done any testing on E110. We read what we
can read.
What we do know from our own experience is
the target of some of these even impurity level
chemistry have large effects. We know from our own
experience that the thermal-mechanical processing at
the tube manufacturer is extremely critical to the
stability of the microstructure and in areas of
corrosion specifically.
That's why we went from 5R to M5. That 5R
microstructure that I was telling you about that has
those banded beta niobium particles, M5's
microstructure is uniform and stable under
irradiation, and that's all a function of that
intermediate annealing temperature.
So I wouldn't say that those two things
don't have something to do with the differences that
we see in E110, but I'm not an expert on E110, and I
don't want to stand up here and talk about it as if I
were.
DR. CRONENBERG: But M5 is not used for
any guide tubes or --
MR. GARNER: Yes, it is.
DR. CRONENBERG: Oh, it is?
MR. GARNER: Yes. We use it for guide
tubes, and we have our first spacer grids in lead test
assemblies hit Davis-Besse right now. So our intent
is to have an all M5 assembly very, very soon.
DR. CRONENBERG: So are you going to show
us the irradiation growth properties of M5?
MR. GARNER: I could. It wasn't part of
this presentation.
DR. CRONENBERG: I was thinking of the
small rod problems that we --
MR. GARNER: Right, right.
MR. ALDRICH: So far the -- this is Mike
Aldrich, Framatome again -- the growth data from the
guide tube material at North Anna and the LTAs that
Garry was referring to earlier at the peak rod burnup
of 46,000 we've seen virtually no growth of the guide
tube material at all.
MR. GARNER: It's not that much different
than the Zirc-4 and that's because the Zirc-4 guide
tubes are also fully recrystallized. So that growth
function, it's not just totally dependent on the
recrystallized versus SRA nature, but it's primarily
driven by the structure of the alloy. Recrystallized
alloys have a lot less growth than do stress relief in
annealed alloys, and that's why the M5 guide tubes,
they do grow a little less for other reason, but just
a little less.
DR. SHACK: The mechanical tests we're
looking at were all done in a single heat of material?
MR. GARNER: Yes, yes. Those tubes were
provided from a single lot at the tube vendor.
DR. SHACK: And how do you then set the
spec on, say, the iron limits? Is it you're checking
the microstructures, that over that fully range -- you
know, how do you test the stability of your
microstructure, since that seems to be your argument?
MR. GARNER: Right. Lots and lots of
tests there. We did a lot of test reactor testing, a
lot of out-of-pile testing, autoclaves and things like
that.
Every time we tweak something like a
sulfur, like an iron, we went through that whole gamut
We wanted to be sure that we weren't buying ourselves
some creep property or some growth property or some
corrosion property at the expense of something else.
So there's an extensive test base behind
those targets for all of those constituents, yeah, and
it's a tradeoff. I mean, when you don't have tin in
your alloy, you have to get creep properties from
somewhere else, and in our case, we've done it with
oxygen, and we've controlled it and controlled its
uniformity with sulfur and these other things, iron.
So, yeah, that was the whole trick with
this alloy. People knew years and years ago that
corrosion was going to be good with a niobium alloy.
The trick was how do you get there and still have
these other properties.
Those ranges were set after lots of
testing. The last mechanical test --
DR. CRONENBERG: What this tells me is
that it has to be a go slow process when you're
talking about these sort of things. When you have
small changes in composition it can affect different
properties in different ways, and so we had had
surprises like control rod insertion problems.
MR. GARNER: Right.
DR. CRONENBERG: The bending of guide
tubes, the irradiation growth, things like that, and
so that's just a general statement.
You're also saying that small changes in
composition can give you a surprise change in
mechanical performance.
MR. GARNER: You bet you, and like I said,
the development -- and we agree with you -- the
development of the alloy was a slow go, and we went
through many iterations before we got to M5. Now all
of those properties are controlled so that one reactor
doesn't get one iron in one oxygen and another guy get
another sulfur. Those are all controlled in our
specification as you would control these things with
any alloy in any specification.
The development of those ranges was slow
go, and now we insure the properties like every vendor
insures its properties, with its spec. And we agree.
MR. ALDRICH: I might also add, if you
were through.
MR. GARNER: Oh, yes, sir.
MR. ALDRICH: As far as the deployment of
the alloy and the fuel surveillance section of the SER
for M5, we are required to take additional PIE data of
things like you're referring to, control rod
insertability, as the burnup of the fuel in reactor
exceeds higher and higher levels up to the license
limit, we are required to take PIE data. So that type
of performance would be verified.
MR. GARNER: We do take an awful lot of
PIE data.
The last post quench mechanical test that
we did was the ring compression test. Again, that's
just the rig, and you can see the sample sitting in
there waiting to be pushed on. And again, the similar
results with Zirc-4, displacement versus weight gain.
The alloys are the same.
DR. SHACK: I mean when I look at these
things, is this really telling me that if I pump the
sort of same amount of oxygen and hydrogen into these
alloys, they act about the same?
MR. GARNER: Yes.
DR. SHACK: And the difference really is
the rate at which you pump hydrogen into it because of
the corrosion properties. When you look at these
things, you sort of see the same thickness of the
stabilized layers --
MR. GARNER: Yes.
DR. SHACK: -- for a given weight gain?
MR. GARNER: For a given weight gain we
do, and that was different than some folks have seen
with other Zirc 1-niobium alloys. We do, and I wanted
to show on that one chart that our oxygen stabilized
alpha and our retained beta were almost identical to
that of Zirc-4.
Now, the reason we're here is Bohmert, and
so we plotted our ring compression tests against the
same variables that he did, the relative deformation
on the left, ECR value across the bottom.
The black line here is sort of the line
through his data, which showed the embrittlement at
the lower temperatures. This is the line below which
you consider the allow brittle. Above 65 you can
consider it ductile, and in the middle it's mixed.
What you can see here with the blue, the
solid blue, the squares and the open blue squares --
the solids are our results for Zirc-4. The opens are
Mr. Bohmert's results for Zirc-4, and you can see that
by and large, with the exception of maybe that point,
we agreed. This told us that his work was probably
pretty good, and he had pretty good control over all
of his test parameters because when we tested an alloy
that we know was like the alloy that we tested, we got
pretty much the same results.
Where we differed was where we compared
his Zirc 1-niobium, which was the alloy E110 of 1992
vintage to our M5, and as you can see, our M5 is right
along on the same curve as the Zirc-4, which our other
data has supported, and where we differed was that's
where E110 came in at 1,100 degrees.
We don't have any inherent quarrel with
Mr. Bohmert's work. What we know is that the alloys
M5 and that E110 that he tested are apparently very
different, and I tried to show you this morning that
they're different in terms of the results that we get,
the measurements on the mechanical tests, the
measurements and the oxides, the oxidation rates, and
even what they look like, the morphology of the
oxides.
These two alloys, while nominally Zirc one
percent --
DR. CRONENBERG: They are not the same.
MR. GARNER: -- they are not the same, and
that's what I showed you.
Now, I haven't got the data to win the
Nobel Prize yet on why, but they're clearly two
different alloys.
So just to conclude just a summary of the
post quench mechanical test, we tested in the Bohmert
range. We tested at that 1,100 degrees temperature
and for the reasons that I tried to explain. That's
where the two alloys, Zirc-4 and M5, are oxidizing at
the same rate so that you can see what's really going
on there with those guys.
We have an order of magnitude less
hydrogen uptake than Mr. Bohmert's 110. He was
getting 400 at 1,100 degrees. We got 20, and I've
showed you that we had a completely different oxide
morphology.
And we had no delaminations in our oxide,
in our mechanical test. We had similar bend test,
similar impact test, similar ring compression test to
Zirc-4, significantly better than E110.
We agree with Mr. Bohmert's conclusions
regarding the Zirc-4, significant different results
though in the two different alloys that we've tested,
his E110 and our M5.
Now, just one last slide to summarize the
entire high temperature, oxidation, quench, post
quench mechanical test results. I hope that I've
demonstrated this morning that the M5 in reactor
operating performance is clearly superior to the Zirc-
4; that our LOCA/post LOCA oxidation rates are equal
to or a little bit slower than Zirc-4 and
significantly slower in certain temperature ranges.
Our LOCA/post LOCA mechanical performance
is equivalent to Zirc-4 essentially.
The performance is acceptable and is equal
to or better than Zirc-4 of events of equal duration.
For a low oxide it takes an awful long time to get to
17 percent ECR. If you had the ultimately perfectly
alloy that didn't oxidize at all, you'd never get
there. So some consideration of time has to be taken
into consideration, and I think everybody does, and
that's why we agree that the 17 percent criterion is
valid, if you consider how long it takes to do that.
And, again, with respect to the E110
alloy, our data is completely different.
So that concludes that I had to say.
CHAIRMAN POWERS: Thank you.
Any other comments for the speaker?
Ralph, do we know more about this E110?
We're going to learn more about E110.
DR. MEYER: In the presentation that I
plan to summarize the meeting that we went to, I have
further information on E110 --
CHAIRMAN POWERS: Okay.
DR. MEYER: -- from other laboratories as
well.
CHAIRMAN POWERS: Okay.
DR. MEYER: And I'll give you what I have.
CHAIRMAN POWERS: Good, good. Well, thank
you.
MR. GARNER: Thank you.
CHAIRMAN POWERS: The next presentation we
have is from Westinghouse Electric Company on the
ductility testing of the Zircaloy-4 and ZIRLO cladding
after high temperature oxidation and steam.
Just for Mr. Garner's benefit we will
acknowledge this as a Garner presentation or the
previous presentation as a Garner presentation.
MR. LEECH: Good morning. My name is Bill
Leech of the Westinghouse Electric Company. I'm also
accompanied this morning by Mitch Nissley, who is
sitting back and has already responded to several
questions.
We're both engineers at Westinghouse. I
am a mechanical engineer primarily in the area of fuel
rod and modeling and data analysis, and Mitch is also
a mechanical engineer with an emphasis on thermal
hydraulics, and his primary emphasis is on LOCA
modeling and methods development.
Our purpose here is to give you an
overview of some of our current work in determining
the properties of both Zircaloy-4 and ZIRLO after high
temperature oxidation and steam. Again, this is an
ongoing program. We started it in late January, early
February as a result of some of the information
discovered by Dr. Meyer. It's an ongoing program. It
has still some time to go to completion, but we do
want to give you an update on what we've discovered so
far.
Now, just some background, and I'm sure by
now you've heard it, but let me repeat it once more.
The ductility measurements on Zircaloy oxidized in
high temperature steam were used to establish the
embrittlement criteria, 10 CFR 5046. And those, in
fact, are the basis of the two criteria, of the peak
cladding temperature of 2,200 and an ECR limit of no
greater than 17 percent.
Now, testing consisted in the early '70s
of both quench tests and ring compression tests.
However, we were aware of the presentation by Mr.
Hache of France, and we went back and thoroughly
reviewed the Commission's deliberations, the staff
evaluations, and agree with him that these were
primarily based on ring compression tests, and quench
tests were simply used as confirmatory data.
And the purpose of the criteria was,
again, to insure cladding would remain sufficiently
intact to assure easily coolable geometry, and as a
practical matter, they met that criteria simply by
assuring themselves that after the transient was
completed, the cladding would retain some ductility.
So basically it's a ductility retention after the
LOCA.
Now, before we proceed, I'd like to talk
a little bit about ZIRLO. ZIRLO is our advanced
alloy. It was developed actually starting about 20
years ago, included autoclave tests, extensive tests
in the BR-3 reactor in Belgium, and reactor
demonstrations here starting in the '80s, and it's up
really now to basically full implementation.
There may be several of our reactors that
don't have ZIRLO, but there are very few, maybe three
or four. Well over 90 percent of our cladding we
manufacture now with ZIRLO. That includes both ZIRLO
cladding, ZIRLO thimbles and ZIRLO grids.
To date, the peak rod burnups that we've
gotten are 70,000. Those are a limited number of rods
at North Anna. We have had four assemblies in the
V.C. Summer reactor with individual rods that have
gone over 66,000.
We have taken extensive in pile
measurements both on the growth, corrosion, creep,
growth, both axial growth of the rods and the
assemblies and lateral growth of the grids. Generally
we find that for equivalent corrosion duties, the
corrosion is probably 60 percent of what we get for
Zircaloy-4. Creep and growth are about half.
So these questions I'm sure you would ask
later if I didn't answer now, and that's what our
experience has been.
So we do consider it in all ways a much
better alloy for normal operation.
DR. UHRIG: One question.
MR. LEECH: Yes, sir.
DR. UHRIG: It's described here as being
low tin content. Do you have a number?
MR. LEECH: It is one percent nominal tin.
DR. UHRIG: What?
MR. LEECH: One percent nominal tin, yes.
DR. UHRIG: One percent.
MR. LEECH: Again, we started licensing
this in 1991. The firm formal licensing process was
initiated, and there was an extensive testing program
that supported that included material mechanical
properties, density, thermal expansion, thermal
conductivity, specific heat, phase changes, high
temperature creep, high temperature oxidation at rod
burst. Plus there was an extensive irradiation
program in the BR-3 reactor.
And our conclusion was there were some
phase change characteristics because of the
composition. The phase change from alpha to beta
takes place at a lower temperature. I don't recall
the exact number. I believe about 75 degrees
Centigrade. So it is a lower phase change.
Other than that, we found that the
mechanical properties were essentially identical.
DR. CRONENBERG: Did you show any changes
in creep with sulfur, too?
MR. LEECH: Creep? That becomes a
complicated question because creep is a function of
both thermal creep and in reactor radiation induced
creep.
DR. CRONENBERG: But I'm just thinking of
the presentation before where he said sulfur affected
their creep.
MR. LEECH: We did not make any attempt to
see if sulfur had an effect on creep. The overall in
reactor creep is lower.
Now, as I say, that gets complicated
because that doesn't necessarily mean the thermal
creep. Out of pile thermal creep is lower. The two
components really interact, and we find must less
irradiation creep.
So the overall in reactor creep rate is
much less.
DR. CRONENBERG: What do you have tech
specs on for trace elements?
MR. LEECH: I can't answer. I don't
recall all of those. I mean there's a long list of
them, but I can't remember them. I can supply them
for you if you'd like.
DR. CRONENBERG: I'm just curious because,
you know, prior indications indicated that they
make --
MR. LEECH: Yes. I simply can't recall
them.
So because we saw that the mechanical
properties were essentially identical during the
licensing process, we argued that because of the close
similarity of Zircaloy, ZIRLO and Zircaloy-4, which
again has been described to others as simply Zircaloy-
4 with a little niobium added, that we thought that
the 17 percent criteria should continue to apply, that
no additional testing was necessary.
The NRC agreed with that, and 10 CFR 5046
was amended to say state that the acceptance criteria
applies to ZIRLO. So that was our licensing history
on ZIRLO.
However, as you know, we got some new
information. We became aware of the Bohmert work in
January. Ralph had done some research in December, I
guess, early to mid-December, discovered the Bohmert
work, several other papers by Griger and the Kurchatov
Institute. There were several references that we
became aware of. Basically in mid-January we became
aware of those, and we did a thorough evaluation of
those.
And just some of the things that we saw in
the Bohmert paper, some of the summaries, that the ECR
to cause complete embrittlement -- this is for the
E110 alloy -- is about one third the value for
Zircaloy-4, and that is, in fact, also consistent with
other work that was done with E110. So it was not
only Bohmert.
However, in looking at that, we also
noticed a number of physical differences in the oxide
layers of E110 and Zircaloy-4, and several of the
things that Bohmert mentioned was E110 displays a
heterogeneous appearance to the oxide layer; that
typically if we look at the oxide layer, there were
two separate oxide layers separated by cracks, and
these tend to play -- multi-oxide layers do tend to
play, and his tests, the Zircaloy-4 always had a
glossy black, firmly adherent single layer, relatively
free from mechanical failures, and he noticed a high
hydrogen uptake -- low hydrogen uptake. I'm sorry.
He noticed low hydrogen uptake only if firmly
adherent, crackless oxide layers were formed.
So there seemed to be a good correlation
between the hydrogen pickup and the condition of the
oxide layer itself.
Our previous history, particularly in high
temperature steam oxidation tests that we had done as
far as the high temperature burst test, showed that we
always had glossy, shiny, adherent, black oxide layers
on both Zircaloy-4 and ZIRLO. So we suspected right
away that there was some difference, and it may have
something to do with the oxide layer.
And let me see if -- however, again, we
thought that in the review of all the papers Ralph had
raised some pretty good points, and we really did feel
that we should do some experimental work and verify
that the 17 percent limit continued to apply.
So we did. Having said that though, let
me reiterate that one other thing we wanted to look at
was clearly to make the point that ZIRLO and E110 are
not equivalent for a number of reasons, and the number
one reason of course is that ZIRLO also contains tin
here at the one percent level, a substantial amount of
tin. It contains iron. The iron level is a tenth of
a percent, and it does contain oxygen. The spec on
oxygen is about .125 percent, or 1,250 parts per
million, whereas in E110 it's typically 700 parts per
million. So there are some differences.
Again, the tin and oxygen are alpha phase
stabilizers, which means that the transition
temperature from alpha to beta is slightly higher when
those are present, or somewhat, just slightly higher,
about 100 degrees or so higher than it would be in a
zirconium-niobium binary alloy. So there are some
differences in the phase change temperatures.
We see simply varying differences in the
structure of the oxide layer.
But we did decide to run some tests, and
we put together a test rig in February. Let me
explain to you what it does. Okay. The main test
section is an Iconel tube here, and inside this
basically are two test specimens. The two test
specimens are a piece of ZIRLO tubing and a piece of
Zircaloy-4 tubing.
So we're putting both tubing types in and
testing them simultaneously. They're held in here.
Basically there's a sheath thermocoupler that goes up
here. It has a small ring on it, and we sit the
samples on top of that.
So here in the constant temperature zone
we have a short piece of ZIRLO tubing, a short piece
of Zircaloy-4 tubing. In alternate tests, we actually
rotate them. So one time one is on the top; one time
the other is on the bottom. So we rotate them.
And basically the objective here is to
oxidize them under identical conditions, and then test
them and see how the results compare.
This is a resistance furnace. It's a
clamshell furnace. We preheat it to about 500
degrees, open it up, and then slide the test section
in, close the clamshell and start the heat up.
We go to final temperatures. We actually
have some thermocouples on the outside of here,
outside of the test section which controls the power
when we get to the final temperature.
We have, again, I said that there was a
main sheath thermocouple coming up through here which
sits in the middle of the tubes. So we have
temperatures -- both two temperatures on the outside
and then the temperature on the inside, and typically
they're within three or four degrees of each other.
So we are getting fairly uniform heating.
Okay. We have basically de-aerated water
from an autoclave. It's pumped through our system.
There's a steam pre-heater. We introduce steam into
the test section. Actually prior to heat-up we run a
purge gas through it, purge gas. There's another line
which is not shown here. Purge the system, heat it
up, and start the steam flow through it.
We run it then through a steam condenser.
The hydrogen is vented out to the atmosphere, and we
actually condense the steam so we know what the steam
rates were and how much steam we run through.
Again, the heat-up rates here. There has
been some discussion of what the heat-up rate should
be. In this apparatus, our heat-up rates are about
one degree Fahrenheit per second. Now, that is --
Mitch, how is that relative to LOCA heat-up rates? I
meant to ask you that.
MR. NISSLEY: For a large break LOCA,
typical heat-up rates would be on the order of ten to
15 degrees Fahrenheit per second. Small break LOCA
might be as low as two or three degrees Fahrenheit per
second. So that is a little low.
MR. LEECH: Okay. So this is somewhat
slower than the actual. It is somewhat significantly
faster than Bohmert used. He used, I think, heat-up
rates of about one third that. I believe he was using
about a third of a degree per second.
The final temperatures when we got the
temperatures ranged from 1,800 degrees Fahrenheit to
2,200 degrees Fahrenheit, which is, I believe, 986
degrees Centigrade to 1,204 degrees Centigrade.
We did run another test. We've run one at
1,700 degrees Fahrenheit, which is 926 degrees
Centigrade, because as we'll discuss later, there was
some concern that there was a temperature range
between 950 and 1,000 identified by Bohmert where he
seemed that the E110 alloy was particularly
susceptible to hydrogen pick-up. So we ran that test.
Okay. We studied those for times ranging
from five to 30 minutes. At the end of the time at
temperature, we opened up the clamshell furnace, let
the section cool by both radiation and convection.
The cooling rates averaged about nine degrees per
second for the test temperature down to 1,000 degrees.
Then, again, Mitch, you had some ranges.
I believe that's reasonable.
MR. NISSLEY: A pretty good cool-down.
MR. LEECH: Pretty reasonable with what we
might actually expect.
We don't quench. We let it cool
completely to room temperature. Now, the objective
here is not to run a quench test to see when we fail
during quench, but to prepare specimens for subsequent
ring tests.
We believe that if anything, this may be
somewhat conservative in that we have a relatively
slow cool-down rates for long periods of time. So if
there is going to be any oxygen infusion to transform
the prior beta phase, then this gives it more time to
occur.
So basically the purpose here is to get
specimens for ring compression tests.
Now, let me just give you the status of
where we are in the process. We have done now --
where my notes are -- I would say we've oxidized about
three quarters of the specimens that we expect to
oxidize. Let's see. Okay. Let me first tell you
what we're going to look at before I tell you how many
we've done.
First of all, the number one priority is
oxide layer characteristics. We believe that of all
the things that we've seen with E110 and Zircaloy-4,
that seems to be the biggest difference, and we want
to take care to look at those. We're doing those by
optical metallography and just general observations.
The next thing would be ring compression
tests to assess the cladding ductility. Those will be
done at room temperature at 275. Two, seventy-five,
I believe, is the official number at which the 17
percent criteria was set up at.
With a tester similar to those performed
by Hobson and Rittenhouse in ORNL report in 1972,
we've attempted to maintain the same length-to-
diameter ratios of the specimens, maintain the same
head speed on the compression rate on the slow
compression rate tests, and these were also similar to
Bohmert, although there were slight differences.
Well, one thing that we did different was
Hobson and Rittenhouse only went to a fixed
displacement and stopped their compression test, where
Bohmert continued to going until he either got clear
indications of a failure or was getting too close to
where he simply couldn't compress them anymore and
backed off. We did that. We thought it gave a little
more information.
There are some other differences. Bohmert
cut his specimens into short sections prior to
oxidizing them, where we oxidize a specimen about that
long, and then we cut the rings out afterwards.
We measure the weight gain of the total
specimen, and then we cut sections out of it, which is
a slight difference, although I don't think it should
make much difference.
Again, we cal look at the oxide thickness.
We're going to look at the thickness of the alpha
stabilized layer and the transformed beta layer. We
will do micro hardnesses across the cladding wall to
assess the oxygen penetration, and then we'll do
measurements for total hydrogen and oxygen
concentrations.
There's some of the matter we've gotten so
far. What this is is a plot of the measured oxide
thickness in microns. This was developed from
metallography, plotted versus the oxide thickness that
would be present if all the oxygen weight gain was
transformed to an oxide layer.
And so there's a couple of interesting
things here. One is that if you look, you'll see that
if all the oxygen had been done into an oxide layer,
then we would expect to go across about -- for a
prediction of 100, you go across and we actually
measured 70, which indicates that about 70 percent of
the oxygen is going into the oxide layer and about 30
percent is going into the metal.
But what we also noticed is that for
Zircaloy-4 and ZIRLO they're identical. There's
really no difference between them, and I think that's
a key difference because in one of the papers, when
they looked at the E110 alloy they said that although
for equivalent weight gains the distribution of the
oxygen could be significantly different. A much
higher percentage of it actually for E110 has ended up
in the metal rather than the oxide layer.
So we believe that's a significant
difference. We don't see any difference here between
ZIRLO and Zircaloy-4.
Anything else I might want to say about
this? No.
Then the next result we have are the
results from the ring compression test. These are the
ones we've done at 275 degrees Fahrenheit. What we've
plotted is the relative displacement of failure.
Relative displacement is the amount of compression
divided by the other diameter of the specimen versus
the measured ECR fraction. Now, this is not
calculated; measured. There's an important
distinction there, and that's the ECR assuming all the
oxygen weight gain is stoichiometrically combined with
the metal.
We see several things. One is we see that
Zircaloy-4 and ZIRLO are for all intents and purposes
the same over the whole range that we've tested. We
see no difference whatsoever.
This is Bohmert's brittle limit. Whether
that's our brittle limit or not, that still needs to
be investigated because we need to look at each of
these specimens and look at the nature of the failure.
Was it brittle, ductile, or partially brittle and
partially ductile?
We know from already that these were
clearly brittle, and some of these actually are still
in one piece, you know. After we bent them down,
they're still in one piece. So they're obviously
ductile, but we have to take some care to look into
this area to suggest exactly what is the ECR at which
we get transition or we are in a position where we're
totally brittle.
Again, one other thing I might mention,
too, which I haven't plotted, haven't shown you.
We're also doing this at room temperature, and we've
looked at some of the preliminary results that we got
for Zircaloy-4 at room temperature, and they're
reasonably in good agreement with what Bohmert got in
his test for Zircaloy-4, which again is another,
probably a second opinion that what he did was really
pretty good work. There was no problem with what he
did. It's just that the E110 seems to be
substantially different than Zircaloy-4 because our
Zircaloy-4 results seem to be consistent with his.
So which I guess is good. It tells us our
Zircaloy-4 results were consistent with his, and our
ZIRLO results are essentially equivalent to our
Zircaloy-4 results, indicating that for ZIRLO-4
there's no reason to think that the 17 percent
criteria doesn't continue to apply.
This, again, is measured ECR. It's not
Baker-Just. Baker-Just probably is conservative by a
factor approaching two. So we don't see a problem.
Again, what did we see? Just comparisons.
Both oxide layers were dark adherent with no
laminations. Both have similar fractions of oxygen in
the oxide layer and in the metal. Ring compression
tests of similar values of displacement of failure
versus the measured equivalent planning reactant. We
believe that the ZIRLO and Zircaloy-4 are just
essentially exhibiting the same behavior. I see no
difference at this point.
Again, we still have some more work to do
on this. We're going to prepare for the remaining
sample preparation. We've got to complete all the
tests. We have got a few more samples to prepare.
We've got some of the -- about a third of the ring
compression tests to still do. The metallography
samples have been made, etched. They have not
necessarily all been evaluated yet.
We want to get all of the data, and what
we really want to do then is get a good independent
review. Those of us working on the project have
reached our conclusions, but we want to bring in
outside people both from in our company and
potentially from outside the company to look at what
we've done, document and review the results.
And our next scheduled meeting to discuss
this with the NRC now is May 16th, I believe. There
will be a review meeting. So we'll give another
update at that point.
That really is what I planned to say
today.
CHAIRMAN POWERS: You mentioned several
times that your Zircaloy oxides showed no evidence of
delamination.
MR. LEECH: Right.
CHAIRMAN POWERS: And the previous speaker
showed some micrographs in Zircaloy-4 that had
evidence of delamination.
MR. LEECH: Okay. Excuse me. One of
those, I believe, was after spalling, wasn't it? Was
that before or after?
After spalling it certainly showed
delaminations.
CHAIRMAN POWERS: I guess my question is,
really boils down to: what causes the delamination?
MR. LEECH: What causes? Obviously it's
a stress and a differential thermal expansion.
CHAIRMAN POWERS: Okay.
MR. LEECH: But what causes one to crack
and one not to crack, I guess I don't -- I don't know.
CHAIRMAN POWERS: Okay.
MR. LEECH: I don't know.
CHAIRMAN POWERS: Any other questions of
this speaker?
(No response.)
CHAIRMAN POWERS: Well, thank you very
much.
MR. LEECH: Thank you.
CHAIRMAN POWERS: Our next speaker has
protested he's hungry, and so I'm going to recess for
lunch, and we'll pick up Dr. Meyer's discussion of his
OECD meeting after lunch.
Thank you.
(Whereupon, at 11:58 a.m., the meeting was
recessed for lunch, to reconvene at 1:00 p.m., the
same day.)
A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N
(1:01 p.m.)
CHAIRMAN POWERS: Dr. Meyer is going to
give us a precis of the OECD topical meeting on LOCA
fuel safety criteria.
DR. MEYER: The meeting was organized by
an OECD related group. Within CSNI there are several
special expert groups, and there's one on fuel, on
fuel safety margins. And it is this group, on which
I am a member, that organized the meeting.
We'd had a similar meeting. A similar
group in OECD had organized a similar meeting in 1995
on the reactivity accidents, very early in the period
where we were looking into that. And it was very
helpful because it brought a lot of people out of the
woodwork and got a lot of information out in public
that could be talked about.
And we decided before the Bohmert paper
surfaced to organize this meeting, but when we learned
about the Bohmert paper, it became sort of the center
of focus of the meeting.
So the meeting really had three groups of
papers: one on post quench ductility, one on axial
constraints during quenching, and one on relocation of
fragmented fuel into the ballooned region.
I have more material in the handout than
be covered in a reasonable amount of time. So I think
I'm going to just focus on this first group here. And
also I'll skip over quickly some things that have
already been discussed.
The first couple of slides in the package
were from an introductory presentation by George
Hache. They go over the ECCS rulemaking hearing and
the fact that the criteria were developed from ring
compression tests and that's been discussed, and I
don't think that's a matter in contention. So I'll
just skip that.
Now, Bohmert is from a research institute
in Dresden, Germany. I did contact him. He was
unable to attend the meeting. But George Hache
presented, among other things, the main slide, the
main figure from Bohmert's report in 1992 that shows
the effect.
Now, you saw a few of these points on
Framatome's slide, where they picked out the ones at
1,100, and they picked those out from Bohmert's slide
and showed them on their graph.
But Bohmert had tested over a wide range
of temperatures, both Zircaloy-4 and the VVER
cladding, E110. And you can see this is the line that
was on the Framatome slide. And you can see it coming
down here around five percent cladding reacted.
I think Bohmert did his tests at room
temperature. And George Hache looking back at all
this says, "Well, it really should have been at 135
degrees Centigrade," so it would be a little higher
than that.
But nevertheless you can see here,
although there is scatter in the data, you can see a
separation between the E110 ductility results and the
Zircaloy-4 data results.
Now, Bohmert is not the only person who's
seen this. This has been seen at four different
laboratories in four different countries, was seen in
Germany. It's been seen in the Czech Republic, in
Hungary, and in Russia.
The Hungarian researcher who did the
confirming work there was present at the meeting and
has a paper and I have a slide from that.
The Czech researchers did not document it
in a public place or in English. They wrote it up in
a agency report in Czech, whatever, in Czech. But we
have contacted them and we may be able to retrieve
that data and get it in an English report.
And then in addition to that, George
Hache, who has this incredible talent to remember
things from obscure places, remembered some meeting in
Varna. I don't even know where Varna is, in 1994
where the Russians presented such results.
And so added to the three that we had been
talking about, the Germans, the Czechs, and the
Hungarians, here is the Bochvar Institute with ring
compression test results and a line that separates the
ductile from the brittle behaving specimens.
And when George -- this handwriting is
George Hache's. He's informal sometimes. When he
goes down this separating line down to the 135 degree
temperature point, and he gets the six percent figure.
So George says, "If you apply Hobson's
methodology to this set of data from the Bochvar
Institute, you get a six percent ECR," which is
consistent with the others that we have seen.
Now, the main presentations on this
subject were given by Maroti from Hungary, Sokolov
from Russia, Lebourhis from France, Bill Leech out
here in the audience, and Hee Chung from our program
at Argonne.
There were actually two papers on the
subject from Russia and I only have a slide from one
of them. The other one was kind of preliminary, and
frankly, I was never able to understand the main
results of that paper and have gone back to try and
get clarification.
So let me just show you a few of the
slides which are fairly easy to grasp, and which I
think will summarize the essence of the material that
was presented at the meeting.
This is the Hungarian work, and I think
it's even cleaner in appearance than the Bohmert work
in terms of seeing the drop-down in the ductility of
the E110 specimens compared with the Zircaloy
specimens.
It's interesting that at least in the
German, the Czech, and the Hungarian work, they always
measure Zircaloy along with their E110 measurements.
So there's a control. And Hee Chung at Argonne has
taken their Zircaloy results and replotted them along
with his own ring test results from the '80s and
Hobson's from the '70s, and they're all consistent,
which is what we heard this morning as well.
So all of these laboratories appear to be
able to make consistent measurements on Zircaloy, and
we get these two sets of differences for the zirconium
1-niobium, and the difference is remarkable. It's not
just a small difference. I mean, from 17 percent to
six percent is a huge reduction.
Now, Sokolov in his presentation included
this figure, and George Hache made interesting
observation from this figure. This is not ring
compression tests, now. These are quench test
results. This is a failure map, and we often plot
failure maps like this where we have the log of the
time, the temperature versus one over temperature, and
show on the plot usually the 17 percent line which
would go on down, but then truncated by the 2,200
degree Fahrenheit curve.
And I'll show you a figure for Zircaloy.
Generally, there is a substantial margin shown above
the boundary until you get to the beginning of the
failures. And you see, you see a margin along here,
but when you get to 1,200 degrees the ductility seems
to start a nosedive, and you have very little to no
margin right here at the knee in the curve.
Now, that was presented -- that figure was
presented at the meeting by Sokolov. George Hache
makes the observation during the discussion and George
Hache -- I don't know if he used these exact figures,
but he pointed me to them and we got them out of our
own reports.
But this is a failure map for Zircaloy
test summarized in a report by Van Houten, but Van
Houten didn't do the work. This work was done at
Argonne.
Okay. The construction lines are not laid
on this figure, but the data points are, and what I'm
going to show you on the next figure, now, is a figure
with construction lines on it and no data points, but
it's the same figure, and you'll see this is Figure 2A
from the reference and this one is Figure 2B from the
reference. And this solid curve here, then, is the
one that bounds the thermal shock failures.
There's some other things on here. And
here is the construction that shows the 17 percent
line and the 1,200 degree limit. And you see quite a
bit of margin, and across here there's a good 100
degree C. margin in this, which appears to be absent
from the E110 plot.
Just an observation that George is saying
is not only the ring compression test that are giving
us this message. There's the quench tests that are
giving us this message.
Okay. Now, I'm not trying to suggest that
this is the same message, but this morning in the
Framatome presentation we did see numbers that were
close to the 17 percent line which don't have a lot of
margin exhibited. I don't know whether that's
significant or not significant, but I point it out to
you.
On the other hand, and you saw both of
these, this one and the Westinghouse figure before,
there is just no difference apparent at all when you
do the ring compression -- when you look at
Framatome's ring compression test and Westinghouse's
ring compression tests. So you saw these slides this
morning.
I ask Labourhis directly at the meeting
what was his opinion as to why there was such a
difference between E110 and M5. And his answer to me
was, "I have no idea."
Now, there's a suggestion that there's a
difference in the material. There are some
differences in the test procedures.
Nothing is apparent at this point. It's
pretty much a mystery.
George Hache makes another observation
which is rather obvious, but kind of important at the
same time, is if it really is a difference in the
material, we kind of ought to understand it because we
may inadvertently move into that material regime. And
it makes a big difference.
Now, you were asking some questions about
the composition, and I have compositions of E110 and
M5 from a couple of sources. The main points in this
table are from a recent Halden report, where they're
testing specimens. I don't know whether they're
coupons or tubular specimens, in some oxidations
tests.
And they have reported these numbers.
These look like -- I would say these look like numbers
that were measured, but I'm not sure about these
numbers here.
Anyway, there are also papers in the open
literature in the ASTM, you know, the zirconium in the
nuclear industry conference that they hold every three
or four years.
There's one with M5 results written by
Framatome authors, and one with E110 results written
by Russians that show these ranges. And you can see
a few hundredths of a percent more oxygen in M5 than
in the E110, and the iron, there's a little more iron.
It's a very small amount.
Both are recrystallized. The E110 is said
to be alpha recrystallized, so it's recrystallized.
It's annealed at a temperature below the phase
transition.
DR. BONACA: Does it show sulfur there?
DR. MEYER: Huh?
DR. BONACA: Does it show sulfur?
DR. MEYER: No, I couldn't find any sulfur
content.
DR. BONACA: We heard this morning
about --
DR. MEYER: Yeah.
DR. BONACA: -- M5, I thought.
DR. MEYER: Did mention the sulfur this
morning, and I don't have any numbers on that.
I'm not sure that the cold work and the
annealing is going to make any difference when you get
into this regime of oxidizing above the face
transition. It just seems to me like it's a soup of
elements at those temperatures, and the chemical
composition is really close.
I simply don't understand it. I don't
have a theory or, you know, a big hunch. It's just
hard to believe that it's the test procedures because
they use controls all along. It's hard to believe
it's the material because the material is so similar.
It's hard to believe that it's the fabrication and
cold work related things because it's a high
temperature process that we're looking at, and I don't
know.
Now, at this point in the meeting Hee
Chung gave a lecture. Bill Shack will understand that
Hee Chung likes to give lectures, and he gave us a
lecture on a post quench ductility of zirconium
alloys. And he repeated a number of things that we
already knew and were talking about.
But he did bring out a couple of other
points. I'm not sure whether all have been verified
or not. But he points out the matter of the hydrogen
induced ductility. And that hydrogen induced -- the
role of hydrogen in reducing ductility wasn't
understood in 1973, when Hobson's tests were done.
It was all thought to be oxygen. The
levels of hydrogen in the specimens at that time were
low, less than 150 parts per million, where it
wouldn't have been above the threshold for some effect
anyway.
But let's see if this is the -- well, I've
got a couple of figures here.
Hee Chung now points out that for
Zircaloy, that there seems to be a threshold around
600 or 700 ppm hydrogen. When you get that much
hydrogen in the specimen, then it also contributes to
the reduction of ductility.
And he has looked at Bohmert's data and
Griger's paper. Griger is one of the Hungarian
workers, and believes that he sees a threshold at a
much lower level, down around 150 to 200 parts per
million. Now, in the specimens that we heard about
this morning, the concentration of hydrogen was even
lower than that. So you wouldn't have been there.
And Hee Chung insists that we have to
consider several factors and not just one. It's not
just hydrogen. It's not just oxygen. It's not just
niobium.
And then he presented this one slide,
which is rather useful, to talk about the three routes
to getting a lot of hydrogen in the specimen and how
we only have hydrogen from one of these routes in the
specimens that we're testing at this time.
You can get hydrogen during normal
operation, and of course, we have not been testing
that because the tests that we've been looking at have
been on fresh tubes.
You can get hydrogen in the high
temperature process. This is what we've been looking
at.
And then there's another process that lets
hydrogen into the cladding associated with the
deformation during ballooning and rupture.
And this, I believe, is the process that
led them to identify the role of hydrogen in
embrittlement because apparently when you get this
deformation, and you now have two-sided oxidation, you
have a stagnant steam environment on the inside and
the hydrogen doesn't get swept away, and the
absorption of the hydrogen locally in that region is
very high.
And so when they -- this work was done at
a couple of -- I guess it was done at Argonne and it
was also done at JAERI, in the early '80's. And when
-- if you took slices near the region of the burst,
took rings and looked at their ductility, they would
not pass the non-zero ductility test related to 17
percent oxidation. So there's a local effect that's
fairly strong.
Well, this slide suggests the importance
of making some measurements on some real fuel rod
material and not just on tubes in the laboratory. And
of course, that's what we are interested in doing in
our research program.
And then I was asked to give a brief
presentation on our research program, and these are a
couple of slides that I used. The first bullet
outlines the program that we have at Argonne at the
present time using Zircaloy. There have been some
adjustments to this based on the PIRT process that was
completed.
And now that we have our Zirc-2 and Zirc-4
in the laboratory and those tests are planned and
ongoing, we'd like to start making arrangements to
obtain some ZIRLO and M5 in this program.
And as I think I mentioned earlier, we
broached this subject with Framatome and Westinghouse
at the meetings that we had in February here at NRC.
I think that if we carry out this full
range of studies with Zirc-2 and Zirc-4 that we may
not need to repeat everything in that menu for the
other cladding types. We might, for example, be able
to skip the integral tests. It's an expensive test.
I'm not sure that we'll be able to, but you might be
able to characterize things well enough from them,
from the simpler tests that they were measuring
mechanical properties.
And so, in particular, we're quite sure
that we'd want to do oxidation kinetics measurements,
probably some sort of thermal shock test, look at the
oxidation and the phase relations and measure the
mechanical properties after running the material
through a high temperature oxidation transfer.
DR. KRESS: Ralph, this looks to me like
more data is going to an empirical relationship. Does
this address Mr. Hache's comment or we need to
understand the effects of small material differences?
I don't see that it addresses that.
DR. MEYER: You don't see it directly, but
it -- we really want to -- I'm not convinced that it's
a small materials difference that's doing this. And
so one of my main objectives is to find out what it
is.
DR. KRESS: Okay. This will do that.
DR. MEYER: Well, it will for it's part of
the equation. The other part of the equation is the
E110 alloy. And what you don't see up here, but it's
buried in one of the bullets on another -- in another
presentation, was that we have this program with the
Kurchatov Institute, and in starting in late 2001,
this year, late in the year, we have them beginning a
series of tests that are designed to shadow this
program in their laboratory with E110.
So we want to look very carefully at ring
compression tests, whether that's the right test or
not. These tests have been criticized in the past.
They're not real precise. They're good screening
tests for some purposes, but maybe an axial tensile
test might be a more precise way of looking at the
ductile brittle behavior.
CHAIRMAN POWERS: When you look at you
specimens, it looks to me like chemical compositions
not going to answer the question for you. They're too
close together.
DR. MEYER: Yeah.
CHAIRMAN POWERS: Now, maybe EDAX on the
distribution of the alloying agents may be different.
Maybe that tells you something, but do you also look
at things like grain size and surface texture?
DR. MEYER: Well, we would. I don't think
we're far enough along to say what we have planned out
in a test matrix, but those are the easy things to
look at, and the kind of things that we would normally
do.
DR. CRONENBERG: Ralph, a couple years ago
you had voted the idea of a 100 calories per gram for
high burnup --
DR. MEYER: Yeah.
DR. CRONENBERG: -- plus a criteria of
retention of residual ductility, that maybe the two
might be the way the regulation should be written up.
Does this flow from that thinking?
Is that thinking still in effect that
there might be a requirement of some residual
ductility rather than hydrogen and oxygen, then oxygen
and hydrogen uptake?
DR. MEYER: It's not really connected,
although you come out at about the same place. The
100 calorie per gram dealt specifically with a rod
ejection type accident. And that's a accident where
the cladding remains at a relatively low temperature,
and where you haven't gone through a phase
transformation and wiped out its fabrication history
and all of that.
Now, the ductility initially when we were
looking at the rod ejection, we were trying to see if
we could use the ductility criterion instead of an
enthalpy criterion.
And the critical strain energy density
method that EPRI and the industry use, and that IPSN
uses and EDF uses, is, in effect, a ductility based
criterion.
But the origin of the two are quite
different because at that time we weren't thinking
about the ECCS hearing and what was done there and
Hobson's results, and so forth.
DR. CRONENBERG: Okay, but I guess I'm
still not clear. Is your thinking still in terms of
residual depility (phonetic) criteria? Is this still
in the background for these experiments?
DR. MEYER: Certainly for the LOCA it is,
definitely. I mean, this is the result of the
hearing,and the philosophy we've been following even
though we forgot that we were following it.
I mean, that was these criteria that we're
using were based on retained ductility.
DR. CRONENBERG: But it's 10 percent
oxidation, not in ductility requirements.
DR. MEYER: Yeah. So we may have to roll
it back to the concept of ductility and look again at
what attribute might characterize that adequately for
us.
DR. CRONENBERG: Okay.
DR. MEYER: Okay. Now, along with the
work on irradiated fuel rods, we'd like to -- well, we
always in our program at Argonne, where we're looking
at irradiated fuel rods, we always look at archive
unirradiated material and do pairs of tests so that we
can tell the difference between the behavior of fresh
material and irradiated material.
There's a lot that we have learned and I
think we sill can learn with the unirradiated tubing,
and so if we can make some arrangement with Framatome
and Westinghouse to work on their materials, we'd like
to get started very quickly on the unirradiated
tubing.
And here was a list of things that we
proposed to do in a program in which we would ask for
their cooperation.
And you see at the top of the list is to
look at ring compression tests and other post quench
ductility measurements to make sure that we're not
using a test that itself has some inherent problems.
And we would propose to discuss this until
we get some agreement on what is -- if the ring
compression test is not the right test to use, what is
the right test, and then to carry this out.
And there's a branch point over here where
the same instructions go to Kurchatov Institute in our
corollary program with E110 alloy.
CHAIRMAN POWERS: The entry on the slide
that I guess I don't understand, it says no mechanical
properties or other testing at this time --
DR. MEYER: Yes.
CHAIRMAN POWERS: -- later in the high
burnup program. I was wondering --
DR. MEYER: Why?
CHAIRMAN POWERS: -- what other program is
there?
DR. MEYER: In the Argonne program, which
we often think of as a LOCA program, we also have a
matrix of regular mechanical properties testing under
low temperature, higher strain rate conditions that
match up with the reactivity action.
So there's a lot of mechanical properties
testing related to rod ejection action and related to
the ballooning process.
This is before you get to the high
temperature and the oxidation. And what we're saying
here is that for the moment we wouldn't enter into
those tests immediately. We would do those in
connection with the high burnup tests at a later time.
It's partly a matter of resources. It's
partly a matter of trying to work with the industry so
that we don't reveal too many things about their
proprietary materials that aren't necessary to reveal
at this time in connection with looking for some
explanation of this LOCA ductility behavior.
So that was put in there to try and be
nice guys.
CHAIRMAN POWERS: No good deed goes
unpunished here, Ralph.
DR. MEYER: And we have a current program
that's working very nicely with EPRI, and we would
just pattern it -- pattern it after that. So I've
said all of these things.
Now, I have a few more slides from the
other discussions. If you don't ask questions, I can
show them quickly or I can just sit down. So it's
your choice.
CHAIRMAN POWERS: Why don't we rely upon
the members to review the additional material and --
because I'm anxious to hear what Margaret and Richard
have to say.
DR. MEYER: Okay.
CHAIRMAN POWERS: And thank you for your
presentations.
I'll comment that the ACRS has made a
suggestion to the Commission that this program be
given additional resources to test additional types of
materials, and it sounds like you very much need it
right now.
At this point, we'll shift gears just a
little bit and move to the business end of the agency.
And Margaret will give us some talk about recent
operational issues and experience with high burnup
fuel.
MS. CHATTERTON: Okay. It'll take me a
minute to get myself organized.
CHAIRMAN POWERS: Oh, yeah, we permit
that. Have you been running lately. That's the
question we want to know.
MS. CHATTERTON: Have I been running
lately? Today was a running day, but there wasn't
enough time. So I ran Monday.
DR. KRESS: We messed up your running?
MS. CHATTERTON: You messed up my running.
CHAIRMAN POWERS: You should have
protested.
MS. CHATTERTON: Two weeks from Monday is
Boston. I will -- did I get this thing on right? --
I will be back at Boston, which I think will be a slow
run, but it will be fun, and that's the major thing.
CHAIRMAN POWERS: That's right.
DR. KRESS: Just as long as you don't that
shortcut.
MS. CHATTERTON: No, I don't take any
shortcuts.
Actually, right now they have a timing
chip. It goes on your shoe. It starts at the
beginning, and they have a map that you run across
every five kilometers.
DR. KRESS: Oh, okay.
MS. CHATTERTON: So they've got your time.
DR. KRESS: They've got you.
MS. CHATTERTON: You can't cheat.
CHAIRMAN POWERS: Well, you can cheat
every three kilometers or something like that.
MS. CHATTERTON: Anyway, I'm here today to
talk about operational issues and high burnup fuel as
we've been using them in the last few years.
And here's kind of an outline to some of
the things that I want to talk about. It's been a
couple of years, I believe, since we talked about
burnup extension activities. So I thought I would
just start off with that, talk a little bit about
where we are on lead test assembly guidelines, some
recent fuel issues, and then talk a little bit about
the current fuel reviews that we're in the process of
doing.
CHAIRMAN POWERS: Okay.
MS. CHATTERTON: So as you probably
remember our basic approach to burnup extension is
that we're working with the industry to develop a
strategy and a plan. It's going to be up to the
industry to do the testing, to come up with the
criteria, and then to justify the criteria.
At that point the NRC will review what the
industry proposes, review their justification, and at
some point, hopefully, be able to endorse their
proposal as a regulatory guide.
We simply do not have the resources to do
the research, to come up with the criteria like we did
in previous times.
So, again, I think you've probably seen
this. Our burnup extension guidelines will be working
with the industry. We've required that they will give
them some advice.
Certainly, it must address the current
licensing requirements, the LOCA, the RAA and the
ATWS. all of those things that are looked at today.
They'll have to give a justification of
why any limit that they decide to use is appropriate
going to higher burnups. And just as a review, what
the industry has said they want to do is to go to
probably 70 gigawatt days for BWRs -- that's the rod
average -- and 75 for PWRs.
We've also told them that some of the
area's can be risk informed. That's going to be their
determination of exactly how they want to handle
different things. And, again, it's all going to be
subject to our review.
DR. KRESS: When you say risk informed, is
that you're thinking Reg. Guide 1.174, risk informed
there?
MS. CHATTERTON: Yes. They will be able
to use some of the guidance that we've given before.
They may look at certain things and decide that they
want to make a proposal that certain things can be
handled slightly differently on a risk basis.
DR. KRESS: See, what bothered me about
that was Reg. Guide 1.174 is based on current burnups,
and if you're going to extend the burnup, then you
have a little bit of a circular argument because then
you have to ask whether 1.174 has the right value of
LERF in it, for example.
MS. CHATTERTON: Yes.
DR. APOSTOLAKIS: But it deals with delta
LERF.
DR. KRESS: It also deals with absolute
value of LERF.
DR. APOSTOLAKIS: Yeah. So I mean I don't
understand what it means that --
DR. KRESS: Even delta LERF is going to be
hard to determine because you're dealing with the
delta fission product maybe. And it's not just
inventory. You can handle that pretty easily.
DR. APOSTOLAKIS: In other words, what
you're saying is LERF might not be the right metric.
Is that what you're saying?
DR. KRESS: That's another issue I have.
That's a separate issue.
DR. SHACK: But I think he was arguing
that acceptance criteria value.
DR. KRESS: Yeah. I was arguing on --
DR. APOSTOLAKIS: On the acceptance
criteria is acceptance criteria. Why should it be any
different?
DR. SHACK: Well, I suppose you could look
at it that way, too.
CHAIRMAN POWERS: Well, if Tom is thinking
the way he has been thinking in the past he says,
"Hold it."
DR. APOSTOLAKIS: Says what?
CHAIRMAN POWERS: He says, "Hold it." You
derived your acceptance value by looking at the
quantitative and health objectives.
DR. KRESS: Absolutely.
CHAIRMAN POWERS: Now, you can't do that
anymore because the derivation path doesn't work.
DR. KRESS: That's right. That's exactly
the way I was thinking.
DR. SHACK: The source term is different.
DR. KRESS: Yeah, maybe. We don't know.
DR. APOSTOLAKIS: The quantitative health
objectives don't change, do they?
CHAIRMAN POWERS: We assume those are are
given to us by God.
DR. APOSTOLAKIS: Working backwards, you
have assumed certain behavior in severe accidents.
And that's what's going to change?
DR. KRESS: Yes.
DR. APOSTOLAKIS: Okay. So the LERF value
then may change.
DR. KRESS: That's what I was saying.
DR. APOSTOLAKIS: Now, the CDF will not
change?
DR. KRESS: No.
DR. APOSTOLAKIS: Because we lowered it by
a factor of ten arbitrarily. Right?
CHAIRMAN POWERS: I mean, I don't know.
I mean, it seems to me --
DR. KRESS: Must have had a reason for
that.
CHAIRMAN POWERS: Maybe it turns out that
things are more susceptible to core damage.
DR. KRESS: Yeah.
DR. APOSTOLAKIS: It's more than a factor
of ten what you be now. I mean there's a problem
somewhere. But anyway, it might be that that --
CHAIRMAN POWERS: Factors of ten are not
our of the question here.
DR. APOSTOLAKIS: Now, you said some parts
may be risk informed. So you have decided that some
parts may not be?
MS. CHATTERTON: No.
DR. APOSTOLAKIS: Okay. Just a figure of
speech.
MS. CHATTERTON: Yes, a figure of speech,
but basically we're letting the industry propose how
they want to handle -- how they think is the
appropriate way and to provide a justification, and
again this will go and do the review.
CHAIRMAN POWERS: I guess, when you raise
the issue of being risk informed, the challenge I see
there has something to do with just what our
discussion was. We typically don't have a great deal
of information on these fuels under accident
conditions, severe accident conditions that will give
you any consequence.
Are you saying that the industry can come
in, but they've got to come in armed with information
on fuel behavior under accident conditions?
MS. CHATTERTON: That might be an option
if the proposal is to go that direction. The main
point is whatever method they decide to take, they
have to provide the justification for why that's
acceptable, with a great deal of emphasis on lead test
assemblies.
That's one thing that we have emphasized
greatly in the last few years, I would say in the last
five years, and that's a result of fuel issues that
we've had in these last five or so years, and I'll be
talking about some of those later, and the things that
we've learned that the fuel -- the lead test
assemblies in the past did not always give us data or
information that was really the most useful.
We've also said that a breath (phonetic)
extension program will also have a fuel performance
monitoring program. Somebody said the fuel
performance monitoring; that's in core. I guess maybe
it's really fuel surveillance program.
DR. BONACA: If you'd just stay with that
slide, I would like to ask a question.
MS. CHATTERTON: Sure.
DR. BONACA: Clearly, the first bullet I
can see that you are concerned about how long the
cycle is going to be or the burnup.
MS. CHATTERTON: Yes.
DR. BONACA: And the issues that we
discussed this morning. At the bottom there, I see
fuel performance monitoring program. Now, currently,
I mean, although it may be a concern to have fuel
failures, the one percent for the fuel assumptions in
analysis allow for 50 pins probably are going to be
failed.
Okay. So I'm curious about what would
this fuel performance monitoring program mean. I
mean, for example, some of the Westinghouse plans have
exhibited at times maybe four or five 17's failed in
some batches. Okay. That's really an operational
concern.
Is it also a regulatory concern right now?
Is that what it's focusing on?
MS. CHATTERTON: This isn't focusing just
on fuel failures.
DR. BONACA: Yeah.
MS. CHATTERTON: This is focusing on
things like corrosion, growth.
DR. BONACA: Okay. I understand.
MS. CHATTERTON: It's focusing on all the
types of parameters that -- I want to characterize it
fairly by saying in the past many times the fuel has
been burned, taken out, put in the spent fuel pool,
and never looked at.
DR. BONACA: I understand.
MS. CHATTERTON: As a result we had some
problems that might have been eliminated had the type
of program that I'm talking about --
DR. BONACA: So for example, oxidation
rates because those also go in estimation of
performance under accident conditions.
MS. CHATTERTON: Yes.
DR. BONACA: Okay. I understand.
MS. CHATTERTON: Yes. And if you're not
measuring your oxidation levels, you don't know if
your inputs to your accident analysis are correct.
DR. BONACA: Okay. Thank you.
MS. CHATTERTON: And that's the main point
in that.
DR. CRONENBERG: Margaret, on that, the
NRC used to publish a fuel performance summary. Every
year PNL used to do the work.
MS. CHATTERTON: Yes.
DR. CRONENBERG: They used to summarize
it. That's no longer in effect.
MS. CHATTERTON: That's correct.
DR. CRONENBERG: Are you going to
reinstitute this type of summary like the PNL used to
do, but EPRI or industry or somebody will be -- will
it be a formal, published monitoring program?
MS. CHATTERTON: I don't think we're far
enough along to really be able too say exactly how
that's going to work. How I envision when we will
come up with a reg. guide will be it listing the types
of testing that needs to be done, giving some ideas as
to the frequency and when.
For instance, if you fuel goes beyond 62,
but it's only to 63, and it's ten years down the line,
it probably doesn't need to be measured again.
On the other hand, the different type of
fuel, the slightly different power history, some of
those, it's going to be difficult to come up with
exactly how we handle this. There's going to be a lot
of thought into that such that it provides enough
data, but it doesn't totally hamper the industry such
that they have to measure everything because that's
not the intent.
It's going to have to be set up with
controls such that after so much data, there's not
need. If, on the other hand, if results aren't
turning out to be good, then you need more. And it's
going to have to have triggers in it for when you do
more results, when you do more testing and also when
you would need to do a hot cell.
Most of the hot cell exams, most of this
is going to be pool site exams. The types --
DR. BONACA: That's what it was. The PNL
was mostly pool site exams.
MS. CHATTERTON: Yes. Oh, yes.
DR. BONACA: But we don't have that data
anymore.
MS. CHATTERTON: We don't have that data
anymore.
DR. BONACA: I would hope that if you're
going to push it to 70-75, that type of program goes
on for a few years until you've had that --
MS. CHATTERTON: Yes. And that's the type
of thing that I think we're thinking about. Yes, I
miss having those reports.
DR. BONACA: Yeah.
MS. CHATTERTON: Those are great.
DR. BONACA: There was a lot of data.
MS. CHATTERTON: Where are we in this
whole plan?
Well, in the last year or so there's been
some progress. I would say not a tremendous amount.
Although the industry is working on it and it's slow,
sometimes it comes in big steps.
We had a draft submittal in March of 2000,
and the staff provided comments, and we had another
meeting with NEI December 6th. They outlined their
approach for RAA, and the staff gave them comments
saying that it looked fairly reasonable.
And basically what they're doing is
proposing a clad failure and coolability limits that
are a function of burnup. They are based on enthalpy
increase, and we've seen the preliminary work on this.
We haven't seen all the details; we haven't reviewed
all the details.
What they presented looks like a
reasonable approach. Again, one it's submitted we
will do a complete review of it.
We expect a submittal late summer. Again,
sometimes work takes much longer than they think.
Originally that was an early 2001 date, and it's been
changed.
CHAIRMAN POWERS: If you get a submittal,
say, in August, when do you think you'd have your
review finished?
MS. CHATTERTON: This submittal I expect
in August will not be a complete submittal. This will
be a partial submittal and it will depend on the
amount that's in it.
I think a submittal like this is going to
take us a considerable amount of time, six months or
so I would say on half of it, possibly a year or more
on the complete package.
If it comes in in pieces, which is I think
the intention, we will kind of review it in pieces so
that, one, we can get feedback that they're headed in
the right direction in a given area. But also so that
we can keep the process moving.
There's going to be a lot of data needed,
and some of this will be actually showing what data
needs to be taken, needs to be obtained so that they
can --
DR. BONACA: Excuse me. RIA stands for
what, rod ejection accident?
MS. CHATTERTON: Yes.
DR. BONACA: Okay.
CHAIRMAN POWERS: Reactivity insertion.
MS. CHATTERTON: Right. I'm sorry.
DR. BONACA: It's more general.
MS. CHATTERTON: Well, I'm sorry. Were
there anymore questions on this one?
(No response.)
MS. CHATTERTON: The next thing I wanted
to get into a little bit was lead test assemblies just
because, as I said, that's been an area that we really
want some emphasis on, and we've stated all along that
we think they should be prototypical, up to the
proposed burnup with reasonable power histories that
are similar to what's being used.
In the past we'd always said we wanted
them in very nonlimiting locations, and it was very
common to burn a lead test assembly to 50 or 60
gigawatt days, but to do it in six, seven cycles. And
then when you put the fuel in and burned it in three
cycles, you may not get the same -- exactly the same
results.
So that's why there's going to be a
real -- we're really emphasizing lead test assemblies,
and we also know the type of cladding makes a
difference, the flow conditions, the water chemistry.
Lead test assemblies need to be characterized, of
course, before irradiation.
And they will need pool site, and or hot
cell exams after. Hot cells exams are probably going
to be relatively infrequent, but there will be some
need. Certainly there'll be -- full site will be
needed certainly after each cycle, final burnup on
assemblies that are designated as lead test assemblies
before they start irradiation.
However, there may be assemblies that
become LTAs after they've had some burnup on them.
And so sort of to address that, to encourage lead test
assemblies, to encourage the gathering of data, we've
taken on a program to try to look for lead test
assembly guidelines, something that we haven't had in
the past.
Sometimes we had a submittal that we
reviewed and approved. Many times there was not an
actual regulation or any restriction. So under the
50.59, under the test parts they were able to do lead
test assemblies. It leads to a lot of things that we
hope by coming out with some guidelines we can
improve.
The purpose, basically, to get a
consistent approach, to get consistent database, to
obtain data. There's a real benefit to the industry,
too, in that they will know what we expect and know
that if they follow these guidelines, that it's
certainty.
I'll give you the outline topics and
things in another slide, but that's basically it.
We've made some progress on this. We met with WOG in
May of 2000. They put a lot of work into it, got the
whole industry together. They submitted a topical
report, which we looked at, and then we met with them
in December.
We gave them our comments on that document
in January, and we expect to hear from them again
soon. Some of the things that are covered and need to
be covered is a definition. Exactly what are we
talking about as a lead test assemblies? What are the
conditions?
Characterization, the type of
characterization of the rods, full site, hot cell;
when are hot cell exams needed?
Characterization will have to address both
pre-characterization and after final burnup.
The guidelines will address the number of
LTAs that can be in any one core. Also the location.
That's what I mean by placement. Location in the
core, what restrictions we think are necessary.
Safety requirements. The biggest thing
here is in almost all cases the LTAs are designed such
that they meet all the fuel design limits that the
current core is meeting.
However, they meet them using a code
that's been verified to 62. If we're now talking
about burnups that are going higher, we're taking a
step in saying that the code is valid beyond.
On the other hand, if you don't get the
data you can't validate the code. So this isn't an
area that we're working on, how to write that up, how
to address it such that it's covered conservatively.
Part of the way that it's covered, of
course, is the few number of pins that would be LTAs
and given the whole number in the core.
DR. UHRIG: There's been reports the last
three or four years of a control rod binding and
sticking, and the general, as I recall, the exposure
was about 43-44,000 megawatt days per ton, in the
vicinity of the assembly.
MS. CHATTERTON: A little higher.
DR. UHRIG: Little higher. What's going
to happen when you get the higher limits here? Are
there going to be more problems of that sort, or is
this something that has been addressed?
MS. CHATTERTON: That is one thing that
will have to be addressed, and you're right. I didn't
have it on the slide. But all the current type
problems that we've seen, like the incomplete control
rod insertion, some of these crud and oxidation type
problems, all of those things are going to have to be
addressed in a program to go to higher burnup,
absolutely.
CHAIRMAN POWERS: When you think about it,
people can go up to the 60 gigawatt days per ton.
Now, somebody comes along and says, "Gee, I want to go
to 70." That's what, 16 percent extrapolation?
It doesn't sound an outrageous
extrapolation to me. Do we have evidence that we
would expect changes in physics of the kind we saw
between going from 30 to 60 when we go from 60 to 75?
MS. CHATTERTON: Do we have hard evidence?
I don't think we have evidence.
CHAIRMAN POWERS: I mean, there are fuel
rods around that have gone up to 75.
MS. CHATTERTON: That's right. There are
fuel rods that have gone around. I know of some in
Europe that have gone as high as 100.
CHAIRMAN POWERS: That's right.
MS. CHATTERTON: Do we have evidence?
No, we really don't. But I think this is
a point that the -- we said there was an extrapolation
at one point in the past, and then there were some
things that happened that maybe weren't thought were
going to happen. And that it's time to stop and
really examine all the criteria before we move or
leave forward.
CHAIRMAN POWERS: I guess what I'm asking
is -- I don't know whether I'm asking -- we're closing
the barn door or we're making up for the sins of the
past on the backs of the people that are guiltless.
And we're talking about relatively small changes here
and asking for a heroic amount of work it looks to me.
And I'm wondering is there really merit in
that? I mean if we sort out the issues in 60 and say,
"Okay, everything's fine here," and that, quite
frankly, looks the direction it's going with these
superior clads. You know, things look like they're
moving along fine. Do we really want to create an
enormous burden?
I mean, clearly moving the lead test
assemblies out of the benign locations and into more
prototypical location, that's something that's been
needed for a long time. But after you go much beyond
that, do we really learn risk significant information
from LTAs?
MS. CHATTERTON: I think we gain a good
deal of information. I think we also gain some
confidence in reproducibility and uncertainty on --
you know, how uncertain are the measurements to take
when you take them only once? You asked the question
and --
CHAIRMAN POWERS: Oh, yeah. Ralph's good
at that. He knows how to do that.
MS. CHATTERTON: And that is -- to me that
is one of the things. This is an opportunity that you
have to do that. That's not a difficult one. You
can't do that on the accidents that Ralph is talking
about. I mean, my goodness, the cost of the tests,
you couldn't possibly do that.
But on this, these are some areas that you
can.
CHAIRMAN POWERS: There are those of us to
take the vote that say you absolutely must do that,
especially because of the test are so expensive.
MS. CHATTERTON: Well, I don't look at it
as -- it sounds like a lot but let me -- maybe I
didn't characterize some of it exactly correctly.
I think there's a lot of areas that the
industry is going to be able to right off very
quickly.
CHAIRMAN POWERS: Okay.
MS. CHATTERTON: With the state -- going
to higher burnup makes no difference and here's why.
We look at this, too, as this will help
not only us, but the industry have a really good
documentation of what is important and, you know, how
things change.
I expect there'll be an awful lot of
things that are written off very quickly, and they do,
too. They're working on the major ones.
DR. CRONENBERG: Then maybe it's not so
small. It's longer burnup, higher burnup, longer fuel
duty times, 20 percent power increases. I thought you
were asking a rhetorical question.
CHAIRMAN POWERS: No, I don't think I was.
I agree with you. Some of the things -- it's more
than just an increment in burnup because we're doing
an increment in --
DR. CRONENBERG: I mean, Commonwealth
Edison has come in on the docket with a 17 percent
power increase, one step.
CHAIRMAN POWERS: There's a lot more going
on here, none of which is really designed to coddle
the fuel at all. It's going to put this fuel under
some pretty heavy stress.
But the question then comes back to is it
a risk significant issue that we're getting into.
They can have all the operational difficulties that
they want to volunteer for, and that's their business.
Is it -- what we're asking about are -- our concern is
more with the risk issues.
And, you know, I think we have to be
careful not to close the barn door and put the burden
on -- that's all I'm concerned about.
MR. CARUSO: I'd just like to make the
observation -- this is Ralph Caruso from Reactor
Systems Branch.
Dr. Powers, you had asked if there was a
regulatory requirement for us to gather this sort of
operational data, and I would make the observation
that we are less interested from a regulatory point of
view in this operational data than in the knowledge
point of view.
One of the reasons why we're encouraging
people to do lead test assemblies is to share the data
with us. In the past they've been reluctant to do
that, but what we're trying to do is we're trying to
make the process easier for them so that they can do
more testing which we believe benefits them.
And by trying to make the process easier
and being a bit less threatening from a regulatory
perspective, we hope that they'll share the
information with us. We'll understand what they're
doing, and we will therefor feel more confident that
they know what they're doing.
So there's quite a bit of working together
on this, and we're not necessarily going to change any
regulations. We're just trying to understand what's
going on. I don't know if that helps any.
CHAIRMAN POWERS: Sure.
DR. KRESS: I see two places where
operational testing could shed light on or that has
risk significance. One of them is on the rod
insertion issue.
And the other one is that it's true that
the iodine spike is due to failed pins, which are few
and far between in a core, but that may be where
that's -- may be where that spike comes from. I would
perceive that if higher burnup increases the failure
rate of those pins, it would increase the iodine
spike, and you might be able to see that during the
operational -- that's where it comes from anyway --
during your operational observations.
CHAIRMAN POWERS: You've got to convince
me that an iodine spike is risk significant.
DR. KRESS: Yeah, it falls more in the
category of design basis accidents.
CHAIRMAN POWERS: Design basis accidents.
I mean I think there are risk -- there are interesting
risk significant features here in the high burnup
fuels. I'm not sure that LPAs get to them.
DR. KRESS: Yeah, that's -- I think that
was your point.
CHAIRMAN POWERS: Yeah.
MS. CHATTERTON: The LTAs do provide you
with the rods you need for something like Ralph's
program and for really --
CHAIRMAN POWERS: Now, there's where you
get it. Now, Ralph's program's got to be extended to
75 gigawatt days per ton; right, Ralph?
MS. CHATTERTON: Actually, we've said the
industry has to then pick up the tab beyond 62.
That's as far as the agency program. We said we do
confirmatory work to 62 and then beyond that --
CHAIRMAN POWERS: I know what you said.
Now, we just won't hold you to it. We'll let you
backtrack on that one.
(Laughter.)
MS. CHATTERTON: The last point on my lead
test assembly guidelines thing is we don't have
reporting in there. Basically, hopefully it would be
a template. It would be very easy to fill out, but it
would give us -- it would provide the data. Then we
would be able to know exactly what's happening with
LTAs.
CHAIRMAN POWERS: Nothing that you are
legally bound to is easy to fill out.
MS. CHATTERTON: I just finished my taxes.
CHAIRMAN POWERS: That's right.
DR. CRONENBERG: You know somebody was --
that wasn't very expensive, that annual -- that kind
of pool site inspections, and I think it was a good
thing, and we don't do it any more.
CHAIRMAN POWERS: Yeah, I mean there's not
question it's a good thing, but the idea that a
licensee is going to have an easy report to fill out,
I mean, it just doesn't exist. There is no report
that the licensee prepares that's easy to do, because
they are --
MS. CHATTERTON: Easier?
CHAIRMAN POWERS: Easier is possible.
MS. CHATTERTON: Okay, easier.
DR. UHRIG: What happens to the lead test
assemblies? Do they remain with the utilities?
MS. CHATTERTON: Yes.
DR. UHRIG: And are they usually sent for
examination in detail or is this just sort of a --
what kind of data comes out of them?
MS. CHATTERTON: At the end, we would
expect an all lead test assemblies to do pool site
exams, and that --
DR. UHRIG: Okay.
MS. CHATTERTON: -- that would consist of
oxidation measurements, probably growth
measurements --
DR. UHRIG: Growth rate, yeah.
MS. CHATTERTON: -- growth rate, visuals,
get an awful lot from visuals. And then depending on
if anything was found, it would determine what
further --
DR. UHRIG: They don't do a destructive
examination though. Metallurgy --
MS. CHATTERTON: No. If something really
is shown, then we would think that a hot cell exam --
DR. UHRIG: Would be in order.
MS. CHATTERTON: A constructive hot cell
exam would be in order.
CHAIRMAN POWERS: How many cells in the
country are available to do full length rods?
DR. UHRIG: One. Two.
CHAIRMAN POWERS: Two.
MS. CHATTERTON: Yeah. A number of hot
cell exams, few and far between.
So moving on, why do we really want a lot
of that?
Well, part of it is because of some of
these recent fuel issues. Oxidation higher than
predicted. We have several cases where, as I said,
the LTAs behave beautifully. If the fuel is burned as
the LTAs were, it behaves beautifully. But if it's
burned at a higher rate, at faster duty, they've
gotten very different results.
I think you're all probably aware of axial
offset anomalies that still tend to be -- that's a
problem that's still not completely understood.
DR. UHRIG: Isn't that pretty much boron
chemistry?
MS. CHATTERTON: It's a chemistry issue,
but it's also a fuel duty issue. And it's a very
strange --
DR. UHRIG: Well, it does depress the flux
in the area and reduces the load on the fuel.
MS. CHATTERTON: That's correct.
DR. UHRIG: But it would force it to be
somewhere else for the same power level.
MS. CHATTERTON: Yes, it forces -- it's
usually the precipitate at the top of the fuel forcing
the power to the bottom. You end up with a shutdown
margin problem. Had one utility that had to operate
at 70 percent power for four of five months as a
result of that, and it's continued through other
cycles.
Several other utilities have seen it, not
anywhere near to that extent.
DR. UHRIG: This is --
CHAIRMAN POWERS: Do we understand -- I
mean this is an inverse chemistry thing. Inverse
solubility issue, and you don't ordinarily think of
that arising with boron. Do we understand why boron
suddenly has an inverse -- boron becomes less soluble
at high temperatures.
MS. CHATTERTON: Actually, it's a sub-
cooled boiling. Basically, what you've done is you've
precipitated some crud onto the control rods, you're
in a region of sub-cooled boiling, and in the process
of sub-cool boiling with the boron, you precipitate
boron into that crud.
CHAIRMAN POWERS: And that's the step I
don't follow.
MS. CHATTERTON: You don't follow.
CHAIRMAN POWERS: Why does the boron
suddenly say, "I want to precipitate"?
MS. CHATTERTON: Well, with the sub-cooled
boiling you've got a mechanism there to -- you want to
give me a little --
MR. NISSLEY: I'm not an expert on this
but some of the theories are that when you have crud
and corrosion in the presence of sub-cooled boiling,
that the boiling mechanism is coming off as pure steam
and leaving the boron behind.
CHAIRMAN POWERS: And then it gets flooded
right back up with water and --
MR. NISSLEY: It's thought to -- it's
sometime referred to as boron hide-out where it's not
on the -- completely on the outer surface. It's
somewhat within the structure of the crud and the
corrosion.
MS. CHATTERTON: You get kind of like
little chimneys in within the --
CHAIRMAN POWERS: This sounds like on of
the things that if you tried to do it, it would be
impossible.
MS. CHATTERTON: Probably so. But it's
certainly been a problem that --
DR. CRONENBERG: But it's real. I mean,
they've measured crud with a high boron content.
MS. CHATTERTON: Yes.
CHAIRMAN POWERS: Well, I'm still asking
why.
DR. CRONENBERG: Yeah, I don't know, but
it's there.
MS. CHATTERTON: Everyone has spent a lot
of time on this issue, and it's still around. We've
had some fuel failures in a couple different plants
due to high fuel duty. Again, a combination of crud
and high fuel duty.
In all these cases, we've seen the effects
of water chemistry, high crud build-up, and we've seen
some accelerated growth of rods in assemblies.
That's much more the IRI issue that is
pretty much under control. I think I could say that
very easily in all plants, or at least appears to be
up until very recently.
The last thing I wanted to talk a little
bit about is some of the current fuel reviews that
we're doing. We have two reviews on cladding types
that are in house now. One is the duplex cladding
developed by Siemans, used extensively in Europe.
That's the one that has rods up to 100 in the Goesgen
plant in Switzerland.
CHAIRMAN POWERS: Wow.
MS. CHATTERTON: The review on that
cladding will be to 62. And we're just beginning that
review right now.
We're also reviewing the use of ZIRLO for
CE plants. We have some CE plants that have fairly
high duty that have been using a low tin Zircaloy
that's not been standing up to quite what they would
like.
And so the use of ZIRLO in those plants
would be extremely advantageous. And that's the
reason the timetable is they really want this by the
end of the summer. So we've got a large review.
The issue is basically a lot of it will be
making sure that the interfaces are done correctly on
the computer codes, that you get the right properties
in, and it's handled in each way. So there's a lot in
there.
And basically that's what I had as far as
issues.
DR. UHRIG: What do you mean by duplex
cladding?
MS. CHATTERTON: Duplex is -- it's a
double type of cladding. It's almost the reverse of
the BWR liner cladding.
DR. APOSTOLAKIS: The barrier cladding.
MS. CHATTERTON: It's got its corrosion
barrier on the outside, and it's a Zircaloy on the
inside.
DR. APOSTOLAKIS: Okay. It's a double
cladding.
MS. CHATTERTON: It's essentially a double
cladding. Very, very --
DR. UHRIG: They're both Zirc?
MS. CHATTERTON: Pardon?
DR. UHRIG: Both are Zirc?
MS. CHATTERTON: No.
DR. APOSTOLAKIS: Different material.
MS. CHATTERTON: The outside is a -- I
have to think.
PARTICIPANT: It's a proprietary material
to Siemans.
MS. CHATTERTON: Yeah.
DR. UHRIG: Oh, okay.
MS. CHATTERTON: The strength part, inner
part is Zirc-4. And as I said, they've had -- they've
used that extensively in Europe. The data from it as
far as corrosion and performance is absolutely
excellent.
CHAIRMAN POWERS: I hope that once you get
through your duplex cladding review, you come down and
talk to us a little bit about that.
MS. CHATTERTON: Sure.
CHAIRMAN POWERS: Because I think that
would be interesting for us to see.
MS. CHATTERTON: Good. Thank you very
much.
CHAIRMAN POWERS: Thank you, and good luck
in Boston.
MS. CHATTERTON: Oh, thank you. It'll be
a slow run. It will be fun, but it won't be a fast
run.
CHAIRMAN POWERS: Don't care how slow it
is. the fact that you're there is just amazing.
DR. KRESS: We're going to look for you on
TV.
CHAIRMAN POWERS: We'll watch for you on
TV.
MS. CHATTERTON: I'll be the last one.
CHAIRMAN POWERS: We're going to switch
gears and Dr. Lee fresh from a vacation of over --
almost a week duration in Italy is obviously going to
be in fantastic spirits to talk to us about the MOX
research program.
Yes, you're in a good mood when you come
back.
DR. LEE: I'd like to briefly tell you
something about the MOX research that Office of
Research undertook, just started last November.
How about now? Thank you.
And you know why our interest in mixed
oxide fuel. In February 2nd, the NMSS team came
before the full Committee and briefed you on the
certification plan and what is the activity related to
your mixed oxide fuel, MOX fuel use in U.S. That is
basically the disposal of up to 33 metric tons of MOX
fuel in using it in our commercial reactors, and the
two plan, four units targeted is McGuire and Catawba.
CHAIRMAN POWERS: What I have never
understood is why ice condenser plants are
particularly suited for using MOX fuel.
DR. LEE: And you were told that they
really did not target ice condenser, remember?
(Laughter.)
DR. LEE: There was two Virginia power
plants that was involved with it, but they drop out of
it, but it happened in the two plants that's left.
There are four units left now, ice condenser plant,
under Duke Power, and I'm sure your concern has to do
with the severe accident issues about fuel dispersal.
CHAIRMAN POWERS: Yep. Comes to mind.
DR. LEE: I think, one, we did the DCH for
ice condenser plan. We found the McGuire plant is
slight a bit higher than the cutoff point that we use,
like .1 conditional failure probability. It came up
to .14, but the Duke Power took issue with us that if
you take into the real design of the plants, that
number will come down.
So when the whole overall evaluation for
MOX use in the McGuire come in, those numbers will
be --
CHAIRMAN POWERS: Yeah, I think I would
have responded by saying, "Yeah, and when I take the
degradation of the containment into account, the
number goes up again."
DR. LEE: Well, the research activities
really focus on supporting a user request that came in
back in late '99, and at that time we didn't have any
budget to address it, but we just had budget this year
to address the technical assistance requested by the
regulations, nuclear reactor regulations.
They are interiors (phonetic) neutronics,
fuel and source terms. The neutronics, they want to
modify the codes that were used for MOX and also, of
course, goes with it all the fuel behavior, monitoring
assessment for the fuel behavior for design based
accidents and under normal as well as design based
accidents need to be corrected before we can use it.
And then in the source term area that we
also need to validate that the source term that we
used for UO-2 fuel (phonetic) is approximate for MOX.
DR. KRESS: How much, what percentage of
the fuel would be MOX in these?
DR. LEE: It's normally one third of the
core would be MOX.
DR. KRESS: Oh, as much as one third?
DR. LEE: Yeah.
DR. KRESS: Okay. I --
PARTICIPANT: Forty percent.
DR. KRESS: Okay.
DR. LEE: Or even more than that. Thirty
to 40 percent.
Now, as I mentioned to you, we started
this activity not too long ago, but at that same time
before that, Ralph Meyer was doing a PIRT on the high
burnup fuel. So since we know the MOX is going to be
coming into play, we attached to ask our experts to
tell us something about what do we have to do for the
LOCA and reactivity accident, and that PIRT has been
completed.
And on that Web site you will see the
reports related to LOCA as well as the RIA accident.
The source term PIRT now is going to be
starting very soon. It's not just for MOX. It's also
for high burnup fuel as well, and we expect to finish
by this year.
The composition for the experts have not
been -- selection not been completed because we're
waiting for a response from French and from Japan and
also from the industry selecting experts to
participate in this panel.
The NRC internal one has suggested some
members, and we're working on that.
Now, in the neutronics area, there are
three areas that we have initiated. The first one is
the PARCS code that we have at Purdue University.
That has been used for many years. This PARCS code is
a neutronics code being interfaced with our thermal
hydraulics codes like TRACK M or RELAP, and we have
used it, and we have used it very successfully for RIA
type analysis.
And we initiated the modification for this
to make it more usable for MOX. That is started in
November, when we initiated these activities.
At this time, we extended the number of
group of energy that can be handled by the code from
two groups to n group because it's very easy to make
it general. One time we can use seven groups, four
groups, two groups, because if the industry comes in
with analysis with two groups, we have to be able to
collapse it to two groups so we can analyze it on the
same base.
The cross-section because of the isotropic
between the UO-2 bundles and the MOX bundles, there
will be very sharp gradients of neutron flux. So we
have to handle the scattering correctly. So we have
expanded the cross-section angle of dependency with P3
approximation. That has been completed as well at
this time.
DR. APOSTOLAKIS: Is there a report where
I could go and find more about these observations
like, you know, why you need to go to the P3
approximation, and so on? The motivation for the
research, in other words.
DR. LEE: I think the motivation if you
look at the Europeans, the way they analyzed the MOX
code, they usually use a high order scattering to do
the approximation.
DR. APOSTOLAKIS: So I should look to
Italy as well to find that?
CHAIRMAN POWERS: Well, I think there's --
DR. APOSTOLAKIS: There must be a report
somewhere.
CHAIRMAN POWERS: Yeah, one of the authors
of PARCS put out a document that went through all of
these things, and it was given to this subcommittee a
couple of years ago, I guess. I can't remember the
exact title, and I mean, I am sure we could find that
for you.
DR. APOSTOLAKIS: Okay.
DR. KRESS: It was a pretty good document
as best I remember.
CHAIRMAN POWERS: It was a pretty good
document. I mean, it raised the scattering and the
group issue. It also raised the delayed neutron
fraction issue.
DR. LEE: Yeah, the delayed neutron
fraction.
DR. MEYER: Was this a Commission paper
that you're referring to?
CHAIRMAN POWERS: No, no. Actually it was
a Purdue report.
PARTICIPANT: It was critical, a bit
critical, right?
CHAIRMAN POWERS: Well, I wouldn't say it
was critical. I would say that he came back and said,
"Look. My PARCS code right now can't do the MOX fuel
because of these things," and he listed down what he
had problems with using PARCS for that.
So I mean if it was critical, it was
critical of his own codes.
MS. SHOOP: This is Undine Shoop with
Reactor Systems.
You can find more detail in the Commission
paper that we wrote. We've authored two of them at
this point. One would be from '99, and one would be
early 2000, and I'm sure we can get copies of them for
you.
DR. APOSTOLAKIS: Good.
DR. LEE: In addition, at this time there
was a researcher from Saclay, is stationed at Purdue
University assay change for about a year, and you can
reduce in the French code CRONOS, and this code has
been benchmarked against many of othe plant data in
France that use MOX code. So we like to compare that
with the developed PARCS that we're going to be using
for MOX code analysis as well.
Tom Downer from Purdue University is the
one who is the PI for this, principal investigator for
this work. He's also working with the OECD and NEA to
develop a theoretical benchmark for reactivity
transient. This is quite a lot of work to do because
now you need to develop an exercise that go from
steady state and looking at some transient, how would
the parts compare with other codes?
Of course, you would be using a Monte
Carlo code calculation, and so forth, but this is a
code-to-code comparison.
We also initiated a very small activities
at Brookhaven under Dave Diamond. He has been helping
us for many years, helping us to do independent
assessment of PARCS, and we intend to use him to
continue this activity.
It provides feedback to code developers,
and we try to make it also more user friendly, too,
because a couple in between the milars (phonetic)
continue to be a problems in setting up the problem,
but we are making it better now.
And then also, in terms of if there is any
technical issues that we require his assistance to
review, the licensee will submit to us and we will ask
them to do so.
At the same time we also initiate a
lattice physics code develop at Oak Ridge. It's a
routine called NEWT, and this is part of the scale
code, the whole suite of codes that Oak Ridge use for
shielding, heating, decay heat, and also analysis.
And this will enable us to generate the
cross-section, assembly-wise cross-section that we can
feed into PARCS and that PARCS can use for steady
state, and as well sa depletion, as well as transient
analysis, especially IA type.
In the fuel area, of course, we started to
update the material properties for the FRAPTRON and
the FRAPTRON codes to be used for the MOX analysis.
Then, of course, we have to assess the experiment
against data.
There is a Halden exercise, blind test,
the CS&I test completed, and at that -- actually the
rig is still inside the reactor. They continue to
monitor, measure the build-up of the fission gas, and
the temperature. So you can get those probably as a
function of burnup.
The exercise they did was allow 14
gigawatt days per ton. At that point they asked all
the participants to do the calculations. That was
back two years ago, and they just finished that.
So those are the information that we would
like to revisit, and of course, in this area, I didn't
mention, of course, the Cabri test for the IRA. We
would like to look at the gigawatt behavior, as well.
In the source term area, we are
negotiating with the France to get VERCORS
experiments. The VEGA from Japan, they will not be
doing any MOX experiment until like 2003.
There's some tests that has been done
already in VERCORS in France, up to like 41 gigawatt
days per ton. It's about three pallets only, and they
are also starting a new facility in Cadarache that we
don't know much about, but this one will have a longer
length rod, about maybe 60 -- six centimeters long.
And we don't know what the test matrix
look like in terms of when the MOX test will be coming
in because this facility is supposed to replace this
town in the near future. So they will shut down all
of the hot cells and those type experiments at
Grenoble in France, and then, of course, research will
assist licensing in terms of review any technical
issues that will be rising.
DR. MEYER: Could I add something here?
DR. LEE: Yes.
DR. MEYER: It's Ralph over here.
I didn't seen Cabri on your slide, but
there are two MOX tests in the Cabri water loop, and
there have been. Did you have that? I'm sorry if you
had it on there.
DR. LEE: No, I didn't put it in here. I
mentioned it in here that we need.
DR. MEYER: Oh, okay.
DR. LEE: I didn't put that on here.
So our activities just started not too
long ago. So we don't have any results to tell you,
but on the source term area, next year, this coming
fiscal year, we will start to initiate the validation
for the codes that are going to be used for source
term analysis, and that's the first one we're going to
do.
CHAIRMAN POWERS: When is an appropriate
time for us to hear about what you're doing with
PARCS?
DR. LEE: I think by May time he will be
able to do some demonstration on using the type of
analysis that he has.
CHAIRMAN POWERS: So maybe some time in
the fall?
DR. LEE: Some time in the fall, yes.
CHAIRMAN POWERS: Yeah, I think the
Committee would be --
DR. LEE: -- MOX calculation was the
difference between UO-2 versus MOX.
CHAIRMAN POWERS: I think it's been a long
time since the Committee has looked at some of these
neutronic things, and since it's an important part of
TRACK M maybe the Fuel Committee and the Thermal
Hydraulics Committee might want to get together and
just focus on that, say, for half a day, just that
particular topic.
MR. ROSENTHAL: Because we're using this
also just plain the UO-2 RIA issues.
CHAIRMAN POWERS: Sure, yeah. I mean,
it's a fairly important code.
MR. ROSENTHAL: Sure.
CHAIRMAN POWERS: I like the way you guys
went about selecting to use it and whatnot. I thought
that was a very analytic process, but when it came in,
there was this list of challenges I would say in
interfacing and shortcomings that the code had for
modern things, and it would be nice to see how it all
came out.
DR. MEYER: By the way, we had a small
task in our program with Kurchatov Institute with IPSN
involvement as well to do some MOX calculations for
the reactivity transients.
MR. ROSENTHAL: And that's really good
because everything we have traces back to NDEF
(phonetic), you know, NDEF E6 or 7, and that's
independent.
Can I just make a summary statement? And
that is that I'm relatively new in the current branch,
just a little bit over a month, and so I go to Richard
and I go to Ralph all the time. In fact, Ralph's
office is next to mine.
And we were talking, and I think it's
important to make the following point. If I go to the
RIA, okay, what we ultimately will discover is that
the speed limit that we thought was appropriate for
decades is probably incorrect and, you know, maybe 280
becomes 100 or 80, some other number, and at the same
time when we do 3D space-time kinetics, we're pretty
comfortable that people will be able to demonstrate
that they can live with a revised lower speed limit.
So you don't have a big safety issue, having done all
that work and recognized that.
And I said, "Yeah, but shouldn't this give
us some humility?"
(Laughter.)
MR. ROSENTHAL: Okay? That here was
something that, you know, we thought of and didn't
question, and now we have a different perception.
And if it's giving us some humility with
respect to the enthalpy deposition, then it's fair to
say, well, what other surprises are there in stock for
us as we go to high burnup or new alloys or your MOX,
and that sense of, well, what other surprises are in
stock for us, and maybe a little humility, leads us
to, in fact, fund fuel work and research as a truly
sensible fraction of the total research budget.
I just wanted to leave you with that.
DR. APOSTOLAKIS: Now, the view for
McGuire and Catawba, are these considered changes in
the licensing basis, the use of MOX?
DR. LEE: Sure, sure. It would have to
be.
DR. APOSTOLAKIS: So what if someone --
DR. LEE: Specific licensee.
DR. APOSTOLAKIS: What if someone decided
to use regulatory guide 174 to argue for or against?
DR. LEE: I think the same question would
arise, that phrase when Margaret was asked about
1.174.
DR. APOSTOLAKIS: The question will arise,
but --
DR. LEE: Yes.
DR. APOSTOLAKIS: -- it says here MOX
research, and I don't hear you doing anything about
it. Why aren't you looking into it?
DR. LEE: I think that is up to the plant,
what they want to do it under the regular 1.1 -- 1.7.
CHAIRMAN POWERS: I guess I'm confused,
George. I mean, if the program includes an
examination of the source term, and so I'm a little
questioned -- I mean, maybe you can say there's some
core degradation work that --
DR. APOSTOLAKIS: If Tom is right and the
left values are not the right ones, you have to modify
them. Shouldn't somebody look into that? Does that
come naturally from this?
CHAIRMAN POWERS: Yes. I mean, that would
be the whole point. If somebody came back and said,
"Look. This" --
DR. APOSTOLAKIS: What does that -- point
to me to that.
CHAIRMAN POWERS: If the source term is
going to be different from that, then once you had
that, that's when you would have to reexamine your
derivation to get from the quantitative health
objectives to get to the acceptance value of worth.
DR. KRESS: They're putting together a
PIRT now just to look at that. You know, they don't
define the program yet. They just want to say what
are the likely phenomena; what are the issues; what
research should we do.
DR. LEE: The source term PIRT is that
we're going to look into what are the issues that we
have to deal with for NUREG 1465. What do we need to
do for that for MOX.
And then in the model developments, we're
going to validate our models. We're going to use --
for example, I'm going to take a core, and I'm going
to have an analysis of all uranium fuel assemblies,
analyze and look at inventories, and I'm going to take
another core which is one third or 40 percent loaded
with MOX, and I look at the two source, and I will do
my consequent analysis, and I would like to compare
what are the consequence, what are the difference from
there.
Now my mother has to be validated
(phonetic).
DR. APOSTOLAKIS: Now, when you say do
your consequence analysis, what do you mean?
DR. KRESS: There's a design basis space
he's talking about.
DR. LEE: For the design.
DR. KRESS: Chapter 15.
DR. APOSTOLAKIS: But LERF was not
developed.
DR. KRESS: No, no. He'll have to do more
than 1465 --
DR. APOSTOLAKIS: Yeah.
DR. KRESS: -- to get to that stage.
They'll have to have more detailed fission product,
release models, and --
MS. SHOOP: This is Undine again.
I would just like to add that as part of
our user need memo we have requested the Office of
Research to look not only into the source term, but
how that will impact the different levels of the PRA,
and I believe that right now that's being looked into,
and I'm sure that when Richard comes back here to talk
about our further research in the future after we're
done with the source term, then we'll be able to go
into more detail on the additional research we're
doing.
CHAIRMAN POWERS: Okay.
DR. LEE: Oh, Dana, one thing that I think
we should also know, that the French is launching a
PHOEBUS 2K (phonetic), which also has a MOX component
in it, and they want to look at is there any sudden
core degradation phenomenon that we don't know about
that is vastly different between UO-2 versus MOX.
And also in the LOCA arena, they are also
looking into doing LOCA as a series of looking at the
loss of cooling accident for high burnup fuel, but I
don't know whether MOX is included in that.
CHAIRMAN POWERS: They're going to have to
jerk their driver core here pretty soon, aren't they?
DR. LEE: Yes.
CHAIRMAN POWERS: Now maybe they're going
to run out of oomph in the driver core.
DR. LEE: I think they need to refurbish
that entire thing. The driver core is only good for
the current series of tests, and after that they
completely have to refuel the whole driver core for
the following improvement.
CHAIRMAN POWERS: So there will be an
examination of the core degradation aspects.
DR. LEE: That's what they would like to
do, yes.
CHAIRMAN POWERS: Right. Any other
questions of the speaker?
(No response.)
CHAIRMAN POWERS: Okay. We have a treat.
Dr. Lyman from the Nuclear Control Institute is here
with us again. Dr. Lyman has spoken to us before.
He'd like to have a word with us.
He didn't tell me what he was going to
talk about, but I'll bet it's on MOX fuel.
DR. LYMAN: Thank you.
CHAIRMAN POWERS: Put it on your tie
probably is a better --
DR. LYMAN: How's that?
CHAIRMAN POWERS: Yeah.
DR. LYMAN: Okay. Well, you're right.
Since the top was MOX fuel and that's one of the main
concerns of my organization, the Nuclear Control
Institute, so I thought it might be a good time to
come back.
Actually I've never spoken to the ACRS
before on MOX. Two years ago I gave a briefing to
interested NRC staff on a study I had done, a
preliminary study which was actually a consequence
assessment, exactly what was just being discussed, of
the use of MOX fuel in light water reactors and
actually a regulatory guide 1.174 approach to how you
might risk inform the use of MOX fuel.
And so I'd like to actually go over those
again. I've since refined the report, and it's going
to be published. I wish I had a final version. This
is a penultimate version, and it should be available
very soon in the Journal of Science and Global
Security, which comes out of Princeton University, and
it will be on their Web site.
So as soon as that's out, I'd be happy to
point people to it if they're interested.
Okay. The title of my talk is "MOX Fuel
Safety, a Need for Research," and I'm very glad that
there's finally money in the NRC budget for doing some
MOX research since there hasn't been for a long time,
even though this program has been coming for a while.
My organization has been very concerned
about the way the Department of Energy has dealt with
the issue of MOX fuel. From the beginning, their
environmental analysis, the whole way in which they
made decisions regarding weapons plutonium disposition
without really looking hard at some of the safety
issues that were going to be coming down the pike with
MOX.
I wish they'd involved the NRC earlier,
and there is still time to deal with these issues, but
it's starting to run out.
So just briefly I'd like to give some of
the overall, the general concerns I have with the way
the MOX program is evolving, including some very
recent developments, and then I'd like to talk about
some of the detailed safety issues that I think are of
concern in this program.
One is the issue of the source term impact
on severe accident consequences and risk, and then the
impact on transience, including the over cooling
accident, pressurized thermal shock, and then RIAs,
and then finally some troubling issues concerning the
MOX qualification plan which has been laid out by the
licensee, Duke, Cogeme (phonetic) with Stone &
Webster, or DCS.
So starting with the MOX program concerns,
I think the question came up before why are ice
condenser plants the best suited for using MOX fuel,
and the answer is they are the only ones that are
willing to do it. There was no real choice for the
mission reactors. There was no real competitive bid
that was worth anything. There were only three
consortia that competed. Two of them didn't even meet
the basic requirements. So they were eliminated right
off the bat, leaving on the Duke Power consortium,
which originally had Virginia Power. They dropped
out, I believe, because they would have had to modify
their control rod systems in North Anna, and they
didn't want to do that.
So for better or for worse, we're stuck
with the ice condensers, and I'll talk about our
concerns about that a little later.
The second great concern we have with the
MOX program is the fact that the timetable is dictated
by international agreement and not by safety
requirements. The U.S. and Russia signed an agreement
last fall or late last summer that commits both sides
to starting to use MOX fuel in light water reactors by
the end of 2007, and our concern, of course, is
because of the political pressure, because this is a
nonproliferation program, that NRC is going to have a
very hard time raising substantive issues that might
cause delays in the schedule, and they run the risk of
being accused of being obstructionist and interfering
with important nonproliferation programs.
And so I feel this is a potential tension
that might influence the ability of NRC to do a fair
assessment of MOX safety issues.
Related to this are the DOE budget cuts
which are impending. The MOX program apparently,
according to news reports, is not going to get the
increases that it expected under a potential Gore
administration, since it was Gore who was shepherding
the bilateral plutonium disposition talks.
And the fact is that a reduction in budget
for MOX is only going to increase pressure that any
safety review for MOX be abbreviated, and that there
will be less DOE resources available for helping NRC
to resolve some of these technical issues.
This could lead to heavy reliance on
proprietary foreign data, which for many reasons our
organization doesn't think is going to be appropriate
or adequate for resolving the issue of MOX use in U.S.
reactors.
And finally, the impending cancellation of
the other plutonium disposition track, which was a
mobilization of plutonium in a ceramic and disposal of
high level waste, this program apparently is being
zeroed out by the Bush administration, and that means
that there will be at least an additional eight and a
half tons of plutonium which will be heading toward
the MOX program for disposition in roughly the same
time period, and it's not clear how DOE is going to
address that at that point, but again, it will
increase the burden on MOX as the only route for
achieving disposition.
So with those political pressure in mind,
I'd just like to review some of our concerns about the
safety of MOX, and the biggest contributor I think to
the enhanced risk of using MOX in light water reactors
is the additional source term that comes mainly from
an increased transuranic inventory in the core.
Now, according to the calculations that I
did using the scale code, you find for the DCS core,
which has a 40 percent MOX core fraction and an
aqueous processing which will remove the americium
that's been building up in the plutonium pits since
they were last recycled; that if you remove the
americium, then at end of cycle I find that you'll
have about two times more of the isotopes like
Plutonium 239, Americium 241, Curium 242.
Plutonium 238 is a little bit less, but
that doesn't have a big safety impact, and also, since
I know the Committee has been interested in ruthenium
lately, incidentally, for a given MOX assembly you
have more than twice the amount of Ruthenium 106. So
an average of the core and into cycle, I find you have
about 45 percent more Ruthenium 106, which might play
a role in events where there's the risk of air
oxidation source term, as the Committee has discussed,
a PTS event, or a spent fuel pool accident.
Finally, after I first put out my study in
spring of '99, DOE revised its EIS calculations,
accordingly did a better job, but there are still
flaws in the values that are outstanding in the
environmental impact statement, and one of those comes
from the fact that they assumed for some reason that
in the reactors in the U.S. you have three or you
divide the core into three equal fractions, and each
burnup interval is an equal burnup interval, which is
not the case in a reactor with an 18 month core
loading like Catawba or McGuire.
So they actually underestimate the burnup
of the second cycle MOX fuel.
So what are some of the impacts on severe
accident consequences from the increased true source
term using the MAX-2 code, suitably revised after I
discover an error in it?
You find that for early containment
failure, for a typical early containment failure
source term, which in this case what I have here
corresponds to about a one percent overall low
volatile release; you find that there's a 25 percent
increase in latent cancer fatalities as a result of
the initial plume. That's not looking at the chronic,
long term consequences, but only what's in MAX-2, in
what's called the early module, and that's because I
don't really trust the chronic module in MAX.
As far as prompt fatalities go, there's a
very small or practically no increase, only about four
percent for early containment failure because the
particular isotopes that are greater in MOX cores
don't really influence that much. Again, the results
will be available in this paper.
Now, I just looked recently at the
possibility of the high ruthenium release that might
correspond to a pressurized thermal shock accident,
and I found that that has a bigger impact on the
prompt fatalities. In that case, this is preliminary,
but there's about a 30 percent increase then in both
latent cancers and prompt fatalities for a 75 percent
ruthenium release.
DR. KRESS: What was the nature of the
error you found in MAX?
DR. LYMAN: It turns out for very high
releases, you could have more cancer fatalities than
there were people.
DR. KRESS: Oh, okay. It was in the dose
consequence.
DR. LYMAN: Right. It was not normalized
properly, and so they fixed that, and it will be in
the next release.
DR. APOSTOLAKIS: Now, when you're saying
25 percent, four percent, and so on, you're obviously
referring to some point value.
DR. LYMAN: Oh, I'm sorry.
DR. APOSTOLAKIS: Is that the mean value
of something or best estimate?
DR. LYMAN: You mean --
DR. APOSTOLAKIS: What does the 25 percent
refer to?
DR. LYMAN: Oh, I'm sorry. Compared to
the exact same source term applied to an only uranium
fuel. So in other words, I --
DR. APOSTOLAKIS: So you did both
calculations?
DR. LYMAN: Right. You look at the
consequence analysis for a particular source term for
a uranium fuel, and then you keep the release
fractions all the same, which may not be a correct
assumption for MOX because there may be greater
volatile releases for MOX fuel, but if you assume all
of the source term, the release fraction is the same.
Then you just look at the impact of the additional
actenites (phonetic), for example.
DR. APOSTOLAKIS: Okay.
DR. LYMAN: But I did it over the entire
spectrum of isotopes.
And again, of course, there are different
release fractions for different accidents. That's a
kind of stylized early containment failure, which was
derived from NUREG 1150.
Okay. So what about the impact on risk?
Well, you can look at a set of a kind of complete set
of accidents leading to a large early release, and
basing on a NUREG report, which binned a whole lot of
severe accident scenarios into a small number. I was
able to do a rough estimate of what is the impact on
the average population risk within one mile, which is
the parameter cited in the quantitative health
objectives.
And so that actually tracks the
consequences pretty well, about 25 percent increase
for the DCS core in average risk to the public within
a mile of the reactor. That's latent cancer fatality
risk.
So then I asked if you wanted to risk
inform, sine it's quite likely that when there's a
submittal for a license amendment for using MOX fuel,
then it will meet all of the design basis
requirements, but the question is: will it have an
impact on risk, which could be something you need to
consider?
And now that the staff has the authority
to use risk information either in a license submittal
that's not risk informed, I thought this might be
something that the staff might want to look at since
this could be one of the biggest impacts. The biggest
impacts of using MOX is not on design basis actions,
but on beyond design basis.
But then this question arises, which the
Committee has discussed frequently, is the 1.174
assumes a particular release, and only looks at change
in LERF, and so the question is: how do you deal with
the situation where the actual frequencies may remain
roughly the same, but the inventory changes?
So I did a quick and dirty -- I'm a former
physicist. So that's what we do, is try to work with
what you've got, and quick and dirty way of using
1.174 was simply to derive what I call an effective
LERF, which is let's say you have an accident, two
different accidents and only the consequences change.
That's associated with a change in risk.
So what's the equivalent change in LERF
that would lead to the same change in risk? And so
it's just a way of using the scale which is provided
in 1.174.
And incidentally, this is also a useful
way for evaluating what's an extended power up rate,
and the issue does arise if you have the 17 percent
extended power up rate. That's going to lead to a
significant increase in consequences from severe
accident, and if that's acceptable, then this increase
associated with MOX will also be.
But inversely, if one isn't, then neither
will be the other. So this could be a way of
addressing at least until the formalism is fixed, to
address this, a way of addressing things like the risk
impact of an extended power up rate.
DR. APOSTOLAKIS: So fixing it probably
will mean not to deal with a LERF anymore.
DR. LYMAN: Possibly. I mean --
DR. KRESS: If you had delta R you
wouldn't need a LERF really.
DR. LYMAN: Right, and that's what this is
just saying. Delta R is the same for both.
DR. APOSTOLAKIS: Because neither the
large or the early change, as you said.
DR. LYMAN: Right.
DR. APOSTOLAKIS: Nor the F.
DR. LYMAN: But if this equation isn't
right, and it may not be because, you know, you end up
with small fractional increases in risk, and you know,
the error bars might be big enough that it washes
those out, but if that's the case, then if this isn't
correct, then the overall 1.174 --
DR. APOSTOLAKIS: So R is the risk.
DR. LYMAN: Right. In other words,
probability times consequences summed over all the
accidents that contribute to LERF.
DR. APOSTOLAKIS: For whatever risk you
have in mind. I mean prompt fatalities.
DR. LYMAN: Right. In this case I looked
at latent cancer.
DR. APOSTOLAKIS: So you do have delta R
then.
DR. LYMAN: Right. You can calculate it
if you know everything.
DR. APOSTOLAKIS: If you had it or you
have it.
DR. KRESS: You have to do some sort of a
PRA. Now, he --
DR. APOSTOLAKIS: But look. Lyman says
that we should use this to define an effective delta
LERF. Therefore, you must have delta R.
DR. KRESS: But he used sort of --
DR. LYMAN: Right.
DR. KRESS: -- an abbreviated --
DR. APOSTOLAKIS: And he did that earlier.
DR. LYMAN: And it's like a Level 3 PRA,
except it's very truncated, and it was based on a
small set of accidents.
There was a study. I don't have the
number with me, but they took, let's say, the Sequoyah
NUREG 1150, and they binned. You know, you have
thousands of different initiators. They binned them
into a small number of accidents with the same source
terms.
So it was manageable. There were three or
four different source terms and frequencies associated
with that. So you could do a kind of very rough Level
3 and get the risk.
DR. APOSTOLAKIS: Now, instead of doing
this, it seems to me since you can do a rough Level 3,
what you could do is take the allowed delta F for
light water reactors that the NRC staff --
DR. LYMAN: Right.
DR. APOSTOLAKIS: -- has declared is
acceptable --
DR. LYMAN: Right.
DR. APOSTOLAKIS: -- ten to the minus
seven --
DR. LYMAN: Right.
DR. APOSTOLAKIS: -- and see what the
consequences of that are with respect to the
acceptable change in prompt fatalities and compare
your delta R with that.
DR. LYMAN: That's actually exactly the
same thing.
DR. APOSTOLAKIS: It's the same thing?
DR. LYMAN: You're just saying it
differently, yeah.
DR. APOSTOLAKIS: It's not obvious it's
the same thing. Is it obvious it's the same thing?
I'm not doubting, but --
DR. LYMAN: Well, I have to think about
it. I think it's the same.
DR. APOSTOLAKIS: I can't see it's the
same.
DR. LYMAN: Because you're just saying
what -- you could rewrite this in that way.
DR. APOSTOLAKIS: In other words, what I'm
saying is, okay, you can calculate the change in
prompt fatalities or cancers and so on, but you don't
know what's acceptable, what delta cancers is
acceptable, but you have a delta LERF that has been
declared acceptable for light water reactors.
Take that and propagate it to the front,
the Level 3, and then compare you delta after that.
DR. LYMAN: Yeah. Do you see where it's
the same thing? Because you're just saying if you
know what the source term is, then you can say, well,
I know what the change in risk is going to be
associated with that change in LERF. Now, if you can
do the Level 3, then you can propagate that through,
and then you would get a delta R, which you would
compare.
This is just doing that backward.
DR. APOSTOLAKIS: And so I guess what
you're saying is after I take the LERF to the left, I
have delta LERF or LERF is delta R over R.
DR. LYMAN: Yes.
DR. APOSTOLAKIS: And there must be some
other duplicative factor there that counts as R.
DR. LYMAN: Right, if the source term is
the same. Right. It's the same thing.
DR. APOSTOLAKIS: Yeah, yeah.
DR. LYMAN: You know, it's a very obvious,
very simplistic --
DR. APOSTOLAKIS: I don't know about
obvious. It took me ten minutes to understand.
DR. UHRIG: Are you contemplating a 17
percent increase in power?
DR. LYMAN: No.
DR. UHRIG: I'm not aware of that.
DR. LYMAN: No. What I'm saying is that
since the risk that I found associated with using MOX
is about, you know, this 25 percent increase. That
could be comparable to the increase in risk associated
with the power up rate.
DR. UHRIG: Well, the power up rate, the
17 percent typically associated with BWR is not PWR.
DR. LYMAN: No, I'm not saying that it's
going to happen. I know Catawba and McGuire are not
planning to. I'm just saying that's another example
where you could use this.
And, again, if those up rates are
approved, then, well, at least it's a way of saying
it. It's a way of saying -- well, let me go on to the
next slide because at least this shows you in the
1.174 context.
Okay. So what's the risk impact of MOX in
ice condenser plants? Now, we know the DCH study that
came out last year concluded that ice condensers are
substantially more sensitive to early containment
failure than other PWRs, and this is precisely the
class of accidents in which you would feel the
additional risk from MOX because these are the
accidents where you would have fuel dispersal and
containment failure.
So that in itself is of concern, but here
I just did -- this is a rough estimate using the
equation from the previous slide where from the
McGuire IPE, which is now ten years old, but the total
LERF, internal plus external, is 4.7 times ten to the
minus six.
So then if you use the delta LERF
effective equation from the last page, you get a
number 1.2 times ten to the minus six that actually
exceeds the reg. guide 1.174 threshold. At least this
crude estimate means that it's in the regime where
changes would not normally be allowed. So that's the
first point.
CHAIRMAN POWERS: Actually, I think it's
in the regime. It simply means it's in the regime
where it gets increased management attention.
DR. APOSTOLAKIS: That's pointed out here.
DR. LYMAN: Well, the actual language is
not normally allowed. It's the top tier. Now, it's
close to the boundary, and nothing is set in stone,
and you also have permission to use other arguments,
you know, quantitative arguments to get out of this
hole, but I would say that at least on the scale
that's proposed in 1.174, this increase associated
with MOX is fairly significant, and I wouldn't write
it off.
Now, going back to that, the McGuire IPE
does not take into account the Sandia finding that the
early containment failure frequency was under
estimated by a factor of seven in Duke Power's own IPE
and PRA, and this, as Richard Lee said, is still a
matter of controversy.
But if you did take into account the
greater early containment failure frequency associated
with station blackouts, just again using the IPE
numbers, you'd end up with a LERF above ten to the
minus five, which is in the regime where no risk
increase greater than ten to the minus seventh would
be allowed. So that, again, would exclude MOX.
Now, I know that the current PRA for
McGuire is about half what it was in the IPE, but I
don't know what the station blackout frequency is now,
and these are not really publicly available, and so I
can't say anything about that. But at least based on
what's public, I'd say, again, that the risk is
significant.
And, again, the implications for extended
power upgrades, I'd say, is one way of looking at if
a 25 or 30 percent increase in risk associated with an
extended power upgrade, this is a way of evaluating
where it fits in the risk informed framework.
And speeding up, now the MOX impact on
transience. This is all pretty well known, but I'd
just like to point out a few other things.
The PTS screening criteria which are now
under review for all plants may not be appropriate for
MOX cores, in other words, the ones that are
appropriate for the LEU may not be appropriate for
MOX, and one reason is the reduced decay heat
immediately after a SCRAM in a MOX core would lead to
a more rapid decline in the temperature in the reactor
coolant system, and therefore, a more rapid entering
into a region, a temperature region where the pressure
vessel might be threatened.
Another aspect, well, again, if you have
an air oxidation source term with greater fuel and
ruthenium releases, then the source term might be more
severe for a MOX core in a PTS event, and a final
point is that because of the greater fast flux, the
embrittlement is going to be somewhat more rapid, and
this is not something that Duke Power is planning to
take into account at its license renewal time limited
aging assessments.
As a matter of fact, Duke made the
alarming statement that, well, license renewal comes
first, and then they'll evaluate MOX, and if there was
a risk that using MOX would impair the ability of
their plant to operate safely to the end of the
license renewal period, then they won't do MOX.
And when I heard that, I wondered if the
Department of Energy knew that was their position, but
considering there's only a two-year, I think,
difference between when they're doing their license
renewal and when they'd have to do the MOX assessment,
it would make sense to do it all at once in my view.
Moving right on in the reactivity
insertion, we all know the increased vulnerability of
MOX to RIAs or potential increased vulnerability as
demonstrated in the REP Na-7 Cabri test is a concern.
And a key consideration is the fuel homogeneity and
the size distribution of the plutonium agglomerates.
And, you know, this has been known, I
think, for decades, and Westinghouse in its
consultant's report to DOE recommended -- this is a
quote -- "adherence to limits on plutonium
agglomerates in the range of 10 to 15 microns should
be required."
And in that context, it's pretty alarming
to learn that DCS appears to actually be relaxing the
existing specification that's in use at the Maalox
(phonetic) plant in France, when they should be going
in the other direction.
And the Cogema MIMAS plutonium particle
distribution that's currently achieved has a mean size
of the distribution of the agglomerates of 20 to 40
microns, and the specification is no more than two
percent of the clusters should be greater than 100
microns in size, and the maximum that occurs is around
140, I believe, while the DCS specification, at least
in the version of the fuel qualification plan, which
we submitted last year, and I understand there's a new
version now; so this may have changed, but they
specify a mean size of less than 50 microns and a
maximum five percent of clusters greater than 100
microns with a maximum size of 400 microns.
So instead of trying to bring this number
down to the ten or 15 range that Westinghouse
suggested, they seemed to be going in the other
direction. I think if this is actually the case that
it's something that they need to be called to account
for.
Now, on the issue of MOX --
CHAIRMAN POWERS: A 400 micron inclusion,
a 400 micron plutonium inclusion would be a fairly
significant inclusion, wouldn't it?
DR. LYMAN: Yeah. I mean, it's about the
maximum. It was the maximum that was set back when
they did those experiments in the '70s or '60s, and
hopefully technology has improved since they were
making this.
CHAIRMAN POWERS: I'm just trying to
understand what the neutronic effects of a 400 micron
-- I mean that's a pretty healthy inclusion, isn't it?
DR. KRESS: It's pretty good.
CHAIRMAN POWERS: I think you would worry
about that.
DR. KRESS: I think you'd see it.
CHAIRMAN POWERS: Yeah, I think you would
see something.
DR. LYMAN: Well, it's right in the fuel
qualification plan if you want to take a look at that
number.
Now, generally speaking, we have a lot of
concerns about the way the fuel qualification is
coming about. First of all, the schedule, I think, is
pretty aggressive. They hope to load the LTAs and
start irradiating them in McGuire in October 2003.
Then they're going to do it for two 18-
month cycles, and so discharge would be around October
2006, and these twice burn LTAs, then they would be
subject to some nondestructive analysis, but the first
reload batch would be a year later.
So that only gives one year really for
doing all of the work that both the licensee and NRC
might want to do on these LTAs.
The other aspect is at least according to
the first version of the fuel qualification plan, they
wouldn't even be burned up to the maximum discharge
burnup that they're proposing for the fuel, but would
fall short, and that's another puzzling aspect.
Then the issue of where the LTAs are going
to be made is still not determined. As you know, Los
Alamos has its contract canceled last year, leaving
the program stranded. So the two bad alternatives now
are, one, the LTA is manufactured in a European
facility, but this raises the issue that they may not
be representative if it eventually comes out of a U.S.
plant, especially if the fuel qualification parameters
are different.
CHAIRMAN POWERS: When you say they're not
representative, are you speaking of the fact that they
did not have weapons grade plutonium in them or --
DR. LYMAN: Well, no. It wouldn't make
sense if they didn't, but where, you know, there had
been talk that it might come from England, you know,
I don't know the details, but it certainly wouldn't be
U.S. weapons grade plutonium that was aqueously
purified according to the plan that we have and
fabricated according to the specifications that DCS is
establishing.
So that has to be looked at. It may not
be that significant an issue, but again, given what
we've heard today about the variability and, you know,
expectations for fuel, small changes in composition,
manufacturing parameters, there seems to be some
sensitivity to these things.
And so I would be more confident if the
LTAs actually were a product of the plant that's going
to be manufacturing them, but the problem with that,
which is the other option, is that clearly it's going
to cause a delay if the U.S. MOX plan is going to be
the source of the LTAs because who knows? They'd have
to establish some sort of a pilot line, I guess, and
who knows if the fuel coming out of the pilot -- I
mean, the first fuel -- is going to be suitable or
representative of a later fuel?
So I think there are a lot of issues that
are not being dealt with adequately here, and because
of this aggressive timetable, NRC's ability to resolve
MOX fuel safety issues, I think, is in jeopardy.
Again, the time for post irradiation LTA
characterization testing is inadequate, forcing a
reliance on proprietary find data, which NRC is not
going to be able to confirm, and I think the M5
experience, however it plays out, should give pause in
this area because whether or not the M5 cladding,
which incidentally is the cladding that's going to be
used on MOX fuel, and Framatome is the fuel designer
and supplier for the MOX program here; whether or not
it turns out to be adequate and meets the existing
criteria, I'd have to say that the behavior of
Framatome since they were aware that they were doing
ring compression tests; they were aware that there was
an issue; they were aware of the results. The Germans
were making them do these tests.
At the same time NRC was reviewing and
approving the M5 cladding without knowing any of this,
and the fact is, you know, they didn't ask the
questions. So maybe they didn't have to get an
answer, but I think if Framatome was completely
forthcoming, they would have notified them.
And so I think it raises questions about
how reliant we should be on foreign data that's not
confirmed independently.
And in this regard, it's especially
frustrating that DOE appears to be uncooperative with
NRC's Office of Research, and you may not be aware
that the Office of Research sent a letter to DOE in
December requesting that access be granted to NRC to
have some samples of the irradiated lead test
assemblies taken to Argonne for NRC's confirmatory
testing.
DOE's answer was basically, "No, thanks.
It's duplicative, and you'd have to work that out with
the licensee anyway," wouldn't have anything to do
with it. It was an evasion.
And this is an example of how I think
things are going to play out especially in the context
of the budget cuts that we're going to see. DOE is
not willing to pay or support any of what it considers
additional research, and I think that's a mistake.
I think that both the timetable and the
staff resources for MOX safety issue resolution should
be based on NRC needs and not DOE needs. You know, in
an ideal world, NRC should design the research program
it thinks is necessary to answer the questions, give
DOE the bill.
(Laughter.)
DR. LYMAN: And then --
PARTICIPANT: In an ideal world.
DR. LYMAN: Right. Well, I'm an optimist.
Cancellation of the immobilization track
is going to increase pressure on NRC not to be
obstructionist in MOX licensing, and I think this path
for MOX approval is not likely to engender public
confidence the way things are going.
So I would like to see a tightening up of
the goals and the objectives and a good research
program addressing some of these concerns.
Thank you.
CHAIRMAN POWERS: Any questions of the
speaker?
That was a great presentation. I think we
appreciate it when you take the time to come talk to
us.
DR. LYMAN: Oh, I appreciate the
opportunity.
CHAIRMAN POWERS: Thank you.
DR. SHACK: Let me. What is your argument
again about why this is appropriate for the power up
rates? You're not arguing that the source terms is
increased in the same way. Are you just saying that
you should consider the change in source term and use
it to modify the LERF?
DR. LYMAN: Well, the source term is
increased not in the same way, but some of the --
DR. SHACK: Okay, but your argument is you
should consider that change in the source term and
modify the acceptance on the LERF. That's what you
mean.
DR. LYMAN: Right.
DR. CRONENBERG: The scale and not the
source term.
DR. LYMAN: Right. I mean, this is
actually discussed here last year where there was some
argument how do you risk inform this if you don't have
a tool that takes into account change in source term,
and I'm saying this is one way to do that.
DR. CRONENBERG: When did the mobilization
-- was that really canceled?
DR. LYMAN: Well, they suspended the
contract. They had had a request for proposals put
out for a mobilization contractor. That's going to be
suspended. That money was zeroed out for the coming
fiscal year.
They don't say it's been canceled, but
everyone I know or what I've heard from people inside
the program is it's dead. People have been
reassigned. The work is over.
Thanks.
CHAIRMAN POWERS: Ralph, we have some time
scheduled for the full Committee tomorrow on this
general area of high burnup and MOX fuel. I'll be
frank. I did not see anything that I felt a burning
need to bring before the full Committee. Is that true
or do you have a different perception?
DR. MEYER: No, I think that's okay. I
was just wondering what you expected the staff to
prepare for tomorrow.
CHAIRMAN POWERS: Well, what I was going
to suggest is, I mean, you've basically given us an
update on where you stand, that you've gone through
your PIRTs. I think that's great.
I was just going to suggest that I'd give
a quick summary to the ACRS and let it go at that.
DR. MEYER: Okay.
CHAIRMAN POWERS: I mean, there's nothing
for us to write a letter about. So I hope you're not
expecting a letter from us.
DR. MEYER: Right, right.
CHAIRMAN POWERS: We need to produce a
letter that says to close out one of the GSIs on this
high burnup fuel.
DR. MEYER: Yes.
CHAIRMAN POWERS: Okay, and basically what
we need to be able to say is everything that's listed
in that GSI is being addressed in the research
program, and I think we're on safe grounds in saying
that.
DR. MEYER: That's correct.
CHAIRMAN POWERS: Okay. So it seems to me
that the only thing we need to do is why don't I just
give a summary of what went on at this meeting? You
guys can go do your work and actually make some
progress.
DR. MEYER: Okay.
CHAIRMAN POWERS: And that's not put -- I
mean, I just don't see a need to have a -- I'm sure
the Committee members would be very interested in
what's going on, but that's all it would be, would
just be technical interest and whatnot, and that's the
job of the subcommittee. We get the fun job.
DR. MEYER: Okay.
CHAIRMAN POWERS: They've got to work
hard.
DR. MEYER: That sounds fine to me. So I
don't have to prepare a presentation tomorrow.
CHAIRMAN POWERS: I don't think you need
to prepare a thing.
Richard, similar I think on the MOX.
You're just getting started. I don't see anything.
I think between Med and I we can take your viewgraphs,
put together a viewgraph that says, "Here's what we
talked about, and our intention is to come back and
look again roughly in the fall."
Because that looks like when things were
coming down both from Margaret's perspective and from
your perspective; is that right, Ralph?
DR. MEYER: Okay.
CHAIRMAN POWERS: I mean that's all I see
to do. I think it was a great update, but I just
don't see anything that the Committee needs to act
upon, except we need to get that GSI out.
DR. MEYER: Yeah.
CHAIRMAN POWERS: But I think that's --
DR. MEYER: That's a separate.
CHAIRMAN POWERS: It's a separate issue
for us, but I think it's -- I mean, I think what we
needed from you is the assurance that the research
program is covering it.
DR. MEYER: The assurance that?
CHAIRMAN POWERS: The research program --
DR. MEYER: Yes.
CHAIRMAN POWERS: -- is taking into
account everything that --
DR. MEYER: It does. It does cover
everything that was said.
CHAIRMAN POWERS: And I think that was all
that was needed.
DR. MEYER: Yeah.
CHAIRMAN POWERS: Okay.
DR. MEYER: Okay. Great.
CHAIRMAN POWERS: Any other comments
people would like to make?
(No response.)
CHAIRMAN POWERS: In that case, I will
adjourn this meeting of the Subcommittee with thanks
to the speakers. All very interesting, and at the
same time somewhat confusing in that there obviously
is at least one variable that I don't understand in
clad behavior.
(Whereupon, at 3:17 p.m., the Subcommittee
meeting was adjourned.)