Thermal-Hydraulic Phenomena/Future Plant Designs - February 14, 2002
Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title: Advisory Committee on Reactor Safeguards
Combined Thermal-Hydraulic Phenomena/
Future Plant Designs: Subcommittee Meeting
OPEN SESSION
Docket Number: (not applicable)
Location: Rockville, Maryland
Date: Thursday, February 14, 2002
Work Order No.: NRC-232 Pages 376-447/485-729
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
(ACRS)
COMBINED THERMAL-HYDRAULIC PHENOMENA/
FUTURE PLANT DESIGN: SUBCOMMITTEE MEETING
+ + + + +
ROCKVILLE, MARYLAND
+ + + + +
THURSDAY,
FEBRUARY 14, 2002
+ + + + +
The Subcommittee met at the Nuclear
Regulatory Commission, Two White Flint North, T2B3,
11545 Rockville Pike, Rockville, at 8:30 a.m., Graham
B. Wallis, Chairman, presiding.
COMMITTEE MEMBERS PRESENT:
GRAHAM B. WALLIS, Co-Chairman
THOMAS S. KRESS, Co-Chairman
DANA A. POWERS, Member
VIRGIL SCHROCK, Consultant
WILLIAM J. SHACK, Member
JOHN D. SIEBER, Member
STAFF PRESENT:
PAUL A. BOEHNERT
MAGGALEAN W. WESTON
MEDHAT EL-ZEFTAWY
ALSO PRESENT:
JOE WILLIAMS
DALE SPENCER
BOB KERESTES
HAROLD CROCKET
FRAN BOLGER
TIM BYAM
LARRY WESTBROOK
JASON POST
KENT SCOTT
ISRAEL NIR
BILL BURCHILL
KEITH JURY
TAD MARSH
JON HOPKINS
GEORGE THOMAS
TONY ULSYS
RICHARD LOBEL
DAN PAPPONE
ED THROM
ZEENA ABDULLAHI
ALSO PRESENT (Continued):
JARED WERMIEL
MOHAMMAD SCHUABI
DAVID SHUM
ANDRZEJ DROZD
JIM LYON
LARRY BURKHART
JERRY WILSON
GOUTAM BAGCHI
DAVID TERAO
RICHARD ORE
MIKE CORLETTI
STEVE BAJOREK
MARINO DI MARZO
WILLIAM BROWN
KATSU OHKAWA
C-O-N-T-E-N-T-S
PAGE
Presentation of AmerGen . . . . . . . . . . . . 381
Discussion of Large Transient Testing, Larry
Westbrook . . . . . . . . . . . . . . . . 389
Evaluation of Risk Impact of EPU on Clinton,
Bill Burchill . . . . . . . . . . . . . . 405
EPU Project Implementation, Larry Westbrook . . 437
Conclusion, Keith Jury . . . . . . . . . . . . . 486
NRC Staff Presentation
Ted Marsh . . . . . . . . . . . . . . . . 488
Jon Hopkins . . . . . . . . . . . . . . . 489
George Thomas . . . . . . . . . . . . . . 493
Richard Lobel . . . . . . . . . . . . . . 499
Review of Phase 2 Pre-application of Westinghouse
AP 1000 Plant Design:
Andrzej Drozd . . . . . . . . . . . . . . 528
Jerry Wilson . . . . . . . . . . . . . . . 533
David Terao . . . . . . . . . . . . . . . 551
Westinghouse Presentation:
Richard Ore . . . . . . . . . . . . . . . 581
Mike Corletti . . . . . . . . . . . . . . 596
Presentation by Andrzej Drozd . . . . . . . . . 598
C-O-N-T-E-N-T-S
Discussion of Scaling:
Steve Bajorek . . . . . . . . . . . . . . 600
Marino di Marzo . . . . . . . . . . . . . 633
Westinghouse Presentation:
William Brown . . . . . . . . . . . . . . 702
P-R-O-C-E-E-D-I-N-G-S
(8:33 a.m.)
CO-CHAIRMAN WALLIS: The meeting will
please come to order.
This is the second day of the meeting of
the ACRS combined Subcommittee on Thermal-Hydraulic
Phenomena and Future Plant Designs.
We will continue to hear from the AmerGen
Company about the application for extended power to
operate at the Clinton Power Station, Unit 1. And
then we'll take a break at about ten o'clock, and then
we'll hear from staff.
MR. BYAM: Good morning. I'm Tim Byam
with AmerGen, and we're prepared at this point to
answer the one question we took away with us
yesterday.
I'd like to turn it over to Joe Williams.
MR. WILLIAMS: Thank you.
Joe Williams, Exelon Nuclear.
Yesterday during the flow accelerated
corrosion presentation we stated that the highest wear
rate location predicted for the power upright flow
accelerated corrosion calculations was 70 mLs per year
versus a 38 mLs per year predicted for pre-EPU
conditions.
The question was asked what the actual
measured wear rate has been at that location. Bob
Kerestes of AmerGen will provide a response.
MR. KERESTES: Good morning. My name is
Bob Kerestes, and we did some checking and the actual
wear rate is 20 mLs per year. We have looked at this
area three times in the past. So we have a good
baseline from which to start.
MEMBER POWERS: Does that imply that the
post EPU corrosion rate will be 52 mLs per year?
MR. KERESTES: I don't think you can
necessarily say that it's going to, you know, double
or be uniform, and that's why we have an
erosion/corrosion, flow accelerator corrosion program
that we look at, and we will continue to monitor this
area.
MEMBER POWERS: Could it be 88 mLs per
year?
MR. KERESTES: Again, you know, we will
continue to monitor this and look at it, and it's
going to change, and we need to look at it closely.
MR. WILLIAMS: We believe the calculation,
the prediction is conservative. We can provide
additional information from our flow accelerator
corrosion specialist.
MEMBER POWERS: Well, I mean, it looks
like it's 100 percent at certain -- when the air range
is equal to the major value.
MR. WILLIAMS: Harold?
MR. CROCKET: Harold Crocket.
In addition to what Bob just said, we do
have inspections ongoing. This particular line has
replacements scheduled for refueling outage ten, which
will be two cycles from now. So we do have scheduled
replacements, and as he stated, it is in our program,
as are the other lines in the steam cycle.
MEMBER SHACK: What kind of a line is it?
MR. CROCKET: It's scavaging steam goes
from the moisture separator reheater to the heaters.
CO-CHAIRMAN WALLIS: It's just pure steam
going through there?
MR. CROCKET: The quality of the steam is
a mix. I don't remember the exact numbers.
CO-CHAIRMAN WALLIS: So they are all
droplets in there. So this is probably why the wear
is so high.
MR. CROCKET: Yes, sir. And these lines
throughout the industry are noted for being fairly
high wear.
MR. BOEHNERT: You're indicating that I
guess eventually these lines are going to be -- all of
them are going to be changed out
MR. CROCKET: Normally we would upgrade
this type of line to a chromolly or stainless, usually
chromolly. But we don't automatically upgrade unless
we see a significant amount of wear either in a sister
train or that type of thing. Normally we would
inspect first.
MR. BOEHNERT: Sure.
CO-CHAIRMAN WALLIS: Anymore questions
about this issue?
MEMBER POWERS: Well, I guess the next
question, I mean, what you're concerned about is some
sort of a surrey event where we get surprised and
suddenly we fry workers and things like that. With
this kind of uncertainty in the calculation, of
course, the next thing you ask is: are there areas
where CHECKWORKS predicts there's no problem and, in
fact, there is a problem?
CO-CHAIRMAN WALLIS: Well, is CHECKWORKS
always conservative?
MEMBER POWERS: Yeah, I mean, that's what
you're asking.
CO-CHAIRMAN WALLIS: What kind of
uncertainty are we talking about?
MR. CROCKET: CHECKWORKS, part of the
planned operation of the code is that you do
ultrasonic inspections to refine the analysis, and you
input your actual measured wear form your UT exams
back into the program, and it has what's called a line
correction factor that refines those predicted wear
rates.
So, yes, they acknowledge that the code is
just a tool.
MEMBER POWERS: Well, I mean, the
challenge you face is that where you do those UT
inspections is where the code tells you you ought to
do the UT inspections; is that correct?
MR. CROCKET: The code is one tool that we
use. The industry interaction has been really
excellent in the realm of fact, and EPRI sponsors
meetings twice a year. We talk about where everybody
is working and what's the right solution, either
material upgrade or changes to the piping system
itself, if you need a larger diameter line to cut back
the velocity or whatever, whatever is appropriate for
the specific location.
CO-CHAIRMAN WALLIS: With this particular
line now, you've changed the steam conditions. You've
got more pressure drop from the boiler to the turbine,
a different kind of wetness, different pots in the
cycle. It might well be the steam scavaging line
would see significant change in dryness fraction. So
it's not clear that it's all that predictable what's
going to happen.
MR. BOEHNERT: Dana, I don't --
CO-CHAIRMAN WALLIS: Is that a true
statement I've made or do you take issue with that?
MR. WILLIAMS: We believe it is
predictable by the CHECKWORKS methodology, correct?
MR. CROCKET: The code has largely proven
very reliable, and there have been few occasions that
they have upgraded the analysis, but they always are
looking at refinements and periodically have new
releases of the code to acknowledge certain things
either in chemistry or different -- where the industry
is learning about.
MEMBER POWERS: CHECKWORKS seems to get
used broadly within the United States. I notice the
Taiwanese developing alternatives to CHECKWORKS. Has
anybody looked at those to see if they're more useful,
less useful?
MR. CROCKET: There is one site in the
domestic utilities that does not use CHECKWORKS, and
that site uses three analysis. So it's proven very
reliable, and there are other foreign utilities that
also use CHECKWORKS, and we have not seen a reason to
switch over from an industry perspective.
CO-CHAIRMAN WALLIS: Does CHECKWORKS put
in something about the flow regime when you have
droplets in the steam?
MR. CROCKET: It recognizes steam quality.
CO-CHAIRMAN WALLIS: It does? It has some
way of handling that?
MR. CROCKET: Yes, sir.
MEMBER SIEBER: That's provided you know
what it is.
CO-CHAIRMAN WALLIS: Mostly you know where
the drops are. When you have a bend --
MEMBER SIEBER: That's right.
CO-CHAIRMAN WALLIS: -- you get subterfuge
and they come off --
MEMBER SIEBER: Yeah, it heats up the end
of the bend.
MR. BOEHNERT: Several years ago, well, I
guess it was four or five years ago, Fort Calhoun blew
out a pipe, and they were using CHECKWORKS. In fact,
we had Mr. Checks in here to talk about it.
MEMBER SIEBER: Yes, yes.
MR. BOEHNERT: And his allegation was or
he said if you use the code properly, it works right.
And as I recall, Fort Calhoun, I don't believe they
were using the code, what they should have done
because they went back and looked at the code again,
and the code, indeed, predicted that area that blew
out was a problem, and they had missed it the way they
were applying the code.
So the code has been shown to be quite
reliable when it's properly used.
MEMBER SIEBER: Yeah, there's another way
that you can get into trouble, which is usually you
don't analyze the whole plant. You, you know, start
with the very large pipes and go down to some level.
Unfortunately, erosion/corrosion will eat away any
pipe, and a lot of the smaller lines are just as
susceptible as the large ones, which leads to the
question of how far down in pipe size do you go to do
analysis for Clinton.
MR. CROCKET: We do all susceptible piping
that's in our program, and if the information is not
there for the analysis, we do inspections that are
commensurate with that.
But there is no size restriction that we
don't go beyond. We have one inch, two inch, socket
weld fittings that we do exams on and replacements,
and it's all in the program if it's in the steam
cycle.
MEMBER SIEBER: Okay. Thank you.
CO-CHAIRMAN WALLIS: Are we finished with
this issue now?
Return to the original plan?
MR. WILLIAMS: Yes, sir, and thank you,
Dr. Wallis.
I would like to introduce Larry Westbrook.
MR. WESTBROOK: Good morning. My name is
Larry Westbrook of AmerGen. I am the Clinton Power
Station senior reactor operator, leading the EPU power
ascension testing team.
I will be discussing large transient
testing.
First, allow me to provide some background
related to large transient testing. ELTR1 specifies
that large transient tests be performed. One
specified test is closure of all main steam isolation
valves at 110 percent of original licensed thermal
power.
The second specified test is a generator
load rejection at 115 percent original licensed
thermal power. AmerGen has taken exception to these
tests as an unnecessary challenge to the plant.
General Electric has concluded that large
transient tests are no longer necessary when reactor
steam dome pressure is unchanged. Supporting this
conclusion are the following.
Clinton Power Station reactor steam dome
pressure remains the same following power up-rate.
Large transient tests will provide no significant ne
information, and existing modeling code adequately
predicts plant response.
Next, I would like to present AmerGen's
methodology for exempting large transient testing.
First, we have reviewed industry power up-rate
experience and determined that similar plants have
demonstrated acceptable performance following large
transients. kkL performed planned transient testing
at up-rated conditions with acceptable results.
Also, an unplanned generator load
rejection at Hatch at up-rated power levels showed
that the plant responded as expected.
Second, performance of plant structures,
systems and components, at Clinton Power Station has
been evaluated at EPU conditions and determined to be
acceptable.
Also, surveillance testing will confirm
that plant structures, systems, components maintain
their required performance capability.
In conclusion, large transient testing
will provide no new significant information. This is
supported by the following.
Plant response to large transients at
extended power up-rate conditions will not change
significantly as demonstrated by acceptable transient
modeling code results and operating experience at
similar up-rated plants.
Based on evaluation and existing
surveillance testing, Clinton Power Station structure,
systems, and components will perform as designed.
Therefore, large transient testing is not required at
Clinton Power Station.
MEMBER POWERS: Can you show us an example
of the plant response to transients before and after
the upgrade?
MR. WILLIAMS: I think your question is
getting at the details of the transient.
MEMBER POWERS: You can just point to it
in the documentation if it's in there.
MR. WILLIAMS: I believe it is. G.
MR. WESTBROOK: Could you repeat the
question, please?
MEMBER POWERS: A key element in this
conclusion slide is the plant response to these
transients as a result of EPU not changed
significantly based on modeling, and I was just asking
is there some demonstration of that equivalency of
response.
CO-CHAIRMAN WALLIS: That's about the only
conclusion that can be tested. Statements such as
plant components will perform as designed is a
statement of faith. What you really mean is that you
think there's a high probability that they will
perform as designed.
And the same thing about tests will not
provide new significant information. You've got a
couple of tests already. One more, that's not a very
big base of tests. So, again, these are probablistic
statements. You think it's not going to provide new
information.
It will provide new information. You just
don't think the information is going to be very
useful.
MR. WESTBROOK: That's correct. I don't
think that the information that would be provided will
be significant or --
CO-CHAIRMAN WALLIS: But if it were to
result in a failure or something, that would be
extraordinarily useful information.
MEMBER POWERS: Without a doubt.
CO-CHAIRMAN WALLIS: So something is
happening here about assessment of the likelihood that
the plant will perform as designed.
MR. WILLIAMS: G.E.?
MR. NIR: Israel Nir, G.E.
What I would like to add is that typically
the actual events at the plant will turn out to be
very benign, and we have discussed this with you in
other occasions. There are control systems in place
and protection systems that actuate specifically to
address this large transient, and the end result is
that typically we see a very small increase in neutron
flux, you know, a pretty minor change in peak
pressure.
And that has been demonstrated
analytically, and that is what is expected from an
actual event.
Now, to actually test the type of things,
I think, that you may have in mind, you really need to
impose some restrictions or come up with a sequence
that may represent some risk to the plant, and it is
just not advisable, such as, for example, have the
bypass, turbine bypass inactivated.
So this type of sequences will not occur
during the actual plant events, and they will result
in a very benign event.
MEMBER SHACK: The prediction is made that
you've done these calculations and comparisons for kkH
and kkL. I mean, all you have to do is essentially
show the comparison presumably.
MR. NIR: Right, and we have -- in other
occasions we have shown this Committee -- excuse me --
last year we have shown you some comparisons for kkL.
kkL happened to be a plant that has 100 percent bypass
capability, and again, the comparisons are good, and
they effects are very benign.
MEMBER SCHROCK: I guess I continue to be
puzzled by what prompted G.E. to require the testing
in the ELTR originally. What has changed that changed
the G.E. position?
MR. NIR: I would say two elements. One,
it was a new program. It was, you know, a major step
forward. We had to consider the entire licensing
process, and that at the time appeared to be the right
thing to do.
The other thing is that we included in the
ALTR allowance for increasing dome pressure of up to
75 psi. That is not the case here. There's no
increase in pressure.
Based on our experience, and now that we
have performed analysis for a large number of EPUs and
actually implemented a few, our position is that if
the dome pressure is not increasing, we don't see the
value for these large transients. Our judgment is
that, again, there is no significant information that
can be extracted from them.
MEMBER POWERS: I guess I'm still
struggling to understand the basis of the second point
here. I mean, the statement is made that we've
modeled before and after and it looks the same. Does
it? I mean, can you show me that this is the case?
MR. NIR: Yeah. That's, again -- I
believe that we already shared some f this information
with the committee, but we can -- we can definitely
review it with you and show you some comparisons.
CO-CHAIRMAN WALLIS: Can you find it this
morning?
MR. NIR: This morning? Yes.
MR. POST: This is Jason Post.
For example, I have an ATWS comparison.
It's not exactly what we're after, but I mean, I have
that that we could show them if that's --
MEMBER POWERS: I mean, you looked at
something in order to write the sentence down on the
viewgraph. If that is an adequate demonstration of
the truth of the second point, then let's see it.
MR. BOLGER: This Fran Bolger from G.E.
When we presented a few weeks ago, I
presented for a AAOO licensing evaluation a low
rejection type transient before and after EPU. Would
you like to see that comparison?
MEMBER POWERS: Well, is it done for the
Clinton plant?
MR. BOLGER: No, it's not.
MEMBER POWERS: Well, then it's not very
useful, is it, to support this statement?
MR. BOLGER: The plants have shown a very
small change of their flux response to a low rejection
type event. All the plants that we've analyzed have
shown a similar type, very small change.
Yesterday when we presented the results of
the transient analysis, the decrease in the calculated
operating limit was presented to you. If that were
looked at in more detail, it would show that there's
a reduction in the integral flux peak before and after
power up-rate, and as was stated yesterday, that was
primarily due to change in the fuel characteristic and
not due to the power up-rate.
CO-CHAIRMAN WALLIS: I think we have seen
these figures before, but I'm trying to just get a
picture in my mind. I think a blue and a red curve or
something with these.
MEMBER POWERS: I simply don't recall any
pictures for Clinton
CO-CHAIRMAN WALLIS: Oh, I don't know for
Clinton.
MEMBER POWERS: Well, I mean, we're
talking about Clinton here.
CO-CHAIRMAN WALLIS: He's saying Clinton
is very close to the other plants.
MEMBER POWERS: I mean the other ones we
were talking about, BWR-3, I think.
CO-CHAIRMAN WALLIS: That may well be
true. That may well be true.
Perhaps you could dig something out to
give us before the --
MEMBER POWERS: There must be something
that existed in order to write the second bullet down.
CO-CHAIRMAN WALLIS: Yes, that's right.
That's the only one which can really be tested by
evidence, isn't it, without doing the test?
MR. WILLIAMS: Yes, we will.
CO-CHAIRMAN WALLIS: Okay. Thank you.
MR. BOLGER: Let me clarify the request.
Would you like to see the Leibestodt comparison?
CO-CHAIRMAN WALLIS: Like to see a Clinton
comparison that supports conclusion two or something
from a plant which you can show is very, very close to
Clinton for some reason.
MEMBER POWERS: The technical basis by
which conclusion number two was reached.
CO-CHAIRMAN WALLIS: And if you have a
code prediction for the same event, that would also
help to reinforce some of these statements. The
existing modeling code adequately predicts plant
response, but there you'd have to go to something like
Leibestodt. We've seen that before.
MEMBER POWERS: That's the first question.
CO-CHAIRMAN WALLIS: But if you happen to
have it, it might be useful to remind us of it.
With that request for information, are
there any other points that the Subcommittee wants to
raise?
(No response.)
CO-CHAIRMAN WALLIS: All right. So let's
move on then. We'll return to this question later.
MEMBER POWERS: We might share with
everyone that part of our challenge that we face here
is that we have a member not present in the
Subcommittee meeting that disagrees with this
conclusion that transient testing is not correct, and
we were unable to persuade him of the wisdom of not
testing based on hand waving arguments.
So more hand waving arguments presumable
are not going -- we're not going to wear him down.
CO-CHAIRMAN WALLIS: Maybe he'll persuade
us this time.
MEMBER POWERS: And he might well persuade
us this time.
CO-CHAIRMAN WALLIS: Okay.
MR. WESTBROOK: The next topic I would
like to discuss is ATWS event response.
The Clinton Power Station extended power
up rate ATWS event response operator actions are
unchanged from pre-EPU conditions. For example,
reactor power and flow response to reactor
recirculation pump trip during an ATWS remains
unchanged from pre-EPU conditions. This is a result
of no change in the flow control line.
The symptomatic conditions for ATWS,
greater than five percent APRM power following a scram
initiation and shutdown criteria not met remain
unchanged for EPU conditions.
The mitigating operator actions remain
unchanged for controlling reactor power, water level
and pressure. These mitigating operator actions
include standby liquid control initiation, lower
reactor water level, and alternate control rod
insertion.
CO-CHAIRMAN WALLIS: The first statement
there, I'm trying to figure out what you're saying.
The reactor power and flow is unchanged from what it
was pre-EPU. Is that what that says? You have the
same power before and after EPU?
MR. WESTBROOK: That's correct.
CO-CHAIRMAN WALLIS: That's a bit strange
if you've got a power up rate.
MR. WESTBROOK: That is following the
reactor resert. (phonetic) pump trip.
CO-CHAIRMAN WALLIS: Once you trip the
pump, the reactor power becomes the same as it was
pre-EPU?
MR. WESTBROOK: That's correct.
As a result of a plant specific analysis,
the minimum allowable standby liquid control boron
concentration was raised to insure that the rate of
negative reactivity addition during an ATWS remains
acceptable.
The table compares analytical results for
EPU ATWS versus original license thermal power ATWS
and design limits. This shows the small changes in
ATWS related peak parameters will occur due to EPU
conditions, and that the peak parameters will remain
within design limits.
The small magnitude of these parameter
changes and the fact that they remain below limits
supports the adequacy of existing operator actions to
mitigate an EPU ATWS event.
MR. BOEHNERT: How much did you have to
raise the concentration of boron?
MR. WESTBROOK: Actually we're not
changing the concentration that we maintain, but we
had a very low minimum level. It was 10.3 percent,
and we essentially made the minimum allowable 10.8
percent.
CO-CHAIRMAN WALLIS: Are these numbers
dependent on proper operator actions?
MR. WESTBROOK: Yes.
CO-CHAIRMAN WALLIS: And they have less
time with the power up rate than they had before?
MR. WESTBROOK: No, that's not correct.
They have the same time.
CO-CHAIRMAN WALLIS: The same time?
MR. WESTBROOK: That's correct.
CO-CHAIRMAN WALLIS: So the operator
actions are all the same as before?
MR. WESTBROOK: That's correct.
CO-CHAIRMAN WALLIS: Do they have to take
action to avoid an instability region or something
like that?
MR. WESTBROOK: During an ATWS, they will
always observe for instabilities and take the actions
as directed by the EOPs.
CO-CHAIRMAN WALLIS: So they could get
into an unstability region during an ATWS event.
MR. WESTBROOK: That's correct. However,
the EOPs are designed to compensate and to mitigate
those oscillations.
In conclusion, Clinton Power Station has
implemented an ATWS mitigation strategy consistent
with the BWR owners group emergency procedure
guidelines. Therefore, EOP operator actions remain
unchanged for EPU conditions.
MEMBER SHACK: Now, is it my memory that's
faulty that in the previous up rates we did see a
change in the operator response time for the ATWS so
that Clinton is different in that respect?
MR. SCOTT: This is Kent Scott from
AmerGen, senior reactor operator.
I believe in the next presentation that
Bill Burchill gives us on probablistic risk
assessment, we'll see that the allowable times for
operator actions are somewhat reduced, but also the
key is that the magnitude of that reduction in time
has no impact on the actions that we take.
So we're taking the same actions. We're
doing them the same way, but as opposed to having 31
minutes to automatically depressurize the unit, now
we've got 28 minutes to do it. They're of a magnitude
where it's easy to do those actions. We can get those
done in that same period of time.
MEMBER SHACK: But when these calculations
are done, they're done assuming the operator does
something at a certain time, and that time is the same
for before and after the EPU.
MR. POST: This is Jason Post of G.E.
That's correct. The example operator
action time are injection of boron and initiating the
RHR suppression pool cooling, and we use exactly the
same operator response time in both cases.
MEMBER SHACK: What is that response time
that was used in the calculation?
MR. POST: For RHR, I believe it's 11
minutes after the beginning of the event, and for
boron injection it's about two minutes after the
beginning of the event, and we use the same in both
pre-EPU and the EPU.
CO-CHAIRMAN WALLIS: So you expect the
action at 11 minutes, but the time available is 28 or
31, whichever?
MR. POST: The time available in the PRA
analysis --
CO-CHAIRMAN WALLIS: In order to success,
to be on the right path of succeed.
MR. POST: Right.
CO-CHAIRMAN WALLIS: So if the operator
waits for 35 minutes, then it might be off on some
other path.
MR. WESTBROOK: Next I would like to
introduce Mr. Bill Burchill, Exelon Nuclear.
CO-CHAIRMAN WALLIS: This is where we need
our PRA expert, keeping a reserve.
MEMBER POWERS: We have on the committee
the classic problem. We have not one PRA expert, but
two or three, and as in all things of this nature,
they simply disagree.
CO-CHAIRMAN WALLIS: Well, Bill is going
to explain it so that even we can understand it.
MR. BURCHILL: Thank you, Dr. Wallis.
Good morning. My name is Bill Burchill.
I'm the Director of Risk Management for Exelon. Good
morning, and happy Valentine's Day. I hope you've all
bought your flowers. Your wives will be happy if
you've done that. Just a little reminder.
I'm here to present the results of the
evaluation of risk impact of the EPU on the Clinton
Power Station. This, as you know, is not a risk
informed submittal, but according to the ELTR-1, risk
evaluation will be done, and as you may also know, an
ELTR-2, a generic evaluation was performed.
That generic evaluation was for a smaller
power up rate than is being proposed here for Clinton,
and we have, therefore, done a plant specific
evaluation.
The presentation that I'll go through this
morning is very similar to one that we discussed a
couple of months ago for Dresden and Quad Cities, but
I'll point out that the Clinton Power Station
experienced much less impact in the risk area than did
Dresden and Quad. The numbers there were small
numbers. Here are even smaller. So we'll discuss
that.
The purpose of the evaluation is to use
generally accepted risk metrics, core damage
frequency, and large early release frequency to
measure the risk impact of the changes being proposed
for the plant.
We've done quantitative evaluations using
the full power internal events, PRA model, and we've
done quantitative evaluations of individual effects
where we had other quantifying tools. Where we did
not, we did qualitative evaluations, particularly in
the area of external events and shutdown operation.
We've identified revisions to the PRA that
would be made to represent the plant following the
EPU. These revisions were represented in sensitivity
studies. During our risk evaluation, they will be
folded into the PRA at its next scheduled update.
The subjects that I'll cover age
essentially the same as I discussed with you a few
months ago, the quantitative, the qualitative, and
then a summary of the risk impact.
The methods that we used were to identify
the plant configuration changes due to the EPU. This
involved examining all of the hardware changes,
procedures changes, operating condition changes, and
set point changes.
The principal change, of course, here is
the power level. As a number of speakers before have
noted, there is really very little change to this plan
in terms of actual hardware and operation other than
the power level. Most of those changes are out in the
balance of plant or in the switchyard.
The primary components' pumping equipment
and so forth that we have available for mitigating
core damage essentially are remaining unchanged in the
EPU state. So there's no change to RCH RHR high
pressure core spray, low pressure core spray, ECCS.
NPH limits remain the same, and the service water
remains the same.
The only change that I'm actually aware of
is that there is a change out of impellers in the
TBBCW, but that has very little impact, and no impact
at all on dependencies.
We used the most recently updated PRA
model. As the staff noted in their SER, the Clinton
PRA has had a lot of attention over the years since
its initial development in the IPE days. They seem to
update more frequently than most that I'm familiar
with, and the most recent was in December of 2000.
I will point out that that update was
motivated by the fact that we were about to submit a
license amendment request to extend the emergency
diesel generator allowed outage time, and that has, in
fact, been approved by the staff.
When I first encountered this PRA, it was
about the time that we merged our companies, and I
wasn't particularly satisfied with what I found. So
we made a number of improvements to that PRA in order
to support that EDG AOT LAR.
CO-CHAIRMAN KRESS: Has your PRA been
through the industry peer review?
MR. BURCHILL: Yes. In fact, the industry
peer review was in August of 2000, and in fact, that
was just before our merger. I had a member of my
staff on that peer review team, and so that was my
first real detailed look at the PRA.
And the peer review results were what
really drove our changing the PRA in late 2000 in
order to support that earlier LAR.
We've also examined all of the elements of
the comments made by the peer review team, and we've
tabulated those for staff review relative to the EPU.
It turns out particularly with the changes that we
made for the EDG AOT extension there were no remaining
significant elements for this particular application.
As you know, in a PRA we use realistic
models. We've talked just a moment ago about operator
response times. I'll come back to that, but we use
realistic limits on thermal hydraulics, and we credit
all available systems, and likewise we compare to
realistic success criteria and limits.
MEMBER POWERS: When you think in terms or
realistic models, what do you do with the frequency of
a pipe break brought on with flow assisted corrosion?
MR. BURCHILL: Well, the frequency of pipe
breaks is represented by industry data of an empirical
nature. We do not do explicit analytical modeling of
fact in a predictive way. We look at industry
experience in like situations, and then we use that in
an actuarial fashion in order to predict those passive
component failures.
MEMBER POWERS: I guess what I'm asking
here is we're coming along. We're making a change to
this plan. One of the pieces of information coming to
us is that we have predictions that some piping
systems are more susceptible to corrosion and a more
rapid rate.
One would presume then that the
probabilities that the pipe would rupture unexpectedly
must go up. I can't imagine it doesn't.
MR. BURCHILL: Theoretically, yes.
MEMBER POWERS: I guess what I'm asking
is: how is that piece of information reflected in the
PRA analysis for the Clinton plant?
MR. BURCHILL: There is no change in those
rupture frequencies within the PRA model to reflect
what you're saying for two reasons.
On is that, again, the PRA model is
looking at a time average, you know, annual average,
and it's not a point prediction. So what we're
looking at is what's the overall impact for the
remaining operating life of the plant.
The second is that the FAC program, I
think, as we've mentioned earlier, does monitor these
pipes and well before they would be expected to
rupture with any probability of significance they
would be repaired or replaced.
So that's taken into account, you know,
industry-wide in the database that we use for the
actuarial data on pipe rupture frequency.
So the simple answer is there's no change
made, but I think there's a good rationale for not
making that change.
The other comment I'd make is the areas
we're talking about, extraction, steam lines, and so
forth, you know, really have very little impact.
You're down the dependency chain in terms of systems
that would respond to a potential core damage event.
You're down the dependency chain about three levels.
So I don't think you' ever actually find
a place to insert that information in the PRA.
MEMBER POWERS: I guess I will concede to
you that I have a poor example here, but --
MR. BURCHILL: Well, we'll work with it.
MEMBER POWERS: -- I'm going to pursue
that poor example because I don't believe your answer,
and that is that if I have a line that is corroding
more quickly, unless I have a monitoring program that
is accelerated parallel to that more quickly eroding
thing, even if I look at an annual average, I must
have a higher rupture frequency with that loss.
MR. BURCHILL: At the instant you're
talking about, yes, you would.
MEMBER POWERS: And your annual average,
I must have a higher value.
MR. BURCHILL: We could debate that. I
agree that it will be higher until that line is
replaced. Once the line is replaced, the rupture
frequency, I think you'll agree, goes significantly
down.
MEMBER POWERS: If I replace the line in
July, from January to December, that average I must
have a high rupture frequency.
MR. BURCHILL: Well, it will certainly
have a higher rupture frequency from January to July.
From July to December if we've replaced it with pipe
that's not susceptible, I would think the rupture
frequency would go down.
So I don't know whether the net would be
plus or minus.
MEMBER POWERS: Okay. Suppose I replace
it in November?
MR. BURCHILL: I think we're in agreement.
MEMBER POWERS: There must be something in
the PRA to reflect this additional information.
You're telling me there's nothing. What I come down
to is this. I mean, what I think I'm seeing in all of
these PRA analyses is that we change nothing, and we
find that the CDF doesn't change. This is not a
remarkable finding to me.
MR. BURCHILL: Hopefully I'll persuade you
otherwise through showing you the things that we
actually did change.
I think the more substantive part of the
answer is the question of what the impact would be of
this line rupture in terms of actually interrupting
any of our --
MEMBER POWERS: I really am using NMA as
an example. I'll agree we're working --
MR. BURCHILL: We're down here.
MEMBER POWERS: -- in subsidiary systems
that actually I'm surprised your answer was we don't
even model that at PRA, but I'm using it as an example
because clearly other things. It must be that
probably if you are rupturing, the core shroud must go
higher. It must be higher.
Do you reflect that in the RPA? I think
you don't.
MR. BURCHILL: Well, actually, as you're
well aware, we've had a recent occasion to experience
some jet pumping failures in another area, and we did,
in fact, explicitly represent that in the PRA and do
a fairly detailed analysis of its impact.
In this particular case, the impact could
be put in the PRA. It's certainly true that we do not
explicitly represent these lines. It would come in
through the initiating event frequency, you know, on
the secondary side, you know, loss of feedwater or
something of that nature that would be created by a
rupture of an extraction steam line.
MEMBER POWERS: Yes.
MR. BURCHILL: Shall I proceed, sir?
CO-CHAIRMAN WALLIS: Please do.
MR. BURCHILL: We did compare the results
of our analysis with the guidelines provided by
RegGuide 1.174. As you know, this RegGuide is for
permanent plant changes, for evaluating risk informed
submittals, but it is a generally accepted guideline
against which to benchmark results. So we did use
that.
The next slide shows how we went about the
review a little bit more detailed. We evaluated each
of the individual PRA technical elements. These, of
course, are the same elements that are evaluated by a
PRA certification team.
We didn't find any change in the
initiating event frequencies. There was nothing that
was indicated in the power up rate because of the lack
of major equipment changes that would create a change
in initiative events, and likewise the same is true
for the success criteria.
The main place where there were changes
were in the operator response times, and I will show
you the examples of the more significant impacts in
that area, but, again, even these changes were small
compared to the required coping time for the
operators.
MEMBER SHACK: Of course, a lot of those
elements were where you've got your two grades in your
PRA peer review.
MR. BURCHILL: Well, that's a good point,
and I'm happy you could bring that observation up. I
mentioned my pleasure with this PRA when I first
encountered it.
The grades on this PRA were somewhat lower
than the ones that I discussed with you before on the
Dresden and Quad situations, but they were a mix of
twos and threes, actually a slight majority of threes,
and some of the areas that they pointed out that they
were uncomfortable with were the ones that we attended
to in late 2000 prior to making that EDG AOT
submittal.
So I'm keenly aware of what you're talking
about in that PRA. I don't think that those detract,
and, again, we gave the staff a detailed tabulation on
this, from this particular application, but, you know,
I can go into whatever detail you like on that.
We also did quite a number of MAP runs to
support this evaluation. The MAP code, as you know,
is a best estimate code for severe accident, thermal
hydraulics and, in fact, releases. The reason that we
did that is that this is a fairly substantial change
in the thermal power of the core, and therefore, the
decay heat level is higher post trip and the heat up
and the boil down rates are considerably higher.
So we did a number of MAP runs in order to
support this evaluation.
CO-CHAIRMAN WALLIS: This is for the
containment, the MAP runs?
MR. BURCHILL: They are -- the MAP goes
from the core all the way out.
CO-CHAIRMAN WALLIS: It is in the core,
too?
MR. BURCHILL: Yes.
The next slide shows the evaluation of
effects in the operating conditions, and the following
slide will talk about systems. I'll preface this by
saying we found no new accident types in the
evaluation. We didn't find any significant changes to
accident scenarios, going back to the previous
discussions on operator response. We found no new
operator responses required, nor were the sequencing
of operator responses changed.
The only thing we did find in that area
was that there is a higher demand on the time of
operator actions in certain situations.
We found no change in system dependencies.
In other words, the front line systems are supported
by the same support systems and essentially the plant
is wired and hooked up the same way that it was
before.
And the risk evaluation revealed no
vulnerabilities produced by the EU.
The principal effect in the operating area
is the increased heat loads, the reduced time to
boiling in both the core and the suppression pool, and
of course, at least theoretically reduce the time to
core damage.
This produces a reduced time for equipment
response and operator actions, and in each case where
those operator actions were shortened and were felt to
be of significance, we evaluated explicitly the
analytical impact of that on the core damage
frequency.
The Atlas does produce increased power
levels and peak pressure. In this particular case, I
think the committee asked me this question a few
months ago on the other plants. ATWS only contributes
about four percent of the total core damage on this
plant. So it's a significantly smaller fraction than
we talked about before, and therefore, changes in this
area would likewise have a small impact.
The biggest difference between this and a
previous discussion is that although the feedwater
flow is increased, there's no change in the normally
operating number of feedwater pumps or condensate
booster pumps. So we still have the same two turbine
driven feedwater pumps and three condensate pumps and
condensate booster pumps that are in the pre-EPU
condition.
MEMBER POWERS: And these pumps are
working harder?
MR. BURCHILL: These pumps are --
certainly they're working harder because you've got
more flow coming out of them, yeah.
MEMBER POWERS: And so does their failure
probability go up?
MR. BURCHILL: The failure probability
could go up, and the way we evaluated that, we did not
do an explicit model of that, but we did a sensitivity
study were we raised those failure probabilities by
ten percent. Well, the initiating event frequencies
associated with them.
MEMBER POWERS: So now the flow in is up
by 20 percent. So you increased the failure
probability by ten percent.
MR. BURCHILL: Of the initiating event,
right.
MEMBER POWERS: Right, and what was the
rationale there?
MR. BURCHILL: yeah, the rationale was we
used -- actually I think we did this more analytically
than is justified -- but we used a bathtub type curve
for the burn-in, and I think we increased the failure
rate in the first year by 50 percent, and then the
second year, I think it was 25, and then we tailed it
down. And then the average came out to be the ten
percent, yeah.
CO-CHAIRMAN KRESS: In your MAP analysis,
do fission products enter in it any way other than
decay heat level?
MR. BURCHILL: There's a source term
represented, correct.
CO-CHAIRMAN KRESS: To determine its
effect on containment?
MR. BURCHILL: Right, and so that was
scaled up, but it was a simple scaling.
CO-CHAIRMAN KRESS: So it was just by the
power ratio?
MR. BURCHILL: Right, correct.
CO-CHAIRMAN KRESS: And that adds a heat
load to the containment.
MR. BURCHILL: Oh, absolutely.
CO-CHAIRMAN KRESS: And that's why.
MR. BURCHILL: Yeah, everything, you know,
starting in the reactor assembly with the boil down
rate of the coolant out to the wet well, out to the
suppression pool, everything, you know, the boil down
rates, you know, go down. The times set shortened.
CO-CHAIRMAN KRESS: What sort of fission
product release model is in MAP? We've never reviewed
that as a committee.
MR. BURCHILL: I'm not going to be able to
answer that question. We can get back to you on that.
I mean, I'm not a fission products person. So I don't
know explicitly how to even answer what the model is.
The next slide addresses the question of
system effects. As I mentioned, there were no changes
in systemic PRA success criteria. Systemic criteria
though is associated with the number of pumps, the
number of buses, the number of instrument air systems,
and so forth. All of those remain the same.
The one area that would be of interest
here would be in the over pressure protection and the
depressurization capability, but the number of safety
valves and relief valves that were required in each of
these cases, as has been previously discussed,
remained unchanged in the EPU state. So that didn't
change at all in the PRA model.
The changes that were made in the BOP and
in the AC switchyard were not represented explicitly,
but rather, again, were represented by sensitivity
studies in which we raised the failure rates to
represent a burn-in period of those new components.
The set point changes also produce a
negligible impact, and we did evaluate those
explicitly in the PRA model. Again, there were no
changes to the equipment itself, but some changes to
set points.
The next slide is really the heart of the
changes that were made. There were 45 operator
actions that were evaluated for impact. Twenty-eight
of these were chosen because of their importance in
the PRA, in other words, their importance with respect
to contributing to core damage frequency, and 17 were
chose because they were the operator actions that had
to be taken in under a half an hour. So we judged
those to be time critical and evaluated them
explicitly.
The remaining operator actions that were
not evaluated, if you sum up all of their
contributions to core damage frequency, it was less
than a half a percent. So we've captured 95 or 99.5
percent of the operator actions that contribute to
core damage.
The first one which has been discussed at
every one of these EPU, you know, applications, of
course, is the failure to initiate ADS. That normally
comes up fairly high on the list, and it requires a
shortening of time from about -- and, again, remember
this is realistic analysis now with best estimate
parameters -- a shortening of time from about 32
minutes to 28 minutes.
So there is some change in the human error
probability, and that does contribute to about a three
percent increase in the core damage frequency, a
fairly substantial increase in relative terms, but a
very small increase, of course, in terms of the
overall PRA result.
The next one is an interesting one. This
is the failure to restart a feedwater pump following
a failure to initially depressurize the reactor
pressure vessel. This is what's called a dependent
failure.
In the reference case in the submittal, we
were assuming that the plant would implement an auto
start on the motor driven auxiliary feed pump, and in
fact, that meant that this situation, this contributor
got better. So that the risk actually went down by
about one and a half percent.
We did do a sensitivity study, however, in
the submittal that showed what would be the effect if
we did not implement the auto restart feature, and
that gave the result that you see on the slide, which
is a plus, about one and a half percent.
And it turns out that at this time, that
auto restart feature is not being implemented at least
initially. So the result that you'll be seeing here
will look like a slightly different figure than you've
read several times in the staff SER, and the reason is
the swing in this contribution from this particular
human error probability.
The next two have to do with the slick
operation by the operator, injecting boron post ATWS.
As you know, there's an early and late period for
that. The early period in this plant's case is
measured by the necessity to inject boron to avoid hot
well depletion, hot well inventory depletion in the
late period. If you miss that one, you're next trying
to save the suppression pool, and so you want to
inject in time to avoid loss of unacceptable
inventory from the suppression pool.
In the analyses we do, we do a set of MAP
runs, and we do these parmetrically so we can see what
the impact is of various operator action times. This
early time, which these next two actions deal with,
decreased from in the case of a two pump injection,
from 12 to nine minutes. In the case of one pump
injection, the second one of these pair on the slide
decreased from nine to six minutes.
So, again, we modeled the effective --
CO-CHAIRMAN WALLIS: Well, that's
different from the 31 minutes we were told earlier.
MR. BURCHILL: Well, we're talking early
injection here.
CO-CHAIRMAN WALLIS: That was for
something else.
MR. BURCHILL: That's right, and this is
just for early injection. The reason I bring this one
up in detail is this is comparable to the two minutes
you discussed earlier. Two minutes is a licensing
number, but if you look realistically at what time is
actually available to the operator, in the worst of
cases only one pump operation, there's at least six
minute available.
And depending upon, you know, the exact
sequence of the event, it could be as much as nine
minutes, and I'll let the operators speak to this
question with respect to their response time if you
wish, but clearly, you know, these are response times
that are well reasonable.
MEMBER POWERS: I mean, what's intriguing
to me in this area is you calculate some probability
that the operator will do things in six minutes or
nine minutes, depending on what he has, and those
numbers, I hesitate to guess off the top of my head,
but I suspect you calculate like ten to the minus
third or something like that.
MR. BURCHILL: On that order.
MEMBER POWERS: And things like that.
And then we speak to the operators, and we
haven't on this particular incident, but I'm just
guessing that if I asked them, "How long does it take
you to do this?" he says, "Well, about 30 seconds."
"What's the probability that you'll fail
to do this?"
And he says, "Not a snowball's chance in
hell that I'll forget to do this."
MR. BURCHILL: That would be pretty close.
MEMBER POWERS: "And I'm tested on this
every two weeks, and it's beaten into my head. I
dream about it at night and things like that. So
there's just no chance."
And I come back and see, you know, you
have this discrepancy, and we talk about realistic
models in PRA, and how do we factor -- how would one
go about trying to factor in what we learn from the
operator testing and whatnot and what we calculate
when we go through Swain or something like that?
MR. BURCHILL: Well, yeah. I'm not sure
I understand the discrepancy you're describing because
what the operator will tell you is that he's tested on
this frequently in the simulator, and that he does it
really fast.
And actually, these are sometimes measured
time-wise, but the really important thing is that
they're responding to symptoms, and the timing isn't
really the deal. It's whether or not they respond to
the symptoms adequately.
Now, the way that the PRA models is this
is using human error probability models, that a big
EPRI program you may recall about ten years ago or 15
did a lot of measurements in simulators and
determined, you know, what was the error rate that was
occurring with various types of actions, either
cognitive or skill based, you know, type of actions.
And so the numbers I'm talking about are
based on that, and I don't think there's a real big
discrepancy. If I say that the error rate is on the
order of ten to the minus three, I think that
correlates pretty closely with the operator saying,
you know, "No way will I miss this, you know."
Because at some point he might miss it, and that's
what that ten to the minus three number is meant to
represent.
But I think that you --
MEMBER POWERS: I think what you're saying
is that, in fact, everyone is pretty bad at guessing
what a ten to the minus third probability is.
MR. BURCHILL: Oh, absolutely, absolutely.
MEMBER POWERS: One chance in 1,000 is
just zero to most of us.
MR. BURCHILL: Guessing it, guessing it,
right.
MEMBER POWERS: To most of us it's zero.
I mean, there's just no chance this will ever happen.
MR. BURCHILL: I will say, I mean, and
there are certainly practitioners in my peer group
that would say this is all voodoo anyway, but the fact
is that it's come a long way since Swain wrote his
handbook, and there are some very sophisticated models
today. I won't try to defend the absolute numbers,
but I think the relative numbers, and the
understanding of the cause and effect relationship
between various performance shaping factors and their
impact on the actual probabilities is pretty good. I
mean, I think we have a reasonable handle on that.
MEMBER POWERS: Yeah. I mean, I think I
agree with that. I also have to inject that I applaud
your use of sensitivity analyses on those things where
it's difficult to find data. We talked about that
earlier yesterday.
MR. BURCHILL: Well, I wouldn't be able to
defend it with you if I did, you know.
MEMBER POWERS: Well, but I mean, that's
the appropriate thing to do. You've got to go find
out is there an effect because all you're trying to do
is see if there's a magnifier.
MR. BURCHILL: Right.
MEMBER POWERS: I mean if I change things
by ten percent and the core damage frequency goes up
by 85 percent, I get nervous about that.
MR. BURCHILL: Right, right.
MEMBER POWERS: And if it goes up by six
percent, then I'm not so nervous about my number.
MR. BURCHILL: Right.
MEMBER POWERS: That's all.
CO-CHAIRMAN WALLIS: Okay. Should we
leave this slide now?
MR. BURCHILL: We can leave this slide.
Let's leave this slide.
MEMBER POWERS: Is it grating on your
nerves?
CO-CHAIRMAN WALLIS: No, I think we've
covered the important items on it.
MR. BURCHILL: Yes. The full power
internal events model in this particular case does
include explicit representation of internal flooding
in both the base case and, as you would expect, in the
EPU case. This is a very small contribution, on the
order of five percent to the total core damage
frequency.
We didn't find any new initiating events
or increases in initiating event frequencies, and so
we found that there really is a negligible effect of
internal flooding on core damage.
The Level II is the actual containment
response and release evaluation, Large in this case or
LERF is used as the measure of merit here, and large
is defined as a greater than ten percent release of
the cesium and iodine inventory in the core, and early
is defined as releases in under six hours.
The methodology used here was an explicit
calculation as opposed to the simplified calculation
I mentioned to you a couple of months ago. This has
explicit PRA model with containment event trees.
We did do a binning of the end states for
the core damage and then put those as the initiating
event frequencies into the containment event trees to
calculate the containment response.
There were very minor changes in the human
error probabilities here, largely because we're
talking very long times, you know, on the order of
hours for response in this case, and these are mainly
conditional probabilities for either repair or
recovery, particularly for earlier actions that were
not successful.
The next to the last bullet should say
that the EPU has no impact on the containment event
trees. I apologize. That was a carryover, the minor
impact. There actually were no structural changes to
the event trees here, and we found that the LERF is
directly proportional, as you would expect, to the
core damage frequency impact, which we'll show on the
next slide.
This is a summary then of the numerical
results from the full power internal events model
evaluation. These are similar to results that have
been reported in previous EPU risk studies. The pre-
EPU PRA has about a 1.4 times ten to the minus five
core damage frequency per year, and the LERF is about
1.4 E to the minus seven, two orders of magnitude
lower, and the reason for that, of course, is that
we're not using a simplified model. We are using an
explicit model of containment response.
The EPU has a very small impact on both,
about six percent. Again, I'll remind you that in the
SER if you've been reading that, it reports a
reference case of about three percent. The difference
here is the swing on that auto start and the motor
driven feed pump.
The absolute numbers are that the core
damage frequency changed by nine E minus seven per
year, and this puts it below the RegGuide 1.174
criterion of ten to the minus six for being classified
as a very small risk impact.
Likewise the LERF was eight E minus nine
per year, which is well below the RegGuide 1.174
criteria of ten to the minus seventh, to be classified
as a very small risk impact.
The other important thing is the
composition of initiating events contributing to core
damage did not change. We did answer an RAI from the
staff that displayed the explicit spectrum of the core
damage due to individual initiating events, and there
was only very slight change, some up, some down. But
essentially the pie chart looked identical.
CO-CHAIRMAN WALLIS: Could you explain,
Bill, on this SER, there's something about sensitivity
number four and talking there about 23 percent
increase in base CBF.
MR. BURCHILL: Sensitivity number four was
the combination of sensitivities one through three,
where we increased those initiating event frequencies
by 20 percent, and we increased the operator error
rates by 20 percent, which was a very high increase,
and to be frank with you, I can't remember what the
third one was.
But we combined all of them. In all total
they gave a 23 percent impact, which I think, you
know, referring back to Dr. Powers' remark, if this
had been an 85 or 90 percent impact, we would have
probably paid very close attention. At 23, percent
giving the bounding nature of that sensitivity, it was
felt not to be significant.
We did look at uncertainties. This is the
following slide. We examined this through the
traditional use of risk importance measures. We did
do explicit sensitivity studies, the ones I just
mentioned, and we did compare this PRA's results to
those reported in NUREG 1150.
We didn't find any sources of uncertainty
beyond those already identified by NUREG 1150, and if
you took the NUREG 1150 outer bound on uncertainty,
which is on the order of a factor of five or six
overall, the results here would still be in the small
risk range in the RegGuide 1.174 risk MAP as compared
to the very small risk.
CO-CHAIRMAN KRESS: Excuse me. How did
you examine uncertainties used in risk importance
measures again?
MR. BURCHILL: In risk importance
measures, what you do is you look at the change from
the pre-EPU state to the post EPU state and see
what --
CO-CHAIRMAN KRESS: Change in what is
important?
MR. BURCHILL: Yes, what is important,
right. And then you can look at whether or not if you
find an important piece of equipment. Is there
anything that's changed about the dependency structure
of supporting that piece of equipment?
Basically you're looking at whether or not
there's anything about the PRA in going from the pre-
EPU state to the post EPU state that would actually
change the impact on that importance for those pieces
of equipment.
CO-CHAIRMAN KRESS: Yeah, I was having
trouble relating that to uncertainty, but I can see
it's a nice thing to have done.
MR. BURCHILL: Okay. The next three areas
are in the qualitative evaluation area. We started,
of course, with the IPEEE. It concluded that none of
the external events had risk significance.
We explicitly evaluated the fire risk. We
do have a fire PRA for this plant. It gives a base
CDF for the plant of about 3.3 E minus six per year,
which is about an order of magnitude lower than the
internal events frequency.
We looked at the dominant scenarios.
These are primarily loss of inventory control or loss
of decay heat removal, and we found no change in the
risk profile produced by the EPU; a minor impact on
decay heat removal scenarios because of the human
error probabilities, but again, most of these are out
late in time.
We didn't find any new initiating events,
nor any increases in fire ignition frequencies. So we
judged that the EPU has a negligible impact.
In the seismic area, a point to make here
is that the IPEEE is based on seismic margin analysis.
That analysis did not reveal any outliers or
vulnerabilities in this plant. We looked at the
seismic margin report, found no reason to believe that
the seismic qualifications of the SSCs would be
changed, and as was previously discussed, there's
really negligible impact to the increase in stored
energy on blow-down loads so --
CO-CHAIRMAN WALLIS: Do you have an
increase in power? We have presumably great
occurrence in part of the electrical system. Is this,
in fact, a fire initiation risk, circuit breakers and
things?
MR. BURCHILL: The only place that I'm
aware that that would be an issue, I think we're
making changes to the equipment. These are mostly out
in breakers in the switch yard, if I remember, and
those were the ones that we covered with the
sensitivity study.
Is that sufficient?
Okay. The EPU has the word "negligible"
that's also appropriate here for the seismic risk. We
could not find any impact here.
In the shutdown risk area, we do not have
a shutdown PRA, but, again, we can evaluate what would
be the impact by principally being guided by the
operator actions required for the dominant sequences
that contribute to risk in the shutdown states. The
shorter times for boiling, the decrease in inventory,
decrease in the inventory loss time does shorten the
time for operator response, and it also, of course,
delays the time when alternative decay heat removal
systems, such as the spent fuel cooling system or the
reactor water clean-up system could be used.
But the times that we're talking about are
measured in hours or even days. So the impact here is
essentially negligible, and is managed through the use
of a configuration risk management program.
So, in summary, we did evaluate the impact
using standard PRA methods, both quantitative and
qualitative. The risk impact is a very small percent
of the current plant risk, on the order of six
percent.
That risk was compared for both CDF and
LERF to the guidelines provided for a permanent plant
change in RegGuide 1.174 and found to be in the very
small risk region, and the risk impacts from external
events and shutdown conditions were found through
examination of the dominant sequences in judgment to
be negligible.
Again, as I will say, as I said a couple
of months ago, the staff review in this area was, in
our opinion, quite thorough, especially that this is
not a risk informed submittal. We had quite a lot of
dialogue with the staff, although not quite as much as
we did on the previous submittal, but they did ask, I
think, quite relevant questions relative to our
submittal.
So if there are no further questions, I
will turn this back to Larry Westbrook, who will
discuss the project implementation.
MR. WESTBROOK: Thank you, Bill.
Good morning, again. I am Larry Westbrook
with AmerGen, and as Bill mentioned, I will be
discussing the EPU project implementation regarding
operator training and EPU testing.
EPU training consists of classroom
presentations and simulator scenarios. Classroom
material was presented in licensed operator
requalification training covering the following
topics:
Technical specifications and updated
safety analysis report changes;
Plant limits and operating conditions
changes, such as core thermal power and generator
limits;
Design changes associated with EPU;
The extended power up rate power to flow
MAP operating procedure revisions;
And Exelon and industry power up rate
experience.
Simulator training is being provided to
operating crews --
CO-CHAIRMAN WALLIS: Could I ask you here
now? Classroom training is presumably a two-way
thing. I mean, you tell them things and they ask
questions.
MR. WESTBROOK: Yes.
CO-CHAIRMAN WALLIS: Were there any
particular questions they raised that you remember?
MR. WESTBROOK: I'm trying to think of
particular questions. The questions generally that
were asked --
CO-CHAIRMAN WALLIS: Did they have any
particular concerns?
MR. SCOTT: Larry, if I could help you out
here, again, I'm Kent Scott from AmerGen, senior
reactor operator.
One of the concerns that I heard with the
guys in training is focused around how we're going to
operate the generator now. You know, we're going to
be limited by the secondary balance of plant side. So
there was a lot of concerns with respect to, okay, how
do the limits on our generator change, explaining the
increase in hydrogen pressure from 60 to 75 pounds.
I think that was the most focus that I had saw, and I
think that was supported a lot by showing the
operators in the simulator the difference in response.
Here's the plant at a higher feed flow.
Here's the plant at a higher steam flow. Here's a
transient. It looks the same.
And they kind of took that away and
focused in on, okay, what's going to be different. So
the balance of plant inside the generator was the
thing that I saw most from them.
MR. WESTBROOK: Okay. Simulator training
is being provided to the operating crews covering the
following: EPU full power conditions, and as Kent
stated, that involved the generator limits and the
plant response.
Normal operations scenarios, such as
raising reactor power along constant rod line and
dynamic scenarios that were selected to highlight both
similarities and differences in plant response at EPU
versus current power levels; these scenarios included
a feedwater pump trip, a reactor recirculation pump
trip, and ATWS.
Operator requalification training is being
covered in two cycles of classroom and simulator
training. The first cycle of training has been
completed, and the second cycle is currently in
progress.
CO-CHAIRMAN WALLIS: I would guess that
there's no real significant change, and the way that
they diagnose events is very much the same as before.
MR. WESTBROOK: That's correct.
CO-CHAIRMAN WALLIS: In terms of the same
-- it's just slight qualitative differences in the
response.
MR. WESTBROOK: That is correct. A slight
difference in response. The way we operate the plant
is essentially the same. If they're -- the most
significant difference, if you will, is that to
maximize Clinton's output will be BOP limited. They
will be looking at and maximizing the output by
focusing on the generator limits, and we won't be as
bounded by our license power level because we'll have
a margin to the license power level. Of course, we
won't be monitoring that.
CO-CHAIRMAN WALLIS: It's like driving an
automobile with a slightly different gear box or
something. I mean, it's not as if many things have
changed. It's just you're going to have a slightly
different response.
MR. WESTBROOK: That's correct.
Also, just in time, simulator training
specific to EPU testing will be conducted prior to
power ascension.
CO-CHAIRMAN WALLIS: Just in time. The
day before?
MR. WESTBROOK: The day before, a couple
of days before so that this training, as I stated, is
specific to the power ascension. It may be specific
to certain testing done during the power ascension,
and it may be specific to the crew that will be
performing the testing.
MR. SCOTT: And again, this is Kent Scott
from AmerGen.
That's a typical process that we do prior
to any start-up, is we have the crew that's going to
start up the reactor take them over there the day
before.
MR. WESTBROOK: Next I would like to
provide an overview of the start-up test program.
The start-up test program employs a
careful and deliberate approach to upgrade power
levels. The program has incorporated previous
successful Exelon power up rate experience.
The La Salle EPU test director is the
Clinton Power Station project manager. Also the
Dresden Quad Cities test procedure developer is also
developing the Clinton test procedure.
Beginning at 90 percent original licensed
thermal power, steady state data collection and
testing will be conducted. Once 100 percent original
licensed thermal power is reached, reactor power will
be raised in two percent increments along a constant
rod line to maximum achievable power level.
Beginning at approximately 70 percent of
original licensed thermal power dynamic testing will
be conducted. These dynamic tests include pressure
control system stability tests, feedwater level
control system stability tests, and turbine valve
surveillances.
This slide shows the testing to be
performed at each power level.
In conclusion, the operator training to
prepare for EPU is extensive and utilizes both --
CO-CHAIRMAN WALLIS: I don't understand
what you mean by testing of core performance, let's
say. Are you just recording how it's performing?
You're not performing any sort of transient tests.
MR. WESTBROOK: That's correct. Core
performance is no transience test. It's monitoring
with our 3D Monocore system that all of our thermal
limits and our inspect and that we are operating on
the power --
CO-CHAIRMAN WALLIS: This isn't so much
testing. It's just monitoring that things are
behaving as expected.
MR. WESTBROOK: Part of the testing is
data collection and analysis and a static, and then
some of the tests, such as pressure control system
stability, feedwater level control system stability,
those are the dynamic tests.
CO-CHAIRMAN WALLIS: Now, presumably you
have guidelines here. I mean, you've got something
like piping vibration data. These folks are recording
piping vibration data. There must be some procedures
written down about what they do if the vibration steps
out of line beyond some magnitude or something. All
of these must be things to look for --
MR. WESTBROOK: That's correct.
CO-CHAIRMAN WALLIS: -- or guidelines
about what to do.
MR. WESTBROOK: The test procedures being
developed by or from the General Electric task report
which selected the tests that need to be performed --
CO-CHAIRMAN WALLIS: All of those checked
things, you've got to check that the vibration has not
increased beyond something.
MR. WESTBROOK: Right.
CO-CHAIRMAN WALLIS: And so on and so
forth.
MR. WESTBROOK: This test has a Level 2
and a Level 1 criteria. Certainly exceeding the Level
2 criteria requires evaluation, and exceeding Level 1
criteria would certainly be more extensive actions
required. Those criteria are being built into the
test procedure.
MEMBER SIEBER: You've run these tests
before. So you have a database --
MR. WESTBROOK: Yes.
MEMBER SIEBER: -- of what the vibration
levels are. So you're really looking for deviations
from that as you go beyond the old 100 percent power
level; is that correct?
MR. WESTBROOK: These tests are similar to
the original start-up tests.
MEMBER SIEBER: Right.
MR. WESTBROOK: As far as the criteria,
we're taking what's established in the task report and
insuring in the course that they are consistent with
the original start-up test.
MEMBER SIEBER: Right.
MR. WESTBROOK: In conclusion, the
operator training to prepare EPU is extensive and
utilizes both classroom and simulator environments.
The testing plan is incremental and comprehensive and
is a careful and deliberate approach to up-rated power
levels.
Therefore, project implementation will
insure that EPU is implemented as designed and plant
response is as expected.
CO-CHAIRMAN WALLIS: When you're doing
this power up rating by two percent, do you have extra
people in the control room?
MR. WESTBROOK: The actual people in the
control room will be very minimal. Clinton Power
Station's control room is physically very small. A
lot of the data collection will be taken off the
process computer and analyzed outside the control
room.
CO-CHAIRMAN WALLIS: So there are other
people who are engineers sitting somewhere else
monitoring what's going on?
MR. WESTBROOK: That's correct.
MR. WILLIAMS: Are there any more
questions for Larry?
(No response.)
MR. WILLIAMS: Thank you, Larry.
Prior to concluding, we are prepared to
answer the follow-up questions on the large transient
testing.
Tim.
MR. SPENCER: Good morning. Dale Spencer,
Exelon Nuclear.
We would like to address the earlier
question regarding analysis of the plant response to
transience. We have provided the staff with the
results of these analyses, but in response to today's
question, we have a couple of things we pulled
together.
A couple of points here. General Electric
has asked that the session be closed. Again, this is
some proprietary information.
The other thing, if you'd please excuse
us, we pulled it together at the last minute. We have
to change computers and get a couple of flemsies up
there so that we have the right information, but
pardon us for just a minute.
And we would ask the Chairman that we
treat this as a closed session also.
CO-CHAIRMAN WALLIS: Okay.
(Whereupon, from 9:57 a.m. to 10:05 a.m.,
the Subcommittee meeting was recessed for closed
session.)
MR. BOEHNERT: Okay. Let's go back to
open session.
CO-CHAIRMAN WALLIS: We're ready to
conclude.
MR. SPENCER: Yes, Dale Spencer, Exelon
Nuclear.
At this time I believe the ACRS had one
pending question for Bill Burchill, and he's prepared.
MR. BURCHILL: This is Bill Burchill,
Exelon.
I wanted to answer Dr. Kress' question
about the radiological release model which is in the
MAP program. I knew it would have some catchy name,
and I couldn't remember what it was, but it's based on
IDCOR data. It was actually a model developed by
EPRI. It's called the IDCOR EPRI steam oxidation
model, and it does detail a phenominological
evaluation of both the release and then the various
partitioning mechanisms as you go out from the fuel
through the coolant, through the steel boundaries of,
you know, the reactor assembly, and on out through the
containment.
So it is a detailed model for radiological
release. Does that answer your question?
Right. It's the old Cubachasie (phonetic)
model, right.
MR. BOEHNERT: Dale, what version of MAP
are you using now?
MR. BOLGER: This was all done with MAP
3B. We are now converting fleet-wide to MAP 4, which
we'll have in place this year, but all of these
results were on 3B, and in this area, there is some
enhancements in MAP 4, but basically the fundamentals
are the same.
MR. WILLIAMS: If there are no more
questions, we would like to introduce Keith Jury to
conclude.
MR. JURY: Thank you.
Good morning. My name is Keith Jury, and
I'm the Director of Licensing for the Midwest Region.
We had to have one licensing guy to talk
here. So I guess I'm the token.
Dr. Wallis and the other committee
members, on behalf of Exelon and AmerGen, we'd like to
thank you for the opportunity to come here and give
this presentation. As always, we appreciate the
insights that we get from you all individually and
collectively.
It's clear at least to me that the team,
including G.E. and Sargent Lundy have done a very
thorough job on this project.
I would like t apologize for the fact that
we weren't able to answer all of your questions
immediately and had to huddle, but I think that given
the fact that the team has taken not only the lessons
learned from our fleet, but the industry as a whole
and have incorporated into this project, that it
speaks to the thoroughness of the effort.
I'd like to bring up the fact that, you
know, I think as time has evolved here and as we've
gone through these up rates, for this up rate, we use
the most up to date and accepted analytical methods
and believe that we've demonstrated that the up rate
will have a minimal impact on the plant from not only
the plant perspective, but the system integrity, and
as Bill so eloquently pointed out, from a risk
perspective as well.
As such, we've concluded that the up rate
in the plant operation at the up rated conditions is
acceptable and clearly safe at these conditions.
And with that, I'd like to thank you again
for your time, and that completes our portion of the
presentation.
Thank you.
CO-CHAIRMAN WALLIS: Well, thank you.
I believe we're ready to have a break, and
so we will break, and we will come back here at 20
minutes past ten.
(Whereupon, the foregoing matter went off
the record at 10:10 a.m. and went back on
the record at 10:23 a.m.)
CO-CHAIRMAN WALLIS: We'll come back in
session, and now we're going to hear from the NRC
staff, and we're looking forward to that.
MR. MARSH: Yes, I have a few opening
comments. Good morning, Mr. Chairman. This is Ted
Marsh, and I'm still the Deputy Director of Licensing.
And in the interest --
MEMBER POWERS: We're working on that.
MR. MARSH: Yeah, I know you are. In what
sense I don't know.
(Laughter.)
MEMBER POWERS: EDO, EDO.
MR. MARSH: EDO. Thank you, thank you.
I appreciate that.
I did want to point out that the agenda
has us talking a little bit this morning about our
efforts to improve safety evaluation quality, and we
did talk about that yesterday, and we also have the
letter to you. So I'm not going to go through that
anymore, unless you'd like.
Let me just begin by introducing Jon
Hopkins, who is the project manager for Clinton, who
will be going through the presentation format and
introducing the speakers.
Jon.
MR. HOPKINS: Okay. Thank you.
As Ted said, I'm Jon Hopkins, NRC project
manager for Clinton.
And next slide.
I'll start with an overview. Clinton's
BWR-6 Mark 3, it's the first one to apply for extended
power up rate. This will be a 20 percent power up
rate. Even after the up rate, which will be to 3473
megawatts thermal, Grand Gulf will have greater power
level and Perry will be slightly greater power level.
It's a constant reactor dome operating pressure, and
to achieve the 20 percent up rate, there are many
balance of plant modifications that need to be done,
and in large part because of that, this has two-part
implementation.
The refueling outage that begins is it
scheduled to start the end of March, early April with
start-up in May, and they expect to achieve roughly
seven percent when they start up this time.
They will also have roughly two-thirds
G.E. 14 fuel when they start up this time, and then
the next refueling outage is scheduled for early 2004,
where they'll make the rest of the modifications and
achieve the rest of the power.
This application mostly follows the
licensing topical reports, the LTR-1 and 2. It was
submitted in June of last year, and as I stated, the
refueling outage starts the end of March, early April.
So we'll be going to full committee in March, and we
will be looking for a letter from the full committee.
There are some exceptions to ELTR-1, too.
It states four of them. The first one is large
transient testing. That exception is similar to
Dresden and Quad Cities, and we found it acceptable
for them not to do large transient testing.
The other three exceptions are in the
reactor systems review area, and they'll be talked
about by the next presenter.
Mainly I'll mention that the other three
exceptions in the reactor system area dealing with
stability, transience and ECCS performance. It's hard
to discuss those without getting into proprietary
information. Our presentation is, you know,
scheduled for an open presentation and everything.
Questions, we may have to go into closed session based
on --
CO-CHAIRMAN WALLIS: I think if things
become proprietary, what you should do is stall them,
and then at the end of the session we'll have
discussion of the proprietary items if we need them.
MR. HOPKINS: Yes.
CO-CHAIRMAN WALLIS: Will that be okay?
MR. HOPKINS: Yes, I'm fine with that. I
just wanted to mention it.
This is a non-risk informed submittal, but
as you just had the presentation, there were
substantial risk evaluations performed. I won't
mention anything more about that.
Experience. This plant is owned by
AmerGen now, which is 50 percent Exelon, and so they
have plant up rate experience specifically with the
Dresden-Quad extended power up rate.
The staff, of course, has that experience
plus Duane Arnold and Hatch and Monticello in boilers.
CO-CHAIRMAN WALLIS: Has Dresden and Quad
Cities actually implemented an up rate?
MR. HOPKINS: I know -- I'm sorry --
Dresden 2 has.
CO-CHAIRMAN WALLIS: Dresden 2 has
influence at the up rate.
MR. HOPKINS: Yes. To continue from our
review to finalize my review of the application, we
have one license condition at this time. It deals
with the feedwater nozzle. That, again was discussed
by the licensee yesterday.
They are doing additional analyses of the
cumulative usage factor in there, and they have to
submit them to the staff. We feel that this is not a
significant item, that we can make it a license
condition and resolve the issue easily during the next
fueling cycle. We don't anticipate that this should
stop them from starting up.
Also, one other item. The staff is going
to perform or, in fact, is performing confirmatory
containment analyses. We just received data from
General Electric to allow us to go do that. The
confirmatory analysis is similar to that performed at
Duane Arnold, and I expect we'll have that done fairly
soon.
MR. BOEHNERT: Who is doing that analysis,
Jon.
MR. HOPKINS: Well, our plant systems
branch.
MR. BOEHNERT: Oh, it's in house?
MR. HOPKINS: Yes.
MR. BOEHNERT: Okay.
MR. HOPKINS: We have no other items other
than those two I mentioned, the license condition and
the confirmatory analysis.
With that I'd like to introduce our first
presenter, who is George Thomas, and he will discuss
reactor systems review.
MR. THOMAS: Yeah, my name is George
Thomas. I'm from the Reactor Systems Branch.
Our review scope, mostly in the standard
review plan, Chapter 4, 5, 6. There is a small
portion of nine, and most of 15. And our review is
mostly about the fuel and the reactor systems, and we
review high pressure core spray, low pressure core
spray, RCA system, CID system, and the standby control
system.
Also we review reactor ore pressure,
transients, LOCA, dose, and stability. So this is our
scope.
As Jon told, you know, basically for all
the ELTR 492, for most of the EPU evaluation, but the
tech. deviations under -- they're in the areas of
transient analysis, LOCA analysis, and the stability
analysis.
Typically they sum it (phonetic) during
the obligation, equilibrium, quote, analysis for the
transients and the LOCA, but this obligation, they
already are very limited analysis for the transients,
LOCA.
MR. NIR: This is G.E.? I think you're
getting into proprietary information which has to do
with scope, and I would like to strike this from the
record if possible.
And, again, we can get into this
discussion in the closed session if necessary.
MR. BOEHNERT: I know no way to strike the
record.
CO-CHAIRMAN WALLIS: I don't understand
what. He hasn't said anything yet which is
proprietary, has he?
MR. NIR: It's G.E.'s position that he's
discussing now some proprietary information, the basis
for the analysis and the scope, which we consider
proprietary.
MR. BOEHNERT: I'm sorry.
CO-CHAIRMAN WALLIS: Well, he hasn't
really discussed anything yet, has he?
MR. BOEHNERT: Yeah, that's --
CO-CHAIRMAN WALLIS: Can we move on? Are
you clear on what he's objecting to?
MR. THOMAS: I'm not going into details.
I'm only just telling, you know, they did only a very
limited analysis. It was not done typically like
before. That's all I'm telling, you know. I'm not
telling the details now.
MR. BOEHNERT: Let me make a comment here
for the record.
There's been some controversy over the
scope of what is G.E. holding proprietary here, and we
need to get this straightened out between the staff
and ourselves because this is becoming a problem.
I don't want to say any more right now,
but I think this has to be straightened out because it
seems that it's become quite broad, and it's impacting
how we're doing business here.
CO-CHAIRMAN WALLIS: The fact that there
are exceptions for transient analysis, LOCA analysis,
and stability analysis is actually in the SER, isn't
it?
MR. THOMAS: Right, right.
CO-CHAIRMAN WALLIS: Which is a public
document.
MR. THOMAS: IT is in the SER.
MR. HOPKINS: Excuse me. The SER is the
draft safety evaluation, which we have sent to the
licensee for their comments on proprietary aspects.
So, in fact, it is not a public document yet.
Now, I will state that this slide was
developed, and we feel this slide is entirely
nonproprietary.
MR. NIR: We agree.
MR. HOPKINS: So we feel that you can say
that those exceptions -- that that's public.
MR. NIR: We agree.
CO-CHAIRMAN WALLIS: Okay.
MR. BOEHNERT: They agree. Okay. Then we
don't have a problem.
MR. HOPKINS: So far, right.
CO-CHAIRMAN WALLIS: Well, you're going to
be watched like a hawk by G.E. now. So --
(Laughter.)
MR. THOMAS: Okay. As part of our review,
we went to D obviously some time in September of last
year, and four of us went there for a week, and we
reviewed their design record files and the past report
and all of the recommendation.
The focus of our audit was all these
deviations. So most of the time we sent there was on
all these three deviations. So the first was about
the transient analysis. With the summit (phonetic) of
the EPU obligation, they only analyze only two
transients. So when the (unintelligible) was done
with the equilibrium core.
So all other transients will be analyzed
as part of the reload analysis, and then --
MR. NIR: This is Israel Nir, G.E.
This is proprietary information.
CO-CHAIRMAN WALLIS: It sounds rather
extraordinary that it's proprietary.
MR. NIR: Well, it is proprietary just to
-- this has commercial implication just General
Electric. The fact that we do or do not do certain
analysis is based on significant experience, power
operator experience, and this is information that is
commercially sensitive, and we would like not to share
it publicly, and we would like --
CO-CHAIRMAN WALLIS: How much of your
presentation is going to run into this problem? Maybe
we should just close the meeting and get on with it.
MR. NIR: We have no proprietary issues
with the slides as they presented. We -- the words
that are written here are not proprietary. We have an
issue with some of the verbal information.
CO-CHAIRMAN WALLIS: So he has said
something which is slightly different from the slide,
which is a problem?
MR. BOEHNERT: Jon, do you want to
comment?
MR. HOPKINS: It might be easier if we
close this session.
MR. BOEHNERT: Okay. That's fine. We can
close the session. We'll close the session. We'll do
that, but I want to put on the record that we're going
to have to decide, you know, where to draw the line
here because you guys in the end make the decision on
what is proprietary. We just go along with it
basically, but I can close the session on the proviso
that if you later decide this is not proprietary,
we'll release the information.
CO-CHAIRMAN WALLIS: Well, let's close the
session and see if anybody leaves.
(Laughter.)
MR. BOEHNERT: So again, anyone who should
not be here to hear G.E. proprietary information,
please leave the room. We're going to close the
session and go into closed session.
(Whereupon, from 10:37 a.m. until **, the
Subcommittee meeting was recessed for
closed session.)
MR. BOEHNERT: And if they jump up again,
I guess we'll see what happens.
CO-CHAIRMAN WALLIS: Yes, if you jump up
again, you'll have to have good reason to do so.
Please go ahead.
MR. LOBEL: Good morning. My name is
Richard Lobel. I'm with the Plant Systems Branch in
the NRR, and I'd like to talk about the balance of
plant review that we did in containment.
This review of plant systems covers a
broad area. The standard review plan, there are some
sections in Chapter 3 that we look at having to do
with high and medium energy line breaks. We look at
Section 6.2 on containment and 6.5 on combustible
gases, and a large part of the SRP Chapter 9 that
deals with heating and ventilation systems, the spent
fuel pool and other miscellaneous topics.
The next two slides are just some examples
of some of the topics that a balance of plant reviewer
would look at, or reviewers, and that were looked at
for the Clinton power up rate. I won't go through
reading them all.
The containment systems review followed
the power up rate topical report. I listed the codes
that were used by G.E. in the calculations for the
containment, and they're standard G.E. codes that have
been used on other containment calculations that are
used frequently. M3CPT and Super Hex are used pretty
much on every containment, G.E. containment
calculation, and LAMB code was used for this review.
It was also used for the Dresden-Quad Cities power up
rate. The LAMB code provides the mass and energy into
the containment. M3CPT is for the short-term
response, and Super Hex calculates the long-term
pressure and temperature in the containment.
CO-CHAIRMAN WALLIS: This is a more
realistic flow-down compared with one which was less
realistic?
MR. LOBEL: It's conservative. The
calculations are conservative.
CO-CHAIRMAN WALLIS: Which ones, the ones
which are less realistic? I'm not quite sure what you
mean by a more realistic flow-down.
MR. LOBEL: Oh, more realistic.
CO-CHAIRMAN WALLIS: More realistic than
what?
MR. LOBEL: Oh, oh, I'm sorry. LAMB is
less conservative, more realistic than M3CPT.
MR. PAPPONE: This is Dan Pappone of
General Electric.
The M3CPT break flow model that we're
using is a simple hand calculation with critical flux,
and then the LAMB code is actually a reactor model.
So that's where we're getting more realistic break
flow rates.
CO-CHAIRMAN WALLIS: It's a hand
calculation because you're using somebody's simple
method?
MR. PAPPONE: That's right. We're looking
at pressure and enthalpy for the up stream source for
the break, the break area and critical flow look up in
a table.
CO-CHAIRMAN WALLIS: Whose model are you
using?
MR. PAPPONE: That's in the M3CPT
approach.
CO-CHAIRMAN WALLIS: Whose model is it?
MR. PAPPONE: It's G.E.'s. Oh, both of
them were actually using the Moody-Slip.
CO-CHAIRMAN WALLIS: Using Moody.
MR. PAPPONE: Right.
CO-CHAIRMAN WALLIS: Which may give you a
reasonable answer, but it's not realistic.
MR. PAPPONE: Right. It's on the
conservative side and over predicts the break flow.
MR. LOBEL: I think the more realistic
refers to the LAMB model.
MR. PAPPONE: That's right.
MR. LOBEL: Correct me if I'm wrong. Is
a more detailed model of the vessel and the core.
MR. PAPPONE: That's right.
MR. LOBEL: Go on?
CO-CHAIRMAN WALLIS: Well, they're all
approved codes. So --
MR. LOBEL: Well, they're approved up to
a point that they're used all the time. We haven't
written an SER on the LAMB code for containment use.
It is an Appendix K code and has been approved by the
staff for local calculations.
Super Hex hasn't been approved. There was
a 1993 letter to G.E. that talks about the staff's
intention with Super Hex, and what it says is that we
don't intend to do a detailed review of Super Hex
because it has the capability to vary the input, and
we didn't see the worth of a detailed review of the
code.
So what we ask is that a licensee that
uses Super Hex compare a Super Hex calculation with an
existing FSAR calculation just to verify that it can
reproduce containment response that it's been
previously approved.
And it also brings in the next bullet of
the confirmatory analysis. One of the reasons for
doing the confirmatory analysis is that these codes,
in general, haven't been approved by the staff, and so
that gives us -- that provides an impetus to us try to
do an independent calculation with a staff code that
we understand and compare it
CO-CHAIRMAN WALLIS: It says "is
performed."
MR. LOBEL: Pardon?
CO-CHAIRMAN WALLIS: It says "is
performed," and that means right now they're doing
this?
MR. LOBEL: It's underway, yeah. We've
just received, I understand, the last of the G.E.
data. The calculation is being done in our branch,
not by a contractor this time, and some preliminary
work has already been done in setting up the model.
CO-CHAIRMAN WALLIS: hat happens when you
get the results? Is that after you've issued the SER?
MR. LOBEL: It may be because we're
looking at this as a confirmatory calculation.
CO-CHAIRMAN WALLIS: That's strange.
MR. LOBEL: Well --
CO-CHAIRMAN WALLIS: It's a kind of "new
speak." I know. I understand, but it's very strange
that you do the test.
What are you going to learn from it if
you've already given out the SER?
MR. LOBEL: If we find a significant
problem, we can always deal with that problem.
CO-CHAIRMAN WALLIS: It would be very good
if you could do these things in a slightly more
anticipatory mode.
MR. LOBEL: I agree.
CO-CHAIRMAN WALLIS: So that the results
were available before you wrote an SER.
MR. BOEHNERT: So this is not an open
issue, Rich?
MR. LOBEL: No, it's not considered an
open issue.
MR. MARSH: Mr. Chairman, let me clarify.
Rich, maybe you could help.
Do we anticipate having the results by the
full committee meeting?
MR. LOBEL: We're going to try very hard
to have the results by the full committee meeting, but
since it's a calculation and we may need to go back to
General Electric or the licensee for some more
information, and we're not sure how it's going to turn
out, I'd rather not make that promise, but --
CO-CHAIRMAN WALLIS: And also if you rush
it --
MR. LOBEL: -- it should be done by then.
CO-CHAIRMAN WALLIS: -- the chance of
making some error is increased.
MR. LOBEL: Right.
CO-CHAIRMAN WALLIS: Are you using your
own codes or are you using the G.E. codes?
MR. LOBEL: We're using Contain 2R code.
CO-CHAIRMAN WALLIS: You're using a
different code altogether.
MR. LOBEL: Yeah. We're using Contain 2,
which is the latest version of the NRC code, and we're
trying to follow the guidance that we've received from
research from a previous user need.
We made a request to research to give us
guidance on how to use the contain code, which is
essentially trying to reproduce the physics that
really occur, how to use that in a design basis
analysis, and they've prepared reports for us for the
different types of containments, and this is going to
be the first time we've tried to use the Mark III
guidance that they've given us.
We did follow the guidance we got from
research for the Duane Arnold calculation, and we
compared very well with the G.E. results.
CO-CHAIRMAN WALLIS: So this has been done
before. They've run Contain for Duane Arnold? I
forget now what happened.
MR. LOBEL: We had very good agreement.
We were within a degree and less than a psi, if I
remember right, with Duane Arnold.
CO-CHAIRMAN WALLIS: It was a different
shaped containment.
MR. LOBEL: Right, and that's why we're
repeating the process.
CO-CHAIRMAN WALLIS: Okay.
MR. LOBEL: And another reason this isn't
an open issue is because we've asked questions about
the containment response and why the results that G.E.
got or the licensee got are what they are, and we
think we understand what's going on.
So we're doing the calculation as a
confirmatory, not as an open item.
CO-CHAIRMAN WALLIS: It's not an open
item.
MR. LOBEL: No.
CO-CHAIRMAN WALLIS: I thought you had
said it was.
MR. MARSH: No, it's not.
MR. SCHUABI: Dr. Wallis, just to clarify,
this is Mohammad Schuabi.
We have reviewed the licensee submittal in
this area, and we find it acceptable. So we don't
have a need for this confirmatory calculation, but we
are doing it anyway. That's why it's being presented
this way.
It's a little awkward. It's not the way
we normally do --
CO-CHAIRMAN WALLIS: I would not encourage
you to kill it just because you don't think you have
a need. Please complete it.
MEMBER POWERS: When you do these
calculations of the pressures and dry wells and wet
wells and places like that, do you consider the dry
well to wet well bypass?
MR. LOBEL: Ed, do you want to answer?
This is Ed Throm for the staff.
MR. THROM: My name is Ed Throm. I'm with
the Plant Systems Branch. I'm kind of the resource
that became available to Plant Systems to do this type
of work. My background is more in the LOCA area, but
I have experience running codes.
As Rich pointed out, we had researchers
set up models for evaluating containment response, and
we plan to use that. What we've basically done is
looked at the Grand Gulf III model, and we are upgrade
-- as a matter of fact, I've upgraded that model to
Clinton, and at this point what I've done is
benchmarked the current USAR calculation for the
recirculation line break, and I missed it by a PSI and
one degree.
So we have --
MEMBER POWERS: Boy, you must have been
embarrassed, huh?
MR. THROM: Absolutely.
(Laughter.)
MR. THROM: And I'm trying to figure out
what the problem is. So it builds up the confidence
in the fact that the analysis techniques are comparing
well, and this is why I think it's really
confirmatory.
The other question, of course, is the
long-term suppression pool, and that presents more of
a challenge to do a calculation in a specific time
frame. That's why we don't want to commit to having
the complete set of confirmatory calculations done
within a certain time frame.
Anyone who runs these codes knows that you
run into unanticipated numerical problems somewhere
along the line. You've got to find the inertia factor
or what's potentially leading to the problem and
resolve it.
But the intent is basically to use the
modeling criteria that research established and
develop guidelines for doing design basis
calculations, and to basically mimic or reproduce the
efforts that we've done on Duane Arnold for Clinton
just to make sure that we're not missing anything in
the Mark III performance.
MEMBER POWERS: My recollection on the
Grand Gulf model is a substantial bypass flow from the
dry well into containment. It does not go through the
wet well.
MR. THROM: That's right.
MEMBER POWERS: And you take that into
account in the calculation, or do you shut that down?
MR. THROM: It's in the model as it is
right now, as I understand it.
MEMBER POWERS: Could you remind me of the
magnitude or do you remember?
MR. THROM: Do you mean area-wise?
MEMBER POWERS: Maybe area.
MR. THROM: Well, I think the area is on
the order of 40 -- in Clinton, I think it's about 42,
43 square feet.
MEMBER POWERS: For the bypass --
MR. THROM: I guess I'm --
MEMBER POWERS: As I recall, it was like
1,000 CFM bypass flow.
CO-CHAIRMAN WALLIS: Forty-two square
feet.
MR. THROM: Well, I can't answer your
question today. As I said, you know, we're using the
models. I will look into that in terms of the way
things are being considered, but as you pointed out,
you know, as a resource, this work is just starting.
As I said, I've been able to benchmark
against the USAR calculation for the recirc. line
break. We did receive the hard part of the data from
G.E. and the applicant on Monday, which is all of the
mass and energies that are necessary to perform this
calculation.
I think I have all the data I need, but
there might be a couple of geometrical inputs, some
flow areas that I might have to go back to the
licensee to get. Sometimes what you think is in the
USAR isn't quite there.
There's no challenging information that's
needed.
MR. LOBEL: The bypass flow area, the K
over squared A number that's in the tech. specs. is
determined from small break calculations and not the
large break because a smaller break turns out to be
limiting for that because you get flow for a longer
time.
We can get back to you on the question.
MEMBER POWERS: Yeah, because the reason
I asked is because it's not clear what's conservative,
to include it or not include it. I mean, if you were
interested in special temperature, then you want to
cut the bypass flow out. If you're interested in
containment pressure, maybe you want to turn it on.
I don't know.
MR. LOBEL: Well, that's another reason
why we tried to do our own calculations, because we
can run sensitivities on things like that, and that
might be a good problem to run.
We did that. Yesterday I was talking
about Arkansas. We did quite a bit of sensitivity
calculations with the Arkansas, independent
confirmatory calculations.
CO-CHAIRMAN WALLIS: As I remember the
table we saw from yesterday, these criteria are met
pretty well, except the dry well temperature briefly
being above 330. Is that the one?
MR. THROM: Yeah, that's correct.
MR. LOBEL: And that was just for a short
time so that it really --
CO-CHAIRMAN WALLIS: Most of the other
criteria seemed to be met by reasonable margin. Is
that true?
MR. LOBEL: Yeah, yeah, and the changes
actually weren't that great. I have a slide that
talks -- in fact, it's the next slide or the slide
after -- talks about the actual numbers.
Yeah, that's true. Where are we? Well,
I think we've talked about that.
The next slide says what you just said.
CO-CHAIRMAN WALLIS: And you're not
concerned about this brief temperature exceeding
criteria?
MR. LOBEL: The SER says that -- we've
looked at that, and that it's brief enough so that the
temperature of the containment structure won't be
changed, won't exceed its limit even though the
atmosphere is slightly above the design temperature.
MEMBER POWERS: Is the concern over the
dry well temperature one of the structure or is it of
the penetrations and the penetration seals?
MR. LOBEL: I believe the limit -- I don't
know for sure. Maybe G.E. can answer that or
licensee.
MR. PAPPONE: This is Dan Pappone with
G.E.
The real concern is the equipment. If the
dry well structure penetration seals, relief valves,
solenoids, things like that, and what we get from the
long-term containment analysis or short and long-term
containment analysis is an EQ envelope, environmental
qualification envelope that then is used for
qualifying all of the equipment there.
And that's where we're looking at the
short term of this excursion above the 330. It was
only for half a second, I believe, and that's where
making the judgment that the response time, you know,
to actually heat up the components that we're
concerned about is so short.
Remember they're starting out at, say, 150
degree temperature, the initial dry well temperature.
So there's a long time lag before they would reach
that 330 degree temperature limit, and this is such a
short pulse before we drop back down.
CO-CHAIRMAN WALLIS: Okay.
MR. LOBEL: Okay. We also look at NPSH of
the ECCS pumps and spray pumps, and there wasn't any
change to the design temperature. Calculations were
done at the design temperature so that the power up
rate didn't affect those.
And Clinton does not take credit for over
pressure. So that wasn't an issue.
Just briefly, some of the other
significant systems that we looked at I'll mention.
The ultimate heat sync was originally designed for two
units, and so there's plenty of margin for --
CO-CHAIRMAN WALLIS: It was also built for
two units, not just designed? It was actually built
for two units?
MR. LOBEL: Built and hopefully designed
and built.
MR. HOPKINS: Yes, the slide should say
megawatt electric.
MR. LOBEL: Oh, yeah.
CO-CHAIRMAN WALLIS: Yes, it's a pretty
small pump.
MR. LOBEL: Yeah, right. The component
cooling water system is an important safety system.
The only loads that increase with power is a small
change to the reactor coolant pumps, and the spent
fuel pool loads, and the temperature increase is
small, I think, just a couple of degrees.
The spent fuel pool --
CO-CHAIRMAN WALLIS: This is sealed
cooling or something?
MR. LOBEL: Yeah, for the reactor coolant
pumps, recirculation pumps.
The spent fuel pool cooling, the peak
temperature is less than 120 degrees with the ultimate
heat sync at its tech. spec. number of 95 degrees, and
one train of cooling. For a normal off load, which is
a portion of the core for Clinton, the guidance is
that the limit has to be met, assuming a single
failure with one train, and the Clinton calculations
show they do that.
An abnormal off load for Clinton would be
a full core off load, and the peak temperature is less
than 140 degrees, with the ultimate heat sync
temperature at 95 degrees, and since it's an abnormal
off load, they can take credit for both trains of
cooling.
CO-CHAIRMAN WALLIS: This is an average
temperature in the pool? This is one 40 degrees?
MR. LOBEL: Yes.
CO-CHAIRMAN WALLIS: The pool mixes
rapidly enough that you don't get concerned about
variations around the pool?
MR. LOBEL: I don't think that's a
problem. Is there somebody else that can answer?
I think that's been looked at and --
CO-CHAIRMAN WALLIS: I think it would have
to be.
MR. LOBEL: -- the decision has been made
that it's not an issue. I can't quite where, but I
remember reading something at one time that that's
been looked at.
MR. SHUM: David Shum from Plant Systems
Branch.
Could you repeat that question, please?
I couldn't hear.
CO-CHAIRMAN WALLIS: Well, we're talking
here about an average pooled temperature. I'm just
asking how much does the temperature vary from, say,
the region of the fuel to the remote region of the
pool.
MR. SHUM: Well, we didn't look at the
temperature at the fuel. From the fuel pool cooling
standpoint, we look at the average in the --
CO-CHAIRMAN WALLIS: Do you assume the
average is good enough?
MR. SHUM: Un-huh, but from another plants
we always look at it, and we have no problem before.
MR. LOBEL: My recollection is that that's
been looked at, studied, and the mixing in the pool
from natural circulation is enough that the
temperature is uniform. But I'll take it as an issue
if you'd like to try to find out where that's
addressed.
CO-CHAIRMAN WALLIS: It would be nice if
someone knew definitely. Yeah, maybe you could find
out.
MR. SHUM: Yes, we will be able to find
that out.
MR. LOBEL: Okay. In conclusion, the
balance of plant containment systems for Clinton
comply with NRC regulations and guidance at EPU
conditions.
CO-CHAIRMAN WALLIS: And this is the end?
MR. LOBEL: The end.
CO-CHAIRMAN WALLIS: Any closing
statement?
MR. MARSH: I have a closing statement
unless, Jon, do you have anything you want to conclude
with up there? No?
MR. HOPKINS: I just wanted to say this
was the end of the staff presentation, and it would
just be questions and if there are questions on other
topics. This was the end of our presentation.
MR. MARSH: I've got some closing
comments, Mr. Chairman, unless you'd like to.
Okay. Thank you.
I'd like to thank the committee for the
time and the opportunity to present to you both
Arkansas and Clinton's incentive power up rate. I
want to emphasize again that the staff has undertaken
extensive reviews of these applications, and all areas
affected by the power up rates have been reviewed and
evaluated.
The staff has critically examined
methodologies and their application for these two
cases. We've concluded that all of the analytical
codes, methodologies used for licensing analyses are
acceptable for these applications. The results of the
deterministic analyses have demonstrated that the
proposed increases in power for Arkansas and Clinton
plants are acceptable and meet the regulatory
requirements.
Now, this concludes the presentation.
Now, we, of course, would be glad to answer any
questions, but I've got a couple other thoughts I'd
like to leave with you.
Based --
CO-CHAIRMAN WALLIS: These are going to
the full committee?
MR. MARSH: next month.
CO-CHAIRMAN WALLIS: Next month?
MR. MARSH: Right.
CO-CHAIRMAN WALLIS: Both of these?
MR. MARSH: Both of them, right. So
any --
CO-CHAIRMAN WALLIS: And both the
applicant and the staff will have considerably less
time.
MR. MARSH: Yes, sir, right.
CO-CHAIRMAN WALLIS: And therefore, you
have to bear that in mind regarding your presentation.
MR. MARSH: Yes, sir. Any guidance you'd
like to give us in terms of what you'd like us to
present would be great.
Let me add one more thing, if I can. Some
of the issues that came out in yesterday's
presentations we're going to respond to. There were
some to do items and things. We'd like to respond in
writing if you don't mind. We want to make sure that
we answered those thoroughly and completely and
formally. Okay?
So we're going to be getting the
transcripts and making sure that we understand the
questions and respond in the right manner.
And as you say, we don't have a lot of
time for the full committee. So any -- we did talk
yesterday about a potential approach for the full
committee. What we talked about yesterday was, of
course, we don't have the time to go through all of
the issues, but you suggested picking a couple of key
topics in each one of these areas and demonstrating to
the full committee some depth of reviews and some
issues that we went through.
So perhaps we'll identify if it's
acceptable to you some key issues in each one of these
plants' reviews, and we will go through with some
detail to the full committee.
MEMBER POWERS: I think you need to be
more cautious with your ANO than you do with the
Clinton because this is effectively the third one.
MR. MARSH: Right.
MEMBER POWERS: One of these, I mean, it
may be a BWR-6 and so it's somewhat different.
MR. MARSH: Right.
MEMBER POWERS: My thought here would be
saying that, look, this is yet another of the BWR up
rates. They've done it a little differently than
others. Here's the exceptions, and discuss the
disposition of your exceptions.
I think you're going to have to discuss
again the transient testing issue.
MR. MARSH: Okay.
MEMBER POWERS: I mean, I just think it's
unavoidable simply because we had added comments to
one of our letters, in which there's a dissent from
the majority opinion discussion.
That's my personal thought on this
subject. I don't think you need -- my point is I
don't think you need to be as careful with this one as
you do with the ANO just because it is a repetition.
And, by the way, I will thank you publicly
for allowing your staff to meet with us this morning,
and we could discuss BWR stability and the mathematics
there and not bore everyone else to death.
MR. MARSH: Good.
MEMBER POWERS: It was a very productive
meeting.
MR. MARSH: I'm glad it was helpful.
When you say "careful," are you suggesting
that we spend more time discussing the Arkansas
application?
MEMBER POWERS: Oh, I think you have to.
MR. MARSH: Okay.
MEMBER POWERS: Because it's new and it's
--
MR. MARSH: It's the first of this type.
MEMBER POWERS: The type, sure, and
whatnot.
MR. MARSH: Okay. We'd be glad to do
that.
MEMBER POWERS: I mean, I personally don't
feel the need to be that cautious and careful with
Clinton. It seems to me that the focus there is
really on the exceptions issue because they're
different, and among the exceptions, of course, is
transient testing.
MR. MARSH: Right.
MEMBER POWERS: And I would try on the
transient testing to really go through your logic very
carefully just because we have a dissenting opinion.
MR. MARSH: We'd be glad to.
PARTICIPANT: We'll continue to.
MEMBER POWERS: Yeah, I mean, I don't know
that you're going to persuade him.
MR. MARSH: But we could at least say it
to make sure that people understand our rationale for
it.
MR. MARSH: Sure, sure.
MEMBER SIEBER: Well, you want it on the
record also.
MR. MARSH: Yes.
MEMBER POWERS: And you want to give us
ammunition to beat him.
MR. MARSH: No, we don't.
MEMBER POWERS: Yes, you do. Trust me.
You do.
MR. MARSH: Whatever you say.
Well, Mr. Chairman, that concludes our
presentation.
CO-CHAIRMAN WALLIS: Thank you.
MR. MARSH: And we thank you very much
again.
CO-CHAIRMAN WALLIS: That's all right.
Anything else that the members wish to say at this
time?
(No response.)
CO-CHAIRMAN WALLIS: Well, thank you for
helping us to finish early.
We will not reconvene until one o'clock,
and then we'll take up a completely different topic.
Thank you all for your contributions to
the meeting.
(Whereupon, at 11:38 a.m., the
Subcommittee 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
(1:03 p.m.)
CO-CHAIRMAN KRESS: This meeting will now
please come to order.
I think this is a new Subcommittee meeting
from what we had this morning. This is combined
Subcommittees on Thermal-Hydraulic Phenomena and
Future Plant Designs.
I am Tom Kress, Chairman of the Future
Plant Designs Subcommittee, and we have with us Graham
Wallis, who is Chairman of the Thermal-Hydraulic
Phenomena Subcommittee. We also have ACRS members
Dana Powers and Bill Shack and Jack Sieber, and our
ACRS consultant, Virgil Schrock.
The purpose of this meeting is to review
the staff's review of Phase 2 of the pre-application
review of Westinghouse AP 1000 plant design. The
Subcommittee will gather information, analyze relevant
issues and facts, and formulate proposed positions and
actions as appropriate for deliberation by the full
committee.
Mr. 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 January 29th, 2002.
A transcript of this meeting is being held
and will be made available as stated in the Federal
Register notice. Therefore, it's requested that
speakers first identify themselves, use the
microphone, and speak with sufficient clarity and
volume so that they can be readily heard and copied on
the transcripts.
We have received no written comments or
requests for time to make oral statements from members
of the public.
I have no introductory comments. It's a
self-introductory subject. Do any of the members wish
to make anything before we start? Any comments?
MEMBER POWERS: This is really just a 67
percent power up rate; is that --
CO-CHAIRMAN KRESS: That's basically what
it is, yeah. It's another power up rate review that
we're doing.
CO-CHAIRMAN WALLIS: It's for a
nonexisting reactor. So it's not quite the same
thing.
MEMBER POWERS: So it's in the imaginary
space.
CO-CHAIRMAN KRESS: Well, we will now
proceed with the meeting, and I'll call upon Jim Lyon,
our friend from the past, to open the meeting for us.
MR. LYON: Thank you, Dr. Kress
I appreciate being here today.
MEMBER POWERS: Well, you didn't. You
left us.
MR. LYON: Well, you know.
MEMBER POWERS: Short exposures are okay.
MR. LYON: Short exposures. A very
helpful one and a very enlightening one.
As Dr. Kress alluded, we're going to talk
about Phase 2 of the AP 1000 design certification
review. This is the final portion of the pre-
application review, and I think just to get us going
I'd like to turn it over to the two project managers
that are working on this at this time.
First, Andrzej Drozd, who is the acting
project manager and has been helping us out since, I
guess, about October time frame.
And Larry Burkhart, who will be the
permanent project manager as Andrzej goes back to his
old duties where he's going to help us a lot doing the
review of this and other advanced reactors.
So with that I'll turn it over to Andrzej.
MR. DROZD: Good afternoon. My name is
Andrzej Drozd, and for the last four months I had the
privilege to and pleasure to coordinate Phase 2 review
of AP 600 pre-application review.
In today and tomorrow's presentation, I
will give a quick background and the status of the
project, and then we'll turn into a technical panel to
discuss Phase 2 issues.
I will do the hopefully simple and
straightforward regulatory exemptions issues, and then
Jerry Wilson, Dave Terao and Goutam Bagchi will
address design acceptance criteria.
After that, testing of AP 600 and its
applicability to AP 1000 will be addressed by Steve
Bajorek, and then Walt Jensen and Ed Throm will
discuss with you and present to you the application of
safety analysis codes.
Just to recall, some time in December
1999, LC certified the AP 600 start-up plant design,
and just about the same time Westinghouse in various
forms expressed interest in doing power up rate, as
Dr. Power says, that is, to apply for a certificate of
a similar, if not identical, plan, but with the power
of about 1,000 megawatts output.
Needless to say, there were a lot of
introductory discussions followed by a meeting in
April when staff and Westinghouse discussed what would
be the best way to approach certification review of
the new plant.
And the three-stage approach seems to be
the most appropriate. The Phase 1 would be or was
just identification of the issues to be reviewed in
Phase 2, and Phase 2 would be review of those issues
and potentially, depending on the results of Phase 2,
the full design certification review as a Phase 3.
In May 2000, Westinghouse requested the
NRC staff to proceed with Phase 1, and after several
discussions and letters and meetings, NRC has provided
six review issues and its estimated how much it would
take to review it.
Based on that suggestion, in August 2000,
Westinghouse asked NRC staff to proceed with Phase 2
and focus only on four issues, that is, applicability
of the test program, APC funded test program to the
new design, as well as applicability of safety
analysis code used for AP 600 to AP 1000.
They asked us to review proposed design
acceptance criteria approach, as suggested by
Westinghouse, and as well, applicability of certain
regulatory exemptions for AP 1000.
As I understand it, by the end of August
we made presentation to you for AP 1000. Therefore,
I don't have any specific presentation on the design
itself. However, I do have some back-up slides if
needed.
CO-CHAIRMAN KRESS: I think we're familiar
enough with the design at this point.
MR. DROZD: Thank you, Tom.
Briefly, status of the review. Phase 1
was complete in July 2000, and Phase 2 is being
completed as we speak. Technical review was done just
about a month ago, and as a result, we are writing two
SECY papers.
The Phase 2 issues were divided into two
packages sort of. One package is a policy oriented,
which is application of design acceptance criteria
approach to design certification, and more technically
oriented or utilization oriented testing of the codes.
Both papers went through internal
iteration, and final drafts are being reviewed and
concurred. Some were in our administrative space.
Today and tomorrow we'll present to you
the results of our Phase 2 review, and eventually in
about three weeks or in about a month, we'll make a
presentation to the full ACRS Committee explaining
what we did and asking for the approval or
suggestions.
CO-CHAIRMAN KRESS: You'll want a letter
in the March meeting.
MR. DROZD: Or asking for a letter of
recommendations.
CO-CHAIRMAN KRESS: When do you go to the
commissioners with this?
MR. DROZD: The week after. The paper is
due to the Commission in March 28.
CO-CHAIRMAN KRESS: Okay.
MR. DROZD: And as we discussed with
Westinghouse our findings, there is a good chance that
Westinghouse will go ahead with full design
certification application some time the end of March,
beginning of April this year.
CO-CHAIRMAN KRESS: Does Westinghouse know
what's in these draft documents?
MR. DROZD: Yes.
CO-CHAIRMAN KRESS: You've already been in
discussion with them?
MR. DROZD: Not the letter of our
findings, but major points were conveyed to
Westinghouse, yes.
MR. BURKHART: We're still working out
some of the what may be policy issues on design
acceptance criteria. There's been some late
developments.
In fact, we just received a letter today.
My name is Larry Burkhart. I'm the new
and it is hoped the permanent project manager for this
project.
So that aspect, the design acceptance
criteria, is kind of in an evolutionary phase. We're
towards the end of it where we still have a little bit
of work to do.
MR. DROZD: Well, if there's no more
questions regarding status of the project, let's turn
to the issues, Phase 2 issues.
One, the first one that I'd like to
address is regulatory exemptions. During AP 600
certification review, Westinghouse asked for several
regulatory exemptions, and three of them are being
suggested to be granted also for AP 1000.
The three are the exemptions regarding the
safety parameters display console, the auxiliary
feedwater system, as well as off-site power sources.
The rationale for granting those
exemptions for AP 600 are summarized right here for
the display console. The reason that the exemptions
for this specific console, this specific part of the
design was that the safety parameter display console
was integrated or is to be integrated with the control
room design. Therefore, there was no need for a
stand-alone requirements for safety parameter display.
CO-CHAIRMAN KRESS: Then the control room
design is a COL item, put off until the COL or --
MR. DROZD: That is put up at the COL,
yes.
MR. WILSON: Dr. Kress, this is Jerry
Wilson.
We'll discuss that in the next
presentation.
CO-CHAIRMAN KRESS: Oh, okay. This is
just an overview of it, and you'll go into that a
little more. Okay.
MR. DROZD: The two other technical areas
that Westinghouse is seeking a regulatory exemptions
are auxiliary feedwater system, as well as off-site
power sources. Both of them are stemming from the
passive nature of safety system and passive nature of
safety injection systems.
CO-CHAIRMAN KRESS: These are general
design criteria, these?
MR. DROZD: The off-site power deals with
general design criteria, and aux. feedwater is -- I
don't remember -- 50, 30 something. I'm sorry. I
don't remember exactly which paragraph it is, but they
were discussed during AP 600 certification review, and
the issues are identical.
CO-CHAIRMAN KRESS: And the basis for
granting approval of these exemptions is that you feel
like the underlying purpose of the rule is met?
MR. DROZD: We feel that the new design
has identical safety system function as AP 600, and at
least the intention of the requirements are being met.
Therefore, we don't see any reason why not to grant
those exemptions.
CO-CHAIRMAN KRESS: It meets the intention
of the rule.
MR. DROZD: The intentions are being met,
yes.
CO-CHAIRMAN KRESS: Okay. We're going to
get into a lot of detail on this testing and codes --
MR. DROZD: Oh, definitely.
CO-CHAIRMAN KRESS: -- and the DAC and so
on. You're giving us the bottom line right now.
MR. DROZD: I'm giving a very high level
summary of our positions and findings.
CO-CHAIRMAN KRESS: Good.
MEMBER POWERS: This is a high level
bottom line.
CO-CHAIRMAN KRESS: That's a little --
MEMBER POWERS: I will think about this.
(Laughter.)
MR. DROZD: Well, if there's no more
questions on the exemptions, let me give you the high
level bottom line position on the design acceptance
criteria. The standardization review, review of
standardized plans are laid out --
CO-CHAIRMAN KRESS: Somebody has fouled up
on viewgraphs, and they didn't give us that one.
CO-CHAIRMAN WALLIS: I have it.
CO-CHAIRMAN KRESS: It's not in my
package.
MEMBER SCHROCK: That's because you're
Chairman.
CO-CHAIRMAN KRESS: Well, I guess so.
MEMBER POWERS: You are supposed to know
all of this stuff.
CO-CHAIRMAN KRESS: I got two copies of
your other slide though, the next one. So I guess
that was -- okay.
MR. DROZD: It was a voluntary gift.
CO-CHAIRMAN KRESS: Okay. Now, you may
proceed.
MR. DROZD: Thank you.
The ground rules for reviewing design
certificate applicants are set out in Part 52, and the
key words, the key approach in staff's view is the
completeness of the design that's being submitted for
review.
The design acceptance criteria approach
were developed and applied during ABWR system 80 plus
reviews, as well as AP 600.
CO-CHAIRMAN KRESS: Now, what exactly does
5247(a)(2) say? Does it ask --
MR. DROZD: That will be covered with --
CO-CHAIRMAN KRESS: That will be covered
later? Okay. We'll wait.
MR. DROZD: Just in a couple of minutes.
CO-CHAIRMAN KRESS: We'll wait a minute.
MR. DROZD: In a couple of minutes.
The bottom line position as of now, as of
today is that based on the experience that we had with
the previous design reviews, we don't see that
approach as the preferred way of regulating, in
general.
And we do agree that there are technical
areas that are affected by rapidly evolving
technologies, like instrumentation and control or
human factors, control room design. Therefore, we do
see good basis for granting design acceptance criteria
approach, because of the nature of the evolving
technologies.
Other areas are less obvious to staff,
whether DEC should be applied or not. I'd like to
make a note that the DEC approach granted or used for
piping, for example, and radiation protection, that
was used in ABWR System 80 Plus, we don't think it
applies to the case of AP 1000, and the rationale for
that would be presented in just a couple of minutes.
Also, we strongly believe that the level
of design details submitted for AP 1000 certificate
should be equivalent, if not the same, as the one that
was submitted for AP 600.
And maybe before you ask anymore specific
questions, I turn over the issue to technical panel,
Jerry Wilson, Dave Terao, and Goutam Bagchi. They
will address all there is to know about DAC approach.
Gentlemen.
CO-CHAIRMAN KRESS: I see they got
together some of the people that reviewed AP 600 for
this.
MR. WILSON: Yes, sir.
MR. DROZD: Sixty-seven percent up rated.
(Laughter.)
CO-CHAIRMAN KRESS: One third.
MR. WILSON: Mr. Chairman, I'm Jerry
Wilson with the new reactor licensing project office,
and I'm joined by my colleagues from the Division of
Engineering that will participate in this
presentation.
Now, as Andrzej introduced, we're here to
talk about the request from Westinghouse to use design
acceptance criteria, and by way of background to
understand our recommendation, I thought I'd give a
little review of how we got to this point.
Now, the origin of this issue comes from
when we were building plants, a lot of plants in the
'70s and '80s, and electric companies got construction
permits based on basically preliminary design
information, and they completed their design while
they were building the plant, which led to a lot of
problems that we're all familiar with.
And so when the Commission set out to
reform the licensing process and issue Part 52, a key
aspect of that was to require design information be
complete before construction was initiated.
So to resolve this problem, we have in
Part 52 and also in the design certification process
in Part 52, a requirement to provide complete and
final design information.
And to get to your question you just
asked, if you looked at page 3 of our handout, you'll
see a summary of that requirement, and basically what
we're looking for is final design information
equivalent to what you would see in a final safety
analysis report for an operating license application
and for essentially complete design information.
CO-CHAIRMAN KRESS: Now, let me ask you
this. What would be the practical implications if you
granted this, the DAC on the piping, seismic loadings
for the -- what would be the practical implications
with respect to your completing your review and making
a safety decision?
MR. WILSON: I'm going to get into this a
little bit more, but to introduce the concept is that
what's being used is design acceptance criteria in
lieu of detailed design information, and so in those
particular areas where it is used, the staff and the
applicant agree on the approach to finalize the design
at a later date.
And what we're looking at is the
desirability to develop design acceptance criteria
that would lead to an acceptable design when the
applicant who references that certification actually
completes it.
Now, picking up a legal aspect of this
deck or a subset of ITAAC and the way ITAAC is set up
in the regulations, an applicant is required to
complete these design information until after they had
gotten a combined license. So you could have a
situation where we would get back to where we were in
the past where an applicant was completing their
design information while reviewing the plan.
CO-CHAIRMAN KRESS: You'd have to review
it again then.
MR. WILSON: Yeah, we'd have to verify
that they did implement the design acceptance criteria
correctly and thereby coming out with an acceptable
design.
CO-CHAIRMAN KRESS: Yeah, okay. I
understand.
MR. WILSON: So when we issued Part 52, we
were initiating reviews of design certification
applications at that time, and the staff did a study
of this issue of what's appropriate and necessary
level of design information to support a design
certification application and gave SECY papers to the
Commission on that subject.
If you skip ahead and look at Slide 4,
you'll see these papers listed here, but, in
particular, SECY 90-241 and SECY 90-377.
The result of that led to a staff
requirements memorandum that had a lot of important
guidance to the staff on how to perform the design
certification reviews, and in there the Commission
said with regard to level of detail, "We meant what we
said when we issued Part 52. Meet the requirement.
Meet the 5247(a)(2) on level of detail."
Furthermore, the amount of information
should be proportional to the safety significance of
a particular system. So more significant safety
systems you would expect to see, more design
discussion.
And finally, they said, "Don't use ITAAC
as a means of resolving this design information."
Now, after that --
MEMBER POWERS: I guess they made it
pretty clear then, didn't they?
MR. WILSON: Yeah. At this point G.E. as
the lead applicant came to the staff and expressed
some concerns they had with actually completing their
design, and they brought up two areas of concern.
The first area of concern, as Andrzej
introduced, was rapidly evolving technology. They
were referring to the digital instrumentation and
control systems and also the human factors design
aspects of the control room.
And what they said is that these areas of
technology are rapidly evolving or concerned if we
completed that design at that point in time; that you
would be locking up a design that if implemented later
could be significantly out of date.
And the other area of concern that they
raised is G.E. claimed that they couldn't complete the
piping and radiation protection design information
because they needed as building, as procured
information to do that.
CO-CHAIRMAN KRESS: So in those two
general areas of a possible DAC criteria, it seems to
me like if you have those conditions, you disregard
bullet three about being able to reach a final
conclusion on any safety question?
MR. WILSON: Yes. What you'll see as I go
through this history is we didn't fully implement the
Commission's guidance in its SRM.
CO-CHAIRMAN KRESS: Right.
MR. WILSON: So because of that, we went
back to the Commission with a new round of SECY
papers, and once again, if you jump ahead to slide
four, you'll see a number of SECYs issued in the '92
time period where we explained these issues to the
Commission and now we proposed to come up with design
acceptance criteria in lieu of design information.
And the result of all of that effort was
that the Commission allowed us to use design
acceptance criteria in lieu of detailed design
information in those first two applications in the
areas of rapidly evolving technology and the as built,
as procured area.
But they admonished us that use of DAC
should be limited and if in the future that staff
planned to use it anymore, we had to come back to the
Commission.
CO-CHAIRMAN KRESS: They didn't prohibit
it. They just said, "Be sure it's limited, and if
you're going to do anymore DACs, come back to us with
it first."
MR. WILSON: That's correct.
CO-CHAIRMAN KRESS: Okay.
MEMBER SIEBER: I presume they
specifically approved the areas that you identified,
right?
MR. WILSON: Yes, and that's what I'm
going to show on the table in the next slide.
So what you see is that for ABWR and for
System 80 Plus, they used design acceptance criteria
for instrumentation and control and human factors,
radiation protection, and piping.
Now comes AP 600.
CO-CHAIRMAN KRESS: Now, he's a case where
piping was included in a DAC.
MR. WILSON: That's correct.
Now comes AP 600, and in AP 600,
Westinghouse continued to use the I and C and human
factors DAC, once again applying the rapidly evolving
technology criteria, but they did provide sufficient
design information for radiation protection and
piping.
CO-CHAIRMAN KRESS: Now, why was it deemed
that, for example, System 80 Plus couldn't provide
sufficient design information for piping whereas AP
600 could? What is different about the two?
MR. WILSON: Yeah, I'm going to ask you to
hold that thought for a minute, and I turn it over to
Mr. Terao and he's our piping expert and it would be
better if he answers that.
MEMBER SIEBER: I have a question about
radiation protection. To me you can split that into
two broad areas. One is the physical features of the
plant, like shielding and ingress and egress pathways,
et cetera, and the other half of that or part of it is
all of this instrumentation, area monitors and
effluent monitors and things like that.
MR. WILSON: And what we're talking about
here is the aspect of the barriers, the shielding and
that sort of thing, protection, and really what
sources are you going to get from various components
and how that would affect the design aspects of the
shielding.
MEMBER SIEBER: And I take it that all
instruments fit into the instrument and control areas,
which would be another DAC.
MR. WILSON: I believe so.
MEMBER SIEBER: Okay.
MR. WILSON: So on AP 600 they will
provide that information. They didn't require design
acceptance criteria. From the staff's perspective
this was an improvement, and we came closer to meeting
the Commission's goals, and our regulations for
complete and final design information.
Now, we come to the pre-application review
for AP 1000, and as you've heard from us previously,
Westinghouse has asked four questions of the staff,
and the one we're focusing on today is can you use
design acceptance criteria on AP 1000 in three areas,
piping structures and seismic analysis.
Well, when we went to look at this, we
looked at it both from a technical perspective and
from the policy perspective. First of all, the
question is, well, is it technically feasible to
develop design acceptance criteria when all three of
these design areas are not complete?
And then the other part of it is where are
we with regard to meeting the Commission's goals.
So what we concluded -- and I'm going to
turn it over here to Mr. Terao to talk about our
ability to develop, for example, a piping DAC when the
structural design and seismic analysis hasn't been
formed, hasn't been completed -- but let me back up
for a minute and say, first of all, in the rapidly
evolving technology, that decision was originally made
in the '92 time period, and so I went back to our
technical staff, and I said, well, a decade has gone
by. Does the basis for allowing the use of design
acceptance criteria for instrumentation and control
and human factors still apply?
And my technical experts said, yes, it
does in those areas. So if Westinghouse requests
that, we are supporting the use of design acceptance
criteria in those areas.
However, from the standpoint of piping
structures and seismic, Westinghouse was able to
complete that at AP 600, we believe, for the reasons
that we started out down this path, that we should
have complete design information in those areas, and
so from a policy perspective, we would recommend that
Westinghouse should provide that information also for
AP 1000.
CO-CHAIRMAN KRESS: You don't happen to
have an example of one of these DACs? For example,
what do they look like? Are they just a set of
criteria that have to be met in terms of general
criteria?
MR. WILSON: Dr. Kress, I've met with you
so many times that I anticipated this question.
(Laughter.)
MEMBER POWERS: You were predictable?
CO-CHAIRMAN KRESS: I hate to be so
predictable.
MR. WILSON: Med, could you give me a hand
here?
I don't have a lot of copies, but I have
some for the reporter and perhaps for the Chairman and
let Westinghouse have one.
What I've done is I've used human factors
as an example probably because I understand it the
least, but first of all, in order to understand how
design acceptance criteria is different, what I've
given you is a handout from some human factors
criteria, and there's two parts of it. The part I'm
going to really talk about is in Section 3.2 there.
First of all, let's talk about what a
regular ITAAC looks like. An ITAAC is a verification
process, and so if you went into the Tier 1
documentation for AP 600, for example, you would see
that ITAAC ware organized by structures and systems.
Most of them are systems based ITAAC. There will be
a high level design description of the system, will
pick out key elements of the system key design
commitments. That will be in your left hand column,
and will then specify whether that verification is
going to take place by a test or analysis, physical
inspection or combination thereof.
So you can see in the first example here,
the design commitment has to do with the technical
support center. We're going to verify it with a
physical inspection, and we have our acceptance
criteria.
This one is pretty straightforward. Most
ITAAC aren't quite that straightforward, but it's good
enough to give you the concept of what you would
normally find for a regular ITAAC.
Now, when it comes to design acceptance
criteria --
CO-CHAIRMAN KRESS: And ITAAC is for
something that's already designed?
MR. WILSON: That's right. That was the
idea.
CO-CHAIRMAN KRESS: And this is just a
verification.
MR. WILSON: You finish the design, and
then this ITAAC sets for the verification process
we're going to use to verify that they built it the
way they committed to do it.
CO-CHAIRMAN KRESS: Right.
MR. WILSON: Now, when it comes to a
situation where the design isn't complete, we need an
ITAAC that has a design process built into it, and
this is just kind of a flow chart of what I'm going to
go through, and I'm not going to show you all of it,
but the idea being that we set up a group to do the
human factors design. They do an analysis and design,
and then they verify it at the end of the day.
And so if you looked at that ITAAC, you
would see a discussion of the process used to do human
factors design, and then we would get into the ITAAC
or in this particular case, DAC, and you would see
things like, okay, you have to perform a TAC analysis,
and it talks about what should be done in that task
analysis, and I'll skip ahead here, and you'll see
that -- make sure that that's done correctly. I'm
trying to find a good example here.
Once the design is complete, then they'll
see some more traditional verifications in the main
control room and key things you're looking for in
there, and acceptance criteria for that.
CO-CHAIRMAN KRESS: And I presume
something equivalent to this exists for piping DAC for
System 80 Plus?
MR. WILSON: Yeah. Now, if you looked at
the instrumentation and control DAC, you'd see it's
very similar to what I've just shown you from human
factors, and it even has a similar flow chart that I
had up earlier, and which I can't put my finger --
yeah, here we are.
So you'll see instrumentation and control
looks very similar to this. So that gives you an idea
that instead of a straight verification of a design
that's completed, you first have to go through a
process where we have criteria on completing the
design, and then once it's complete, you would verify.
CO-CHAIRMAN KRESS: There was something
thought about piping and structures and seismic that
prevent you from developing such a DAC high tech?
MR. WILSON: Yes, and with that question,
I'll turn it over to Mr. Terao.
MR. TERAO: I'm David Terao. I'm with the
Mechanical and Civil Engineering Branch in NRR, and
before I go into my presentation, just to answer your
question, Dr. Kress about piping DAC, each of the DAC
actually has slightly different format, and for
piping, we did have ITAAC, but the DAC were actually
the design and acceptance criteria that were spelled
out in the FSAR by G.E. and System 80 plus, and they
were in different parts of the FSAR.
Since the FSAR doesn't have a section
called piping, they have a section called seismic and
mechanical design and relief valve designs. So the
DAC were just spread out in all different parts of the
SSAR, the standardized safety analysis report.
What the staff then did was to be
consistent in addressing DAC, we wrote in our safety
evaluation for ABWR and System 80 Plus. We had a
specific section, Chapter 3.12, which was devoted
solely to piping and piping DAC and piping ITAAC.
And in those sections what we addressed
were all of the different design and acceptance
criteria that we reviewed and devaluated as part of
our Part 52 reviews, and there are actually seven
different areas that we looked at in piping design,
and these are some of the areas.
They're broken into different sections.
For example, we looked at the codes and standards. We
looked at the analysis methods. We looked at the
piping modeling techniques. We looked at the pipe
stress analysis criteria. We looked at the pipe
support design criteria. We looked at energy line
break criteria, and we looked at leak before break
criteria.
And each of those had many subissues that
we addressed in our safety evaluation. So there are
many issues that we had to review specifically for
ABWR and System 80 Plus.
MR. WILSON: Now, those two slides aren't
in your handout, but we'll provide you copies.
Do you want to go to this one next or the
next one?
MR. TERAO: Yeah. Actually since we got
diverted a little with this discussion of DAC, let me
try to refocus where we are and try to put my
presentation in perspective again.
Staff was requested to review the
acceptability of the DAC approach for AP 1000 for
seismic analysis, structural design, and piping
design. And I stress that because, first of all, this
is a pre-application review. This is Phase 2. So we
aren't reviewing the actual DAC themselves, as I put
up on my slide. We aren't at that level.
That would be done in Phase 3 when
Westinghouse submits its application for design
certification. So right now we're looking at the
acceptability of the DAC approach.
And I guess the two big picture questions
that we asked ourselves were, one, since seismic
analysis and structural design DAC are new, they've
never been used before; can we accept that approach?
And then the second question was on the
piping DAC which had been used before on ABWR and
System 80 Plus. We said, well, why wouldn't it be
acceptable for AP 1000.
So those are the big picture questions.
That's what I addressed today.
CO-CHAIRMAN KRESS: Well, what I was
fishing around for is is there something about piping
structural and seismic that actually makes it
extremely difficult for you to do a DAC or is it just
a policy question.
MR. TERAO: Those are the details I'm
going to get into right now.
CO-CHAIRMAN KRESS: Okay.
MR. TERAO: Yes, for each of the areas.
So in answering the question of what is
the acceptability of using the DAC approach, the first
question that came in our mind is: well, what
criteria do we use for establishing acceptability of
DAC?
And this is why we went into a lot of
detail about the policy issues, because we reviewed
the DAC for three different areas:
One, does it meet established policy
issues as described in the SECY paper?
Two, does it make technical sense? In
other words, by that what I mean is when we look at
the seismic design of a plant, we start with the
ground motion, the earthquake ground motion
accelerations. We look at those ground motion
accelerations onto the building. We perform a
building analysis. The building analysis generates
the amplifications, the accelerations that are used
for the piping analysis.
The piping analysis then is performed and
the results are used to calculate break.
CO-CHAIRMAN KRESS: Which is something you
can do if you have a site, and the site described.
MR. TERAO: I'm sorry?
CO-CHAIRMAN KRESS: Well, you can do all
of that if you have a site, and the site
characteristics.
MR. TERAO: yes.
CO-CHAIRMAN KRESS: But if you don't have
a site, what then do you do? Use a bounding analysis
for those things?
I mean, I can see how you can do all of
that for a plant where you have a specified site.
MR. WILSON: The way we handle the site
characteristics and design certification is the
applicant specifies what we call site parameters, and
then we review the design to the site parameters.
So, for example, the applicant may specify
that he's designing the plant SSE of .3 G. He'll
specify ground motion associated with that. He's
specify flood levels and those sorts of things.
So part of the definition of the design
includes that site parameter.
CO-CHAIRMAN KRESS: Okay. So the design
then would -- I mean, if someone built one of these
things, it would be limited to the site, to at least
fall within reference categories, and that would be an
ITAAC? You would have to --
MR. WILSON: No, it's just part of the
review at the combined license stage to verify that
that particular applicant is referencing AP 600. We
checked the site they planned to put it on and made
sure that that site fits in with those parameters that
Westinghouse specified when they did their design.
MR. BAGCHI: My name is Goutam Bagchi.
I just wanted to give a perspective as to
what constraints we had. AP 600 design layout is the
basis for AP 1000, the same footprint, the same
structural element, except everything is higher. More
mass, more of everything.
We know that while doing our AP 600
reviews the site parameters were varied over a wide
range, and certain sections became quite critical, and
it is our concern trying to locate AP 1000 under all
site conditions my not even be feasible. That was the
concern.
So we could have perhaps in a theoretical
sense used the structural design acceptance criteria
had it not been constraining based on AP 600.
MR. TERAO: Okay. Coming back to my slide
-- thank you, Goutam --
MEMBER POWERS: Just to make sure you
don't get through this slide --
(Laughter.)
MEMBER POWERS: You've discussed over and
over again and the slide discusses, you know, what
makes things acceptable, and I'm saying, gee, I wonder
why you don't also look at what makes things
unacceptable. That is, why don't you include going
back to 52.47 and say why is it that the Commission
would impose such a rule.
And in the introduction, I thought a
pretty good synopsis was given, that they put this
rule in because when you didn't do this, you got a lot
of on-the-fly design activities that resulted in just
an enormous number of headaches.
MR. TERAO: That's a very good point, and
we certainly did consider it. And that's the third
point that I'm trying to make if I can get to it.
(Laughter.)
MEMBER POWERS: Oh, no. That's our job,
to try to keep you --
MR. TERAO: You saw that, didn't you?
MEMBER POWERS: -- to keep the flow of
your presentation --
MR. TERAO: So the three areas, the policy
issues we looked at. We looked at technical issues,
and we also looked at the safety issue on 5247, and I
believe Dr. Kress asked what does 5247 actually say.
These are the relevant words: application
must contain the level of design information
sufficient to enable the Commission to reach a final
conclusion on all safety questions associated with the
design before the certification is granted.
CO-CHAIRMAN KRESS: There's that word
"sufficient" in there.
MR. TERAO: Sufficient.
CO-CHAIRMAN KRESS: That's a judgment
call.
MR. TERAO: That's a judgment call.
That's correct. But the key words I'd like to focus
on are before the certification is granted because
there is certainly a timing issue involved here on
when the staff can resolve any safety questions or
potential safety questions.
And that's a crucial issue here because
with the Westinghouse DAC approach, they propose in
their topical report to provide certain information
after design certification, and that leads me into my
next slide.
CO-CHAIRMAN KRESS: But aren't the DACs --
MR. TERAO: I'm sorry?
CO-CHAIRMAN KRESS: Are the DACs that have
already been granted for the other types sort of a
precedent for that?
MR. TERAO: Well, yes. We did establish
a precedent for I and C, human factors and piping, and
RAD protection, but seismic and structural, we have
not.
CO-CHAIRMAN KRESS: I know, but precedents
are established by doing it.
MR. TERAO: Yes. The precedents were set,
and what I'm saying is that Westinghouse is not using
the exact approach. They're proposing something
different than the DAC approach that were used in
these other plants.
CO-CHAIRMAN KRESS: Is there some reason
why they have to be exactly like the others?
MR. TERAO: Well, that's what I'm trying
to point out. They may not be. It may be acceptable,
but what I'm trying to point out here is where they
deviate from the DAC approaches in the other plants,
where do they not meet established policy issue ,
technical issues or safety issues? That's what I'm
trying to point out here.
CO-CHAIRMAN KRESS: Okay.
MR. TERAO: Okay? Thank you.
What this slide shows, what I'm trying to
show in this slide is what design information will be
provided and when, and for three different areas: for
seismic analysis, for structural design, and for
piping design.
And the second column shows the
information that would be provided at the design
certification stage, and the third column shows the
information that would be provided at the combined
operating license stage.
CO-CHAIRMAN KRESS: Now, this is
Westinghouse's proposal.
MR. TERAO: Yes, yes. This is in their
topical report.
CO-CHAIRMAN KRESS: Okay.
MR. TERAO: And so at the combined
operating application stage, that's just prior to
construction, and then the last column shows what
information would be provided once the plant has been
licensed and construction has started.
So for seismic analysis, Westinghouse is
proposing to develop stick models for the AP 1000 at
the design certification stage, and then perform a
more finite element analysis for AP 1000 at the COL
stage.
We didn't see any major problems with
that. You go down to the next area, and what
Westinghouse did here is they broke down their
proposal between rock sites and other than rock sites,
and for purpose of simplicity, we'll just call them
soil sites. But soil sites means other than rock
sites.
So for rock sites, the seismic analysis is
actually rather simple because you don't get the soil-
structure interaction, and so they can -- Westinghouse
is proposing to perform a fixed base seismic analysis
and generate the amplified response spectra for the
piping design, and they would also perform the
overturning and stability analysis for the plant. And
we agree with that.
But when you get to the soil sites,
Westinghouse proposed to not do the analysis, but
rather, they proposed to develop the seismic analysis
DAC, and that would tell you how to do the analysis,
say, to generate the soil structure analysis, how to
generate the amplified response spectra, and how to do
the overturning of stability analysis.
And that analysis, we believed, is very
complicated. There's a lot of uncertainties. The
amplifications can be much larger, and in fact, let me
from here jump to the next slide, and I'll come back
to this slide.
For the seismic analysis, as we're just
discussing, it is a first time DAC approach, and it's
applicable to only non-hard rock sites, but you would
see that the applicant would be required to complete
most of the seismic analysis, including the soil
structure interaction, the amplified response spectra,
overturn the stability analysis.
And we found that this was inconsistent
with the policy issue of level of detail that's in
SECY 90-377. And specifically what it said in SECY
90-377 is that the level of detail should be
equivalent to or no less than an FSAR at the operating
stage for a recently licensed plant.
So at that stage obviously the seismic
analysis has been completed. So we identified this as
a policy issue, but more than that, in the next
bullet, we believe that there are some safety concerns
or safety questions that will not be answered until
after design certification, and as Goutam Bagchi has
described briefly, the AP 1000 plant is taller and
heavier than the AP 600 plant. So there is a question
about the overturning stability and you would have
higher seismic amplifications on soil sites.
CO-CHAIRMAN KRESS: But it's not a safety
concern until the plant is built and turned on, is it?
MEMBER POWERS: Can't one say that just
about anything in the design?
CO-CHAIRMAN KRESS: Well, my point is you
have another chance to say no.
MR. TERAO: Yeah, but, no, we don't.
MEMBER POWERS: Don't you do that on
everything?
CO-CHAIRMAN KRESS: Can't you at the COL
stage -- can't you say, "No, this has a safety
problem. That's not sufficient"?
MR. WILSON: Let me get in here and
amplify what Dave said, is what we're laying out are
ITAAC, in this particular case certain types of ITAAC
that's set forward process and acceptance criteria for
doing the design and then verifying the design.
So if we wrote an ITAAC and they met the
ITAAC, then we would have to find that acceptable, but
if we didn't write one sufficiently robust that would
lead to what we would normally refer to as an
acceptable design, if we were reviewing the design at
that stage, then, yeah, we would have a significant
problem now.
Because in effect, the way the process is
set up is if they met all of the ITAAC we're obligated
to give them an operating license. And so the burden
would be on us then to say --
CO-CHAIRMAN KRESS: Yeah, so what that
means i you'd have a real burden trying to anticipate
all of the things that might be safety related and
being sure you have a DAC and ITAAC written in such a
way that it covers those.
MR. TERAO: That's exactly right. In
fact, what I was going to say is that was the
difficulty when we established the DAC, the piping DAC
for ABWR. I was intimately involved with that. What
we had to do is identify what type of issues,
anticipate what type of issues can come up in piping
and to assure that we have designed an acceptance
criteria that will address how those issues are to be
analyzed at the COL stage.
And it required us to go back and look at
all the bulletins, generic letters. At the time there
were issues with Comanche Peak on piping and pipe
support designs. We did quite a -- and plus, the
five Stone & Webster plants that were shut down.
So there were many issues that we tried to
incorporate in addressing that.
CO-CHAIRMAN KRESS: Well, let me ask you
another question about that. If you're doing the
design certification at the design certification
review stage, don't you have to do that anyway at that
time?
MR. TERAO: I'm sorry?
CO-CHAIRMAN KRESS: Don't you have to know
what the issues of the safety problems with the
seismic are and be able to review the SAR and looking
for those issues and going back and checking all of
the letters?
MR. BAGCHI: This is Goutam Bagchi again.
Dr. Kress, if we had the freedom to change
the base size to anything the applicant wants, then we
could do that. However, we are constrained by the
size. We are constrained by the design. That is the
certified design. The Commission has certified that
design.
MR. WILSON: Also, Dr. Kress, in your
question, you said at the design certification stage.
So understand we don't have that design information at
that stage.
CO-CHAIRMAN KRESS: If you had that
information thought, you would have to do exactly what
he said.
MR. WILSON: That's correct.
CO-CHAIRMAN KRESS: So it would seem to me
like the question is do you do that now or do you do
it later?
MR. WILSON: Yeah, but can you write the
design acceptance criteria to appropriately pick up
all of that information so that the output product is
acceptable, and that's what he's talking about.
CO-CHAIRMAN KRESS: And I think that's a
legitimate question there, is can you do an ITAAC or
a DAC robust enough to cover everything at that time.
MR. TERAO: So I guess the bottom line
here is that we felt that there were too many
uncertainties or unanswered questions that would
remain at the design certification stage for us to say
that this was a viable approach to use seismic
analysis DAC.
Even though it's technically feasible, we
may be able to come up -- and I say technically
feasible with uncertainties here because this is a
first time approach, and we don't know what we're
going to encounter or whether we can anticipate all
the type of problems for seismic analysis.
CO-CHAIRMAN KRESS: And all you're left
with is you have to make a judgment, and it's a
judgment call.
MR. TERAO: That's right, and also
remember that PAR 52 means design finality. So once
the design is certified, this staff -- and this is to
assure regulatory stability and predictability so that
the staff cannot ask anymore questions on that part of
design that's certified at the COL stage.
CO-CHAIRMAN KRESS: But I view DAC as an
exception to that.
MR. TERAO: No. Oh, no, as a matter of
fact. It is not because once we approve the DAC, then
the CO applicant only has to complete the piping
design and verify it using ITAAC.
MEMBER SIEBER: And the only set of rules
come from the DAC.
MR. TERAO: That's correct. That's
correct. The only set of rules come from the DAC.
MR. WILSON: Are you ready for the next
slide?
MR. TERAO: Let me go back to this one.
So we discussed seismic analysis. The
second item here on structural design, what
Westinghouse is proposing is to provide some
preliminary assessment of the key structural elements
for both soil and rock sites at the design
certification stage, and then they would establish
structural design DAC that would be used at the COL
stage to complete the piping structural design.
And then at the COL or after the plant
started construction, then they would verify the as
built structural design using ITAAC.
MR. WILSON: Ready for the structural one?
MR. TERAO: So to address the structural
DAAC, again, this is the first time DAC approach. So
this is a new policy issue, and this is appropriate
for the use of the structural DAC is to be applicable
to all sites, hard rock and soil sites.
The CO applicant would then complete the
structural design. So a substantial amount of design
would remain for the CO applicant to complete, and we
found that the level of detail here again is
inconsistent with the policy issues in SECY 90-377,
where a complete design -- in other words, it's not
consistent with the type of information that's in an
FSAR at the operating license stage for a recently
licensed plant.
And obviously we will lose benefits of
standardization if we don't know what the final design
of the buildings look like.
The approach is technically feasible
though. We believe it can be done, especially for
hard rock sites. For soil sites, we think it's
technically feasible, but there are some uncertainties
that we might anticipate. It gets a little bit more
complicated when you have not completed the analysis
up front.
So we anticipate that there could be some
problems later on for other than hard rock sites.
And finally, for the piping DAC, the
piping DAC is similar to ABWR and System 80 Plus
approach. The only difference is that you can see in
the third column there on the COL application what
Westinghouse is proposing is that the analysis for
leak before break qualified piping be provided after
design certification, and this is different than what
we accepted on System 80 Plus.
For ABWR, obviously they didn't use the
leak before break approach.
Let me address the piping DAC in two
slides. One, we'll discuss the policy issues with
piping DAC and the second slide will discuss the
safety issues.
So the policy issues here, piping DAC was
used in ABWR and System 80 Plus, and both of those are
evolutionary plants. AP 600 completed its piping
design, and obviously it's a passive plant.
So the question was: how was Westinghouse
able to complete the piping design for AP 600 while
G.E. and ABB Combustion Engineering were not able to?
And staff would note that in a passive
plant, there are significantly fewer number of safety
related piping subsystems that have to be modeled, and
most of those, if not all, are inside the containment,
and that the passive plant by nature does not have
large motor operator valves with offset actuators.
They don't have pumps, heat exchangers where the pipe
nozzle is dependent on the vendor design.
So because of the simplicity of the
design, I mean, you do have check valves, but check
valves are relatively easy to make assumptions on its
weight. It doesn't have offset center of gravities.
So you can make reasonable assumptions to complete the
piping design.
And that is exactly what Westinghouse did
for AP 600, but not for AP 1000. The diameters have
changed, and Westinghouse is proposing to use the DAC
for the AP 1000 piping.
Now, the question came up and we've had --
yes?
MEMBER SIEBER: In the AP 1000, the number
of components and the type of components would be the
same. It's just the size that's different. Is that
true or not?
MR. TERAO: That's correct. You may have
a slightly larger check valve.
MEMBER SIEBER: That's right. Larger
diameter piping, larger heat exchanger, larger pumps.
MR. TERAO: But they wouldn't have pumps
and heat exchangers in either passive plants.
MEMBER SIEBER: Right.
MR. TERAO: Right. So the question is:
why could we not let Westinghouse use DAC approach for
AP 1000?
And, again, this is more of a policy issue
because -- what I'm addressing now -- because we have
to go back to the basis for using DAC in ABWR and
System 80 Plus, and the basis for using the DAC or a
rapidly evolving technology, which doesn't apply here,
but the other basis was that as built and as procured
information was insufficient to complete the design.
So for evolutionary plants when you had the motor
operator valves or heat exchangers or pumps and didn't
know where the nozzles were going to be on these
vendor specific equipment, it wasn't reasonable for us
to have General Electric and Combustion Engineering
route the piping when the piping routing can change
significantly when they actually purchase the
equipment.
So the bases was as built land as procured
information is insufficient to complete the piping
design, and that we felt did not apply here with the
passive plant; that the information is sufficient to
complete the design.
Now, can it be done? Yes, it can be done.
We've proved that it can be done, but unless we raise
to the Commission that there's a different basis here
for approving the use of piping DAC for AP 1000, we
felt that this is a policy issue that needs to be
brought to the attention of the Commission.
CO-CHAIRMAN KRESS: Could be done, but it
would have to be on a different basis, and then you've
done it before.
MR. TERAO: That's true, and we asked
Westinghouse if they can provide us that type of
basis, that rationale for using piping DAC since the
basis of as procured information being insufficient
does not seem to apply in their case.
A separate issue, but nonetheless just as
important, is that the leak before break approach is
inconsistent with SECY 93-087. On my previous slide,
I showed that they would provide the leak before break
analysis at the COL stage.
In SECY 93-087, it states that leak before
break may be used by establishing bounding limits
using preliminary analysis results during the design
certification phase and verified by ITAAC at the COL
phase.
So that's the one difference that we felt
that for leak before break on AP 1000, that it was not
consistent with the SECY 93-087 policy.
And there are some safety issues, as well,
which I'll discuss in the next slide that's related to
leak before break.
Piping safety issues. We found that
there's actually three areas that we identify that's
related to safety, that identify safety issues. Some
of these we've discussed with Westinghouse, and
actually some of these we have not because they
actually came out as a result of our discussions on
developing our internal position on these issues.
So I apologize to Westinghouse that they
may not have heard some of these issues, but the first
issue on leak before break, when we approve leak
before break, it is usually for an operating plant,
and for an operating plant, you know the seismic
stresses. You know the piping materials. So for a
standard plant, for a paper plant, the staff had to
address three critical areas to assure that leak
before break can be used.
And those three areas were the margins on
load, the margins on flaw size, and the margins on
leakage rate. And what we approved for System 80 Plus
was System 80 Plus established bounding surveys for
those three areas of loads, flaw size, and leakage
cracks.
And then what System 80 Plus did is they
used preliminary stress analysis at least for those
leak before break lines or the critical size leak
before break lines to verify that there was sufficient
margin available so that when the plant is finally
built and they get the final stress analyses and the
material properties, that we felt that there was
enough margin to not have any questions raised at that
time.
But Westinghouse is proposing to do this
analysis at the COL stage. So what we feel is that
these margins will not be confirmed at the design
certification stage. So we believe that that's one
issue that does not meet 5247(a)(2).
Another issue, the second issue here
emerges as a result of the leak before break
uncertainty, and this is an issue on subcompartment
pressurization and flooding because without knowing
what is the minimum size line that we will approve for
a leak before break, the question is: well, what is
the pressurization one would use? What size pipe
break would one use to establish your subcompartment
pressurization for flooding effects?
And in fact, this was an issue that was
raised by the ACRS back in June, in an ACRS letter
dated June 16th, 1992. So we were well aware of ACRS'
concerns at that time of compartment pressurization
and flooding of DACs, that it be addressed as part of
the DAC.
And Westinghouse has not addressed this
issue. It's not say that it cannot be resolved, but
at this time, we have not discussed with Westinghouse
how thy plan to address those issues.
The third issue is a relatively new issue,
and this deals with that the passive piping design
needs to be finalized because the thermal hydraulic
characteristics of the passive safety systems are very
sensitive to the piping volume.
And so when they change their piping
sizes, we need confirmation that those final piping
sizes are going to be the final piping size, and they
will meet the seismic stresses and thermal stresses
and the routing.
So without that confirmation, it does
interface with the thermal hydraulics area.
So in summary, just summarizing what I
talked about, so the seismic analysis DAC, what we
found was there were safety issues involved and policy
issues involved, but technically it's a feasible
approach with some uncertainty.
With a structural design DAC, it's mainly
just policy issues, and we feel that it's technically
feasible even though there are some uncertainties
since this is the first time we're using this
approach, and for the piping design, we find that
there were policy issues, safety issues, but we
believe it is technically feasible.
So that concludes my presentation.
MR. WILSON: So, Mr. Chairman, in summary
then, similar to what Mr. Drozd pointed out, what the
staff is recommending is that we not accept the use of
additional design acceptance criteria on AP 1000.
And with that we'll turn it over to you.
If you have anymore questions in this area before we
turn the presentation back to Westinghouse.
CO-CHAIRMAN KRESS: Well, I think you've
answered most of my questions. Do any of the other
committee members?
MR. LYON: This is Jim Lyon from the
staff, and before Westinghouse comes up or, you know,
we're making this shift, we presented this information
to Westinghouse a couple of weeks ago. We've been
having some conversations with them, and one of the
things that we had raised to them is that in the
seismic area especially, seismic area, that doing a
design acceptance criteria for the seismic area for
the soil sites could be alleviated by limiting the
plant to a hard rock site, providing us the type of
information that they were going to provide us, you
know, and then providing us the structural design on
that.
Westinghouse has been talking to us, and
they actually sent us a letter this morning that talks
about, you know, one of their ways of maybe providing
limiting AP 1000 to a hard rock site, then a plant
that would be built at an other than hard rock site
would then have to be reviewed at the COL stage for
that other than hard rock site. It would become
basically a site parameter. One of the site
parameters would limit it to a hard rock site.
That's something that then could be
reviewed at the combined license stage for an
applicant that wants to reference the AP 1000, but put
it at a site other than a hard rock site with all of
the site parameters.
CO-CHAIRMAN KRESS: What that does for you
other than having a DAC is it gives you a handle to
say no?
MR. LYON: Right, and what it allows us to
do is resolve all of the issues at the combined
license stage prior to issuing --
CO-CHAIRMAN KRESS: You can make a safety
review then.
MR. LYON: Right.
CO-CHAIRMAN KRESS: And see if it
satisfies your criteria, and if it doesn't, you can
say no.
MR. LYON: Right.
MR. WILSON: In that situation, similar to
what you asked before, then they would -- we would
have an opportunity to review the actual design before
the granting of the combined license.
CO-CHAIRMAN KRESS: And that's what I was
misinterpreting the DAC to allow you to do, but
apparently it doesn't, but this process would allow it
to --
MR. LYON: That's right, and so we're
considering -- and I think in that scheme, then if you
take the piping DAC by itself, we may be able to work
with that, and that's something that we're kind of in
discussions with Westinghouse on now to try and move
forward.
So I think that's probably a good, you
know, lead-in to segue.
CO-CHAIRMAN KRESS: A good segue into the
Westinghouse.
MR. LYON: Into their discussion.
MEMBER SIEBER: I wonder about that a
little bit though. If you limit this to just hard
rock sites, I don't think there are that many hard
rock sites around, are there?
If you look at river sites, you have to
have cooling water somehow. If you look at river
sites, every one of them is -- or not every one, but
a log of them have flood plains there. So that's not
hard rock. In fact, soils liquification is pretty
high on those kinds of sites.
Building them on pilings is not good
because that effectively just raises the height of the
plant, which seismically makes it worse.
Lake and ocean front sites, those are
pretty tough, too, and so it may solve the licensing
problem and the technical problems that you have, but
then you end up with a certified design that nobody
wants to buy.
MR. WILSON: Well, what it would mean it
would be more likely if the plant is by reference,
that there would be additional design work that would
have to be reviewed during the combined license review
process.
MR. BAGCHI: As I recall though, Seabrook
site is hard rock. Catawba is hard rock. There are
hard rock sites.
MEMBER SIEBER: There are some sites, but
most of them are not, it seems to me.
CO-CHAIRMAN KRESS: I presume Westinghouse
is sufficiently sagacious to design their system to
make use of the sites they want to make use of even
though they may have not put an end to this SAR yet.
They probably would design it to make use of the
various sites.
MR. WILSON: Why don't we let Westinghouse
answer that?
CO-CHAIRMAN KRESS: With that we'll turn
it over to whoever is taking the floor for
Westinghouse.
MR. ORE: Good afternoon. My name is
Richard Ore. I'm with Westinghouse. I have
responsibility for the seismic work, for the
structures, for the piping.
I would like to start by thanking the
staff for sort of an excellent summary. It allows me
to jump through some of my presentation very quickly,
but I would like to sort of make a few points from it.
Firstly, --
MR. CORLETTI: One second, Richard.
This is Mike Corletti.
His presentation starts on about page 15
of the handout.
MR. ORE: And the first slide I'm going to
use is number 11.
Firstly, to address the --
CO-CHAIRMAN WALLIS: What is this? Excuse
me.
MR. ORE: We did indeed --
CO-CHAIRMAN WALLIS: Excuse me.
CO-CHAIRMAN KRESS: We're still trying to
find it.
CO-CHAIRMAN WALLIS: This thing leaps from
slide five to slide 14.
MR. ORE: There should be a package that
is entitled AP 1000, approach to design acceptance
criteria.
MEMBER SIEBER: Yeah, that's slide 15.
MR. ORE: And that was slide 15, and in
the handout package, four back, is one that is
identified as slide 11.
CO-CHAIRMAN KRESS: N, that's not in ours.
MR. CORLETTI: Richard, you added that.
I'm sorry. I didn't put that in there.
MR. ORE: My apologies. Would you arrange
for a hard copy for them?
We had submitted a request in Phase 2 to
review the approach of using design acceptance
criteria on both structures and seismic analyses. We
had one meeting with staff early last year, and we
have had another meeting recently and additional phone
calls and resulting from that we transmitted a letter
yesterday that said that in design certification, in
the design control document that we propose to submit
at the end of March, we will request design
certification including seismic analysis and
structural design for a hard rock site.
Now, in response to the question earlier,
on AP 600 we did a review of 22 existing nuclear power
plant sites, and we found 50 percent of them --
MEMBER SIEBER: Are hard rock.
MR. ORE: -- the nuclear island is founded
on rock, yes. Sometimes it's down at the -- our
excavation depth is 40 feet. I think there were one
or two cases where the rock was down 50 or 60 feet,
where perhaps I've taken credit for a rock site
because you would go down to it with your foundations.
MEMBER SIEBER: Yeah, i think I know where
that one is.
MR. ORE: Fifty percent of the existing
plants are on rock.
MEMBER SCHROCK: Is there any potential
ambiguity about the definition of a hard rock site?
MR. ORE: I don't believe so. There is a
site parameter that is established in the first
section of the design control document. For AP 600,
we requested and obtained certification for sites
where the shear away velocity of the soil exceeded
1,000 feet per second.
For AP 1000 we have changed that to a
shear away velocity exceeding 3,500 feet per second.
In the AP 600 we did a series of analysis. One of
them was at 3500 feet per second. In order to
demonstrate there that soil structure interaction
effects were negligible.
In fact, in one of the previous revisions
of a standard review plan I believe that that cutoff
of 3,500 feet was given then for whether you needed to
look at soil structure in traction or not.
CO-CHAIRMAN KRESS: So given that comment,
basically what Jim Lyon suggested, it looks like it's
going to come about.
MR. ORE: Yes.
CO-CHAIRMAN KRESS: And all we have to
deal with then is the piping deck.
MR. ORE: That's correct. So now I would
like just to talk about the piping deck unless there's
any questions related to the seismic analyses or
construction design.
CO-CHAIRMAN KRESS: No, I think we're h
pretty good shape there. So we can go straight to
the --
(Laughter.)
MR. ORE: Now I will continue on the
piping, and this now starts at slide 22 of the handout
package that I believe you have. And this will be
fairly brief because it's been covered by the staff
extensively.
The DAC and ITAAC approach was permitted,
was used for ABWR and for System 80 Plus. The way
it's done is in the design control document we agree
on the design acceptance criteria and for the piping
design criteria and the analytical methods and the
acceptance criteria.
The implementation is verified in the
ITAAC, and basically the verification is that there is
a series of ASME design reports on every piping line,
and that the scope of that is clearly defined in ASME.
It includes any reconciliation of as built changes,
and in addition, we have an evaluation against the
acceptance criteria for leak before break to
demonstrate that those lines that we intended to apply
LVP to were qualified to the acceptance criteria.
CO-CHAIRMAN KRESS: What about this
comment that small changes in the piping layout or
sizes could have significant safety implications
because you're highly driven by passive systems that
rely on natural convection driving forces?
MR. ORE: We don't believe so. I've got
some additional slides. The AP 1000 -- let me just --
I think I prefer just to defer that if I could get to
it, if I may.
MR. CORLETTI: Richard, this is Mike
Corletti. If I could just speak to that a little bit.
In regards to changes in the pipe routing
and how it could affect the passive system performance
from a thermal hydraulic safety point of view, those
characteristics are already captured in the ITAAC and
are set in other portions of the ITAAC. So
information that we had a range of piping resistance,
L over D information, for the inlet and the outlet of,
say, the core make-up tank lines.
The elevation of those core make-up tanks
are also set. So from the things that would affect
the safety analysis from like elevations and
resistances, those are already covered in the ITAAC.
MR. ORE: As described by NRC, for AP 600
we did extensive analyses of all safety related
piping. I believe it was all piping larger than two
inch in diameter. In doing those analyses, we made
certain assumptions on vendor data, particularly
things like valves, valve weight, valve operators.
These calculations were audited by NRC and
found acceptable. The question we ask ourselves
though is: what have we gained when we end up with a
verification process, the ITAAC, that is virtually
identical to the ITAAC required of the other vendors,
of the other plants?
This now goes to slide 25. What we gain
from our AP 600 work, very definitely we gained
confidence that all of the line routings work in the
audit. So with NRC staff we got agreement on the
methods and the implementation particularly of all of
those methods.
We believe on AP 1000 we gain significant
benefit from what we did on AP 600. Our line routings
are substantially the same obviously, with a slightly
bigger diameter pipe, with some slight changes, but
nothing significant.
The analyses will use the same methods as
AP 600, and the design specifications, transient
conditions, the operating conditions are all
substantially the same as AP 600.
Now perhaps I can go back to the question
of the LBB lines. Mike had addressed the response on
thermal hydraulics may be different. Clearly with the
larger diameter there's going to be slightly different
stress conditions, but we do not believe that there
will be significant changes required.
On AP 600, in the ITAAC that we have on
each system, we have a requirement that we meet all of
the acceptance criteria, and in the extreme, if we had
to reroute pipe, we would have to still meet all of
the acceptance criteria, which includes subcompartment
pressurization issues, flooding, all of the sort of
pipe rupture hazard type issues if you have to reroute
a line.
We recognize that there are some sort of
issues related to subcompartment pressurization that
we don't want to defer until after the plant is built.
Therefore, we have included in the design control
document and a preliminary mock-up of this was
actually informally given to the staff month ago, and
we specifically say combined licensed applicants
referencing the AP 1000 certified design will complete
the leak before break evaluation, and this is sort of
at the time of the combined license application, and
they will demonstrate that they meet bounding curves
that are provided in the AP 1000 design control
document.
An example of this, which will be in the
submittal, it was in the preliminary one a month ago
and will be in the one at the end of March. We define
for the piping analyst when he is qualifying the
system a bounding curve that shows if he's at an area
below the curve, designated here as .1, and this is a
relationship between the stresses under operating
conditions to the stresses under the safe shutdown
earthquake condition; if he's below that curve, he
meets to leap before break acceptance criteria. If
he's above it, he doesn't.
These bounding curves include all of the
margins required by NRC. There are margins on load
and margins on leak rate built into these curves. So
if the combined license applicant can demonstrate that
he meets these curves, then his piping layout
satisfies the leak before break requirements.
MEMBER SCHROCK: Could you restate the
meaning of those terms on the axes of that graph?
MR. ORE: Certainly. On the X axis the
normal stress. This is the stress, the membrane
stress in the pipe, membrane plus bending, and normal
operating conditions.
CO-CHAIRMAN KRESS: The internal pressure?
MR. ORE: No, this is -- well it's
internal pressure and it's dead weight. It's normal
thermal, and under a given condition, he identifies
what his stress in the pipe is, and he can then go up
to the bounding curve and see this is the maximum
stress then that I can accept when I combine it with
the safe shutdown earthquake, the faulted conditions.
And if, indeed, he's above this curve, he
does not meet the criteria for leap before break. If
he's below this curve, then he does. And this is
similar to the approach that was used on the CE
plants.
CO-CHAIRMAN KRESS: The only difference
between the two axes is the loadings you'd get from
seismic?
MR. ORE: Pardon?
CO-CHAIRMAN KRESS: The only difference
between these two axes is the loading you'd get from
seismic?
MR. ORE: One of them is normal operating.
The other is faulted. I think the primary difference
is just seismic, yes.
MEMBER SHACK: To be certain, to make sure
that you have enough stress to get a measurable leak
compared to the size of stress it's going to take to
actually fail the pipe.
MR. ORE: That's correct.
MEMBER SHACK: Which is sort of what
you're comparing.
MR. ORE: And that's why actually if the
operating stress is higher, you can go to a higher or
safe shutdown.
MEMBER SHACK: It's better news, yeah.
MR. ORE: It doesn't sound right, but
effectively you're stressing the pipe enough under
normal operating conditions if there is a crack, it
will be a visible leak.
CO-CHAIRMAN KRESS: And why do these
things flatten off at the ends? Is that just a
safety --
MR. ORE: What we do is we take sort of
two points actually sort of as a lower extreme and an
upper extreme based on sort of typical practice. I
think the upper extreme is the sort of highest that
you can go under ASME.
CO-CHAIRMAN KRESS: So you know you're not
going to be --
MR. ORE: You're pretty much up to yield
at that stage.
CO-CHAIRMAN KRESS: Okay. I understand.
There's just not any significance to those lines
really since you're not going to be in there.
MR. ORE: You will never be able to
qualify a line outside these ranges.
So sort of in conclusion, we intend, we
are proposing to follow a path that we believe is
consistent with the approach used on ABWR and System
80 Plus. I was very pleased to see the overhead from
Dave Terao that said he believes that such an approach
is technically feasible.
Yes, there may be some sort of discussion
on it, but we believe the overall approach is
appropriate.
CO-CHAIRMAN KRESS: How do you argue
against or debate their point that passive systems,
piping is not nearly as complex as, say, System 80
Plus piping, and therefore, it shouldn't be that
difficult for you to really lay out a similar thing to
AP 600 that you did?
MR. ORE: We have effectively laid it out
the same as AP 600. Now, systems are clearly simpler
because of the passive systems. Unfortunately, we do
still have certain sort of valves and things like
that, and Part 52 specifically says you do not have to
select vendors on things like that.
So you cannot do final analyses anyway,
and the question really is: well, why do very
elaborate initial analyses for review by NRC and then
have to redo them at the as built stage?
CO-CHAIRMAN KRESS: So you think it might
fit that second criteria there. You really can't do
it until you get the as procured and as built?
MR. ORE: You are not going to be able to
have final confirmation until you have done the as
built.
MR. CORLETTI: As you all know -- this is
Mike Corletti again -- as much as I'd like to agree
that we only have check valves in our plant, we do
have quite a bit of motor operator valves. All would
have to be procured. All with end to end dimensions
and centers of gravity.
So whereas on AP 600, we did do analysis
making assumptions, we learned, I think, the hard way
that we still had to come back to that at the end and
redo it again with all of the final, as procured valve
information.
CO-CHAIRMAN KRESS: If you were to build
an AP 600, is there some chance that you'd find
yourselves in a bind because you said one thing in
your FSAR or your certification rule and you find out
that when you get ready to build it that your
assumptions were bad?
MR. CORLETTI: We certainly hope not, but
again, our DAC approach and the fact that we've gone
through that on AP 600 gives us significant confidence
that even for AP 1000 we're not going to have to face
that same problem.
MR. ORE: Okay. That concludes my
presentation.
MR. BURKHART: Mr. Ore, this is Larry
Burkhart, the project manager.
You say that it's similar. Can you
quantify the differences between the approach compared
to the System 80 Plus?
You're saying it's similar. Is it exactly
the same? What are the differences?
MR. ORE: Our intent is that it be exactly
the same, yes, that we would, indeed -- we have got
all of the design criteria in the design control
document. We will provide a similar reference in the
CE approach in Chapter 1, I believe it is.
There's a reference to all of the
subsections containing design criteria for piping. We
would have a similar table and that all becomes Tier
2 stuff.
MR. BURKHART: So it sounds like your
approach is the same, not just similar. I just
thought there --
MR. ORE: The intent is that it be the
same.
MR. BURKHART: Okay. I just wanted to
know --
MR. ORE: Without going line by line, I
did not want someone to pick me up and say, "Well,
something is slightly different.
MR. BURKHART: I just wanted to know if
there were any differences that you knew of.
MR. ORE: No, the intent is that it be the
same.
MR. CORLETTI: And this is Mike Corletti.
It is certainly our intent to follow that
precedent, and we'd be very interested to work with
the staff in regards to places where we perhaps have
deviated. We don't believe we have, but we certainly
look forward to working with them to resolve that.
CO-CHAIRMAN KRESS: Any other questions on
the DAC from the members?
(No response.)
CO-CHAIRMAN KRESS: At this point in time
I propose we take a break. We were supposed to be
scheduled for one at three, but we can start a little
early, and I'd say be back from the break at three,
and we'll start on the applicability of the scaling.
Is that all right with everyone?
Okay. Let's recess for a break.
(Whereupon, the foregoing matter went off
the record at 2:43 p.m. and went back on
the record at 3:03 p.m.)
CO-CHAIRMAN KRESS: Let's reconvene.
Before I turn this over to you to continue
with the tests and scaling, I think Mike Corletti
wants to make a few words for Westinghouse, if you
don't mine.
MR. CORLETTI: Yes. I just thought I'd
just take two minutes to brief you on really where we
think we are in the overall program for AP 1000.
As you know, we started the pre-
certification review. Really it began last December.
We've been answering the RAIs with the staff, the
staff RAIs. We believe we're just about to the end of
this phase.
As part of that, we've just recently
submitted to the staff a red line and strikeout, what
we call our highlight and strikeout version of our
DCD, of the AP 600 DCD, that really illustrates the
differences between AP 600 and AP 1000, and this is
one of the books.
There's 20 volumes, and it's being
provided to the staff now as they plan the design
certification review, as they get their estimates to
see what they think it will take to certify AP 1000.
CO-CHAIRMAN KRESS: That's that little
line with the pointed end on it
MR. CORLETTI: Yes, here. We're right
here in the number three, pre-certification review,
and we are preparing and we are prepared to submit our
application at the end of March. March 28th is our
scheduled date when we will be submitting the complete
design certification document and the PRA.
What we've submitted so far is everything
except for Chapter 15 and except for the PRA.
So our plan is to submit March 28th. The
other two lines you see there is our projections for
where the rest of the marketplace is as far as
utilities are now getting ready to submit early site
permits, and hopefully a COL application, which we
believe is a critical path item.
Our goal is to make AP 1000 design
certification not to be critical path, so to be
finished prior to the utilities having their early
site permits and, therefore, AP 1000 would be a viable
product for them.
CO-CHAIRMAN WALLIS: How many utilities
are involved in this?
MR. CORLETTI: There are several that are
considering early site permits at this time, at least
two or three that we know of.
CO-CHAIRMAN WALLIS: Haven't made a
downpayment or anything.
MR. CORLETTI: No, not to our knowledge.
(Laughter.)
MR. CORLETTI: But we're working on it.
So that's all. I can pass this around.
We have this on a CD, which we could also make
available to the ACRS. I can get you copies if that's
appropriate.
Paul we could do that.
CO-CHAIRMAN KRESS: Well, I think we
definitely want that.
MR. CORLETTI: Okay, great. And you can
look for the complete edition coming March with.
CO-CHAIRMAN KRESS: Okay.
MR. CORLETTI: So thank you.
CO-CHAIRMAN KRESS: Thank you.
Okay. With that I'll turn it back to you,
Andrzej.
MR. DROZD: The last two remaining groups
of issues from Phase 2 review are testing or a more
precise applicability of AP 600 testing to AP 1000,
and the applicability of safety analysis codes that
were used in AP 600's review into 1000.
Today, this afternoon Dr. Steve Bajorek
and Dr. di Marzo will address testing and scaling
issues that were analysis and applicability and its
applicability to AP 1000 design. I just want to
briefly summarize major points that staff will be
including in a proposed SECY paper to Commission on
safety and testing.
First of all, we will notice that there
are no new phenomena identified from what we see.
That is, what other questions and problems there were
with the design during AP 600, the certificate review,
we see the same issues being addressed or needed o be
addressed during AP 1000.
Therefore, in general, in general, we do
see that whatever was used to analyze and/or support
codes during AP 600 certificate review is applicable
to AP 1000 review.
That doesn't mean that we do not have
certain concerns. First of all, we notice that some
physical models that are the basis for various
subroutines and various analysis used in analyzing
safety behavior of the plant needs to be either
modified or verified or both, and specifically liquid
entrainment model is one of the outstanding issues
that we think needs to be verified and/or modified
because it's a basis for many other applications, for
transient analysis or small break LOCAs.
CO-CHAIRMAN KRESS: It's the same
entrainment model they had for AP 600?
MR. DROZD: That is the same issue.
CO-CHAIRMAN KRESS: And we found AP 600
acceptable because of what?
MR. DROZD: Well, Dr. Bajorek will show
and prove the application for this particular model
may increase -- the application of this model to AP
1000 increases certain uncertainty of behavior.
Therefore, we need some improvements.
CO-CHAIRMAN WALLIS: There's one liquid
entrainment model or it applies everywhere liquid is
entrained?
MR. DROZD: Might as well jump in.
(Laughter.)
MR. BAJOREK: This is Steve Bajorek from
Research.
There are two models that we're going to
discuss. I believe that in NOTRUMP and in other codes
that Westinghouse would use there are different models
and correlations in that code which would try to model
both of these processes.
MEMBER SCHROCK: Isn't the question where?
I mean, there are different entrainment problems in
the whole system. Which one was we talking about?
CO-CHAIRMAN WALLIS: Yeah, there's
entrainment from the core. Let's say there's gas
coming out, vapor coming out the core. Does the
entrain drop?
Then there's the question of whether or
not drops are entrained into the ADS-4 line. From a
pipe there's maybe a completely different problem.
MR. BAJOREK: There's two different
processes. Let me start and get to it.
CO-CHAIRMAN WALLIS: Did I force you to
go? Did I drive our friend away?
(Laughter.)
CO-CHAIRMAN WALLIS: As soon as I
mentioned simple hydraulics.
MEMBER POWERS: The man has the good sense
to evacuate the scene.
(Laughter.)
MR. BAJOREK: Okay. There are two
packages coming around. The first one will say 81,000
top down and bottom up scaling evaluation, and there's
another one coming around that we'll get to in a
moment.
We'll start on the top down/bottom up
scaling, though I'm going to switch the first couple
of overheads, and I want to explain exactly what the
role of research was in the review.
We were asked to participate by reviewing
the PIRTs that Westinghouse had supplied, look at the
test programs that they applied to the primary system,
and determine to what extent Westinghouse's contention
that all of the test programs that were submitted as
part of the AP 600 still were valid and could be
applied to the AP 1000 conditions.
So we started off by taking a look at the
PIRT. We basically agreed with what we saw there at
Westinghouse. They increased the rankings of some
processes; others stayed the same. By and large, we
did not have any major objections there.
We asked them to add condensation induced
water hammer and deal with that, and that was pretty
much what we ended our analysis of the PIRT.
And then we looked at scaling for the
primary system tests. We looked at AP 1000. We
revisited AP 600 because we wanted to make sure that
as we did the evaluation, we were consistent with the
decisions, and we were seeing the same types of
problems that we saw in the AP 600.
We also took a look at SPES APEX and we
threw ROSA into the mix as well. To make sure that I
clarify what we looked at, we stuck primarily to the
test data and the test programs. We did not get
intimately involved in looking at the codes or the
analysis as part of our review.
We did not look at the containment issues.
These were things that NRR decided that they would
prefer to handle themselves.
So I'm going to look primarily at the
test, the range of conditions. Is that data
applicable now so that you can take a NOTRUMP, a LOFT
TRAY (phonetic), and a COBRA TRAC, whatever code they
might apply, and use these experimental data for
validating --
CO-CHAIRMAN KRESS: You're looking
strictly at the integral effects test and not any of
the separate effects?
MR. BAJOREK: We looked at the separate
effects test. I'm not going to talk about those today
primarily because the separate effects tests were
either unchanged relative to how they would be
applied, meaning ADS-1, 2, or three was not changed in
AP 1000. The conclusions from those tests still
apply.
Westinghouse did produce some scaling
rationale and some arguments for the separate effects
test. They showed that they were very close or within
the range for AP 1000.
In some cases, for example, the PRHR, if
you remember, that was a test, three vertical tubes
and they developed a heat transfer correlation. To
comment on those tests was a bit difficult for the AP
1000 because approval really hinged on their ability
to model the PRHR component in the ROSA facility.
So we're starting to get away from the
applicability of the tests, even in AP 600. So the
focus of our work was on the integral tests.
How we're going to structure our
presentation this afternoon is along the lines of the
three main elements that we perform for scaling. I'm
going to talk initially about top-down scaling and how
we evaluated ROSA, SPES, APEX and compared that to AP
600 and AP 1000.
Dr. de Marzo is going to talk about some
supplemental calculations that he performed to take a
look at the transient process in going from ADS 1, 2,
3 through the IRWST.
What I will be doing for the top-down
scaling looks at many of these processes in their
steady state performance mode. What Dr. di Marzo will
take a look at is the transient effects in going
through some of these important changes in the
transient.
Once we get done with that, we're going to
focus on some of the bottom-up scaling issues, and
that's where we're going to spend more time talking
about entrainment. So what I'd really like to do is
to go through the top-down type of scaling in a
relatively rapid fashion so that we can get to some of
the things that I think are going to be more
contentious.
I think on the first page, I just wanted
to lay out a few of the things that as we go through
the top-down scaling those parametric changes in the
AP 1000 that tend to affect most of or many of the
scaling parameters, certainly the power.
Other things that come into effect are the
larger volumes in the AP 1000. The steam generator is
larger. We've got larger CMTs, and a larger pressure
rise.
So we have a larger primary system, larger
pressurizer.
One of the things that was done to the AP
1000 is a change to the resistance of several lives.
In particular, the ADS-4 line resistance was reduced
till it was only about 28 percent of what it had been
in the AP 600.
That's good. The idea is to try to get
the system to blow down faster.
They also made it easier to get water into
the system by reducing the line resistance from the
CMT through the DVI line. The thing I want --
CO-CHAIRMAN KRESS: Don't they do that by
changing the diameter?
MR. BAJOREK: The diameter, and also there
is an orifice change in the CMT itself to make it
easier. The thing I want to note here as we start
going to the top-down scaling, well, in the tests we
had whatever the AP 600 was, or as close as they could
get to it in the facility, but now we've made change
to make it easier to get water out than it is to get
water in, and that's going to cause a little bit of a
problem in one part of the -- in one period of the
transient.
Where are we going to lead with this? I
think as Andrzej was starting to mention, by and large
we're going to find that the tests that were run for
AP 600, SPES, ROSA, and APEX largely have many of the
characteristics that show that they are acceptable for
use with the AP 1000. That's good. That covers much
of the transient.
However, we find two major exceptions, and
we feel that those are the things that we're going to
have to deal with in a lot more detail in Phase 3 of
the review, and those are the entrainment processes as
they affect the flow quality leading to the ADS. A
look of these separately in terms of entrainment that
occurs in the hot leg, something that might be
important for one or two inch breaks where we have
levels up into the hot leg, and this is our primary
mode of entrainment that carries liquid into the ADS-
4.
For the DVI line break, which leads to the
minimum vessel inventory and is the most severe
accident for AP 1000, the process that we're going to
be most concerned with is this upper plenum pool type
entrainment where we have a two phase level above the
core plate. Gas is bubbling through this liquid
carrying droplets along the streamlines, again, out
through the ADS-4.
As we go along, we're going to talk about
these processes. What Marino is going to take a look
at is what's the effect of this flow quality as it
gets to the ADS, and how does that affect the
transient nature of AP 1000?
Top-down scaling. There were several
different methodologies --
MEMBER SCHROCK: I guess --
MR. BAJOREK: Sorry?
MEMBER SCHROCK: I wonder if this is the
right time or maybe it shouldn't be left until later,
but the modeling of this in APEX experimentally showed
that you get a sloshing back and forth in this line.
It doesn't look like the picture that you're showing
here, but is something quite different from that.
MR. BAJOREK: We're getting ahead a little
bit.
MEMBER SCHROCK: Well, okay. So you don't
mean literally that what you're looking at is a steady
flow that has this kind of configuration?
MR. BAJOREK: No, no. We'll have some
water in the hot leg. It will have a different
regime, and I'd like to talk about that when we get to
it.
MEMBER SCHROCK: A pulsating discharge.
MR. BAJOREK: Yes.
MEMBER SCHROCK: Okay.
MR. BAJOREK: There were several
methodologies that were proposed for doing top down
scaling for the AP 600. Westinghouse proposed, too,
that they used for the AP 600 one that I would prefer
as a multi-loop top down scaling, another a single
loop.
In the W cap that was submitted in support
of AP 1000, they submitted another scaling methodology
different from the first two. So I would characterize
that as the third methodology.
INEL produced a scaling methodology for
the NRC, and in addition, there was work done by
Wolfgang Wolfe to come up with another scaling
methodology. Each one has different ways of looking
at the system. Each one has their good points and
their bad points.
For doing an independent scaling analysis
for this, we decided we wanted to stay independent.
So we stayed away from the Westinghouse methodologies.
We looked at INEL and the work that was done by
Brookhaven and decided that it was clearer but faster
to use the INEL scaling methodology and went down that
path.
That methodology takes AP 600 or the AP
1000 transient, divides it up into five major periods,
eight subperiods. We divided the AP 1000 transient up
in the same type of detail and looked at two different
scenarios, double ended guillotine break of the DVI
line and also the one inch break, and developed the
scaling parameters for each one of those scenarios.
CO-CHAIRMAN KRESS: Now, with the top
down, you mean you write the equations for energy
conservation and momentum and continuity, and you non-
dimensionalize those, and you end up with your pi
parameters.
MR. BAJOREK: That's exactly it.
CO-CHAIRMAN KRESS: Based on reformulating
for some figure of merit of interest, like pressure or
level or something. That's what's meant by top down.
MR. BAJOREK: Yes. Yes, sir.
The INEL looks at primarily the energy
equation and the mass conservation equation to develop
a pressure rate of change in a tank level equation, as
they call it, basically a mass balance in the vessel,
which is one of the things I liked about that
methodology, because it focuses on mass in the vessel
through much of the transient.
CO-CHAIRMAN KRESS: I don't like reopening
old sores and things, but I don't think I ever got a
good response to my question for AP 600, and I'd like
to raise it again.
Just why is it you feel that that ratio of
pi groups at range .52 is an acceptable range? What
is it about that range that tells you it's acceptable?
Is that a judgment or do you have some technical basis
that says that if your pi groups are in that range,
you don't have significant distortion of the equation
so that you're outside of some limits?
I don't think I've ever received a good
response to that question yet.
MR. BAJOREK: In starting this, I looked
for that answer in the previous documentation, and I
was unable to find that. My use of it here is
primarily historical.
CO-CHAIRMAN KRESS: Yeah, that's the way
it was used before.
MR. BAJOREK: Now, what I've tried to do
here is at least make it scrutable and say, well, this
is the criteria that as I step through the scaling,
I'm going to look for numbers to fall within this
range, say that they're okay. I keep an eye on things
being close.
CO-CHAIRMAN KRESS: You've got to have
some criteria for these pi groups to say whether
they're acceptable or not. You know, why not .5 to
1.5 or .75?
You know, the question is why this choice.
MR. BAJOREK: From the top down scaling I
really don't have a good answer for that, although I
want to add that when we start to look at the bottom-
up scaling, my opinion is that this is insufficient.
You need to not only look at this relative change in
some number, but also if you're headed to some type of
a cliff, if you're simply looking at a froude number,
well, this will give you a nice, warm feel that it's
in about the same type of area of a flow regime map,
but doesn't tell you if you just inched over a
boundary or not.
CO-CHAIRMAN KRESS: Right, right.
MR. BAJOREK: So in the bottom-up scaling,
I try to scrutinize this a little bit more that --
CO-CHAIRMAN WALLIS: That's because of
CCFL. I mean, if the criterion is one and you happen
to be two, that makes a tremendous difference.
CO-CHAIRMAN KRESS: Yeah, yeah.
MR. BAJOREK: Right.
CO-CHAIRMAN KRESS: That leads me to my
other question about this, and maybe you can file it
away, Professor di Marzo, and that is the pi groups,
in my mind, are kind of coefficients on partial
derivatives of these overriding equations, and there's
no reason in my mind that each partial derivative
should have the same influence as the other partial
derivatives.
So why should I have one range that I say
is acceptable? It seems like that range ought to be
specific depending on the pi group I have. And also
carrying on there, it looks like one could make
sensitivity studies with pi groups, and you know, if
you had the right way to do it, and you could find
out. You could actually pinpoint some range that
would give you the same change, percentage change in
the endpoint thing.
You know, I don't see any of that in this
at all, and that's what's bothering me about it.
MR. BAJOREK: Okay.
MEMBER SCHROCK: I had one other comment.
That is that I think the INEL and Brookhaven scaling
began with the HEM equations.
MR. BAJOREK: They did, yes.
MEMBER SCHROCK: And this leaves open the
question of the adequacy of scaling decisions where
non-equilibrium effects are of significance. I never
heard an answer during the AP 600 discussions that
resolved that question, and I don't know if we're
headed towards one now or not, but I guess I might
keep asking it.
MR. BAJOREK: Okay. I won't have an
answer for that today.
MEMBER SCHROCK: Okay.
MR. BAJOREK: But just to point out how
we've addressed top down scaling in this review, very
similar to what we saw for the AP 600. We stayed
towards this acceptance criteria between a .5 and a
two.
When things fell out of that range, we
asked ourselves is the distortion conservative or non-
conservative. If it was a conservative distortion in
the test that was deemed acceptable,, if it was a
distortion in one of the parameters that on an order
of magnitude basis was a no never mind quantity --
CO-CHAIRMAN KRESS: Okay. So you did take
that into consideration in your judgment.
CO-CHAIRMAN WALLIS: Let's look at a
specific phenomenon. The fangs in the CMT or the
condensation sucked the water back into the CMT. This
was later concluded not to apply to full scale plant
because the hydrostatic heads were so much bigger than
something.
Did that get revealed by these pi groups
or by someone's wisdom from something else or what?
I mean, shouldn't it have shown up in the pi group
somehow or would it be a bottom-up scaling?
MR. BAJOREK: I think a bottom-up scaling
is a place where that would have to be addressed. The
top-down scaling tends to homogenize the entire system
and wash out sharp discontinuities.
CO-CHAIRMAN WALLIS: Well, this was a
balance between them, some parts were dropped due to
condensation and some pressure difference due to
hydrostatics. It's a balance between the things. You
would think it would show up as a pi group.
MR. DI MARZO: This is Marino de Marzo.
Specifically to that phenomena there, it
wouldn't be captured in top down because you don't
have the description to that detail of components
level that would -- so you have a PBL line and a CMD.
CO-CHAIRMAN WALLIS: No, if you look to
the equations, you find you have the hydrostatic term
and the pressure term that you --
MR. DI MARZO: Not at this level of
detail.
CO-CHAIRMAN WALLIS: Not at this level?
Okay.
MR. DI MARZO: So it is left to be
addressed at bottom up.
MR. BAJOREK: This example of the top-down
scaling, I just want to show you the way that we
arrived at some of our results. The example is what
INEL referred to as the intermediate Sub-phase 3.
That refers to this period. You don't have this
figure in your package. I'm sorry.
This is when the ADS 1, 2, 3 system has
been triggered. The water in the CMT is below 80
percent. The CMT is injected for a period.
Accumulators are starting to come on. The system is
depressurizing. There's voids in the upper head,
upper plenum, hot legs. As a result, the steam
generators have become ineffective.
The PRHR is operable, but is under a
degraded condition at this point. Lots of things
going on during this period, which makes it
interesting from a scaling point of view.
CO-CHAIRMAN KRESS: When you talk about
pressure, pressure varies around the system. Is that
pressure in the upper head or where is that pressure?
MR. BAJOREK: Generally --
CO-CHAIRMAN KRESS: Oh, it says upper head
pressure at the top.
CO-CHAIRMAN WALLIS: Yeah, it does.
CO-CHAIRMAN KRESS: Okay, okay.
MR. BAJOREK: It was either that or --
well, the pressurizer wouldn't be good for this.
CO-CHAIRMAN WALLIS: Well, until you get
on later on, it doesn't matter too much. Five mega
Pascals is quite a big pressure compared with the
hydrostatic.
CO-CHAIRMAN KRESS: Yeah, did you consider
pressure a very definitive figure of merit to worry
about?
Rather than say water level in the core,
of course, is very important. Pressure may be, but
I'm not sure I can see it.
MR. BAJOREK: Well, it makes a difference.
It's used very frequently in defining the reference
conditions in each one of these subphases.
CO-CHAIRMAN KRESS: Okay.
MR. BAJOREK: Okay. For the intermediate
subphase, INEL defined roughly a dozen dimensionless
groups to characterize the depressurization flows into
and out of the system.
To apply the methodology, we got geometric
information for the AP 1000, as well as for the test
facilities and for the AP 600. In this case, APEX
doesn't get thrown into the mix because it wouldn't be
applicable for early parts of this period.
It took into account the larger area side
in the ADS system. It gives us a larger flow rate for
the AP 1000, lower resistance in the CMT, which
increases the CMT flow rate; larger volumes in the CMT
and the primary system
Recalculated all of the various pi groups;
looked at the ratios of those pi groups for AP 600 to
the AP 1000, ROSA and SPES, comparing those
dimensionless groups. In this case ROSA comes through
clean.
CO-CHAIRMAN KRESS: If your criteria of .5
to two --
MR. BAJOREK: Point, five to two. If you
decrease that to about 1.5, it still wouldn't make it.
So it --
CO-CHAIRMAN KRESS: Right. And the closer
those are to one, the more you feel --
MR. BAJOREK: The closer they are to one
they are the better. As they start to get to 1.99 or
two or .5, it's at least an indication that it's
starting to get out of that acceptance criteria that's
been used before.
Now, in the case of SPES, there were two
groups that did fall outside of that range for the AP
1000.
CO-CHAIRMAN KRESS: So when you
reformulate your equations, undimensionalized, and you
get these pi groups from the top down, you need to
focus on the dependent variable that you're going to
come out with.
Now, these pi groups are for pressure?
MR. BAJOREK: These two are for pressure,
those two that like I say I have indicated. The si
are for pressure. The five or six down below are from
the mass equation.
CO-CHAIRMAN KRESS: Okay. So you've got
different ones in there.
MR. BAJOREK: Yes. They're both in there,
but for SPES we find two that are potentially
distorted. We looked at both of those. In both cases
the distortion is a conservative one, meaning that
performance in the plant should be better than what it
was in the facility, and they're both in groups which
are of relatively minor importance in the comparison
of the pi groups that had the maximum, the largest
values.
CO-CHAIRMAN KRESS: And how did you
determine the importance? How did you determine the
importance?
MR. BAJOREK: Oh, that's done by an order
of magnitude.
CO-CHAIRMAN KRESS: I see. You estimate
the order f magnitude.
MR. BAJOREK: Yes. In general, when you
look at each one of those coefficients, two, sometimes
it's only one; maybe it's two or three stand out and
are an order of magnitude larger than the other ones,
which say those terms should be dominating the
process.
CO-CHAIRMAN KRESS: I understand.
MR. BAJOREK: So the critical thing was to
make sure that those were scaled relatively well. If
they were of minor importance, you kind of have to
keep in mind that one of these facilities -- I think
the idea is that you should get most of the things
right, but it's virtually impossible to get all of the
things simultaneously scaled correctly.
CO-CHAIRMAN KRESS: Yeah, we all know
that.
MR. BAJOREK: As long as it was in a
parameter of minor importance, we deem that as being
acceptable.
CO-CHAIRMAN WALLIS: Well, probably three
might be of minor importance because there's pressure
change due to change in specific energy of the sub-
cool fluid from heat transfer. Now, that's probably
not a big contributor to pressure change. So slide
three probably isn't all that important, is it?
MR. BAJOREK: I believe in both of these
parameters the comparison between AP 1000 and the test
is closer than what AP 600 and the test had been.
So --
CO-CHAIRMAN WALLIS: Someone had some
foresight.
(Laughter.)
MR. BAJOREK: There was some foresight
there.
And as I mentioned earlier, in a lot of
the periods and their subphases, we see much of the
same story where the scaling groups stay within this
acceptability range; the distortions are of two minor
groups; or the test is more conservative for that
process than what you would expect in the plant.
There is one exception, however, that
starts to get our attention, and that was in the ADS-4
blow-down phase where we start from a relatively high
pressure. The ADS-4 system opens. Entrainment starts
to pick up during this period. The break flow or out
the break, but the flow also leaving the system is
fairly large. We may not be getting much flow from
the CMT and the IRWST has not started to inject at
this point.
So it's a period where we are losing a lot
of inventory from the system, not necessarily getting
much back in.
CO-CHAIRMAN WALLIS: This is a critical
part in the process.
MR. BAJOREK: It is the critical path.
CO-CHAIRMAN WALLIS: The flow-down and all
of that, really don't care.
MR. BAJOREK: Right.
CO-CHAIRMAN WALLIS: But when you get to
the point where you have to lose pressure without
losing water, that's when you worry about it.
MR. BAJOREK: So everything else up to
this point has been a bit preliminary, but when we
start to get to the ADS-4, this is where we've taken
it very serious, and everything we talk about from
here on out is really pertaining primarily to this
critical period, especially the double ended
guillotine break of the DVI line where you don't have
as much mass coming into the system as you would for
the one or the two inch gold leg (phonetic) break.
Westinghouse contends that during this
period both APEX and SPES are scaled acceptably to the
new conditions in the AP 1000. As we go through and
redo the scaling, we find that SPES doesn't look all
that bad. In fact, it looks more like the AP 1000
than it did the AP 600.
We disagree, however, with APEX. The
mismatch in those resistances between the ADS-4 line
and the DVI line creates a distortion for APEX while
the pressures are high and if you make the assumption
that the flow is critical during the early part of
that ADS-4 period.
Once the pressure drops towards the end of
that period, the flows diminish. APEX starts to scale
acceptably. So we're --
CO-CHAIRMAN KRESS: APEX was meant for
that period.
MR. BAJOREK: APEX was meant for the low
part of the period. So we're saying close to the same
thing in different ways.
Westinghouse in their report feels it's
appropriate through the whole period, most of it if
not the whole period. We aren't willing to go that
far. We say it's only after you depressurize, the
flow has gone noncritical, then we can start to
believe APEX.
If you're going to use these data for code
evaluation during the blow-down period itself, stay
with SPES. Don't bring APEX into the picture.
Now, the problem does show up in the
dominant dimensionless group that you get out of the
scaling rationale. I'll refer to this one as pi 16.
It's the relative flow rate between the CMT and a
reference flow rate, which is chosen as the ADS-4 flow
for this period.
Physically it represents how easy it is to
get flow into the system versus flow leaving the
system. So if I calculate --
CO-CHAIRMAN KRESS: And to get that pi
group, don't you have to have an entrainment model?
MR. BAJOREK: To analyze it, yes, you
would, and that's -- when INEL did their evaluation,
okay, one of the things you could see in the report,
and I did talk with some of the people who had done
that evaluation, and they explained that one of the
difficult features that they had was estimating the
flow quality in the ADS-4.
In their report, they looked at the core
exit quality, basically on the low end of things, and
the flow quality of 1.0. They looked at AP 600 for
both of those limits, and they found that the AP 600
scaled well for both APEX and SPES.
I found a couple of problems in some of
the numbers that they used with regards to the CMT,
fixed those, redid that for the AP 600 and still
agree. Both APEX and SPES would fit those
susceptibility criteria for that wide range anywhere
you pick that flow quality.
AP 1000, however, it's going to be more of
an important criteria because I think we've already
moved APEX during this period. So it has potential
distortion. As we talk about entrainment later on and
if we want to revisit top-down scaling, if we go away
from INEL's assumption that the flow quality in the
text and the plant are similar, and we make one have
a low quality and the other one have a high quality,
then we start to see distortions potentially even in
SPES through this period.
CO-CHAIRMAN WALLIS: So we change our M,
zero, DOT.
MR. BAJOREK: yes.
CO-CHAIRMAN WALLIS: The rest of it is the
CMT gravity driven flow, right?
CO-CHAIRMAN KRESS: Yeah, you get that
pretty accurately.
MR. BAJOREK: So as I've been proceeding
right now, I've tried to stay true to the assumptions
and the methodology that INEL put together. I didn't
think this was the best time to invent anything new.
CO-CHAIRMAN WALLIS: Now, this should show
up in the code. I mean, if the code is modeling SPES
and the code is modeling APEX, this distortion should
show up in different predictions if it's important.
if you just run the code with these numbers, I mean,
M, zero, DOT is predictive by the code, and delta Y is
predictive by the code and all of that.
MR. BAJOREK: Well, I think the code will
predict a certain MDOT and a certain flow quality.
The question --
CO-CHAIRMAN WALLIS: But then what
happens? The way the transient develops will depend
on the ratio of these two things up here.
MR. BAJOREK: But the question we would
have is --
CO-CHAIRMAN KRESS: How good is that
predictor?
MR. BAJOREK: -- how good is that
entrainment.
CO-CHAIRMAN KRESS: And then you have to
turn to the test.
CO-CHAIRMAN WALLIS: We could do a
sensitivity test. You could vary M, zero, DOT in some
ways, see if it makes any difference.
CO-CHAIRMAN KRESS: Yeah, you could do
that.
MEMBER SCHROCK: You talked about the flow
resistance of the ADS line in the beginning of your
discussion. That discharge flow is determined by
several things, one of course being the valve at the
end of that line.
It isn't clear in my memory what the flow
-- can you explain that in terms of this pi group?
It's unrelated to this pi group, I guess?
MR. BAJOREK: Well, not exactly. For the
CMT we used the resistance and the head that was
apparent in either the one inch break or the double
ended break to get the driving head, the resistance of
the CMT line, to get the CMT flow rate.
For the ADS-4 flow, followed what INEL
did, used a homogenous HEM critical break flow model,
calculated the flow rate with that. Okay? The flow
rate increase or the mass flux with the HEM model, and
then calculated the flow rate for the larger sized
flow area in the AP 1000.
MEMBER SCHROCK: And what is the source of
the information on flow quality approaching the break
for that? How does it relate to the entrainment that
you described at the beginning of the discussion?
MR. BAJOREK: At this point it doesn't
have a relation to the entrainment. Since we didn't
know what the entrainment was, we said, well, it has
to be somewhere between 1.0 and the core exit quality
and looked at both ends of that range. We did not try
to pick a flow quality that might be based on an
entrainment model or some other estimate.
That's one of the questions that we're
going to have for ourselves at the end of the top-down
scaling, and we're going to try to address.
MEMBER SCHROCK: This parameters varies
during the ADS operation, and at any point in time
during the ADS operation it depends upon the amount of
entrainment. There are other aspects that enter into
it, such as whether or not there is a role for
thermodynamic non-equilibrium in that critical flow
process.
I don't see that this parameter addresses
the question of what is the amount of entrainment, and
certainly it doesn't address anything that.
MR. BAJOREK: No, it ignores
nonequilibrium both in the flow and also in the energy
of the flow itself. The top-down scaling essentially
homogenizes everything. So those details don't come
out on the top-down --
MEMBER SCHROCK: Well, homogenizing
everything gets you off the hook on the nonequilibrium
question, I suppose, in a sense, but it doesn't do
anything for you in determining what is the fluid
which is flowing in the ADS-4 line. That's the
central issue, central question to get to.
MR. BAJOREK: Yeah. That's why when I
finished the top down, which I think might be the
next --
CO-CHAIRMAN WALLIS: This is a bit like a
PRA, the case where non-risk informed application.
You do this like a PRA. It reveals there are certain
things. You now need to go back and think about some
more.
MR. BAJOREK: yes.
MEMBER SHACK: But is that the conclusion
I get from slide 11? You have the two critical flows
at the two limiting qualities, and you still say it
scales okay for either one of those. That's where
you're going to get to basically.
MR. BAJOREK: pretty much, although I
think I would have gone to slide number 12 to
basically say that. Slide 11 is just a summary of
where we wound up with those pi groups and is just an
indication that when we looked at APEX, we're seeing
numbers outside of that acceptability range.
It doesn't matter what I assume for my
flow exit quality. So whatever kind of entrainment I
go back and try to estimate, I'm still from a top-down
scaling saying it's still out of range, and it doesn't
matter whether it's the double ended break or the one
inch break. I still have a problem on there.
SPES seems to fit things more acceptably.
CO-CHAIRMAN KRESS: But, once again, you
know, when you look at the APEX one and say it's out
of the range, that doesn't necessarily invalidate it
to me because we've fixed that range sort of
arbitrarily, and so it just tells you, well, maybe we
ought to think about this one a little more, look at
it a little more.
MEMBER SHACK: Two, point, oh, two, you're
--
CO-CHAIRMAN KRESS: Yeah, at 2.02 you're
not good, right.
CO-CHAIRMAN WALLIS: It's not necessarily
no good. If you run the code and if it's APEX, it
tells you something about the code. And if you run
them on the code for AP 1000, then you've got some
more confidence even though these numbers are
different.
You're still modeling the same phenomenon,
but you haven't modeled exactly the same kind of a
trace versus time.
MR. BAJOREK: That's true.
CO-CHAIRMAN WALLIS: The balance of
things.
MR. BAJOREK: But when this number gets
larger and larger, it tells me that maybe you aren't
modeling the same transient anymore. Okay? If it's
two or three --
CO-CHAIRMAN WALLIS: Then you should
perhaps run the code with modeling the real APEX and
then a virtual APEX which has a bigger ADS-4 so that
it comes back to one or something. See if it makes
any difference.
You can do an awful lot of things with the
code with numerical tests, and you learn from that.
MR. BAJOREK: The conclusions that we get
out of taking a look at the top-down scaling as a
whole.
One, by and large, in many of these
periods the tests are still very valuable, and it can
be used to help benchmark those codes for AP 1000
usage.
The ADS-4 period gives us some concern,
and it does go back to what flow quality do you have
in that ADS-4 line, and if you're off or if you were
wrong for some reason, what's its potential effect on
the AP 1000 itself?
In looking at these scaling groups, okay,
varying the quality, we didn't do it explicitly, but
if we varied the break model, varied other things, one
of the things is that they are sensitive to some of
your assumptions.
Quality was one of those. We could jockey
those numbers around, depending on what that quality
assumption was. So even though top-down scaling left
us feeling like this isn't too bad, it left us a
couple of questions saying that quality in that line
and things that would affect that quality in the line
need to be looked at in more detail.
To do that, we said, well, we need to set
up a simple model where we can make some parametric
studies, varying quality, other things in the system,
to try to get on a first order approximation on AP
1000 and how it gets affected by some of these
parameters.
Dr. di Marzo is going to show you this
calculation. It's valid from the beginning of the ADS
1, 2, 3, and transitions us into the IRWST.
Do you want to sit here in the middle?
MR. DI MARZO: This is Marino di Marzo.
What I have tried to do here is to
basically look at the single node type system, and
first I formulated this model, if you wish.
MEMBER POWERS: It's not a handout. It's
different.
MR. DI MARZO: And then I had to build
some confidence in myself that this thing had
something to do with what I was trying to describe.
So I picked the DVI guillotine break as a start, as
Steve mentioned to be the most severe transient.
And then I used the ROSA AP 600 test of
that particular transient to see if I was close.
Again, it's one node. We'll see with fixed parameter.
So it doesn't even -- the qualities do not change
during the transient like Dr. Schrock alluded.
And so I made a judgment there, and you
will see the result as to how good that is.
Then I asked several questions. The first
question I wanted to ask was: of all the non-
dimensional groups that are used in the model, which
one is relevant and which ones are not relevant? I
mean, what is very important during this portion of
the transient.
Then I compared AP 600 and AP 1000, which
is kind of similar to what was done top down, and
finally I start to run some sensitivity analysis, and
the first one I run was changing the ADS quality and
see how that impacted minimum vessel inventory, which
has been the figure of merit throughout the AP 600
certification effort.
Then I'll draw some conclusions from what
I'm looking at.
Now, this here is the set of assumptions
used in deriving the model. The first thing is that
the quality is fixed at each port. It can be
different at each port.
Remember you have an ADS 1, 2, 3 and 4,
which is considered one part. Another port is the
break vessel side, and the third one is the break of
the CVI line on the DVI line side. So you have
essentially three outlets for the system. Each one
has its own quality, and that's kept constant
throughout the process.
The reference of the system is at the
enthalpy at the average temperature during your
transit, and you will see from the analysis how this
is justified.
The specific heat of the liquid is
artificially doubled to account for metal masses that
are in contact with the liquid. This is pretty
standard. We've done that on many, many occasions.
The accumulator clearly comes in as a soup
cool leaking (phonetic), and that represents a sync
(phonetic) to the system. You will see that in the
equation. PRHR also is a sync and will be modeled as
a function of time because it was impossible to take
it as constant as it decays, as you go through the
transient.
All parameters unless otherwise specified
are taken as constant at this average value in order
to make this analysis simple enough that you can
scrutinize the effect of each variation on the answer.
Now, here are the submodels that have been
used to characterize mostly the flows. The AF is the
ADS flow. It's the flow that goes out ADS 1, 2, 3,
and 4. Obviously as the transient progresses, the
cross-sectional area of that port keeps increasing
when the activation is timed in this particular
transient. So at each time that an additional valve
opens, the cross-sectional area changes.
The flow is critical. Quality is
considered to be equal to one to start, and in that we
use the Henry Fauske, and then when the quality is
considered to be less than one, we use basically the
homogeneous HEM model.
The break vessel side, VB, is basically
considered vapor because the data so indicates from
ROSA facility, and therefore, we used the Henry Fauske
there.
The break on the DVI side is very complex.
If you look at the data, this break goes two phase
initially. Then it becomes fully liquid because the
accumulator is dominant, and then becomes fully vapor
because then the accumulator stops injecting.
So throughout the transient it goes all
over the place so that we don't have any idea what
that might be, and so we will see that that particular
term will use it as a tuning parameter in achieving
our result.
The DF is the DVI flow on the impact side,
and that is basically gotten from the specification of
the test. Essentially we know the flow rates, and we
input that as boundary condition to our single node.
Decay heat similarly is known. PRHR is modeled as a
function of time from the data. Then there is the
system volume that we are going to use that's going to
be different we're going to do for ROSA for AP 600 and
for AP 1000.
According to the size, initial pressure
also is gotten from the test, and initial inventor is
gotten from the test and from the other facility
according to older plant and facility according to the
size.
These are the references where we got the
information from. Yes?
MEMBER SCHROCK: You've got a quality of
one. You have single phase steam. I don't understand
choosing Henry Fauske to calculate the critical flow
of single phase.
MR. DI MARZO: Yeah, that's no problem.
Basically it's critical flow, steam critical flow. I
can give you the --
MEMBER SCHROCK: I don't think I'd go to
Henry Fauske to do that for --
MR. DI MARZO: Well, you can use --
MEMBER SCHROCK: -- saturated steam.
MR. DI MARZO: Well, you can use a number
of type of situations there, and you will see that you
could also use the HEM model or other models. It
doesn't really make much of a difference in the end
what you get, but that's what I've been using in a
number of occasions, and it worked kind of reasonably
well.
You will see the variation there is in
terms of using a different model for that. So I
picked that as it is no problem implementing a
different formulation.
CO-CHAIRMAN WALLIS: What is that si band
reference?
MR. DI MARZO: That's basically the
original reference of the homogeneous equilibrium
model. It was a cascade of references. At some point
I finally found that this is the first guy that put it
down somewhere. It's on some remote things.
MEMBER SCHROCK: I don't think that you
could argue that he's original.
MR. DI MARZO: I don't know if he's
original, if there's someone even before him, but
that's what I basically landed.
MEMBER SCHROCK: You could turn to checks
on thermodynamics 20 years prior to that.
MR. DI MARZO: Before to that, right.
Yeah, that's the one I landed with. That's where I
stopped myself.
CO-CHAIRMAN WALLIS: -- and Prague and
people like that --
MR. DI MARZO: Yeah.
CO-CHAIRMAN WALLIS: -- have homogeneous
models.
MR. DI MARZO: So this is the formulation
of the model. We start with the consideration of
mass, and V is the inventor in the system. Again,
this is the inflow from the intact DVI side. This is
the break on the vessel side of the DVI. This is the
break on the DVI side, and this is the ADS flow.
CO-CHAIRMAN WALLIS: So V is the symbol
of?
MR. DI MARZO: V is the symbol of the
amount of liquid, the volume of the liquid in the
system, not in the vessel, actually in the overall
system because this is a single node.
CO-CHAIRMAN WALLIS: The volume of liquid.
MR. DI MARZO: The volume of liquid.
Conservation of energy says that basically
you have core power and stored heat, and these two go
into making vapor. They go into the PRHR, who is a
sync, and they go into heating up the injection.
CO-CHAIRMAN WALLIS: You have been very
wise in that you have no momentum equation.
MR. DI MARZO: Right, no momentum.
(Laughter.)
MR. DI MARZO: It's a single node.
Nothing goes anywhere. There is not much going on
there.
MEMBER SIEBER: It's a straight line.
MR. DI MARZO: So in terms of the vapor
generation, you have the first three terms, which are
basically from the equation of state. They just are
the variation of the amount of mass that's in the
vapor due to the changing pressure, the changing
volume because the liquid recedes.
CO-CHAIRMAN WALLIS: It's a perfect gas.
MR. DI MARZO: It's not a perfect gas. I
use R star, which should be corrected for
compressibility. Okay?
And then -- which is kind of fixed at some
MOD (phonetic).
And then I have the amount of --
MEMBER POWERS: In thermal hydrolysis, the
chemists go to all of this effort to get you a
universal gas constant, and you --
MEMBER SIEBER: And then you don't want to
use it.
MEMBER POWERS: Right. Decorating it with
all of these correction factors.
(Laughter.)
MR. DI MARZO: It's some sort of a -- then
you have the three flows with the qualities associated
with them. This is quality one, as we discussed
before, the flow from the vessel side break.
We use Clausius-Clapeyron in this form,
this being a constant as you will show later. So if
you go to the equation of state that's like that, and
as I said, I used the compressibility factor at the
average temperature.
CO-CHAIRMAN WALLIS: It looks upside down.
MR. DI MARZO: It looks upside down? No.
That's --
CO-CHAIRMAN WALLIS: No, the bigger the
pressure, the bigger --
MR. DI MARZO: Yeah, it's upside -- well,
a misprint, okay? I mean vis-a-vis and then I wrote
-- you're right.
The depressurization equation is basically
the energy equation. So if you take the consideration
of energy, the vapor generation, Clausius-Clapeyron,
and the equation of state and put everything together
and solve with respect to the pressure, that's
basically what you get.
CO-CHAIRMAN WALLIS: This looks like one
of those things in Moody's book.
MR. DI MARZO: Right, that's correct.
The bottom part is what we -- it's
basically what we have been referring to as the
compliance of the system. It has two parts, a thermal
part, which is associated with stored heat, which is
basically this part here, and another part which is
associated with the vapor space, which is this part
here.
And up here you have basically all of the
energy contribution. In other words, core power,
PRHR, injection, and then you have thermal dynamic
group, which comes from the gas law, and then you have
all of the dischargers that carry away latent heat.
Now, if you take the depressurization
equation and put in the consideration of mass, you get
this thing I call the trajectory equation, which is
very important because it basically relates how much
liquid you lose compared to how much pressure you
lose, which is what Dr. Wallis was alluding before.
It's essentially your problem. You want to lose
pressure without losing too much mass, too much
liquid, so to speak.
So this trajectory equation is key, and
that's basically what we are going to then use.
So you want to have dv/dp small in terms
of a safety concern. If dv/dp becomes large, you're
in trouble.
So we take this equation, and we non-
dimensionalize it in the following manner. We
identified four nondimensional groups, and the
interesting thing is that each group is directly
related to a scale of the system, to a physical scale
of the system.
The first group is basically related to
the net in-flow. So that relates to all of the
boundary conditions to the system, the flow that goes
in and out, establish the boundary conditions.
The second group is the energy associated
with that net in-flow.
The third group relates to power, and
obviously the PRHR has been lumped in there.
And finally, the last group relates to the
size of the system essentially. It relates to the
total volume that's available in the system.
And then there is a last group which is
the ratio of the density of the vapor and the density
of the liquid which I just left out, but, I mean,
that's one term that you also want to consider.
Now, if you assume that your system is
independent of scale, in other words, if you assume
that all these groups are small with respect to one,
look at this formulation here. Basically they're
compared to one are small, and you assume also that
the quality of the ADS-4 is one. Actually the ADS-1,
2, 3 and 4 is one.
Therefore, you may call these terms small.
You get basically a self-similar (phonetic) solution.)
CO-CHAIRMAN WALLIS: It looks like an
adiabatic --
MR. DI MARZO: Yeah, it's basically
popping up in adiabatic, which should be this equation
here. You would get a coefficient of this kind.
So whatever the system introduces, it
basically makes this part of the equation different
from one, and will basically affect your transient.
So now you can ask the question after you validate it.
You can ask the question which term is doing this with
respect to -- so how far are you from this type of
situation? So who affects; which scale affects the
system performance? So that's the key answer.
Incidentally, if you look at the early
part here for the Clausius Clapeyron, this term is
equal to nine, which basically justified the
assumption of measuring the properties of a constant
average temperature because a minimum variation in
temp. will give you the maximum -- a tremendous
variation in terms of pressure.
If, incidentally, you also take this and
compare it to the date on the steam table, you
basically get an excellent agreement between 7,000
kilo Pascals and 100 kilo Pascals if you use that kind
of relationship. So that's good.
So now let's look at how reasonable the
result might be with this. So I took ROSA, the AP-DV-
01. I considered the transient between ADS-1 and
IRWST injection, which is in that time frame. I
analyzed what initial inventory was and the final
inventory were in this particular transient.
You have to consider that on the DVI
broken, DVI side, both the accumulator and the CMT
will discharge through that line. So you have to put
them into the inventory of the system, and basically
on the other side, on the other DVI line, we consider
that s an injection.
So you have to play a certain number of
adjustments in order to fit the transient, and that's
all reasonable and justified.
And all of the parameters are set in
accordance to what we have from the test. The only
parameter that remains open is the quality of the DVI,
of the break side on the DVI line, DVI side, and the
reason for that is that that flow is very much
changing from total liquid to total vapor throughout
the duration of the transient.
So I said: okay. Let me match the
initial and final inventory. So the initial inventory
is the initial condition. Let me adjust the quality
until the amount of liquid in the system matches the
one in the test. All right?
And that was found for one third.
Obviously if you make the quality higher, you have
more liquid. If you make the quality lower, you have
less liquid. That's fairly simple.
So adjusting singly that parameter, then
I calculated what would be the pressure trace, and I
compared that with the pressure trace in the test. In
other words, if the single node evolves, what is the
pressure path that it will trace and how does that
compare to the test?
So if these two traces are somewhere on
the same page in some ways, I have certain confidence
not as a predictive tool, but at least as a
comparative tool so I can make my sensitivity studies
and compare one facility to the other on the basis of
those parameters, and that's basically the result that
I get.
MEMBER POWERS: All this without the
momentum equation.
MR. DI MARZO: Right. Well, that's in the
breaks. The answer to my question is in the breaks.
CO-CHAIRMAN WALLIS: That's why it doesn't
matter what you have for a momentum equation.
MR. DI MARZO: Look. Remember this is a
very fast transient. The system is basically all
saturated. It's flashing all over. So it makes kind
of sense that you're not too far off.
All right. So then I have two slides
which I'm going to skip in the interest of time, which
basically give you all of the numbers that I used. So
if you want to basically do it, the code is a one-
pager. So it's not a big deal. It runs on Quick
Basic, which is an archaic form of computing.
But what I think more important is to
first ask the question how -- let me see. How do I do
this? I would need -- we are referring to this
equation here. Okay? The original equation here, and
we are now looking at how the different governing
group affect the answer.
So remember they are all compared to one.
So if they are less than one, far less than one, like
.01, it doesn't matter. If they are .1, they are ten
percent of the answer. If they are one or above, they
are really affecting the answer significantly.
In order to do that, I plot them as a
function of time, as the transient evolves, each one
of them. I took the logarithm of them. I first took
the absolute value. I took the logarithm, and so to
give you orders of magnitudes compared to the value
one. Okay?
So if the number is down here, it means
that that particular group is not very important.
It's the number that drifts up towards one that
affects significantly. It's a scale parameter that
really is important.
Now, on the first plot, this one here, I'm
concentrating on the terms in the denominator of your
trajectory equation. So this is the terms in the
compliance. This is the density ratio, and the
density ratio is really very much non-important. We
knew that from the beginning. Basically it's rhov
over rhol. It's a term that's very small compared to
one.
The other two terms are the power, which
basically goes to zero, is first negative. The shaded
one that's not very well seen on the viewgraph, but
better on your overhead; the shaded part is when the
original function was negative. Remember I took the
absolute value.
So it's negative, goes to zero, becomes
positive, but during the transient remains less than
ten percent. So power is not a big issue, right?
The controlling factor in the compliance
is the stored heat, which basically is what you see up
here.
Now, let's analyze what happens in the
transient, and this is a good figure to do so. This
is the activation of the accumulator. This is
activation of ADS-2, ADS-3, ADS-4, and this is when
the accumulators stop injecting.
CO-CHAIRMAN WALLIS: So the purpose of
these ADS-1, 2, 3, 4 is to keep everything around .3.
MR. DI MARZO: Right. Yeah, but that's --
you understand the goodness of the thing. In other
words, ports are scaled properly from here. You
understand it. To give you a very nice, gradual
depressurization. That's basically what this is
telling you.
All right. Now, let's look at the top
part of the trajectory equation. I left alone the
density ration because that wasn't that important. So
I took the other two and added them up together, and
then I look at the term associated with the 19th
floor, which is F sub-G, which is actually the
dominant term, which is up here.
And then we looked also at the sum of the
two, which is, again, doing this, and then the last
one that we looked at is the actual volume of the
system. Remember the AP 1000 if far larger than the
AP 600, but that term doesn't make much of a
difference.
CO-CHAIRMAN WALLIS: FG is an in-flow?
MR. DI MARZO: FG is the net in-flow. So
it is basically -- let me show you again.
CO-CHAIRMAN WALLIS: In-flow minus out-
flow.
MR. DI MARZO: Yes, but it's normalized
with respect to the ADS flow.
CO-CHAIRMAN WALLIS: Okay.
MR. DI MARZO: That's what it is. So it's
what comes in through a DVI line, intact one, minus
what goes out of the broken two sides over what goes
out of the ADS-4.
Now, the bonus that we get out of this
kind of analysis is that we kind of understand what
each thing does with respect to the figure of merit,
and that's what this slide over here is about.
So if you have a negative term, PG, EG, PG
plus EG, that decreases the trajectory slope.
Remember trajectory slope is dv over dp. So a small
dv over dp means that you're losing pressure without
losing liquid, which is desirable.
MEMBER SCHROCK: V here means specific
volume? What is V?
MR. DI MARZO: V? V where? V? V is the
volume, liquid volume in the system.
CO-CHAIRMAN WALLIS: Volume of liquid.
MR. DI MARZO: Inventory, if you wish,
liquid inventory.
MEMBER SCHROCK: Lower case V.
MR. DI MARZO: Lower case V. High case V
is the total system volume.
CO-CHAIRMAN WALLIS: So lower case V is
specific volume.
MR. DI MARZO: No, lower case V is
unfortunately the volume of the liquid in the system.
CO-CHAIRMAN WALLIS: Volume of the liquid
in the system. Okay.
MR. DI MARZO: Inventory. I could have
written I, but then I is kind of, you know, even more
cryptical (phonetic) than small V.
So now the power group EG is negative when
the PRHR is removing heat. Remember the PRHR removes
much more than core power in the beginning, and that
later on turns positive because the PRHR degrades. So
you can see that. That's why this turns positive.
And then you can analyze this thing. You
can read it again. I don't know how my timing is, but
basically there's a description of how each term
affects the bottom line.
The key important point is that the net
in-flow is the main term. It is negative. So is this
term here. It is negative, and therefore, clearly it
affects this in a positive fashion.
If it's negative, it basically means
you're losing water, and so it affects this in a
negative sense. It makes this term bigger, and so it
means that you're losing more liquid than pressure.
Now, if FG --
MEMBER SCHROCK: Marino, is there
something in here about conservation of liquid? I'm
not quite with you yet. I mean as you have flashing
during this flow-down process.
MR. DI MARZO: Yes, right.
MEMBER SCHROCK: You're generating steam,
and now you've diminished the quantity of liquid in
the system because of the flashing. You also diminish
the volume because liquid goes out the break.
Is there any other way you diminish the
liquid?
MR. DI MARZO: Let me show you. You
diminish the liquid basically because the liquid goes
out of the various hole and because vapor is
generated. That's the only way you can lose -- ever
even lose liquid.
CO-CHAIRMAN WALLIS: Marino, I think both
FG and PG and EG are negative. So the VDP is
positive. So as the pressure goes down, the mass of
liquid also goes down
MR. DI MARZO: No, if these two are
negatives --
CO-CHAIRMAN WALLIS: They are.
MR. DI MARZO: -- that's good.
CO-CHAIRMAN WALLIS: The numerator and the
denominator are both negative.
MR. DI MARZO: They are both negative.
These two are negative. That's a good sign.
CO-CHAIRMAN WALLIS: And the numerator is
negative. So the VDP is positive.
MR. DI MARZO: So this fraction becomes
smaller.
CO-CHAIRMAN WALLIS: Is positive.
MR. DI MARZO: Yeah, but it's smaller.
CO-CHAIRMAN WALLIS: Well, it's positive.
MR. DI MARZO: Now, wait a minute. If FG
becomes larger than one, which it's trying to do that,
you get refill.
CO-CHAIRMAN WALLIS: Yes.
MR. DI MARZO: Which in ROSA, for example,
you do. Okay?
In other words, if this here --
CO-CHAIRMAN WALLIS: You've got to refill.
At the end you get refill.
MR. DI MARZO: Right.
CO-CHAIRMAN WALLIS: At the beginning
they're both --
MR. DI MARZO: Right. If this curve here
crosses this line you get refill.
CO-CHAIRMAN WALLIS: It does, yes.
MR. DI MARZO: Now here it's all negative.
So you don't get anything. Actually you get a
tremendous loss of liquid.
CO-CHAIRMAN WALLIS: Yes.
MR. DI MARZO: So that's all there, but
basically you cannot make sense of all that is
happening in detail. So it's in the handout, and you
can look at it. Obviously it's going to be right
up --
MEMBER SCHROCK: I'm afraid that I'm just
too slow in assimilating all of the unfamiliar
notation, but I don't know how to interpret the
conservation of mass equation.
MR. DI MARZO: Conservation of mass. I
should have taken more time in going through this.
Yeah, here, right?
CO-CHAIRMAN WALLIS: It's the conservation
of liquid volume.
MEMBER SCHROCK: Conservation of liquid
volume. It's got four different terms in it.
MR. DI MARZO: Right.
MEMBER SCHROCK: We talked about two
things that change it. I don't understand what the
four terms are here.
MR. DI MARZO: Okay.
MEMBER SCHROCK: The top equation,
conservation of mass.
MR. DI MARZO: This is the amount of
liquid that comes in through the DVI line.
MEMBER SCHROCK: Liquid entering?
MR. DI MARZO: Right. This is the amount
of liquid that goes out from the vessel side of the
break. Remember it's double ended. So from the
vessel side of the break. All right?
This is the amount of liquid that goes out
from the opposite side of that break, and this is the
amount of liquid that goes out from the ADS-4.
Now, the amount of liquid that becomes
vapor, which I think is what you're alluding to, is
not considered because it's very small.
MEMBER SCHROCK: Huh?
MR. DI MARZO: It is not in this equation.
MEMBER SCHROCK: It's not in this
equation. It's small during the ADS phase?
MR. DI MARZO: Because of the flashing
part, yeah.
MEMBER SCHROCK: I'd buy that maybe a
little earlier on, but --
MR. DI MARZO: So I could basically block
in here the term associated to this term here with the
density. This is what I basically neglected from that
particular, but I recycled this thing into there
again, into the energy equation because what I'm
interested is the amount of energy associated with
that rather than the actual physical amount of mass
that goes. I made that approximation. Yeah, that's
correct.
MEMBER SCHROCK: Okay.
CO-CHAIRMAN WALLIS: But if you were just
boiling a pot of water with vapor going out the -
MR. DI MARZO: Obviously it won't work.
CO-CHAIRMAN WALLIS: It won't work.
MR. DI MARZO: It won't work, but remember
that the discharges here are very, very large compared
to what's happening here in terms of flash.
CO-CHAIRMAN WALLIS: Well, the amount of
liquid entrained is important.
MR. DI MARZO: Right. Now, we'll get to
that.
CO-CHAIRMAN WALLIS: Yeah, but it is
because it's the only term you've got.
MR. DI MARZO: That's right.
(Laughter.)
MR. DI MARZO: Now, when I compare the two
traces, they are basically the same. The notable is
ADS-4, which is at this point here.
What happens there is that you have to
realize that the ADS has been scaled with power, and
power is not the dominant term. The dominant term is
the discharge. So if you make the hole bigger, you
lose more water, and so that's basically what's
happening there and why these two are different.
However, as you lose more water, also the
pressure takes a dive. So if you are in terms of
trajectory scale, it doesn't really matter.
CO-CHAIRMAN WALLIS: What you're showing
is you think that AP 1000 will, relatively speaking,
lose more inventory than AP 600?
MR. DI MARZO: ADS-4, but it will get to
RWST faster. You see, it's a race, and both terms --
in the end, if I were to plot P over view of P against
V, you won't see any difference. I have to plot in
this way to make you see.
CO-CHAIRMAN WALLIS: When does the IRWST
come in? Pressure has to go down to a certain value.
MR. DI MARZO: Yeah, right, at that
pressure, at this pressure here it comes in.
CO-CHAIRMAN WALLIS: But the pressure
terms, the pressure curves look the same, don't they,
or does --
MR. DI MARZO: No, no. This --
CO-CHAIRMAN WALLIS: -- the end at RW --
MR. DI MARZO: They end at slightly
different positions.
CO-CHAIRMAN WALLIS: Okay, okay.
MR. DI MARZO: So basically you open a
bigger port. You lose more water, but you pressurize
faster. That's all.
CO-CHAIRMAN WALLIS: So you go down to the
same inventory where one recovers a little earlier
than the other.
MR. DI MARZO: That's exactly. ADS-4, AP
1000 results to be a little faster. That's all.
Now, so this is good in the sense of
saying I can play with this in any number of ways and
variations and whatever, but so far so good. I can
say the top-down scaling pretty much -- am consistent
with that, but here comes the punchline.
What if the quality of the ADS-4 is
different? What is the impact of that on the figure
of merit?
Now, this is very deceiving because of you
look at the pressure trace, and you compare the
pressure traces for different qualities, they
basically are the same curve. They're not going to
change much, and this is due to the competing effect.
It took me a while to figure it out, but
basically what happens here is that if you were to
decrease the quality, your pressure stays aloft more
in the initial part because you lose liquid rather
than vapor. But then you have lost basically all of
the liquid. So basically your stored heat goes down
the drain.
At that particular point the system
pressure drops because there is nothing to hold it up,
and so in the end, they all loop in the same type of
trace.
But if you look at inventory, meaning how
much shorter you are left with for injection, that's
a completely different answer. There there is a
tremendous impact.
Now, again, this is not a predictive tool.
It's just a comparative --
CO-CHAIRMAN WALLIS: No, you're saying
it's very important. You carry over --
MR. DI MARZO: But what I'm saying it's
tremendously important.
CO-CHAIRMAN WALLIS: It's so obvious. You
carry over a liquid to --
MR. DI MARZO: You carry over. So that
puts a tremendous importance on how well you will know
what entrainment is. In other words, the uncertainty
on entrainment cannot be large to draw a safety
conclusion. You have to have an uncertainty on
entrainment that's pretty -- that's basically all that
this says.
Now, there are other considerations. I
haven't taken all the facility. I could run all of
the facility against this and see how effective they
are in reproducing this transient for code variation.
I could do a number of things with this scheme, but
that was the point that we wanted to make, and so I
stopped there.
CO-CHAIRMAN WALLIS: I think this also
shows up in the code. You put in different amounts of
entrainment in a big code. You should get something
very similar.
MR. DI MARZO: That's right.
CO-CHAIRMAN KRESS: Because all he does is
represent the code the center way.
CO-CHAIRMAN WALLIS: Now, are you folks
able to run -- you are able, but are you in a position
realistically to run a system code, like RELAP or
whatever?
MR. DI MARZO: On this kind of thing?
CO-CHAIRMAN WALLIS: For AP 1000.
MR. BAJOREK: Well, as part of the review,
and I think NRR will discuss that tomorrow -- we did
RELAP calculations for AP 1000.
CO-CHAIRMAN WALLIS: You did. Okay.
MR. BAJOREK: Yes.
CO-CHAIRMAN WALLIS: But you have no table
to run the Westinghouse code.
MR. BAJOREK: NRR, I think, send some
people up to Pittsburgh, but I don't believe we had
access to the codes.
CO-CHAIRMAN WALLIS: They didn't come back
with the code. So you have to run your own code. You
have an independent code.
MR. WERMIEL: This is Jared Wermiel, Chief
of the Reactor Systems Branch.
No, Dr. Wallis, we did not exercise the
Westinghouse codes. We ran our own independent
analyses. We did a code review of the documentation
at Westinghouse, and we'll talk about what we actually
did tomorrow.
MR. DI MARZO: So to summarize basically,
it 's a very simplified approach that are obviously
sweeping approximation all over the places, but it's
used to give you a sense of who's playing what and
what's the net impact of everything on the end result.
It clearly doesn't mean to be accurate or predictive
or any of that. It just has to be a variational type
process that you're doing. If I change this, is it
more or is it less? That's the kind of thing you
want.
Now, one last thing. Again, having said
that, to put in another frame, point three was the
point. Point three, when 30 percent of the liquid was
there, was the level where in a specific standard we
had said if it gets to .3, we're going to get core
results. That's just to give you a sense of what it
means. And then, again, this is very cursory, but if
you are down somewhere in here, you're at the position
where you may experience some core results.
MEMBER SIEBER: Now, do you have an
estimate of where it would be for AP 1000? The same
number?
MR. DI MARZO: Well, we don't know the
entrainment. We have no ways, you know. That's an
area that now Steve Bajorek is going to go into in
great detail.
MEMBER SIEBER: But do you know what level
of inventory --
MR. DI MARZO: Right.
MEMBER SIEBER: -- would cause that?
MR. DI MARZO: Basically this says
inventory is crucial, and now that's where he's going
to come in and say what do we have to --
CO-CHAIRMAN WALLIS: But there are ways to
predict entrainment. If you use the homogeneous
model, presumably entrainment is 100 percent, and then
it depends on how you define entrainment perhaps, and
if it's --
MR. DI MARZO: Entrainment 100 percent,
yeah. It depends on what you call entrainment. What
do you mean? Call it zero
CO-CHAIRMAN WALLIS: And there's no fade
separation.
MR. DI MARZO: Okay. So you take core
exit policy.
MEMBER SIEBER: But that changes with
time, the amount of entrainment.
MR. DI MARZO: Yeah, that's the point that
Dr. Schrock did at the beginning. All of this quality
changes with time. This is just a sweeping
approximation to take them constant, and you go with
that.
MEMBER SIEBER: Well, that could give you
some conservative answer.
MR. DI MARZO: That gives you a sense, but
that's all it gives you.
MEMBER SIEBER: Right.
MR. DI MARZO: Okay?
MEMBER SCHROCK: Have you got this written
up?
MR. DI MARZO: I'm five hours a week.
(Laughter.)
MR. DI MARZO: I am five hours a week.
MEMBER SCHROCK: I think I could buy all
of your arguments better if I could sit down in front
of the fire and --
MR. DI MARZO: And look at it, and then go
page by page through everything.
MEMBER SCHROCK: I know.
(Laughter.)
MEMBER SIEBER: There's no place to sit.
MEMBER SCHROCK: I know where to put it.
Discouraged.
MR. DI MARZO: It seems front of the fire
is a kind of a consequential.
CO-CHAIRMAN WALLIS: Core exit quality is
pretty low, isn't it?
MR. DI MARZO: What?
CO-CHAIRMAN WALLIS: I mean, if you use a
homogenous model, core exit quality is pretty darn
low. Every bubble carries up a lot of liquid.
MR. DI MARZO: See, in the INEL study, the
core exit, what they use was .3. I don't know why
they use .3, but that's what they basically use.
MR. BAJOREK: Around it. It varied from
.1 to --
CO-CHAIRMAN WALLIS: If you use a
homogeneous model for the core though, you carried out
liquid out at no time at all.
MR. DI MARZO: Yeah, that is what the big
differences are. I would like Steve to go into it,
and you will see what the big differences are.
CO-CHAIRMAN WALLIS: When you try opening
a champaign bottle when you've shaken it up.
MR. DI MARZO: Yeah, but this is a much
narrower champaign bottle than the one before.
MEMBER SCHROCK: I guess you didn't
comment at all about the point that I made earlier,
that what you saw in the experiments at Oregon State
so far is pulsating.
MR. DI MARZO: He's going to do that.
CO-CHAIRMAN WALLIS: He's going to come
back.
MEMBER SCHROCK: You're going to do it
later. You're not going to take it into consideration
here.
MR. DI MARZO: Me? Impossible. Setting
this, this is one pot.
MEMBER SCHROCK: thank you.
MR. BAJOREK: No, what we've been trying
to do is basically build a case that as we look for
top down scaling, we see concerns in what our exit
quality is in the ADS. What Marino, we think, has
shown is that when we do simple calculations, yes, we
verify to ourselves that getting this ADS-4 quality
correct is going to be crucial in determining if we
have uncovery (phonetic) or heat-up in the AP 1000.
That leads now into the bottom-up scaling.
Now, I'm going to spend most of the time looking at
entrainment, but I just want to let you know that as
part of the bottom-up scaling exercise, we looked at
hot leg regimes, cold leg regimes. We looked at
flooding in the serge line, used the Yei correlation
to look at core exit void fractions.
As we go through that, we apply our .5 to
2.0 criteria. They fit within that. When we looked
at the hot leg flow regimes, we still stay within the
same two phase regime, although we're closer to a
boundary now. We'll see that in a few minutes.
But the big concerns now are hot leg
entrainment, how it was scaled, what the data says
about that process, and upper plenum pool entrainment
-- and we'll get to that in a second.
Okay. The first thing I want to talk
about is entrainment in the hot leg, and as Dr.
Schrock pointed out, sine there is a couple of ways we
need to examine this flow regime or intended flow
regime in the hot leg, most of our thinking on this,
at least in AP 600 was that this was a smooth,
stratified level that was fairly low in the pipe, and
we could use horizontal stratified correlations to try
to predict the entrainment and the onset of
entrainment.
We have seen work at APEX and in another
separate effects facility, supposedly well scaled for
the AP 1000 and the AP 600, which would suggest that
that level may be higher. It may be in an
intermittent or another one of the flow regimes, and
that this correlation and this process that we have
assumed may not be appropriate.
CO-CHAIRMAN WALLIS: What matters is
whether or not that liquid gets back into the vessel,
spills over from the hot leg. If it gets in there and
it can't get back into the vessel, then no matter what
it's gone out as far -- eventually. So whether or
not it can get back in is what matters.
MEMBER SIEBER: There's nothing to draw
out of there.
CO-CHAIRMAN WALLIS: It just builds up and
eventually it goes out the ADS-4 if it can't run back
into the vessel.
MR. BAJOREK: Hot leg scaling or scaling
for the onset of entrainment in the hot leg.
Westinghouse used an approach that was used in the AP
600. It essentially takes a modified froude number
and uses the correlations that have been developed to
say the modified froude number should be equal to some
constant times a ratio of the free space in pipe, the
region where vapor is fee to flow, H sub B, ratioed
with small D, which is the branch line diameter, and
that should be a small B in the froude number as well.
Now, Westinghouse used this correlation,
scaled the process for entrainment in the hot leg,
found that it was acceptable. We looked at it, and we
see two problems with it.
CO-CHAIRMAN WALLIS: Excuse me. How do
you get hb?
MR. BAJOREK: Well, what Westinghouse did
is they just said hb was equal to capital D, as a link
scale, put that into this expression.
CO-CHAIRMAN WALLIS: Oh, so hb is D.
MR. BAJOREK: Capital D, and when you take
the ratios in that context, I think it basically says
that you're okay as long as you don't double your
branch line superficial velocity.
Well, since the power only went up by 75,
76 percent, yeah, it has to be well scaled then.
Now, we don't have --
CO-CHAIRMAN WALLIS: I'm sorry. Jg-3 is
what?
MR. BAJOREK: That's getting the velocity
in the branch line.
CO-CHAIRMAN WALLIS: But that can't be hb
over D to a power M. I Mean, that doesn't make any
physical sense. That's not an entrainment. That's
the onset. Now, that's the onset.
MR. BAJOREK: This is the onset.
CO-CHAIRMAN WALLIS: That's just the
onset. You've still got to say once it onsets, once
it sets on, whatever the word is, what happens then?
MEMBER SCHROCK: Well, that requires a
separate correlation.
MR. BAJOREK: Right. You need a --
MEMBER SCHROCK: The point I made
originally sort of makes this discussion meaningless,
it seems to me, and that is that this is not a
configuration that exists at the time of concern, and
therefore, what relevance has it in determining the
flow of liquid out the break?
MR. BAJOREK: Let me show you how we tried
to look at this from the scaling. I'll agree with you
that this physical situation is probably not relevant
for what we see in the hot leg. We don't have a lot
of other models and correlations to go on at this
point.
So I want to take a look at it. Let's
assume that we do have a horizontal stratified flow,
but then --
CO-CHAIRMAN WALLIS: But again, I get back
to the question: how do you k now hb? It isn't going
anywhere. So how can you calculate horizontal
stratified flow? It's just sitting in this hot leg
sloshing around, waiting to be entrained. There's no
stratified flow. It's just a pool of liquids.
MR. BAJOREK: It's a pool of liquids.
This correlation --
CO-CHAIRMAN WALLIS: And what flows back
is what matters.
MR. BAJOREK: What this correlation would
say is that gas velocity is sufficient to entrain if
you have --
CO-CHAIRMAN WALLIS: I'm saying you don't
know that. Once your vessel level goes below the hot
leg, there's nothing to hold that liquid in the whole
leg, is there? Doesn't it just drain back into the
vessel?
MEMBER SIEBER: Or go out
MR. BAJOREK: It would depending on
horizontal CCFL at the nozzle.
CO-CHAIRMAN WALLIS: Well, if it doesn't
draw back into the vessel, is held up there, it
doesn't really matter whether it's gone out the --
MEMBER SIEBER: If it goes out ABS.
MR. BAJOREK: At that point, for that
situation, I don't think we would care about this. We
would be more interested in upper plenum pool
entrainment at that point.
CO-CHAIRMAN WALLIS: Yes,
MR. BAJOREK: Okay.
CO-CHAIRMAN WALLIS: That's what matters.
Once you get a little hot leg.
MR. BAJOREK: Right, but the only
consideration that we've seen for entrainment process
is in the hot leg, was Westinghouse used this
correlation.
CO-CHAIRMAN WALLIS: Well, I think what
they predict though is that it's okay, and as soon as
it gets below the hot leg, it shuts off. Isn't that
what you predict? As soon as the level gets below the
hot leg, the mechanism shuts off.
MR. BAJOREK: This would shut off, and --
CO-CHAIRMAN WALLIS: Right. So that the
level hops just around the hot leg.
MR. BAJOREK: Or some low level.
CO-CHAIRMAN WALLIS: That's not serious.
MR. BAJOREK: If we had a high enough
level and it were stratified the way we would say that
you should apply a correlation like that is to take a
look at the gas velocity that you do have and what
would be the hb or, better yet, the hb over D value at
which you would expect entrainment in the AP 1000 or
in the test facilities?
Then hb varies around depending on the gas
velocity. Now, as I think it's been pointed out, that
type of a correlation in the regime that we do have up
there needs to be taken with a lot of distrust. Okay?
We're not sure what that regime is.
That's why I want to look at it. Let's pretend it's
horizontal stratified, but I'm going to look at a
different regime, a few overheads from now, to take a
look at it from a different point of view.
Even if it is horizontal stratified, we
see some problems in trying to scale the data using
this correlation. Principally this was developed from
existing flow type solutions. It ignores roll wave
entrainment, viscous effects, entrainment, the
shearing of droplets from the top of this level.
In addition, it's almost universally based
on data where the small D, the branch line to main
line diameter was very small, a soda straw off of a
one or two inch pipe, as opposed to the ratios that we
would see for AP 1000, which is a little bit larger
than .5.
So from a geometric scaling, we're out of
founds from where this correlation had been developed.
Now, if we let hb float, and we try to
calculate what is that dimensionless ratio at which we
would expect entrainment, if this correlation were
correct, and if we had a horizontal stratified pool in
this hot leg, we find that for the AP 1000, that
ratio, hb over D is larger for the AP 1000 than the
applicable test facilities, which would be SPES for
the high pressure periods of the transient and APEX
for the low pressure periods.
It really doesn't tell us if it's scaled
well or not, but we see it as an indication that the
AP 1000 will see entrainment for a wider range of
depths in the hot leg than we would in the AP 600 or
in any of the integral tests.
So onset is more likely in the AP 1000.
We can't make a judgment on if it's distorted at this
point, one, because we find it a little bit difficult
to apply a .5 to two on a number that can only range
between zero and one. It basically says if you have
a level in the middle anything between .25 and 1.0 is
fine.
CO-CHAIRMAN WALLIS: Now, they're doing
experiments at APEX. So one could see if this .232
whatever it is is actually happening or not. I'm not
sure that there's any confirmation of that number from
the APEX facility, but at least you can check it.
MR. BAJOREK: Okay. It gives us something
to go on, but --
CO-CHAIRMAN WALLIS: Again, we visited
APEX, and our impression was that the flow regime was
not nicely horizontally stratified.
MR. BAJOREK: it was well from it.
When we looked at the hot leg froude
number, one of the things that we found is that
depending on the regime and the qualities that we were
assuming in the hot leg, going through the hot leg to
the ADS, we were in the wavy flow regime, but we're
finding ourselves fairly close to the boundary between
wavy and annular flow.
Now, in this particular figure, I've
picked a condition at low pressure where the quality
coming out of the core was fairly low, and that jams
it over here very close to the transition point on the
Taitel Duckler map. Other cases tend to be further
over to the left, but in the wavy regime.
But this transition boundary is not a very
sharp transition, but it's a more gradual transition
from what Taitel and Duckler described as annular flow
in an annular wavy regime around this line.
So our interpretation is that, well, if
it's wavy or stratified, it's starting to look very
much like annular flow or interfacial shear and
viscous effects are going to be important in the
droplet entrainment.
For annular flows --
CO-CHAIRMAN WALLIS: Excuse me. Don't
you have a co-current (phonetic) flow? I mean, going
back to this, you can't have a co-current flow because
there's nowhere for the liquid to go into the steam
generator. But do you have a counter --
MEMBER SCHROCK: Well, it won't be a flow.
CO-CHAIRMAN WALLIS: Well, I know, but
then this map is for --
MR. BAJOREK: Right.
CO-CHAIRMAN WALLIS: So I don't quite
understand what you're doing.
MR. BAJOREK: What I'm doing is I'm trying
to show that this really can't be interpreted as a
stratified type regime, and that for whatever reason,
even if it were co-current, I'd expect a lot of waves,
and I expect this to have viscous effects so that
mechanisms for entrainment similar to annular flow
should be looked at.
CO-CHAIRMAN WALLIS: This flow regime map
doesn't really apply. You have a short L over D. You
have a steam generator 1M, which is blocking the flow
so that you cannot have a co-current flow there. You
have this inlet at the end, which is giving you a non-
fully developed flow. There's flow around the bend.
Everything is very different.
So it really has to be looked at as a new
problem.
MR. BAJOREK: Okay.
CO-CHAIRMAN WALLIS: You can't just borrow
something from the literature like this that doesn't
apply. That's a no-no. And that is not acceptable.
MEMBER SCHROCK: Well, it won't lead to
success.
CO-CHAIRMAN WALLIS: It's not acceptable.
It's not professional engineering practice when you
know something else is happening to apply something
like this just to sort of invoke the names of Taitel
Duckler. That's religion rather than engineering.
MR. BAJOREK: We agree. It's not --
CO-CHAIRMAN WALLIS: I'm sorry. I'm
beginning to sound like Novak Zuber, but I mean,
just --
(Laughter.)
CO-CHAIRMAN WALLIS: Someone has to do it.
MEMBER POWERS: Well, do we have a map
that's appropriate to this?
CO-CHAIRMAN WALLIS: I don't think so.
MR. BAJOREK: No.l
CO-CHAIRMAN KRESS: We have an APEX text
that would be appropriate for it.
CO-CHAIRMAN WALLIS: A test.
CO-CHAIRMAN KRESS: If you could interpret
it.
MR. BAJOREK: We have an APEX test that
show that there's a lot of oscillations.
MEMBER POWERS: Well, I'm trying to
understand the criticism a little bit here. I mean,
as I understood what you're doing here is you were
saying would I think that there's different physics
applied here than what was assumed when this
correlation was derived.
MR. BAJOREK: Yes.
MEMBER POWERS: And you said for this
particular flow regime, yeah, there's -- just on the
map you say you're close to the boundary, but in fact,
when you looked at the particular paper, you found out
that boundary was funky.
Okay. You don't have a map that's
particularly appropriate to this.
MR. BAJOREK: No.
MEMBER POWERS: This is the best map you
can possibly look at.
MR. BAJOREK: This is the only one that I
can come up with.
MEMBER POWERS: And that led you to say,
well, it's entirely possible that there's different
physics here, and that's the only conclusion you drew.
MR. BAJOREK: That's correct.
MEMBER POWERS: And I guess I'm trying to
understand. How do you fault him for that?
CO-CHAIRMAN WALLIS: Well, I'm trying to
think of something that you would understand, Dana.
(Laughter.)
CO-CHAIRMAN WALLIS: It's like saying --
MEMBER POWERS: We don't have that long.
(Laughter.)
CO-CHAIRMAN WALLIS: We have some chemical
reactions with sulfuric acid, and we have this
correlation or this understanding of that. So we'll
just assume that this applies to nitric acid or --
MEMBER POWERS: We do it all the time.
CO-CHAIRMAN WALLIS: -- or something else.
It's not without any scaling at all.
You're simply taking something that applies to one
thing and apply it to something else that doesn't
apply.
MEMBER POWERS: Well, as I see what the
question is he's posed is not whether he can come up
and use this for some --
CO-CHAIRMAN WALLIS: You can use this. It
says it doesn't apply, I guess.
MEMBER POWERS: I mean, he's not using
this to say, "Ah-ha, here's the answer." He's using
this to say, "Ah-ha, I'd better go get the answer."
CO-CHAIRMAN WALLIS: But it's like
applying, say, a lamina flow method to a Togalin
(phonetic) flow regime. It's a different situation.
It's not appropriate.
MEMBER POWERS: But what are you telling
him, to throw up his hands and say, "I can't tell what
we need to do"?
CO-CHAIRMAN WALLIS: Well, he's saying
we've got to get some more information. That's
essentially what he's saying.
MR. BAJOREK: I'm going to conclude that.
(Laughter.)
MR. BAJOREK: I take it the thing that we
did see of ATWS is that this is a very chaotic new
flow regime. I want to use what I've got rather than
trying to invent a new regime at one of these
meetings. We may need to do that to resolve the
problem.
MEMBER POWERS: And we'll have no trouble
with your reasoning. Sulfuric acid can tell you
something about how nitric acid behaves.
MEMBER SCHROCK: Dana, Graham mentioned
something that should have been highlighted more when
this was first reviewed at OSU, and that is that what
was seen occurring in that apparatus was neither
stratified -- and that's what I seized on and spoke
very strongly about -- nor is it co-current flow, and
both of those things are needed for these correlations
of the form that Steve is showing to have any
relevance whatsoever.
It's the lack of the possibility of co-
current flow --
MEMBER POWERS: But I guess what I'm
asking --
MEMBER SCHROCK: -- in the system at that
time.
MEMBER POWERS: I'm not looking to use the
correlations for anything quantitative. I'm asking
are there transitions in physics that occur in this
flow map, and he says, well, the only flow map he has
is the one he puts up.
He didn't have one for the particular
situation, and he says, "Yeah, they occur there."
It doesn't seem to me a terrible leap of
bad judgment to say I bet there are transitions in the
flow regime if I had the original map.
You're never going to use those numbers.
You aren't going to use those numbers for anything, as
far as I can understand.
CO-CHAIRMAN WALLIS: Well, I don't know.
MR. BAJOREK: The point I want to make is
that we have a very chaotic regime. We think there
was a lot of waves. We saw a lot of waves. We saw a
lot of chaos in this flow. I only have a few
correlations that I could pull out of the literature
that are applicable to known things that I can apply.
I don't have them yet for this physical situation.
If I use the closest I think I can get at
this point at least --
CO-CHAIRMAN WALLIS: There's still a
problem.
MR. BAJOREK: -- there's still a problem.
If I look at entrainment for a co-current annular
flow, okay, which says that, hey, I'm shearing off
droplets from waves --
CO-CHAIRMAN WALLIS: You don't have a co-
current flow.
MR. BAJOREK: -- you don't have a co-
current flow. But we're shearing droplets from waves.
That's as close as --
CO-CHAIRMAN WALLIS: I don't understand
how you get an entree in to the figure because there's
an X, which is the ratio of flow rates, and if you
don't have a co-current flow, you can't even go into
the figure.
MR. BAJOREK: Well, I did that the same
way as the top-down scaling does. It assumes that you
have a co-current flow up and out the ADS-4.
CO-CHAIRMAN WALLIS: Well, let's not dwell
too much on this because there may be other things
like this going on with some of these codes.
MR. BAJOREK: Okay.
CO-CHAIRMAN WALLIS: Of a similar nature.
MR. BAJOREK: Now, as I look at the onset
of entrainment, assuming that I have a sump split in
the system, two thirds, one third based on a single
failure assumption in the ADS and I look at that gas
velocity where you would get entrainment for a co-
current annular flow, I see something kind of
interesting drop out.
It tells me that I would expect
entrainment for that type of a flow in an AP 1000
situation. I would not get it for any of the test
facilities or the AP 600.
Now, it says for AP 600, not getting too
excited on entrainment may have been the right thing,
but it's not anymore for the AP 1000. Not knowing the
flow regime and now seeing from newer tests that we
have a flow regime that's different from horizontal
stratified or co-current annular only gives us more
evidence to say that we don't understand entrainment
in the hot leg for the AP 1000.
CO-CHAIRMAN KRESS: Or the AP 600.
CO-CHAIRMAN WALLIS: Or for any of these.
MR. BAJOREK: Or for the AP 600.
CO-CHAIRMAN WALLIS: Any of these tests or
for any of these geometries like this.
MR. BAJOREK: So our conclusion at this
point is we can't say that hot leg entrainment is well
scaled in these tests relative to the AP 1000 for
several reasons.
Our conclusion at this point is that
Westinghouse has not demonstrated that those processes
were adequately present in the test facilities for the
range of conditions that would apply to the AP 1000.
So we're saying for Phase 3 this is something that we
have to investigate in more detail.
MEMBER SCHROCK: See, a part of your
problem is that the inappropriateness of this was just
as great for AP 600, which is already approved using
a code that imagines the physics as you were trying to
describe them here.
MR. BAJOREK: That's what those numbers
say. AP 600, if you look at it --
MEMBER SCHROCK: And now you've got to
deal with AP 100, where this tradition of not
challenging previously approved concepts comes home to
haunt.
MR. BAJOREK: We think a more critical
entrainment process, however, is this upper plenum
pool entrainment. If we're entraining and we have
these high levels and these intermitting sloshing
regimes in the hot leg, well, that's also a clue
there's still an awful lot of water left in the
system, and we're a long way from core uncovery
(phonetic).
When that liquid is gone and there's not
much of a level in the hot leg and the mixture level
has gone into the upper plenum, we're now looking at
the situation where gas, steam being bubbled through
the core plate, through a diminishing pool, picks up
the droplets, sends them out the ADS. There's a trace
of liquid in the --
CO-CHAIRMAN WALLIS: And your previously
work suggested that if it gets into the hot leg it's
gone.
MR. BAJOREK: Yes, yes. So some might be
entrained, but our assumption is if it's entrained
here, it's gone.
Now, this also comes about from tests that
were run in the APEX facility following most of the AP
600 work. These are what we would term the no reserve
tests. They were beyond design basis tests that they
basically shut off the accumulators, the injection
flows to the system, drained it down to the bottom of
the hot leg, opened up the ADS-4 starting from 100 and
200 psi initial pressures for a range of pressures.
What they found -- and this is power down
here at the bottom versus pressure on this figure --
in some cases it was sufficient to blow out all of
that mass in the upper plenum.
All of the tests as you look at them in
the hole suggest that upper plenum entrainment is
real. There's a large amount of it, and
Westinghouse's reranking of that process and the PIRT
from the medium to a high was correct.
I think what Dr. Schrock had noted maybe
in AP 600 that should have been a high and should have
been looked at in greater detail.
CO-CHAIRMAN KRESS: Let me understand.
You had this whole system closed up at temperature and
at pressure and with a certain water level, and then
you opened up the ADS-4. So you get a flashing
process going on.
Now, that's not exactly what happens in
the AP 600 when you get down to that level. You've
flashed a long time ago, and now you're boiling,
aren't you?
MR. BAJOREK: These tests are not
indicative of whether you should get core uncovery or
not, but they're showing that when you are having some
flashing, power generating steam in the bottom of the
core, that you are generating the type of gas
velocities in the upper plenum that's sufficient to
entrain a lot of fluid.
It doesn't necessarily mean you're going
to get core uncovery because purposely in those tests,
they've shut off IRWST, cumulators and other things
that would help replenish that mass.
CO-CHAIRMAN KRESS: Sure, I understand,
but my point was that you'll never get to that stage
in the AP 600 because once you open ADS-4, you will
finish your flashing process long before the water
level gets down to where you worry about this process,
and then the steam that's entraining is steam coming
from the decay heat, and I just don't think --
CO-CHAIRMAN WALLIS: But as long as the
pressure is dropping, you've got flashing.
CO-CHAIRMAN KRESS: Yeah, but not --
MEMBER SCHROCK: The most important
flashing is in the flow path from the low mach number
regions into the critical flow zone, and there there's
a considerable amount of flashing.
CO-CHAIRMAN WALLIS: But that's not in the
core. It's not in the upper plenum, is it?
CO-CHAIRMAN KRESS: No, that's outside.
MR. BAJOREK: Here, again, I'm going to
look at this in terms of a steady state process. I'm
not going to worry about the flashing, but I'm going
to look at the gas velocities through the upper plenum
that is due to the steam that is being generated in
the core.
So we're going to throw away the flashing
component. Even though there were tests in that APEX
no reserve that started at fairly low pressures that
I would say were indicative of the end of the ADS-4 --
CO-CHAIRMAN KRESS: That might be more
indicative. I'll go along with that.
MR. BAJOREK: Okay. Here, again, we find
ourselves in looking for correlations that may not be
applicable to this situation. I've listed several of
them up here, and I'll talk about why in just a
second.
But what I want you to note is the way we
would look at entrainment in a non-dimensional fashion
is this Efg parameter, which is the ratio of the
entrained flux to the gas flux that enters a certain
region.
Now, several works have been done on this.
They have looked at bubbling gas through relatively
large diameter pools, not complicated with guide tubes
and upper plenum structure. Okay? So there are
atypicalities of them.
Several of them had been proposed, an
earlier one by Rozen, some Russian workers. I can't
pronounce this guy's name.
Most recently, and perhaps the best work
in this country, by Kataoka and Ishii, who did some
studies in the mid-'80s where they took data from
previous investigators and developed some non-
dimensional, more mechanistic type of correlations
based on what they had.
One of the things that you want to note
from the equations is that they depend primarily on
the gas exponent or the gas velocity to some power,
three to four depending on how you define your regime,
and these people did it in different ways, and how far
you actually have to carry the droplet before you're
up and out of the system.
So H or distance enters into there in some
format, either in the exponential or in a
dimensionless form the way Ishii treated it.
The important thing to note right now is
that the sensitivity between the entrainment and the
gas velocity, third to the fourth power. Okay. Well,
let's assume everything was scaled fine for the AP
600, that we're down at about the same pressure. So
we don't have to worry about the H or the H star. We
don't worry about geometry. We can throw away the
delta rho over rhos.
It tells me that entrainment should scale
something like J sub g to some power, which ranges
between three and four.
Well, that's a direct relation to power.
Throw the power in there, and without a whole lot of
work, it tell us that entrainment in the AP 1000 is
going to be five, maybe ten times what we see in the
AP 600, and presumably the tests, if they were as well
scaled for that.
Now, in looking at this, we can look back
at AP 600 and look at the AP 1000 documentation. No
one had ever scaled that before. This is something
that went through the cracks.
CO-CHAIRMAN WALLIS: No one had ever
scaled one of these very important phenomena?
MR. BAJOREK: This I couldn't find it. I
asked Westinghouse, and they told me no. It hadn't
been looked at.
So we took a look at the test facilities,
APEX, SPES, ROSA, and of the correlations which are
available, I think that the Kataoka and Ishii is the
most complete set of work that's available. So I
said, "Let me look at that."
What they do, and this isn't in your
package, is they break entrainment up into several
regimes. If you have a level up near the hot leg,
you're in what they would call a near surface regime,
which means any kind of a gas flux entrains virtually
everything. It doesn't exist for very long.
As you entrain more and more from a pot or
a pool or an upper plenum, you go into what he refers
to as a momentum controlled regime, and that depends
on the gas flux. Okay?
The exponent increases. Okay? It's three
for this intermediate flux regime, which I just showed
you. It's up to seven in his report or 20 when you
get to a high gas flux regime.
And then eventually as you drain this
level to a low enough, you enter the deposition
controlled regime, which I really interpret as being
a no man's land. It says you can't analyze it because
deposition has a bigger effect.
Now, as I looked at the facilities, the AP
600 and the AP 1000, I find that we're in the momentum
controlled regime. We don't get down to the
deposition controlled regime, and we remain in this
intermediate gas flux regime in all of the facilities
and the AP 1000.
So that Jg to the three correlation that
I showed you on the last page I think would be the
most typical one to use. So I'd define a scaling
parameter, this pi sub up, entr as being this upper
plenum pool entrainment parameter, and I'd say, well,
let me define this as the package of terms from
Ishii's correlation Jg over H to the cubed times the
hydraulic diameter, one and a half, ratioed that from
the test to the AP 1000.
And I said, well, let's assume that we had
the same pressures. That lets them get rid of the
star terms on there. It leaves me with this
expression at the bottom of the page that I leave in
terms of the core power area available for flow in the
upper plenum.
Delta Z, the distance between the bottom
of the hot let and the top of the active fuel. And
when I scale, when I pull out numbers from INEL,
numbers today obtained from Westinghouse for areas,
lengths and so forth, and the power factors, I wind up
with this table.
Now, the scaling ratio that I defined is
on the next to the last one, but I think an easier way
to look at this is the one over pi up, entr, which
gives me the relative entrainment in AP 1000 to what
I saw in the test facilities.
AP 1000 should have about 18 times the
entrainment that we see in ROSA, 156 times in SPES,
only six times APEX. There's a saving grace here.
APEX was a one quarter height scale.
So where it doesn't have the correct J sub
g, it makes up for that because there's less height
you have to carry fluid out of the upper plenum. We
might call this a compensating error if we look at how
much time it would take to train the facility. It is
saying that Apex may not be unusable.
CO-CHAIRMAN WALLIS: Excuse me. This H is
the height from the hot leg down to the core?
MR. BAJOREK: Top of the core, yes.
CO-CHAIRMAN WALLIS: When you're above
that, you're entraining much more rapidly.
MR. BAJOREK: Oh, yeah. Now, if I just
defined the scaling parameter, I would use this to
demonstrate the problem and our concern.
Now, since I put these numbers together,
I have taken a step back and used this initial form of
this scaling parameter with the dimensionless terms in
there. That allows me to look at different pressures
and play games with heights.
Even when I do some of those
sensitivities, I still get these entrained -- I still
get this parameter out of range. It improves a little
bit. This .16 might go to .2, maybe a .25.
CO-CHAIRMAN KRESS: When you take your
equation or the Efg which you think is most
appropriate -- I guess it's this bottom one -- and
plug in some numbers or say AP 1000, what do you get
for an absolute value rather than these ratios?
MR. BAJOREK: We took that out of the
overhead yesterday. I'm sorry.
CO-CHAIRMAN KRESS: Because that's going
to tell you the importance. It's an importance
measure if you know the absolute value. You know, if
you're not --
MR. BAJOREK: Well, not now because I'm
not comparing it to anything else like I was in the
top-down scaling.
CO-CHAIRMAN KRESS: Well, yeah, I mean if
you assumed this model is correct and you plug in the
things for AP 1000, it's going to give you an idea of
how much liquid is entrained to go out with the gas,
and you know or you've got an idea of the gas flow.
So it's an importance measure in my mind if you had
that absolute number.
MR. BAJOREK: Well, the way I've tried to
look at this, and I don't have those results and could
spend half a day looking at this, is in terms of how
long it would take to drain the upper plenum in an AP
1000 and an APEX facility, and how does that time
compare to the transition time from ADS-4 to IRWST.
So far I've been able to convince myself
that APEX drains at about the same time, but
preliminarily, it still tells me that I could
potentially deplete the mass in the upper plenum
before I've completed that transition.
I'm not far enough along on that to say
whether we've got an uncovery or whether there's
plenty of water. I'm very comfortable, however,
looking at these numbers and concluding that upper
plenum entrainment is something that we definitely
need to look at in more detail in Phase 3 of this
review.
Keep in mind that this is a dominant
process in the most critical small break that we would
see for an AP 1000, the double ended guillotine break
for a DVI line where you would expect the two phase
level to be somewhere in that upper plenum. So
entrainment, okay, is going to be higher in the plant
than in the test, and our question is: if we're off
in however this is modeled in the core -- in the code
-- excuse me -- are we potentially claiming there's no
cover uncovery in heat-up when, because of test data
not being in the right regime, we may actually see
some type of a heat up?
Now, in looking at the analysis, the
RELAP, looking at the type of entrainment on here, I
think we can say at this point if we do get a core
uncovery, it's probably not a very deep one, and it's
probably not a very prolonged one. It still gives us
the appearance that there's a lot of water in the
system that has to be swept out in addition --
CO-CHAIRMAN WALLIS: Well, once you begin
to uncover the core, you don't have a pool anymore in
which you've got entrainment. You've got little
channels in which you've got entrainment.
MR. BAJOREK: Right.
CO-CHAIRMAN WALLIS: And the stuff is
being pushed along the channel wall. One might wonder
if it's actually worse entrainment because it's in a
little tube, and it's got a launcher for its droplets
instead of being in a pool.
But I'm not sure.
MR. BAJOREK: That's one of our questions.
At what point when that level drops in the upper
plenum can you really consider it a pool anymore and
you have to start looking at localized effects and
jets in various regions of the upper plenum?
CO-CHAIRMAN KRESS: Will given my past
experience with rod bundles and the opening in there,
with other things that are similar to this, you can
treat it as a pool. Just forget the rods, but you
know, that may not be true, but that's my experience.
MR. BAJOREK: So our conclusion with
regards to upper plenum pool entrainment is that we
think that this represents a distortion between AP
1000 and the test data. We see a nonconservative
distortion in that we would be losing more mass out of
the AP 1000 than any of the test facilities.
At this point we haven't seen a scaling
rationale from Westinghouse, and we don't see evidence
that this test data or other test data that you might
want to consider for an entrainment effect is
appropriate for the AP 1000.
So we think that in Phase 3 this is
another area that we would need to look at, and I
guess I would have to say I would consider this one a
more critical than the hot leg entrainment.
CO-CHAIRMAN WALLIS: Do you know what
equations Westinghouse uses to predict entrainment?
MR. BAJOREK: In COBRA TRAC, but not
NOTRUMP.
CO-CHAIRMAN WALLIS: You could make that
comparison between Ishii's -- use an Ishii's model.
They use something else. They must use something, and
presumably you can find out what they do use and make
that comparison.
CO-CHAIRMAN KRESS: I thought I remembered
that they use test data from the 2D3D program, but I
may be thinking about another code.
MR. BAJOREK: Well, in the core model I
know we used or Westinghouse used a model that was
benchmarked or had a very close relation to one of the
Ishii correlations. It's something different than the
upper plenum. They used the upper plenum test
facility to try to look at that for a large break.
We're looking at small break, and we'd
have to look at NOTRUMP for that.
To wrap up and to give some conclusions,
I don't think we want to lose slight that a lot of the
test data, a lot of this integral effects data is
still pretty good for the AP 1000. It still has a lot
of use.
We see a couple of exceptions. Hot leg
entrainment, we're not sure how we should treat it.
We don't know what the flow regime is. We think that
we're probably at a situation where we would expect to
see that onset in the AP 1000, but not in the test.
For the upper plenum pool, we think
there's going to be a lot more entrainment in the AP
1000 than was observed in any of these three integral
test facilities.
CO-CHAIRMAN KRESS: Do you know how much
entrainment was observed in the integral test
building? Have you gone back and extracted that
information?
MR. BAJOREK: We haven't yet, but that's
one of the missions right now, yes.
CO-CHAIRMAN KRESS: Because it was a small
value in the first place, and even though AP 1000 may
be considerably more of a small thing, it can still be
a small amount.
CO-CHAIRMAN WALLIS: It may be zero. One
hundred times zero is still zero.
(Laughter.)
CO-CHAIRMAN KRESS: Good point.
MR. BAJOREK: Anyway, that's where I'd
like to conclude and wrap up, with the idea that it's
entrainment processes that we need to look at in a lot
more detail.
CO-CHAIRMAN KRESS: At this point we have
on the agenda to hear from Westinghouse or the other
alternative is to have break and then hear from
Westinghouse. But I'd ask how long Westinghouse might
be.
MR. BROWN: We're going to be here
tomorrow. So --
CO-CHAIRMAN KRESS: Okay. Well, it would
be a good time to comment on what you've already heard
now rather than to wait.
MEMBER POWERS: Rather than to let you sit
on it.
CO-CHAIRMAN KRESS: Rather than to let you
sit on it. So the question is about how long would
that do you think take you.
MR. BROWN: Well, do you want to take a
break and come back? Is that what you're thinking?
CO-CHAIRMAN KRESS: Well, that's what I'm
trying to decide. If you're going to take -- if it's
not going to take you long, well, we'll go ahead and
hear you.
MR. BROWN: He has quite a bit
presentation.
CO-CHAIRMAN KRESS: Well, let's take a ten
minute break.
CO-CHAIRMAN WALLIS: Could I say I really
appreciate this sort of discussion from the staff?
And it's a real breath of fresh air compared to --
CO-CHAIRMAN KRESS: Oh, yeah.
CO-CHAIRMAN WALLIS: -- literally that
they've met the requirements and everything is okay.
In thinking about what really happens, it's a breath
of fresh air.
CO-CHAIRMAN KRESS: Yeah, we appreciate
that.
(Whereupon, the foregoing matter went off
the record at 5:11 p.m. and went back on
the record at 5:23 p.m.)
CO-CHAIRMAN KRESS: We can start again.
MEMBER SIEBER: Sort of.
CO-CHAIRMAN KRESS: Scaling.
MR. BROWN: Some of this, most of this is
repeat. So I'll try to go over this pretty quick.
As you all know, we're talking about no
new phenomena. We found that entrainment certainly
ranked higher for AP 1000.
We submitted a WCAP, AP 1000 curb scaling
assessment, and we also answered quite a significant
number of RAIs associated with that. The one thing we
did add, you know, additional work, was we also scaled
the ROSA facility. Originally we scaled SPES and
APEX, and we also added ROSA on top of there.
We also had some additional work here
trying to address some of the ACRS comments,
specifically yours, Dr. Wallis, from before.
CO-CHAIRMAN WALLIS: You did a
cylindrical, symmetrical CFD model or something
instead of a slab?
MR. BROWN: We did a pie shape.
CO-CHAIRMAN WALLIS: Okay, pie.
MEMBER POWERS: Oh, it's truly 3D.
MR. BROWN: Well, you could say that.
We had two areas of importance which we
just went over. Of course, we're talking about upper
plenum and hot leg. The main area I wanted to focus
on here was obviously back in the upper plenum, and we
had done some work. Obviously Steve has discussed
that already before about some of the stuff from the
hot leg, but I had a little bit of maybe a little
different way of slightly looking at it as far as the
upper plenum. So I wanted to go into some of that.
MR. BOEHNERT: Did you use the same
correlation?
MR. BROWN: Yes, I did. Yes, I did, but
I guess I don't need to get into it too much, but one
of the things I did want to mention a little bit is
that when you look at the Kataoka-Ishii work and he
talks about the near surface region and momentum and
he talks about the near surface region and momentum
control region, the near surface region was found at
least in their work to be independent of Jg and H, and
it essentially correlated with density ratio.
So it seemed to me that for events in
which we had a level in the hot leg where we didn't
really have to lift the droplet very far, whenever the
facility would get into pressure similitude with the
plant, then from that standpoint it would seem like
there was not really a serious scaling issue there.
It really is whenever you get levels where
you drop below the hot leg is where you really get
interested because that's really where the Jg and the
height and so on really come into play.
So I think for the majority of events in
which we get these smaller breaks and so on as far as
the upper plenum is concerned, entrainment, I don't
think there's a serious scaling issue there. I think
the issue is with, for example, the DVI break where we
actually start to drop below.
Now, I think Steve would agree with that.
MR. BAJOREK: I agree with that.
MR. BROWN: Okay. I could skip on a
little bit there. I guess it was good to see at least
Steve and I must have made the same error or got to
the same place here as far as scaling is concerned,
but going back to looking at this regime where you're
below the hot leg and so you have the height and Jg
dependency to the third power, it essentially just
shows that you got to the same places trying to scale
the entrainment ratio.
And when I went through the exercise, I
struggled a little bit with trying to come up with a
reference value for velocity, quite honestly, because
you're going through the upper core plate for the
upper core plate holes, and then you are going to move
into the main part of the upper plenum where you have
all of these guide tubes there.
And so trying to pick, you know, a
characteristic velocity is, I think, pretty tough. So
what I did was essentially try to look to sort of try
to see if I could hit the extremes, and one was where
you went into looking at the flow area through the
core plate itself, and then finally when you went
through all of that and then get to where you've just
got the guide tubes to contend with as sort of a bound
on that.
CO-CHAIRMAN WALLIS: Don't the guide tubes
de-entrain some of these things, the droplets, too?
MR. BROWN: Yeah, they sure do. I think
that's probably even less well understood than
entrainment unfortunately unless you really want to go
there.
CO-CHAIRMAN KRESS: The core plate is how
thick compared to the length of the guide tubes?
MR. BROWN: Thick? It's not very thick at
all. I mean, compared to the length? You said
compared to the length of the guide tubes? Yeah, it's
pretty small.
CO-CHAIRMAN KRESS: You know, that would
to me say it's more appropriate to use the guide tube
velocity
MR. BROWN: Yeah, but the only thing I was
thinking is that depending on perhaps where the level
was, if you were talking about a level that was maybe
you would -- if you were to think that there might be
an event where there was a level that was approaching
the guide tubes, you might be -- I mean approaching
the core plate, then maybe you'd be closer to the core
plate and vice versa.
CO-CHAIRMAN KRESS: The percentage of time
it's there is going to be small.
MR. BROWN: Let's hope we don't get there,
but anyway, when I did that --
CO-CHAIRMAN WALLIS: Well, excuse me.
These facilities like APEX, do they have simulated
guide tubes in them?
MR. BROWN: Yes, they do.
CO-CHAIRMAN WALLIS: They do?
MR. BROWN: Yes, they do.
CO-CHAIRMAN WALLIS: Does SPES have that?
MR. BROWN: I don't know.
CO-CHAIRMAN WALLIS: I think APEX does,
but I'm not sure the other facilities do.
MR. BROWN: Well, the one I think that
concerns me the most is the ATWS facility that you've
been discussing here a bit earlier before, looking at
some of the entrainment in the hot leg, is the fact
that you know, I don't know enough about it so you
guys certainly know more than I, but when I hear
about, you know, a wave bouncing around back and forth
and I've at least seen some papers on that and talked
to Dr. Reyas out there, it doesn't sound like there's
any upper internals, and I begin to wonder about the
boundary conditions on this thing even being
applicable to an AP 1000.
But, yes, the APEX facility itself does
have upper guide tubes in it, yes.
MEMBER SCHROCK: Yeah, I think you're
right. There are a lot of reasons that's not a very
similar.
MR. BROWN: Yeah, it's hard to say what
its relationship is.
MEMBER SCHROCK: Only we're initially
pulling the gas through the pool that's air-water, and
they're blowing the air through the pool, and Graham
suggested they try putting it in the upper plenum to
see what happens. I guess they did try that, didn't
they?
MR. BAJOREK: They did try that.
This is Steve Bajorek from Research.
The oscillations that were seen injecting
from the porous injectors from below were also
apparent changing that.
MR. BROWN: Were you suggesting to blow it
horizontally, vertically downward or --
MR. BAJOREK: Well, this would go through
the top of the vessel.
MR. BROWN: The top of the vessel?
MR. BAJOREK: But those tests are -- were
intended to take a look at entrainment and flows in
the hot leg.
MR. BROWN: Right.
MR. BAJOREK: They were never really
intended to take a look at the pool entrainment.
MR. BROWN: Yeah. It just did look like
there was quite a bit of an influence potentially
there, but to answer your question, APEX is probably
as prototypic, I guess, as you can probably attempt to
do in an integral effects test facility in that
regard.
But anyway, as you can see, certainly you
would expect from this here that we're going to get a
significant amount of entrainment in AP 1000 relative
to OSU, and I certainly would agree that there's
distortion here, and I don't think that we can claim
that this is necessarily from looking at this type of
analysis that this is, you know, well scaled, but this
is how I came at it. So --
CO-CHAIRMAN WALLIS: So you're
substantially in agreement with --
MR. BROWN: yes. I think though that I
guess the question is, which goes back to what Dr.
Kress has asked, and I don't think anyone has given a
good answer, and I think even early on when APEX --
I think Dr. Reyas was asked this -- is, you know, well
what do you do when something like this is a factor of
two, is a factor of three?
Because Dr. Wolfgang Wolfe from
Brookhaven, when he was doing scaling in AP 600, he
used a factor of three. Westinghouse used a factor of
two. Looking in the FSER, the NRC never really came
out and said what the criteria is. They sort of just
reiterated Westinghouse's criteria.
So I don't think anyone has given a
satisfactory answer.
CO-CHAIRMAN WALLIS: What's in your code?
I would think what matters is what's in your code. Do
you use this Ishii-Kataoka?
MR. BROWN: I don't believe we use it
explicitly, no.
CO-CHAIRMAN WALLIS: So what is --
MR. BROWN: Probably if you want to know
with COBRA TRAC, Dr. Katsu Ohkawa of Westinghouse here
could answer your question with respect to COBRA TRAC.
MR. OHKAWA: Katsu Ohkawa of Westinghouse.
I can speak for COBRA TRAC, and Steve
mentioned earlier, and he was correct, that we use the
form very similar to Kataoka-Ishii in the core. In
the upper plenum we only looked at EPTF 3D, the
sterility type (phonetic).
CO-CHAIRMAN WALLIS: Kataoka-Ishii in the
core is for an annual flow regime. Do you have
another correlation for annual?
MR. OHKAWA: No. The core behave much
more open than the annual.
CO-CHAIRMAN WALLIS: So you use a pool
correlation for the core?
MR. OHKAWA: As a starting point, and then
we modified it, and then checked against FLECHT.
MEMBER SCHROCK: And what does that code
use for the ADS break flow, ADS-4?
MR. BROWN: The hot leg ADS-4, we use the
froude number type at inception.
MR. OHKAWA: Currently we use the froude
number correlation. Okay?
CO-CHAIRMAN WALLIS: So what's your
conclusion?
MR. BROWN: Well, we're going to get lots
of entrainment in AP 1000, and we certainly see more
in that than we expect in --
CO-CHAIRMAN WALLIS: Well, it's not clear
that you get lost. You get more than an APEX, but
does it matter?
CO-CHAIRMAN KRESS: That was my other
question.
CO-CHAIRMAN WALLIS: Maybe it doesn't
matter.
MR. BROWN: Well, I think as long as we
have levels in the hot leg, I think we don't care too
much. I think we're concerned about if it goes too
low.
The next thing I wanted to move into,
looking at this thing with this liquid entrainment,
may be, you know, a lot of salt here needed, but it
may be a little bit to address your question, Dr.
Kress. I had the same interest of, well, okay,
everybody has been scaling this entrainment ration,
and you know, we're getting back into, well, is it a
factor of two, a factor of three, but how big is it?
How important is it? What does it look like?
Well, I made a crude attempt in the short
time I had to try to do that, and I'm not going to
claim here that this is an extremely rigorous model,
but I made an attempt here.
I said, well, what if I have a level here
that's dropping at the hot leg or below and I looked
at the region from the core plate up to the bottom of
the hot leg and said, well, what if I had some
situation where I have a two phase mixture in this
region.
I looked at the time approximately when
ADS-4 would initiate and said, okay, what if I were to
take this entrainment upper plenum correlation where
we were in this momentum controlled regime here, and
I just applied enough liquid to satisfy core decay
heat. What would happen as a function of time with
this level? And just kind of see where this would
fall out for AP 1000.
So I started off with just, you know, a
very simple conservation, a mass equation where I had
a two phased mixture in the upper plenum region from
the core plate up to the hot leg initially, and I've
got the core flow in and I've got the steam passing
through this, and I've got an entrained liquid at the
surface from this Kataoka-Ishii correlation here. And
I reference my two-phased mixture level to the top of
the upper core support plate.
So I used the simplified Yei correlation
to estimate what the void fraction would be in this
region. Then I, of course, was using the entrainment
model using Kataoka-Ishii and this regime which has
this form which Steve showed you before, and I
applied, even though this certainly is still a
transient, but I applied just to see what would
happen.
At the peak period of decay power when
ADS-4 would go off, what was, you know, the mass flow
through the core? And, again, I neglected things that
might help like subcooling and so on to try to
maximize the amount of steaming I might get in there.
So I took those with the initial condition
of starting from at the top of the upper plenum near
the hot leg, that is. What I found was this thing
really immediately grabbed liquid and took it out of
there quite fast, very, very rapidly and dropped it to
-- this is roughly, I think, -- is roughly a meter in
AP 1000 from the upper course of the core support
plate to the bottom of the hot leg, and as you can
see, within a few seconds it seemed to settle out to
kind of a steady state of --
CO-CHAIRMAN WALLIS: That's because it
goes down to H cubed.
MR. BROWN: Right, H cubed. So you can
see while Jg is killing you, H is helping you. If you
can get away from that it can also stabilize. So I
think, again, probably from our perspective it's --
CO-CHAIRMAN WALLIS: Well, I've thought
about that. There must be a certain Jg which is big
enough that H no longer matters, and once you entrain
everything and carry it up, you're going to carry it
over. If the droplets can't fall down, it doesn't
really matter what H is. They just keep on going.
You need a moisture separator or something
to get them back.
MR. BROWN: Yeah, although I guess if you
look at the Kataoka-Ishii work, you know, and I think
Steve put up the curve before, yeah, I mean, you see
though that basically with H you're dropping.
CO-CHAIRMAN WALLIS: There must be some
range though with this correlation. You can't go up
to J cubed forever because, again, you know, you
predict for that amount huge amounts of entrainment
for that.
MR. BROWN: Well, you do get the large
amounts as you go back up to the near surface region.
CO-CHAIRMAN WALLIS: But I mean eventually
if you took the correlation too far with a big Jg,
you're going to get entrainment which is bigger than
homogeneous flow, and it's physically impossible.
MR. BAJOREK: This is Steve Bajorek.
I think what you're referring to is it
comes up in this type of a gas regime. There are some
criteria on how you switch from the correlation that
you see there to other ones, and I think what you're
referring to is the situation that when you do get a
gas velocity that's so high, it just sweeps everything
out.
CO-CHAIRMAN WALLIS: That's right. The
gas loss is bigger than the terminal speed of the
biggest drop you could have. t hen everything is gone.
MR. BAJOREK: So even if you have a large
H, it still sweeps it out.
CO-CHAIRMAN WALLIS: That just be a Jg
star because that's what Jg star is. If Jg star is
bigger than a certain amount, everything goes.
MR. OHKAWA: This is Kat Ohkawa from
Westinghouse.
Yes, there's a flooding limit, and it
should go back all the way to that new surface,
according to the paper, and that's probably
appropriate.
MR. BAJOREK: Right.
CO-CHAIRMAN KRESS: So your basic
conclusion is that this is more or less self-limiting.
MR. BROWN: Yes. I think that's what I
thought was interesting to me to share with you, that,
you know, you can look at Jg and at first you can get
pretty startled by it, but as you, you know, begin to
look at it a bit further, you realize that it's a big
limiting, and again, looking at AP 600, many of the
tests tended to have an oscillatory level up and down,
almost sort of the self-correcting type of behavior
where as you get a level up into the hot leg, of
course, you get a hellacious amount of entrainment,
which, you know, knocks it back down, but then as you
get away from the hot leg, it tends to reduce
significantly, and then you can catch up with the
injection again and fill it back up, and it would go
back and forth.
MEMBER SIEBER: At the same time you're
moving an awful lot of heat, which should tend to
bring the entrainment down.
MR. BROWN: Down. That's right. In fact,
you think about the other end of this is we look at
imagining, well, if I get a pretty good slug out
there, then shouldn't I also be slowing down in my
entrainment because I've essentially plugged or
temporarily blocked my outlets?
So you can't just look at these things
from a separate effect, and we all get spun off on
this here, as I see us in this conversation, but you
know, I look at it as, you know, you can't really get
too carried away with that because you recognize
what's -- you have the H in here, which we're
restoring you. Plus you also have the effect that you
have on the -- I agree that the entrainment is
important, but also it's going to effectively increase
the resistance significantly on the outlets and now
you can't entrain as much. So the whole thing slows
down again until you catch up.
CO-CHAIRMAN WALLIS: Well, there's
something very strange. I can understand Jg star. Jg
star is the ratio of the gas flux to the terminal
speed of the biggest drop you can have.
MR. BROWN: Yeah.
CO-CHAIRMAN WALLIS: That's what it is,
and H star is the ratio of the height to the size of
the biggest drop you can have.
Now, the biggest drop you can have is
going to be on the order of millimeters. So I'm
rather bothered about giving a height of, let's say,
a meter by a millimeter size droplet. It doesn't seem
to make sense to me.
So the scaling of H is weird. The scaling
of J makes a lot of sense. Physically the way H is
designed is H star. The characteristic dimension of
the droplet scaling the size of the vessel doesn't
seem to me quite right.
Well, Jg star is what you get in a thunder
storm. If you get a big enough up draft, you carry
the raindrops up instead of down.
CO-CHAIRMAN KRESS: Well, you have to
think about that Graham because what I envision is you
impart -- the gas flow determines some sort of
spectrum or sizes for your droplets that you're
entraining, and they have a certain momentum that they
get started with.
Now, that momentum may or may not be
enough to carry that droplet up far enough to get it
out of there, and so the H related to the droplet
size, which is somehow related to the momentum through
the velocity might be an appropriate scale, although
it's --
CO-CHAIRMAN WALLIS: That is true. If you
look at firing particles into a gas --
CO-CHAIRMAN KRESS: Yeah.
CO-CHAIRMAN WALLIS: -- or the rain -- the
distance they go is scaled and total flow linearly
with the size of the particle.
CO-CHAIRMAN KRESS: Yes. So it might make
some sense.
CO-CHAIRMAN WALLIS: It's a transient like
that.
CO-CHAIRMAN KRESS: Yes.
CO-CHAIRMAN WALLIS: They get thrown out,
and then they go up in the trajectory and fall back
down, and you're right. But if they're carried out
completely --
CO-CHAIRMAN KRESS: Then it wouldn't
matter.
CO-CHAIRMAN WALLIS: -- because they're
just lifted up --
CO-CHAIRMAN KRESS: That's right.
CO-CHAIRMAN WALLIS: -- then it doesn't
matter anymore. So this has got to be a regime where
it just doesn't matter anymore, but I don't think
you're -- do you know how much -- how big is your Jg
star? What's the number? That would be very
revealing.
MR. BROWN: Offhand I don't know.
CO-CHAIRMAN WALLIS: That would be the
first thing I'd calculate, would be the ratio of the
velocity of the vapor to the terminal speed of the
biggest drop I could put in there, and if it's above
that, then it's all gone no matter what you do.
MR. BROWN: and I thought this was kind of
instructive to --
CO-CHAIRMAN WALLIS: Did you know, Steve,
how big the Jg star is?
MR. BAJOREK: I can get those numbers for
you tomorrow.
CO-CHAIRMAN WALLIS: That would be good.
MR. BAJOREK: I'm thinking that they're on
the order of one and a half to two, but I'm -- this
scaling, you look at a lot of numbers, and I'd have to
check, but I'll get those for you tomorrow.
CO-CHAIRMAN WALLIS: If it's one and a
half to two, then I would think you're in real
trouble.
CO-CHAIRMAN KRESS: Yeah. That's what --
CO-CHAIRMAN WALLIS: But still maybe it's
two times ten to the minus two or something. That's
more like what you mean. You'll get that tomorrow
anyway. That would be very good.
MR. BROWN: Anything else on that one?
CO-CHAIRMAN WALLIS: Well, again, I'm very
suspicious of a correlation which has a coefficient of
E to the sixth.
(Laughter.)
MR. BROWN: We can change it if you like.
CO-CHAIRMAN WALLIS: It indicates to me
that some of the physics is wrong.
MR. BROWN: We can --
MEMBER POWERS: You must have real
troubles with homogeneous nucleation then.
MEMBER SIEBER: Just need another
offsetting.
MEMBER POWERS: It has surface tension
cubed in the exponent.
CO-CHAIRMAN WALLIS: Yes.
MEMBER POWERS: Well, I mean, that's even
more than your sixth.
CO-CHAIRMAN WALLIS: No, these are the
sixth -- the number, the dimensionless number
represented by this correlation is 5.417 E to the
sixth. That's very suspicious.
CO-CHAIRMAN KRESS: Yeah, yeah.
CO-CHAIRMAN WALLIS: Unless there's some
kind of units. No, there isn't a units problem here,
is there?
CO-CHAIRMAN KRESS: No, those are all
dimensions.
CO-CHAIRMAN WALLIS: Anyway, this is the
best correlation we've got.
MR. BROWN: This is the best I'm aware of.
I was told you did some work earlier in your days,
but --
CO-CHAIRMAN WALLIS: Yes, but that was --
MR. BROWN: One of the reasons for
thinking this is the best we've got, yes.
(Laughter.)
MR. BROWN: Dr. Powers paid me for that
one.
(Laughter.)
CO-CHAIRMAN WALLIS: That was when I was
as student, you mean?
MEMBER POWERS: Well, maybe it's better
than we thought.
CO-CHAIRMAN KRESS: Well, this at least
suggests what I wanted to see, is how important is it.
MR. BROWN: Yes, yes.
CO-CHAIRMAN KRESS: It's a way to address
it.
MR. BROWN: It is not quite a top-down
scaling, but it was sort of an attempt to do that and
sort of put it in perspective a little bit, and I
think as I said before that you can't lose sight of
the fact that as you begin to entrain such a huge slug
where everybody is concerned about this, then you have
to look at, well, this meant this probably temporarily
sort of degraded the vent path a bit so that your
entrainment is going to slow down.
CO-CHAIRMAN WALLIS: I'll tell you the
problem is probably that MUG, depending on viscosity
of the gas, is a small number, which is why you need
this huge number to multiply it by, and I'm wondering
if it should be there. Because even if the gas had no
viscosity, you still would be entraining droplets. So
it's something weird.
Again, I'd have to look at the paper.
Could we have this paper? Could you folks get
Kataoka-Ishii and send us a copy?
MEMBER POWERS: It's two topical reports.
They're relatively thick topical reports.
MR. BROWN: I have a bad copy of the paper
that they submitted if you want.
CO-CHAIRMAN WALLIS: Do you have it,
Steve?
MR. BAJOREK: Yeah, I have the NUREG
upstairs. It's less than an inch thick.
CO-CHAIRMAN WALLIS: An inch thick?
MR. BAJOREK: It's less than that.
MEMBER POWERS: It's substantial.
MR. BROWN: Yes, it is.
MEMBER POWERS: And I don't know. There
are about what, 38 operational pages in it? But it's
a substantial document.
There is a paper, but I don't know that
the paper is -- I mean, I prefer the topical.
CO-CHAIRMAN WALLIS: But there isn't a
published paper in a journal or something?
MR. BROWN: Yeah, there is.
MEMBER POWERS: There is.
CO-CHAIRMAN WALLIS: Well, that can't be
an inch thick.
MR. BROWN: No, no, it's not.
MEMBER POWERS: But the topical is more
useful. What you guys publish in journals is too
terse for me.
MR. BAJOREK: I have a copy I could
probably give you.
CO-CHAIRMAN WALLIS: Topical. You mean
Kataoka-Ishii is a contract?
MEMBER POWERS: They were contracted by
the NRC.
PARTICIPANTS: It's a NUREG.
MEMBER SCHROCK: Steve has got a copy.
CO-CHAIRMAN WALLIS: Okay. I guess if you
want to give it to me, I could carry it.
MR. BROWN: Okay. Moving on, I guess
there was some issues before in a previous meeting
that we had in which Dr. Wallis had a number of
questions, again, similar to the question you brought
up, Professor Schrock, on the use of homogeneous,
which we all seem to do in the AP 600 as well as
probably AP 1000 scaling.
And specifically, you asked with respect
to its impact on two phase natural circulation of the
pressure, and you specifically directed me, even
though you may criticize it in a moment, for actually
trying to actually put some of this down on a flow
regime map, and this was during long-term cooling,
closer more towards sump injection as opposed to IRWST
injection.
And I also went back and just used more of
a slip model in the equations and just to check the
core exit quality scaling.
Again, at the risk of revoking my license
or something like that I'll put up these --
(Laughter.)
MR. BROWN: -- flow regime maps here. You
can have my PE stamp at the end of the talk.
But one of the reasons, I guess, for
showing this, I was interested in a bit, is we had
been doing some COBRA TRAC calculations, and during
the sump injection phase, and COBRA TRAC seems to so
far show two situations in the hot legs and the ADS-4,
and it seems to sort of bounce back and forth between
a counter current flow and the hot leg, for example.
Any other time it's showing that it's in some kind of
a stratified regime intermittently between the two.
It spends the majority of the time so far
in this -- it predicts it's using a horizontal flow
regime map, and it's predicting that it's stratified
most of the time, and the other times it's going
through a counter current flow back to the core.
Looking in the --
CO-CHAIRMAN WALLIS: So presumably if it's
a co-current flow, it's all going out the ADS-4, is
it?
MR. BROWN: I would think so, yes.
The other, I looked at the ADS-4 pipe, and
again, similar to COBRA TRAC, the vertical and then
also next, the horizontal sections appear to be an
annular flow.
CO-CHAIRMAN WALLIS: This is in the flow
regime map for the hot leg?
MR. BROWN: No, this next is the ADS-4
pipe, the vertical section.
CO-CHAIRMAN WALLIS: Oh, the vertical
pipe, okay.
MR. BROWN: Yes, and then also for the
horizontal section for the ADS-4 pipe, and both of
them show that they're in --
CO-CHAIRMAN WALLIS: And UGS at 100 meters
a second at these pressures is pretty high velocity.
MR. BROWN: Yes, it is.
CO-CHAIRMAN WALLIS: It carries everything
out.
MR. BROWN: Well, the interesting thing to
note, too, is that, again, in situations like the hot
leg in which we didn't -- we didn't change the size
from AP 600 to AP 1000, you know, you can see that AP
1000 has a bit, you know, drifted off a little bit
relative to AP 600, but in situations where we did
change the size of the ADS-4 piping, you can see that
we, in fact, were maybe a little bit closer toward OSU
with respect to that with the piping that we did
modify for the AD 1000.
And then the last thing that I got
involved in to move into was the core exit quality
scaling during this same low pressure phase here. Dr.
Wallis asked me to look at it again, and I just used
the slip model. I just simply went back to using a
quality and density ratio based slip model for the
pressure drop and so on.
And I did find that the results that I got
for the core exit quality did change significantly
because of that. I mean, they really did, but, again,
as far as ratioing between the plant versus the test,
the same conclusion, not exactly the same ratios, but
they hovered around one as you can see.
So it did have a significant effect on the
actual answer, but again, as far as scaling them, it
sort of came to the same conclusion.
And that's about all I had.
CO-CHAIRMAN KRESS: Good.
CO-CHAIRMAN WALLIS: Good. Thank you.
CO-CHAIRMAN KRESS: Thank you.
Well, I guess we will -- is there anything
else we need to do today? We're going to continue
this tomorrow.
CO-CHAIRMAN WALLIS: I think we all have
a homework assignment to figure out the importance of
what we learned today.
CO-CHAIRMAN KRESS: Yeah. So we'll go
home and ruminate on it. So at this point I'm going
to recess until tomorrow morning at 8:30. We'll see
you all then.
(Whereupon, at 5:50 p.m., the Subcommittee
meeting was adjourned, to reconvene at 8:30 a.m.,
Friday, February 15, 2002.)
Page Last Reviewed/Updated Monday, July 18, 2016