Thermal-Hydraulic Phenomena - March 15, 2001
Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title: Advisory Committee on Reactor Safeguards
Thermal-Hydraulic Phenomena Subcommittee
Docket Number: (not applicable)
Location: Rockville, Maryland
Date: Thursday, March 15, 2001
Work Order No.: NRC-112 Pages 1-136
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
THERMAL-HYDRAULIC PHENOMENA SUBCOMMITTEE
+ + + + +
WESTINGHOUSE PROPOSED APPROACH TO ADDRESS
AP1000 T/H ISSUES
+ + + + +
+ + + + +
MARCH 15, 2001
+ + + + +
+ + + + +
The Subcommittee met at the Nuclear
Regulatory Commission, Two White Flint North, Room
T2B3, 11545 Rockville Pike, at 1:00 p.m., Dr. Graham
B. Wallis, Chairman, presiding.
GRAHAM B. WALLIS, Chairman
THOMAS S. KRESS
WILLIAM J. SHACK
I N D E X
AGENDA ITEM PAGE
Introduction by Chairman Wallis. . . . . . . . . . 3
Westinghouse Presentation: Proposed Approach
to Address AP1000 T/H Issues
A. AP1000 Pre-Certification Review Overview
M. Corletti. . . . . . . . . . . . . . . . . . . 4
B. AP1000 Plant Description, Differences
From AP600 Design, M. Corletti . . . . . . . . .12
C. AP1000 Passive Safety Systems Design
and Analysis, T. Schultz . . . . . . . . . . . .29
D. Review of AP1000 PIRT and Scaling
Approach (Closed), W. Brown. . . . . . . . . . .92
E. Approach to the Application of Analysis
Codes to AP1000 (Closed), Include Discussion
of Need for Uncertainty Assessments,
J. Gresham . . . . . . . . . . . . . . . . . . 105
F. Concluding Remarks, M. Corletti. . . . . . 119
NRR Staff Presentation: Comments on
Westinghouse Approach, Schedule Milestones,
Potential Problem Areas (if any), J. Wilson. . 134
CHAIRMAN WALLIS: The meeting will now
come to order. This is a meeting of the ACRS
Subcommittee on Thermal-Hydraulic Phenomena and I am
Graham Wallis, the Chairman of the Subcommittee.
The other ACRS Members in attendance are
Thomas Kress and William Shack, and we expect Mario
Bonaca to be present within about an hour and we also
may have a consultant, Dr. Novak Zuber present.
The purpose of this meeting is for the
Subcommittee to review the Westinghouse Electric
Company's proposed approach to address
thermal-hydraulic issues pertaining to its AP1000
passive 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. Paul Boehnert is
the cognizant ACRS staff engineer for this meeting.
Portions of this meeting will be closed to
the public to discuss Westinghouse Electric Company
proprietary information. I'd ask Westinghouse to let
us know when the meeting should be closed.
The rules for participation in today's
meeting have been announced as part of the notice of
this meeting previously published in the Federal
Register on March 1, 2001.
A transcript of the meeting is being kept.
And the open portions of the transcript will be made
available as stated in the Federal Register notice.
It is requested that the speakers first
identify themselves and speak with sufficient clarity
and volume so that they can be readily heard.
We have received no written comments or
requests for time to make oral statements from members
of the public.
I'd like to say that it's a pleasure to
welcome Westinghouse back. We haven't seen you in a
while. We've spent some time with your competitors in
the last couple of years, but we're glad to have you
here again and we look forward to your presentation.
MR. CORLETTI: Good afternoon. On behalf
of Westinghouse, my name is Mike Corletti. Thanks for
the warm welcome.
Today, we're going to be talking about the
precertification review of the AP1000 and I'd like to
start with the purpose of today's meeting and the
Today's meeting is going to be an
informational meeting. It's really the first time
we've been able to present this to the ACRS and we
wanted to concentrate on the thermal-hydraulic -- the
two major issues that this Subcommittee would be most
interested in. We're going to really outline what the
objectives of the AP1000 precertification review are
and review our proposed approach to resolution of the
two key issues.
Those two key issues we'll talk a lot
about later, but really they are whether our desire to
use the AP600 test data in support of Design
Certification for AP1000 and the applicability of the
AP600 analysis codes.
Throughout the meeting we would be looking
for feedback on our approach. I understand you've
just received our deliverables probably last week, so
we have no expectations that you have reviewed it
completely and thoroughly.
CHAIRMAN WALLIS: We found all the
MR. CORLETTI: If you found them all, that
would be good, so we can correct them very quickly and
But really, we'd like to get the feedback
on our general approach to these two key issues and
then talk about the expectations for our future
You see the agenda that's up there. We're
going to have myself speaking about the objectives of
the review and some of the plant description overview.
We're going to have Terry Schultz talk about the
AP1000 passive safety systems, the design that we've
done in some of the analyses that are included in our
plant description, an analysis report. That's the
first report that we've sent in.
Then we're going to have Bill Brown talk
about our approach to scaling. We've just recently
last week sent in the PIRT and scaling assessment
report and Bill is going to talk you through his
approach to scaling and what the conclusions that
we've drawn from that report. And then Jim Gresham is
going to speak to our plan for the use of the approved
AP600 analysis codes and how we would plan to apply
those. He's going to be talking about a future report
that we're still working on.
We went to the staff last year to talk
about AP1000 and really introduce the staff to the
AP1000 and what our major design objectives were.
Really, we're taking the AP600 design that went
through very extensive design and analysis and
starting with that, but trying to increase the
capacity of selected systems to increase the power
output. To have a plant that would have an overnight
capital cost in the range of $900 to $1000 per
CHAIRMAN WALLIS: Now isn't the AP1000
designed to compete in the U.S.? What was the AP600
MR. CORLETTI: The AP600 was also designed
to compete in the U.S. and at the time that it began
which was in 1988, the target economic goals were
something on the order of $1500 a kilowatt. AP600
meets those goals, meets the original design, the
goals that were set forth.
What's happened since 12 years passed, I
think the deregulated market has come upon us and
basically AP1000 is in response to that.
CHAIRMAN WALLIS: So there wasn't some
physical reason why it was smaller? There wasn't some
sort of design constraint?
MR. CORLETTI: It was not a design
constraint, but it really was a sampling of the
industry and of the utilities and it was a size
selected that they felt would be what they'd like to
CHAIRMAN WALLIS: It seemed sort of a
strange economic decision where most others are
thousands --anyway, we should move on.
MR. CORLETTI: So the realities of the
AP600, we had invested quite a bit of money both in
the design and licensing of AP600, some upwards of
$400 million had been invested by Westinghouse and the
industry as a whole.
We knew that going forward we would not be
able to start with a clean sheet of paper, but we did
want to -- but if we could design the AP600 with --
the AP1000 within the space constraints of AP600, we
would be able to have a plant design that would meet
the economic targets that were necessary.
When we say the space constraints, this is
what we mean here. Basically, the AP600 and AP1000
side by side, the structural design, the --
CHAIRMAN WALLIS: Looks like one of those
things you have in high school or something, to spot
MR. CORLETTI: There are a few.
CHAIRMAN WALLIS: There are a few --
MR. CORLETTI: That's true and the
structure -- you can lay them on top and they'd be
exactly the same.
You'll see some of the differences, the
much larger steam generators that we have there. And
that is -- besides that, it's probably tough to see.
But we also wanted to retain credibility
of proven components. We don't want to redesign all
the components that -- our design approach is being
still using proven components to the extent possible.
This gives us the advantage then that we can retain
the cost basis of the AP600 which we have a very good
handle on and by doing the changes, only those
necessary to increase the power up of the plant --
CHAIRMAN WALLIS: You say proven
components, you mean components previously accepted by
the NRC? They haven't been built.
MR. CORLETTI: We're going to get to that,
but like the steam generators are based on the ANO
steam generators that we just --
CHAIRMAN WALLIS: Okay, some of them have
been really proven.
MR. CORLETTI: Yes, that's right.
CHAIRMAN WALLIS: Thank you.
MR. CORLETTI: I think you're going to
find that most all of them been actually.
But also we set out to retain the AP600
licensing basis. What did that mean? Well, AP600 met
the regulatory requirements with large margins, large
safety margins. We knew it would be unacceptable to
come in with a plant design that did not meet those
large safety margins.
Except all the policy issues that we
fought so long and hard with you and us and the staff
and we came to resolution, we basically want to accept
all those policy issues for AP1000.
So when we went to the staff, we explained
that based on our -- the available resources that we
saw that we thought we could pursue a Design
Certification under certain conditions and basically
we had -- those conditions were such that the staff
thought it was a good idea that we review those first
and get an agreement on those first before both the
staff and us invest a large amount of resources in the
whole Design Certification effort, that while we're
improving the efficiency process, and basically, we're
trying to leverage AP600 Design Certification. The
NRC, we just completed that review, not but two years
ago it was at the time and we felt that staff had a
good understanding of the design and why it was
acceptable as did we and now is the best time to go
forward with getting into some of these tough issues.
The issues that we are trying to -- we
actually identified six issues. ACRS also provided
their insights and guidance. I think the insights and
guidance that you gave us are really not necessarily
focused just on this Phase 2 review, but really for
the Phase 2 and the Design Certification.
Two of the items that were identified, we
did defer to Design Certification. We will do them at
that time, that being what percentage of the SAR we
could retain from the AP600 for AP1000 and also the
issues with regards to the AP600 PRA. Those have
totally been deferred to Design Cert.
The four issues that this phase will be
dealing with, as I said, the sufficiency of the AP600
test program to meet the requirements of 10 CFR Part
52. We had a very extensive, I think most of you were
familiar with the AP600 test program. We're going
through a systematic review of that and in our report
we've tried to show how we would plan -- how that test
data, we believe is applicable to AP1000 and really
what it means and I think that will be the subject of
The second item is the applicability of
the NRC approved analysis codes, how we can use the
approved codes for AP1000, what were the major issues
that we had to resolve for AP600 and the way we
resolved them, is that still applicable to the AP1000?
The two other issues are probably --
they're not the subject of today's meeting, but in the
third one is how we would use design acceptance
criteria in lieu of detail design engineering in some
of the selected areas such as piping, structural and
seismic. And the fourth issue was the applicability
of the exemptions that were granted on AP600 and how
we'd be able to use for AP1000.
Just to give you a status of the review,
we submitted the Plant Description and Analysis
Report. In that report it gives a comparable
description of the AP1000 design features and compares
them to AP600. We include a safety systems margins
assessment which Terry is going to be talking about
today. Really, it's the first principles comparison
of the key passive safety features.
And then we did safety analysis
assessment. We basically took the AP600 analysis
codes and revised them all to reflect the AP1000 and
ran those codes, really as a way to characterize the
performance of the AP1000, not as the final Chapter 15
safety analysis, but just to give everyone an
understanding of the phenomena that we would be --
that the AP1000 would exhibit compared to AP600.
CHAIRMAN WALLIS: Now when you made the
design choices, I noticed that you retained nearly the
same -- some things are the same, right, and some
things are different.
Some things are scaled up a little bit and
some of the pipes are bigger --
MR. CORLETTI: Right.
CHAIRMAN WALLIS: But the reasons given in
this blue document here are very simple sort of
reasons. It would seem you would have to actually run
some codes to figure out if they're really the right
MR. CORLETTI: And those codes are in the
first report that we did. We ran -- and Terry will be
speaking of those in the colored -- it's the book that
CHAIRMAN WALLIS: It was something like
-- I guess the accumulators are the same.
MR. CORLETTI: Right.
CHAIRMAN WALLIS: Those are some of the
things that are the same.
And one would think that the optimum
accumulator for a bigger reactor would be different.
MR. CORLETTI: Terry's going to
CHAIRMAN WALLIS: He's going to address
MR. CORLETTI: That's right and I think
that takes us really to the next part which is the
Plant Description and Analysis Report.
MR. BOEHNERT: Mike, before you get off
that last slide, when is the Analysis Report, the Code
Applicability Report going to be available?
MR. CORLETTI: We're working on that and
our schedule is to submit that in April.
Our approach is really -- you can't decide
on the code -- where the contents of the Code
Applicability Report until you really resolve the test
data that you used to validate those codes is still
appropriate for AP1000. That's been our approach and
that's really the reason we've performed them in the
order that we have.
The first slide is the comparison of some
of the key selected parameters.
CHAIRMAN WALLIS: The thing that strikes
one the most, to me anyway, is the heat rating has
been upgraded considerably?
MR. CORLETTI: Yes, it has.
CHAIRMAN WALLIS: So you're going to have
to convince someone that that's okay?
MR. CORLETTI: Right.
CHAIRMAN WALLIS: You're making quite a
demand on cooling, both in normal operation and in --
MEMBER KRESS: That was accomplished by
increasing the enrichment.
MR. CORLETTI: Yes. The focus of the --
I'll get into the basis for the fuel, but it's -- we
haven't brought our fuel people with us today
because it's really going to be part of the Design
Certification and the review of the fuel design at
that point. This meeting is really the test and
analysis, but I can give you a little basis of the
fuel design and I don't want to go too deep into it,
if that would be okay.
The reactor power is 3400 megawatts
compared to the 1933 of the AP600, so as you said,
we've increased the number of fuel assemblies from 145
MEMBER KRESS: See, that's a change of --
MR. CORLETTI: Twelve fuel assemblies.
MEMBER KRESS: That's a change of what, 74
percent in power?
MR. CORLETTI: Yes. We did it in two
ways, increasing the number of fuel assemblies and in
lengthening the fuel rods.
Our three loop course, our three loop
plants typically have 157 fuel assemblies. In fact,
our 3XL which has the same fuel, 14-foot core, are
operating Doel 4 Tihange 3 in Belgium, so really the
fuel assembly design, the reactor vessel design is
essentially the same as those units.
CHAIRMAN WALLIS: So you bring it into
line with something that already exists?
MR. CORLETTI: Yes. And the core power
density has been increased over those units. Those
units are 3,000 megawatt units. We've increased the
core power density to be the same as our operating
three loop plants that have 12 foot fuel.
MEMBER KRESS: Does this give you any
problem though of too much fluence on the reactor
MR. CORLETTI: No, it really doesn't.
With the materials that are selected today, you can
essentially radiate them --
MEMBER KRESS: You can almost get away by
MR. CORLETTI: Yes. With good materials,
you really can almost show infinite irradiation and
you can still meet the 60 year design life.
MEMBER KRESS: This is for 60 years?
MR. CORLETTI: Yes. This is 60 year
CHAIRMAN WALLIS: Infinite irradiation?
MR. CORLETTI: Not exactly.
MR. CORLETTI: Very long and essentially--
CHAIRMAN WALLIS: The fluence did go up?
MR. CORLETTI: Absolutely.
MEMBER KRESS: But with the right
materials you can stand it?
MR. CORLETTI: Right, that's right. The
hot leg temperature, you'll notice, has increased from
600 to 615, but again, it's still well within the
operating range of most of our plants.
MEMBER KRESS: Your RCS materials are all
MR. CORLETTI: Yes. Again, the 17 by 17
fuel assemblies. The number of control rods has
increased. We've got 8 control rods, really filled up
the available space.
MEMBER KRESS: Now with the increased
length of the fuel by two feet, does that mean you
have to increase the length of the control rods also?
MR. CORLETTI: Yes sir.
MEMBER KRESS: By the same amount?
MR. CORLETTI: By the same amount and then
it's really the same as what we have in the South
Texas Unit also which is also 14 foot.
MEMBER KRESS: You already have
MR. CORLETTI: Yes, we do, yes.
MEMBER KRESS: Okay.
MR. CORLETTI: Both in South Texas and in
Doel and Tihange.
MEMBER KRESS: Okay.
CHAIRMAN WALLIS: Which are fairly old
MR. CORLETTI: They've been around. Yeah.
I don't thing -- South Texas is one of our newer old
CHAIRMAN WALLIS: But the Belgium plant --
MR. CORLETTI: I think they're the same
vintage as the South Texas plants.
The reactor vessel did not change and
basically it's the same -- the AP600 we started with
a larger vessel to begin with. We started with our
three loop configuration vessels.
The steam generator, you'll see a big
change in the surface area there. We have some slides
to show you the relative size of the two.
MEMBER KRESS: Did you put in more tubes
or longer tubes?
MR. CORLETTI: More tubes and we wanted to
have a low pressure drop steam generator so that we
could minimize the impact on the reactor coolant pump,
so we have many tubes.
It's similar, I guess, it's not -- the
last -- what happened when we started AP600, Delta 75
was our replacement generators for our Model Fs. So
that was the generator we picked when we began AP600.
Since that time we've been supplying
replacement steam generators and at the time we
started AP1000, we were just finishing the design and
we were actually finishing the construction of the ANO
replacement steam generators which are about the same
megawatt rating as this unit.
Now subsequent to that we have merged with
Combustion Engineering and they have even more
experience with large steam generators like this. We
really have had a collaborative effort on the Delta
125. They've been working, the two design teams have
worked together to get the design of this --
MEMBER SHACK: Do you use egg crates?
MR. CORLETTI: No.
The reactor coolant pump is another that
I'll talk some more about. Reactor coolant pump flow
rate was increased to accommodate the higher core
power, the inertia is increased to accommodate longer
flow coast down to meet the DNB requirements.
MEMBER KRESS: You just make them bigger
with a bigger motor on them?
MR. CORLETTI: Well, you'll see it is a
higher capacity. It is a bigger motor. Didn't get
that much bigger. The hydraulics are different and
I'll show you a little sketch of that.
CHAIRMAN WALLIS: It's interesting you
have a table of these parameters, whereas the other
parameters that are really the key, the passive like
the CMTs and all of that, you really ought to compare
those. I don't see a table of that comparison.
MR. CORLETTI: That will be in Terry's on
the passive -- I was just trying to set the stage for
the main reactor coolant system.
CHAIRMAN WALLIS: The message I get from
this is that you've been there before with other
reactors and there's nothing unique about AP1000.
MR. CORLETTI: I can move on then.
MEMBER KRESS: With the exception of that
last one on there --
MR. CORLETTI: Which one is that, sir?
MEMBER KRESS: Containment height.
MR. CORLETTI: Containment height. We
have increased the containment height. I don't
believe it's taller than the AP600, but I don't
believe it's taller than -- I don't have a good
comparison with operating plants.
MEMBER KRESS: I mind the aspect ratio, it
seems like it's unusual compared to containments I'm
used to, the lift diameter.
CHAIRMAN WALLIS: It's nowhere near in
proportion to the power, is it? It's gone up a little
bit, 10 percent.
MR. CORLETTI: We've done a couple of
things. We've increased the design pressure. We've
made the wall thicker. And we've increased the
CHAIRMAN WALLIS: But the height is 10
percent, the volume is 10 percent bigger and the
diameter is the same.
MR. CORLETTI: I think it's about 20
percent actually. I have -- Terry has a slide that
gives the percentage of change.
It's 12 percent.
MEMBER KRESS: So you got about that much
more surface area to take out 74 percent more heat?
MR. CORLETTI: The mass and energy aren't
-- yeah, I think Terry has -- Terry is going to get
MEMBER SHACK: You really kicked up the
MR. CORLETTI: We increased the design
MEMBER KRESS: Which makes a substantial
difference in terms of heat transfer.
MR. CORLETTI: Again, I think we've
covered most of this. When we did this upright, we
basically started with a proven fuel design, again
fuel core that we've had experienced with before.
Core power density has been increased.
We then went about -- the reactor vessel
was similar to Doel 4 and Tihange 3. Basically, we
fixed the elevation of the hot leg and cold leg
pipings and we let the bottom of the vessel drop to
accommodate the longer fuel.
I think the steam generator, we talked
about that. Reactor coolant pump, I have a slide on
Again, you'll see, this is a comparison.
We've added three fuel assemblies on the periphery.
That's the main difference. Again, that looks -- that
is basically the same as our three loop operating
This shows a comparison of the overall
length and as you were saying not only does the vessel
get longer, but the integrated head package, to be
able to pull the control rods out gets longer also.
And this shows the relative dimensions of
the steam generator.
One of the items -- I know the steam
generator is a lot bigger, the mass energy is larger,
but the flow restricter is the same diameter, so
really the rate of the discharge from a steam line
break, for instance, is the same.
The reactor coolant pump, AP600 was based
on a proven motor design at the time that we began.
But due to the higher power density of AP1000 we
required really a longer flow coastdown, which means
we had to increase the inertia. We also had to
increase the pump head and flow to get sufficient flow
through the core, to accommodate the core power.
MEMBER SHACK: People used depleted
uranium before for that?
MR. CORLETTI: As part of AP600, we did
for the AP600 pump. And we ran a test, we built one
of the flywheels to demonstrate that we could. I'm
not sure, no one uses them -- the Navy doesn't use
them in these kind of pumps and we typically use shaft
steel pumps, so that was a new feature of AP600.
CHAIRMAN WALLIS: Just to make sure, if it
flies apart, it's unstoppable?
MR. CORLETTI: Right, that was part of
-- that was a lot of what the review on AP600 was, was
the flywheel integrity.
One of the things we did to minimize the
impact on the pump motor size is we have added a
variable speed controller that allows the pump to
operate at low speed during cold conditions in the
reactor coolant. So as you're heating up the reactor
coolant system, they typically heat up the loops on
pump heat, until the system has come up to operating
temperature, then we basically disengage the variable
controller and it's really locked out. So it's really
only operating at shutdown condition.
Another difference is the hydraulics.
It's really -- hydraulics, different hydraulics. It's
one that we've designed and actually built for the
Saruga plant that we're working on. We built a test
Here, you see some of the operating
conditions. The pump flow, the head has been
increased and the inertia has been increased from
5,000 to roughly 15,000.
MEMBER KRESS: Your test model, is it a
MR. CORLETTI: Of the hydraulics, yes.
MEMBER KRESS: Of the hydraulics.
MR. CORLETTI: And you see, we've
increased the motor rating. It still is significantly
CHAIRMAN WALLIS: The megawatts is 6000?
MR. CORLETTI: 6000 horsepower. Our AP600
pumps were roughly 6 megawatts for the four of them.
CHAIRMAN WALLIS: So this is maybe 11 or
12 or something?
MR. CORLETTI: Yeah, that's right, 12
CHAIRMAN WALLIS: Energy efficiency.
MR. CORLETTI: Yes. The thing with these
hydraulics are slightly more efficient than the AP600,
As you see, the interesting thing, the
overall size is not increased that much which is
important to us because we had to make sure we could
take the pump out for pump replacement which was one
of the key limiting design criterias that we placed on
EMD, our pump designers as we have been doing this is
make sure you can replace, pull the pump out.
CHAIRMAN WALLIS: Presumably things are a
big tighter, things are bigger in the same containment
and it's tighter.
MR. CORLETTI: Yes, it is. Things are
We did in the Safety Analysis Report, we
performed a loss of flow analysis. The same way we
did AP600 really.
MEMBER KRESS: Is that a pump coastdown?
MR. CORLETTI: Basically, you lose all
four pumps simultaneously.
MEMBER KRESS: Yes, but they coast down?
MR. CORLETTI: And they coast down, that's
And the next slide here really shows the
analysis results for that. And you'll see with the
higher efficiency AP1000, at the time of minimum DNB,
it's a significantly higher flow than we had for
AP600. You need that to meet the DNB requirements for
the higher power core.
This is presented in the first report, the
Plant Description and Analysis Report.
My final slide is one that really just
shows the increase in the pressurizer. Again, we did
not increase the diameter. We did increase the volume
by raising the height, so it can fit in the same
pressurizer compartment. We didn't want to change the
structures around the pressurizer. It's larger to
accommodate pressure transients associated with the
CHAIRMAN WALLIS: What's the major
criterion for pressurizer design?
MR. CORLETTI: It's really for loss of
load, you want to prevent --
CHAIRMAN WALLIS: In-surges?
MR. CORLETTI: Yeah, in-surges and
minimize pressure -- basically, you want to limit
design -- the pressure to 110 percent of design
pressure -- and you don't want to over-pressure above
110 percent of design pressure. So it works in
combination with the safety valves to really accept
the pressure transients. There's many different ones,
but that tends to be the limiting one.
With that I'm going to move to the passive
safety systems and Terry is going to talk about that.
MEMBER KRESS: Now that volume is changed.
It's not the same ratio as --
MR. CORLETTI: No, but basically you look
at the in-surge and really the pressure outside. It's
dependent not only on the power, but really on the
overall reactor coolant system volume and the changes
MEMBER KRESS: You didn't change it.
MR. CORLETTI: That's exactly right. We
didn't change that 70 percent.
CHAIRMAN WALLIS: So it's just thermal
expansion of the water?
MR. CORLETTI: That is -- yes. So
basically you would have a -- the real sizing they do
a trip without inserting the rods and you have the
expansion of the water and you see whether prevent
filling it up.
MEMBER KRESS: To anticipate a transient
without a scram is --
CHAIRMAN WALLIS: Just like losing a fan
belt on a car.
MR. CORLETTI: That's how we do our design
calculations, yes. Okay. Thank you very much.
CHAIRMAN WALLIS: We are ahead of time,
MEMBER KRESS: We'll be challenged on that
MR. SCHULZ: Thanks, Mike. As Mike said,
there are several items that I am going to try to talk
about here related to the passive safety systems
design for AP1000.
I would like to talk about the design
changes that we've made and try to give you some
insights and understanding as to how we went about it
and how we arrived at the sort of curious some things
are bigger, some things aren't.
MEMBER KRESS: What's an Advisory
CHAIRMAN WALLIS: A smart one.
MR. SCHULZ: It's an official category of
engineering at Westinghouse.
MEMBER KRESS: It's one of their official
CHAIRMAN WALLIS: It's like an Advisory
MR. SCHULZ: I wouldn't --
MEMBER SHACK: He's eligible to become a
MEMBER KRESS: Where I went to school we
never had a category called Advisory Engineer.
MR. SCHULZ: The second thing that I'm
going to talk about is a margins assessment that we've
done. These are very simple --
CHAIRMAN WALLIS: I was very struck by the
simplicity. They're extraordinary simple. I would
think a thing as expensive as this and as important as
this would be something a little more sophisticated.
I mean it seems to be very, very crude --
MEMBER KRESS: Actually, I was quite
pleased with that approach myself.
CHAIRMAN WALLIS: You liked that. It's
very good for a start.
MEMBER KRESS: What's why --
MEMBER SHACK: That's what the designer
really did and then he went off and did the analysis
after he had it.
MEMBER KRESS: Yes.
CHAIRMAN WALLIS: I would have thought
you'd optimize it or something -- computer codes and
say is this really the best we can do?
MR. SCHULZ: It is an iterative process.
We already in this short time that we've been working
on AP1000, we've gone around several times on should
we make the accumulators bigger or not? What happens
if we run them faster? What happens if we try to make
them bigger? What are the costs of making them bigger
in terms of the plant impact? We've been doing that.
What you see here is more of the end result of the
iterations we've made.
There is a role. The margins assessment
is -- a lot of this comes out of the design process.
What you see in the report is more of an end
assessment of where we ended up.
In addition, as Mike mentioned, as I
already showed you some, we've used the AP600 SSAR
safety analysis codes and we've made some analysis on
AP1000 for the purpose of assessing where we think we
are in terms of the design changes and do we think the
design is adequate.
MEMBER KRESS: Now if Hal Lewis were here,
he would point out to you that principals run schools.
MR. SCHULZ: Yes.
MEMBER KRESS: But I'll refrain from doing
MR. SCHULZ: Thank you. Mike has already
talked about the design approach on the plant level
and of course it's very important for us in terms of
the economic viability of what we end up with and the
resources it takes to get there that we minimize the
changes to the plant. But at the same time we need to
make the plant safe and have adequate margins and in
doing that margin thinking, we've been looking at both
deterministic, which I'm going to talk about mainly
today and also the probabilistic area.
Now when I'm talking about probabilistic-
wise is more the T & H success criteria, how many ADS
valves do we need to prevent core melt kind of thing.
And we have done some looking at that already in our
process in trying to check whether the systems are
adequately sized for AP1000.
And also keeping in mind as we did in
AP600, where there is uncertainty in testing and
analysis, we're providing margin in the systems
CHAIRMAN WALLIS: Your hand calculations
struck me as being sort of independent. You assessed
each part independently, but as I remember an analysis
of AP600's behavior, it's the interaction between
these systems which is pretty key in an accident and
the balance between them as hydrostatic heads and
Novak Zuber's bathtubs and things and you cannot
really look at one by itself and say well, just look
at how that performs because it affects the whole
transient which changes when the next one comes on and
how they interact and all that.
I think you'd have to run computer
programs quite a bit in order to iterate to a good
design. I was very surprised that the hand
calculations seemed to work out or they did work out.
That was your real basis --
MR. SCHULZ: I think we are not making
these hand calculations in a vacuum. We have done a
lot of analysis on AP600, a lot of testing and from
that we've gained insights into what are the limiting
points in a transient? What are the limiting events?
And then with that understanding we say
okay, if we take a snapshot in time when the RWST is
just starting to inject, this is the delta P we have
and if we take that same delta P which is an
assumption, a critical one, and apply it to AP1000,
how much more flow do we get and does that seem to
make sense? So yes, it is very much separate effect
on a system, but we're using our experience and
judgments on what we learned on AP600 analysis and
testing to try to focus in on what we think will be
the limiting situation for AP1000.
Now it's not enough to just do that.
That's why we've already exercised a computer code as
a good check on the integrated effects.
MEMBER KRESS: Are you going to give us
details on the sub-bullets, the deterministic criteria
and this PRA success criteria later?
MR. SCHULZ: As I go through each feature,
I will mention what we've looked at. I won't really
be presenting any analysis. We have done some map
analysis which is what -- what we would do to at least
initially assess what the success criteria is. In
fact, I don't think we've reported any in that report
that we gave you that we did mention that we had done
that. I will mention a couple of features where that
-- where we did some work in that area.
So if I don't say enough about that, I'm
sure you'll remind me.
So what I intend to do now is go through
feature by feature and talk about what's different,
what's the same and then go through the margins
assessment and some safety analysis for each.
So in the passive RHR, the configuration
is identical in terms of where it connects to the RCS,
where the heat exchange is located, where the water
returns to the steam generator, the valving
arrangement is all identical. The elevation of the
heat exchanger is the same.
The changes that we made were increase the
pipe through the system from 10 inch to 14 inch, so we
made a significant increase in the pipe size. Now
originally, when we started AP1000, that's all we did.
We left the heat exchanger alone.
But then through some of our iteration
process, into a computer analysis, that wasn't enough,
so then we looked at making the heat exchanger a
little bit bigger. And so the other change that we
made to the design is to increase the surface area of
the heat exchanger about 20 percent.
When we did that by adding a few tubes, it
turned out on AP600 on the tube sheet there were about
9 tubes on the top and bottom rows that were left out.
We left them out because we just didn't need that
So on AP1000 we said well, that's an easy
way of getting a little bit more surface area, so we
filled in the tube sheet on the top and bottom rows
and we also extended horizontal portion of the heat
exchanger 3 feet. Of course, on the top and the
bottom. And the total effect of that was to give us
about 22 more percent more surface area.
CHAIRMAN WALLIS: Now are the flow
patterns in this heat exchanger are important, I mean
the stratification and that sort of stuff in the pool
and everything, does that make a difference to the
MR. SCHULZ: One of the things that we got
out of the testing that we've done in both the passive
RHR, especially the passive RHR testing, we
investigated and tried to quantify the mixing in the
tank and what we learned from that is the tank does
mix well horizontally. The heat exchanger is kind of
like a pump and it pumps water and keeps the tank
The tank is not getting bigger, but we are
adding a little bit more water on top by adding,
putting in some more accurate low instrumentations, so
we're gaining a foot or so of water level in the tank,
but the tank is not getting significantly larger. So
it will heat up a little faster and it will start
boiling a little sooner, but that doesn't seem to be--
CHAIRMAN WALLIS: That's a well-mixed tank
in both cases?
MR. SCHULZ: Yes, it's a well-mixed tank.
So early on when you're trying -- worried about
mitigating the design transient or loss of flow of
feedline break, the tank is going to be subcooled.
MEMBER KRESS: This tank is half moon
MR. SCHULZ: Yes.
MEMBER KRESS: The heat exchanger is on
the flat side?
MR. SCHULZ: It's on one end. I don't
know that I have a slide.
I've got one that shows --
MR. CORLETTI: Page 42 of the slides.
MEMBER KRESS: I see where it is.
MR. SCHULZ: This is the tank here.
CHAIRMAN WALLIS: Where is the tank
MEMBER KRESS: It's that solid line.
MR. SCHULZ: This is actually the
operating deck so you can't really see the tank, but
it's under this part here. The heat exchanger is
actually under this hatch.
CHAIRMAN WALLIS: That sets up currents
which mix the whole tank?
MR. SCHULZ: Mix the whole tank, yes.
MEMBER KRESS: That's why I was --
CHAIRMAN WALLIS: That doesn't change --
SPECIAL AGENT WHITE: That's the reason I
was asking. Over there in the corner. You ran a test
on a model of this with 3 or 4 tubes?
MR. SCHULZ: Yes, yes. And those tubes
were baffled off from the rest of the tank to try to
give us a way of assessing the mixing.
This is the corner -- this is where the
pressurizer is, so it comes up through here and the
tank wraps around.
CHAIRMAN WALLIS: Do you get boiling from
these tubes to empower the transient?
MR. SCHULZ: As the transient goes on, you
do. Early on, you get essentially no boiling.
CHAIRMAN WALLIS: And it's the same kind
of scenario, I mean the boiling progresses and is
there anything different about the two power levels in
terms of boiling?
MR. SCHULZ: We don't think so in the
short term. In the longer term, the AP1000 tank will
heat up quicker, but in both cases, we're talking
about more than an hour before you start boiling in
the tank, before the tank really reaches saturation.
We'll get some local boiling a little before that, but
until you get really vigorous boiling around the heat
exchanger will be well beyond an hour.
MEMBER KRESS: Now how did you know this
tank was pretty well mixed? Is this a calculation
that you did?
MR. SCHULZ: Well, it's based on this
MEMBER KRESS: Based on that testing?
MR. SCHULZ: Yes. We took temperature
measures around the tank and could see that -- I mean
the top of the tank heats up a little faster than the
bottom of the tank, but there is this, around this
heat exchanger there's a strong circulating current
driven by the heat that you're putting in there, so it
keeps putting the colder water from the bottom of the
CHAIRMAN WALLIS: Is steam being involved
MR. SCHULZ: Eventually, yes.
CHAIRMAN WALLIS: Presumably, there's more
steam being involved in an AP1000?
MR. SCHULZ: Be more and it would happen
a little sooner, yes. And that gets vented to the
containment when that starts occurring the passive
containment cooling comes into play.
CHAIRMAN WALLIS: How is this formed? Is
there sort of a great evolution of steam and level
swell in this corner of this tank and --
MR. SCHULZ: Yes. I think that fluid to
the heat exchanger --
CHAIRMAN WALLIS: Will be some kind of
bubbling and frothing and swelling.
MR. BROWN: Dr. Wallis, Bill Brown here.
We noticed that's really more -- the real level swell
we saw in the test was really more associated when the
automatic pressurization system goes off. So if
you're thinking about swelling and things in the tank
the real challenge to that is when the ADS 123
discharges in here. This is more of a bulk boiling
situation. So level swell is really associated with
CHAIRMAN WALLIS: Now this pipe goes into
the bottom of the steam generator, it's a different
MR. SCHULZ: Yes.
CHAIRMAN WALLIS: I think there's
something about mixing in the bottom of the steam
generator which they used a stratification in the cold
leg or didn't we go through that with APU 600?
Didn't we worry at one stage about mixing
it in the -- where your pipe comes out of the steam
MR. SCHULZ: I believe there was some
notice of that, especially in OSU. When the heat
exchanger starts working, there's still steam being
released from the steam generator which means that
there's still a flow going through that, through the
loop, through the tubes of the steam generator and
it's not just flow through the passive RHR circuit.
As long as that continues and it will
continue for some time, depending on the particular
transient, it may be half an hour or it may be an hour
before the steaming is terminated in that steam
generator and passive RHR has -- its capacity matches
the decay heat. Now after that point in time there
will tend to be more of a stratification in the loop
than before them when there is still circulation flow
through the steam generator.
CHAIRMAN WALLIS: Is this the same as in
600 or do you run into some new phenomenon of
stratification that's different.
MR. SCHULZ: I don't think it's the same
phenomenon. Things may occur in a little bit
different timings and because the flow through the
passive RHR is a little bit greater. The temperatures
are not lower coming out of the heat exchanger or are
not significantly different, so and the flow is a
little bit greater because of the bigger pipes through
But I think the phenomenon is the same
that there would be colder water running around the
bottom of the pipe, primarily after you get to the
point where you match decay heat.
The margins assessment again as we
mentioned to start with, I had mentioned to start with
tends to be a simple first principles calculation. In
this case, we tried to calculate the natural
circulation heat removal from the heat exchanger.
Now this is not quite as simple as some of
the other calculations so actually you're exercising
the same correlations, heat transfer correlations that
we use in our safety analysis codes that in a stand
alone spreadsheet type of program which gives fairly
close agreement with the safety analysis code.
So by putting some boundary conditions in
terms of the RCS temperature and the temperatures we
picked were based on the steam generator safety valve
setpoints which are a little bit higher in AP1000 than
AP600. In exercising that code, we end up calculating
did we get about 170 so percent more decay heat
removal which is not quite equal to the increase in
power, but very close to it. So that's a comforting
Another thing that we checked in this is--
CHAIRMAN WALLIS: In thinking about this,
isn't the limiting heat transfer resistance until you
get boiling, it's only RWST side, it's a natural
convection in the RWST which is much less effective
than this stuff that's whipping around your 14-inch
pipe and going around.
MR. SCHULZ: Yes.
CHAIRMAN WALLIS: So that's the limit and
that isn't going to be increased much by increasing
the velocity or decreasing the flow path resistance.
MR. SCHULZ: In terms of heat transfer
coefficients, yes, but by having more surface area and
CHAIRMAN WALLIS: There isn't 72 percent
MR. SCHULZ: No, there isn't. There's 22
percent more surface area, but by reducing the
resistance on the primary side, we can allow more
water to flow with a given density difference and you
can go through that using the same correlations which
are more limiting on the tank side, but you still
can't improve that heat transfer by reducing the
resistance on the primary site. Not so much from a
heat transfer effectiveness point of view, but by
giving a certain density difference, you can get more
flow to circulate and of course, there is interaction
there between how much flow is circulating and what
the delta T is on the heat exchanger.
CHAIRMAN WALLIS: So you've increased your
heat flux by 30 percent, I see.
MR. SCHULZ: Yes. But we think we're
still comfortable versus critical heat flux in the
The other thing that is an important
consideration here is as I mentioned when you first
turn the heat exchanger on it can't match decay heat.
CHAIRMAN WALLIS: I don't know why you
worry about heat flux anyway because it's going to get
cooled somewhere else eventually. It's not that
critical, is it if you get critical heat flux? It's
not as if it's called a heat source which kind of --
any type of heat aspersion --
MR. SCHULZ: Right, it's not critical from
that point of view, no, no.
MEMBER KRESS: That's why the design
MR. SCHULZ: It's just that if you got
less heat transfer than you thought you were getting
from a portion of the heat exchanger, the overall heat
transfer of the heat exchanger might be a little less
than you thought, but what tends to happen if you ever
got into critical heat fluxes the hot part of the
tubes gets a little less heat temperature reduction
and the heat moves on to the heat exchanger and these
tubes are relatively long so eventually you get most
of that heat out, but it would affect the overall heat
transfer somewhat, so it's something that we look at
and try to keep in mind in the design.
The other factor is how much of the steam
generator secondary side water you boil off in an
event and these percentages are a little confusing,
but basically if you look at the amount of water in a
steam generator per megawatt of core power. The
AP1000 has 36 percent more mass at the time of trip
than AP600 does per megawatt.
And that when I say final water this is
when you would calculate that you terminate steaming
with the passive RHR. So you've decreased the water
level some, but in AP1000 basically you end up with
twice as much water in the steam generator at the end
of a transient than AP600.
CHAIRMAN WALLIS: Does this make up for
some of your other capacities not being increased in
your accumulators at the same CMTs aren't that much
bigger, but you have steam generators with a lot more
water in them.
MR. SCHULZ: It doesn't help accumulators
of core make up tanks very much. It helps the passive
RHR a little bit. Initially, we had a passive RHR
that was maybe 25, 30 percent more capacity than the
AP600 and this story on secondary site mass was more
important than it is now because we've almost got
parity with power in passive RHR. So this is not so
important as it was once when we were playing around
with smaller passive RHRs.
CHAIRMAN WALLIS: I guess it's only energy
balance you care about in this phase?
MR. SCHULZ: Yes.
CHAIRMAN WALLIS: It's not mass --
MR. SCHULZ: In fact, this -- we've toyed
with, is it good to have more, is it good to have
less? When you get to the containment, having more
mass here challenges the containment mass energy
input, so it's not so good from that point of view,
but it is better from say a transient mitigation point
So that's the margin story. We've also
done some safety analysis, transients again to assess
the AP1000 and the selection of the passive RHR heat
exchanger changes. Same methods as we used on AP600,
conservative inputs and models. We selected several
limiting transients. We've looked at all four of
these events for AP1000.
Due to time limitations we chose to show
you what the loss, feedline rupture looks like and the
curves are a little confusing here, but basically the
criteria is subcooling and maintaining subcooling and
if you look at this dotted line versus the solid line,
the dotted line is saturation temperature and the
solid line is the hot leg temperature, so you see the
hot leg temperature is dropping down to a comfortable
level below the saturation temperature.
In both cases, the transients look a
little different. AP1000 doesn't cool down quite as
much, as fast. So in both cases there's a comfortable
margin. If you actually look at one of these
transients for an operating plant, you would see the
subcooling become much less in time, it tends to drop
down, but then come back up and the margin is a lot
less. So AP600 and AP1000 both have a lot higher
subcooling margins than most operating plants.
And so our conclusions from that is that
from our assessment of these computer calculations is
that the AP1000 has comparable behavior and margins to
AP600 and so we're -- we feel comfortable with the
sizing of the passive RHR.
CHAIRMAN WALLIS: Margin, what do you
mean, your measure of margin?
MR. SCHULZ: In this case, like
CHAIRMAN WALLIS: Like subcooling. Yeah.
It's not a direct measure of some sort of safety
margin. It's an indirect measure -- it's better to
have more subcooling therefore as an indication. But
it's not a very quantitative measure.
MR. SCHULZ: You can go through this
transient side about so many degrees minimum.
CHAIRMAN WALLIS: Yeah, but I'm not quite
sure what that means in terms of --
MR. SCHULZ: Real safety?
CHAIRMAN WALLIS: More definite measure of
MR. SCHULZ: The acceptance criteria for
feedline break is keeping the reactor subcooled. Now
does something real bad happens when it becomes
saturated? Not necessarily.
So it's more of a licensing criteria than
a safety criteria.
I'd like to now move on to the passive
safety injection. And again, the configuration of the
AP1000 in terms of numbers of tanks, the valving
arrangement is identical with the AP600. The
accumulators are the same size and we'll be talking
about each of these features in a few minutes. The
core makeup tank is 25 percent bigger and we increased
the orifice that controls the CMT flow so that's flow
rate is also 25 percent bigger. We've increased the
pipe sizes from the IWST to the direct vessel
injection line. We've also increased the
recirculation piping and we've also increased the
stage 4 ADS piping and valving. The other piping has
not been changed.
CHAIRMAN WALLIS: On the ADS 1, 2 and 3
are still the same?
MR. SCHULZ: Exactly the same.
CHAIRMAN WALLIS: Which I found a little
surprising, but maybe there's a good reason for it.
MR. SCHULZ: There is a reason. We think
CHAIRMAN WALLIS: You have less mass of
water per megawatt?
MR. SCHULZ: Yes.
CHAIRMAN WALLIS: But your hand
calculations emphasize flow rate. You have bigger
pipes and all that, so you try and duplicate the flow
rate, but there's less mass. So in terms of a mass
balance you might think there's a less margin, less
water to cool to use the RWST?
MR. SCHULZ: You could say that. We think
that when you look at a bit more mechanistic in terms
of specific situations that we have sufficient margin
and I think if we try to go through in our discussion
here, you can see what you think.
CHAIRMAN WALLIS: Okay.
MR. SCHULZ: In some cases yes.
MEMBER SHACK: You have adequate margin,
but by and large, it does it smaller, for things like
peak clad temperature.
MR. SCHULZ: For large LOCA yes. That's
not necessarily true, we don't think it's true for
like long-term cooling. We think we're at the same or
better margin because we've increased things much more
than we did in other areas. In other words, we were
rather selective in where we increased things and how
CHAIRMAN WALLIS: But you haven't decided
that you want to stay with the same PCT as AP600, for
instance. You're going to get closer to the
requirement, the regulatory limit.
MR. SCHULZ: That's right. Again --
CHAIRMAN WALLIS: So you're going to see
how far the staff will let you go in terms of raising
MR. SCHULZ: I wouldn't put it that way.
We're not playing a --
MEMBER SHACK: There's a known regulatory
MR. SCHULZ: That's right.
CHAIRMAN WALLIS: Yeah, but presumably
somebody came in here several years ago and said 1644
was great, now you're going to tell us 1940 is great.
MR. SCHULZ: When we originally selected
the accumulator sizes for AP600, we had no idea what
the PCT was going to be.
CHAIRMAN WALLIS: It seems very strange to
me. I think it worked backwards. So you have a
design criteria and worked back to whether cumulative
size needs to be to meet it.
MR. SCHULZ: I'm sorry?
CHAIRMAN WALLIS: You seem to have picked
an accumulator and then see what you get the PCT and
then said that's okay. I think it worked backwards.
You designed the accumulator to achieve or the PCT you
MR. SCHULZ: In this case, we realize what
we're starting from a design, the AP600 equipment and
we're saying how big does it need to be to provide
adequate margin for AP1000. So it's a little bit
different than starting with a clean sheet of paper.
CHAIRMAN WALLIS: But the adequate margin
apparently is a higher PCT?
MR. SCHULZ: In this case, yes.
CHAIRMAN WALLIS: Why is that so? Why is
it now acceptable to have a higher PCT than before?
MR. SCHULZ: It always was acceptable. We
didn't design to a particular PCT.
CHAIRMAN WALLIS: But some designers
decided 1940 is now okay as opposed to 1644. These
are much more important numbers than choosing the
accumulator as the same as before or something. If
you had twice as big an accumulator, maybe you could
bring that down to 1644, I don't know.
Someone has decided that these are okay
MR. SCHULZ: Basically, all the operating
plants in the United States have temperatures that are
CHAIRMAN WALLIS: So you decided to be a
little less conservative than you are with 600?
MR. SCHULZ: That's right.
MEMBER KRESS: If that number had been
2100 would you have felt "iffy" about it?
MR. SCHULZ: Yes, I would have felt iffy.
If we had gotten to the point where these numbers were
higher than most operating plants and were getting
close to the limit --
MEMBER KRESS: How do you know how close
that is to 2100?
MR. SCHULZ: 2200.
MEMBER KRESS: Well, I said 21 because I
didn't want to get all the way up to 22.
MR. SCHULZ: Okay.
MEMBER KRESS: Because I was seeing the
uncertainty on the number.
CHAIRMAN WALLIS: It's with uncertainty.
MR. SCHULZ: With uncertainty.
MEMBER KRESS: That's with uncertainty?
MR. SCHULZ: With uncertainty. We're down
at 1670 without, approximately.
MEMBER KRESS: What's that, two sigma of
MR. SCHULZ: This is the -- I don't know
what -- maybe Bob Kemper can help me out, but these
are the uncertainties that were calculated basically
for AP600 for both co-retract and the PCT uncertainty.
So the plant parameters and the models both are
accounted for in here.
MR. KEMPER: Bob Kemper here. What Terry
is showing as the reflood PCT with uncertainty would
be our assessment of the 95th percentile value as
obtained with our large break LOCA EE methodology.
And that's the assessment we have at this time for
MEMBER KRESS: So 99 percentile might have
been about 2200? Why is 95 percentile acceptable is
MR. KEMPER: That is the basis for the
large break LOCA methodology with best --
MEMBER KRESS: Using best estimate.
MR. KEMPER: COBRA track best estimate.
MEMBER KRESS: So this was a best estimate
calculation rather than Appendix K?
MR. KEMPER: The AP600 was done with the
approved methodology for large break LOCA.
MEMBER KRESS: Which was Appendix K?
MR. KEMPER: No. It uses the uncertainty
determination and that would be essentially best
estimate decay heat and then the decay heat
uncertainty is rolled in with the other uncertainties
in CSAU type methodology to obtain the result.
MEMBER KRESS: Okay. That's very helpful.
CHAIRMAN WALLIS: So it looks as if
somebody said let's go with the same accumulator and
see what we get for PCT?
MR. SCHULZ: Yes, and there are several
considerations here. One, we actually studied taking
the same accumulator and readjusting the flow orifice
to get more flow. And by doing that we can reduce the
large break LOCA PCT.
However, there are other events that the
accumulator plays a role in, small break LOCAs and in
particular small LOCAs that are involved in multiple
failure PRA-type events where say the accumulator --
not accumulator, the core makeup tanks have been
failed due to some common mode failure that would be
considered in the PRA. And having accumulators empty
quicker in those kind of events is not good, so it's
So we in the balance between those two
events, the large break LOCA which is a very low
probability event and in fact some day maybe
eliminated from consideration, we felt that the safety
was better balanced by maintaining the small break
LOCA-type performance where the accumulator runs a
little bit longer.
Now we also could have made the
accumulator bigger. However, it's buried in concrete,
basically, surrounded by concrete. It's already a
sphere, so we've maximized the volume in the space
available. So it would have been rather disruptive to
the plant design and structures to make the
CHAIRMAN WALLIS: So now you've got a
sensible argument. It seems from a spherical
perspective, it's not a very expensive thing
comparatively speaking. The obvious thing would be to
make the accumulator bigger, but if they've got some
good reason for space constraints, then that makes
some sense, but in terms of
thermal-hydraulics, it seems to be very arbitrary to
say well, choose the same size.
MR. SCHULZ: Yes, yes.
CHAIRMAN WALLIS: So there is some reason
MR. SCHULZ: The tank itself is not that
expensive, so we have had space to make it easily
bigger, that would have been a different story.
In the core makeup tanks, we actually did
choose to make the tank bigger. This tank does have
constraints vertically. It's located on a concrete
floor so we can't go down very easily. The operating
deck is fairly close to above the tank so we can't
really make it bigger. We did choose to make it a
little fatter, 25 percent increase in volume.
This is a very expensive tank. It's a
full system design pressure, Class I vessels. It's a
very expensive tank, so it's something that we have to
keep in mind. Now the cost of the tank is not so
significant relative to the cost of the plant. I want
to imply that, but it is an expensive tank.
Why 25 percent? That's kind of a curious
number. Let me try to explain what we did.
MEMBER KRESS: You knew we would ask that,
MR. SCHULZ: Yes. We asked ourselves too.
In our looking at the range of accidents that the core
makeup tank deals with, we think one of the more
limiting events is a direct vessel injection line
break which eliminates one of the tanks, in fact, half
of the injection system.
And at the point in time where the
accumulator is empty which of course happens first in
a transient, the accumulators start running and when
they run fast as the plant depressurization occurs,
the core makeup tanks don't inject because of the way
the systems interact, the tanks interact and when the
accumulator empties then the core makeup tank is
basically relied upon to maintain core cooling. That
occurs in about 10 minutes after the accident.
Now if you look at that period of time and
take decay heat and some instrument of sensible heat
coming out of the fuel in the reactor vessel, you can
estimate a heat that you should be removing and back
calculate a safety injection flow.
Now it turns out that when you do that
AP600 had a significant margin relative to that flow
requirement, basically 38 percent margin.
MEMBER KRESS: And you want to remove that
amount of heat over some time period? There's a
window in there, I seem to remember --
MR. SCHULZ: I'm looking at an instant in
SPECIAL AGENT WHITE: You were looking at
an instant in time.
MR. SCHULZ: I'm looking at an instant.
MEMBER KRESS: But you want to continue
that removal, have enough water to continue that
CHAIRMAN WALLIS: That is why the flow
capability and the volume are up by 25 percent?
MR. SCHULZ: Right. There's two questions
here. I'm addressing right now the flow rate question
and then I should speak separately about the volume
and duration question because we actually separately
looked at those two.
In thinking about the flow rate question,
it turns out that if you take the AP600 core makeup
tank, it would just about match the AP1000 calculated
requirement, but it would have very little margin to
it. And we were uncomfortable with that. So that's
part of where the 25 percent increase in flow rate
The other place where that comes from is
looking at some multiple failure scenarios where we
don't have any accumulators and in that case having
some more flow rate out of the core makeup tanks is
CHAIRMAN WALLIS: Now these numbers, this
requirement per flow is based on an energy balance, do
you know the core, the decay heat --
MR. SCHULZ: Decay heat --
CHAIRMAN WALLIS: There are very few
uncertainties in those numbers?
MR. SCHULZ: Decay heat there's very
little uncertainty. There's a little more uncertainty
maybe in the sensible heat that's also
CHAIRMAN WALLIS: Those are pretty hard
numbers, not much uncertainty, so you don't need a
huge margin. You need some margin.
MR. SCHULZ: That's right. I would agree
CHAIRMAN WALLIS: So the rationale thing
probably would be to match your margin to your
uncertainties in some logical way so it would be
explainable to a committee like this one?
MEMBER KRESS: Where have I heard that
CHAIRMAN WALLIS: That is to make Dr.
MEMBER KRESS: Yes. I would be ecstatic,
MR. SCHULZ: I am not sure I could do
The other thing that we factored in here
in terms of volume is that we did not want the tank to
have a shorter duration of injection. That plays into
ADS sizing and RWST injection capability which we feel
that that is an area where we really want to keep
margin and not reduce margin versus AP600 and by --
with the power increase that we already need to deal
with, shortening the CMT injection would add to the
burden that ADS and RWST injection would have to
So one of our sort of internal guidelines
was not to shorten that initial cut in requirement for
when RWST injection should start. And by maintaining
CMT duration of injection that helped us in that
mission. That's kind of a soft requirement, but
that's the way we kind of approached it and why we
kept the CMT, increased its volume when we increased
the flow rate.
What I'm going to go through is the ADS
margin assessment and then RWST margin assessment and
then show you a small break LOCA which kind of looks
at the integrated effects of those three elements.
In ADS we increased the size of the four
stage and again the ADS four stage is obviously one of
the most important features in getting to the low
pressure where you can get RWST injection. So we feel
that that is a very important feature. It's important
to have this adequate capacity.
CHAIRMAN WALLIS: But did you increase it
because you didn't want to change 1, 2, 3 -- 1, 2 and
3 are the same so you're getting less depressurization
from 1, 2, 3 proportionally or the bigger system? In
order to catch up, then you have a bigger ADS fall?
Is that the way that I should think about it?
MR. SCHULZ: There's a couple of reasons
why we didn't change stages 1, 2 and 3. One of them
is 1, 2 and 3 is not very useful at the RWST cut in
point because you end up, according to the testing and
analysis with water in the pressurizer which severely
limits the amount of flow that you get through the
pressurizer. And so making 1, 2 and 3 bigger we
didn't think would help very much. So it was much
more effective and probably necessary to increase four
CHAIRMAN WALLIS: What about 1, 2, 3 and
MR. SCHULZ: To basically get you to --
CHAIRMAN WALLIS: To depressurize and to
depressurize along the same sort of curve, you
presumably need a bigger 1, 2 and 3 for a bigger
MR. SCHULZ: If you wanted to depressurize
along the same curve.
CHAIRMAN WALLIS: Do you depressurize a
little slower at the beginning and faster later on, is
that what you --
MR. SCHULZ: Yes, and where you notice it
is at the lower pressures. At the higher pressures,
it doesn't make a lot of difference because you
depressurize fairly rapidly when you're say above 100
psi or something.
CHAIRMAN WALLIS: This doesn't quite
follow the same scenario time-wise as an AP600. You
would have actually increased the size of 1, 2 and 3
MR. SCHULZ: If you wanted to scale it up,
CHAIRMAN WALLIS: Right.
MR. SCHULZ: Exactly.
CHAIRMAN WALLIS: So you've changed the
sequencing of things a bit, presumably, with this
MR. SCHULZ: A little bit, yes. And we
have even thought, if we started with a clean sheet of
paper, which we aren't, that maybe we'd even take out
the third stage because it's not so useful. One of
the things we've learned in all our testing and
analysis is it's much more effective in terms of
getting to the lower pressures to have hot leg venting
than pressurizer venting because there's things that
happen in the pressurizer that interfere with its
effectiveness at low pressures, not high pressures,
but low pressures.
MEMBER KRESS: I would have thought when
you depressurize this vessel that the most important
thing was the stored energy in the water and how much
water is in there and the volume and not the power.
So the 76 percent, various size, bothers me a little
because I don't think the total volume of water in the
primary system changed that much and I think don't
it's internal energy at the start changed that much.
So it seems to me like -- I'm disagreeing
with what Graham said because I would have thought the
thing would have just depressurized about the same as
the AP600 with the 1, 2, 3 as it's sized. And even
with the 4 as it's sized. I may be wrong there --
CHAIRMAN WALLIS: It's a bit bigger, but
it's not bigger proportionately.
MEMBER KRESS: No, it's not 76 percent
bigger internal energy. So I'm still not quite sure
of the choice.
MR. SCHULZ: I think there's an element,
if you think about large LOCAs, I think that's much
more true, when you're blowing down the system very
rapidly, how much you're starting with is very
Our EDS is sequenced and blows -- it's a
more protracted blow down.
MEMBER KRESS: If you're right -- you're
MR. SCHULZ: Probably more important.
MEMBER KRESS: Yes.
CHAIRMAN WALLIS: Anyway, it's all going
to be clarified by some computer runs of how they
really work, not just the hand calculations.
So you have a new ADS 4 valve that you're
going to test at full scale?
MR. SCHULZ: We need a new ADS stage 4
CHAIRMAN WALLIS: Are you going to test it
at full scale?
MR. SCHULZ: When we build the plant we'll
-- when we build the valve, we will test the valve.
CHAIRMAN WALLIS: Yes.
MR. SCHULZ: In a test facility.
CHAIRMAN WALLIS: You don't need the plant
to test the valve.
MR. SCHULZ: Right, in fact, it's a little
difficult to do that. Right.
MR. GAGNON: What I provided Terry is
basically the depressurization response. I'm Andy
Gagnon from Westinghouse.
That's just a depressurization response
for an inadvertent ADS small break LOCA situation. As
you can see, the depressurization characteristics are
very similar with the common ADS 1 to 3.
MR. SCHULZ: And you're right, it does
come down a little slower, but still, this is pretty
rapid depressurization. I think is this where the
fourth stage opens up?
MR. GAGNON: Yes.
MR. SCHULZ: So we come down here and hold
until the fourth stage finally opens up near the end
-- later in the transient.
CHAIRMAN WALLIS: Now do you put on there
things like CMT training, would they occur about the
MR. SCHULZ: I think so, as I recall.
MR. GAGNON: For the inadvertent ADS, yes,
it's approximately the same time.
MR. SCHULZ: And about the same duration
as it was designed to do.
The other assessment that I wanted to talk
about before we showed the small break analysis was
the IRWST. And we basically did two things here to
improve the injection capability. One is to raise the
initial water level. Now we didn't raise the maximum
water level. What we did is compress the operating
band and a lot of that operating band was given up to
errors in level measurement because we only had a wide
range level measurement in the tank and it's like 30
feet. So that ended up giving us some significant
possible errors in measuring the level and what we are
doing at AP1000 is putting in a narrow range level
instrument that will cut that error down quite a bit
and end up giving us one and a half more feet initial
The more significant thing was increasing
the line size, basically that's the injection line
size from the IRWST from 6 to 8 inches which
significantly reduces the line resistance and if you
run through the numbers, assuming some RCS pressure,
containment pressure and with the initial hit of water
in the tank, you'll get like an 84 percent more
injection flow which is a little bit more than decay
heat. So again, this is an area where we think it's
important to maintain AP600 margins without reducing
them, and in fact, increasing them slightly.
CHAIRMAN WALLIS: Now again you're talking
about flow rate and not just the overall capacity.
The IRWST is not much bigger?
MR. SCHULZ: It's not much --
CHAIRMAN WALLIS: Same amount of water.
MR. SCHULZ: Well, you get a little bit
CHAIRMAN WALLIS: Little bit.
MR. SCHULZ: But not much.
CHAIRMAN WALLIS: And it's going in
faster, it's going in much faster.
MR. SCHULZ: Right, now that will affect
when you get to recirculation --
CHAIRMAN WALLIS: Yes.
MR. SCHULZ: Which is long term cooling
which I'll get to in just a little bit.
Now if you take those effects and then
look at a small break LOCA transient and then again
using for assessing purposes the AP600 SAR analysis
and this is NO TRUMP, Appendix K type approach, we
looked at several different events, a 2-inch curve leg
break which is sort of a reference one we look at.
DVI break which tends to be challenging from
accumulator core makeup tank early injection
capability and inadvertent ADS which tends to be more
challenging in terms of ADS and IRWST cut in a little
This shows you the core mixture level and
tries to compare AP600 to AP1000 for DVI line break.
You see initially AP600 has a dip and AP1000 doesn't.
And the main reason for that is this is the same break
size in both plants and it's limited by a Venturi in
the nozzle to about 4 inches ID and so since AP1000 is
a better plant with the same break, it has a little
bit less of a blow down effect early on. So that's
the main reason why AP1000 actually doesn't have this
Later on, the trends are pretty similar.
There's some minor variations of AP1000. This is when
IRWST is starting to inject and shut off and inject
and shut off until out in the 2800, 2900 seconds RWST
comes on and stays on and the level goes up a little
CHAIRMAN WALLIS: What's the top of the
core on this?
MR. SCHULZ: This dotted line.
CHAIRMAN WALLIS: That's the top of the
MR. GAGNON: The vast line is the top of
active fuel and they've been offset to represent the
difference in vessel lengths.
CHAIRMAN WALLIS: So the zero for the
level is different than the two?
MR. GAGNON: Yes, that's correct.
MR. SCHULZ: And from this, we basically
conclude that from the three basic transients that we
looked at, we didn't see any core uncovery, the
behavior of the plants were similar. We didn't see
any phenomenon that was different.
CHAIRMAN WALLIS: Now this is a mixture
MR. SCHULZ: Yes.
CHAIRMAN WALLIS: The collapsed level is
somewhere else, so there's some kind of a two-phased
flow level which says how much the expansion due to
the presence of the vapor is that we can rely on?
MR. SCHULZ: Of course.
MR. GAGNON: That's the core fraction
MEMBER KRESS: The collapsed level would
be a lot level for --
CHAIRMAN WALLIS: There's a good
verification of this, whatever the model is for the
level swell, whatever you call it? Because some of
the models they use are not very good, over rise
glossing or whatever it is that governs the level
swell. Maybe yours is.
MR. GAGNON: It's the same model as in our
approved valuation model.
CHAIRMAN WALLIS: That's for AP600.
MR. GAGNON: Yes.
MR. BOEHNERT: Is this calculated using NO
MR. GAGNON: Yes.
MR. BOEHNERT: Both of these?
MR. GAGNON: Yes. Same methodology.
CHAIRMAN WALLIS: The staff has this code?
MR. GAGNON: The staff does not have this
code at this point.
CHAIRMAN WALLIS: Is the staff going to
get the code?
MR. GAGNON: Uh --
MR. SCHULZ: Mr. Gresham will address that
CHAIRMAN WALLIS: It's interesting, I
think, for people to run sensitivity studies on these
sorts of questions about how is the collapse level
related to the mixture level and so on, what's the
sort of certainty with which you can make these
predictions and what are the uncertainties and what's
the sensitivity to some assumption of its own.
MEMBER KRESS: Particularly with respect
to this item here, because that's one of the things
that gave us comfort was that the level never got down
below the top of the core, except in one little period
there it was, but --
MR. BOEHNERT: There was also a problem
using NO TRUMP from the standpoint the staff said, I
think, said you could only use it on AP600 and there
was an issue there about applicability to AP1000.
CHAIRMAN WALLIS: All the more reasonable
why it needs to be tested in some way.
MR. BOEHNERT: I would think.
CHAIRMAN WALLIS: Does the staff have any
intention to get this code and run it and then look at
the kind of sensitivity questions?
MR. WILSON: Jerry Wilson, NRR. We made
a request of Westinghouse to have the codes as part of
our review. And we expect -- you'll notice earlier
there was a discussion about an additional report to
be submitted and that's the code report and we expect
to discuss this issue when that report is submitted.
CHAIRMAN WALLIS: It's just going to be
something you insist upon? Just discussing it doesn't
give me a good feeling. Are you going to insist upon
getting this code?
MR. WILSON: I don't want to prejudge it.
As I said, we made a request and we'll see that
Westinghouse has to say.
CHAIRMAN WALLIS: Would it help if the
ACRS said that you should insist upon it?
MR. WILSON: It always helps to hear from
MEMBER KRESS: Do they teach you guys
MR. WILSON: It's part of the job.
CHAIRMAN WALLIS: Okay.
MR. SCHULZ: The last part of the safety
injection system that I want to talk about is that
involving the long term cooling or the containment
recirculation part. And so what we have here is again
a margins assessment looking at this aspect. The line
resistance has been significantly reduced by again
making pipe sizes bigger.
Another factor in that that is important
is when do you get to recirculation? And we have, as
you noticed, kept the RWST about the same volume.
We've increased the injection lines which is helpful
in terms of getting water into the reactor, but it
also increases the spill rate, if you have a direct
vessel injection line. So if you look, for example,
this DVI case without the RNS as the pumped RHR system
which can interact in this event, so we have to keep
account of whether it's operating or not. But if you
look at the case without RNS, this is just with the
gravity passive systems working, in AP600 you would
get to recirculation in 4.7 hours. With that same
event for AP1000, you get there soon, 2.6 hours and
the main reason for that is that you've got a bigger
line that's spilling so it helps drag down the level
and get you to recirculation sooner.
CHAIRMAN WALLIS: And you've got more
decay heat then because it's sooner.
MR. SCHULZ: And you've got more decay
heat then, so we have to deal with that.
Now it turns out on AP600 that the
limiting case was not the gravity case. It was, in
fact, the case where these RHR pumps were running.
The operators in the plant are instructed that if you
have a LOCA and the ADS goes off, turn on these RHR
pumps, even though they're not safety, they can help
you. Well, if you have a DVI line break, it actually
can hurt you in some respects and that is instead of
having recirculation occur in four some hours, it's
What we analyzed in AP600 SAR is the
limiting recirculation cut in or initiation time was
the 2.1 hours.
Now if we take that same event for AP1000,
it would even be short than 2.1 hours. And we didn't
want that to occur. So what we ended up doing is
changing the RHR pump design so that initially it will
not take water out of the RWST. It takes water from
outside of containment, like a more conventional
MEMBER KRESS: There's another tank out
MR. SCHULZ: Yes. We're going to take
water out of the spent fuel, CAS loading pit or
something. We will not take water out of the RWST.
MEMBER KRESS: So you've essentially added
another source of water to your system?
MR. SCHULZ: Yes, so if the nonsafety
system works, it cannot make the event worse. And in
fact, will make it better because it will put extra
water in the containment.
So now for this event, the limiting case
is, in fact, without the RNS working because that adds
extra water and we end up with additional margin, in
fact, more than twice as much flow as AP600.
MEMBER SHACK: Isn't there something where
you change that alignment in some situation?
MR. SCHULZ: Yes. What would happen as
this event would continue is the event that's talked
about here is you run the RNS at pump until you get to
recirculation and then it magically fails. That's the
worse design basis deterministic type assumption. So
this case here is not what the RNS pump continuing to
run indefinitely, it runs until recirc. and then it
Now in the AP600, once you get the system
lined up and you start it running, it just -- it just
keeps going in the same mode of operation. You don't
have to realign it.
For AP1000 when the outside water supply
gets depleted, we have to switch to the inside water
supply of the RWST. We'll have valves to do that and
it will be manual operation, but it's again, not a
safety. It doesn't have to be done to make the plant
safe. It's an extra level of defense.
So we've done some things to increase the
water level in terms of the initial water level in the
tank and avoiding flooding of the refueling cavity.
We've changed RNS alignment so it takes
water from outside. We've increase the line
resistance. So we've done a number of things that
significantly improve the situation.
We've also done the long-term cooling
calculation again using the SAR methodology which is
COBRA/TRAC in this case.
CHAIRMAN WALLIS: Your ultimate heat sink
MR. SCHULZ: The passive containment
cooling and the air.
CHAIRMAN WALLIS: Containment.
MR. SCHULZ: Yes. Just like in the AP600.
Now if the RNS pumps were running, which
they may be, they have big heat exchangers and if the
CCW is running that's where the heat will go, or most
of it. But again, that's not safety. We don't really
on that to work, but it is another level of defense.
So in this case we've looked at this
limiting DVI break case for AP1000, running the same
methodology as we did with AP600.
CHAIRMAN WALLIS: This nonsafety system
that's pumping water in from outside, it's not
recirculating, it's just pumping it in?
MR. SCHULZ: Initially, pumping it in.
When the outside water supply gets depleted, we'll
realign it to inside containment, much like you do in
CHAIRMAN WALLIS: But you're not taking
another path for taking contaminated water out with
MR. SCHULZ: If you continue to run the
system, you will.
One of the restrictions on running this
system in terms of the operating procedures is that
the activity levels are not very high in containment.
If you get to a situation where you've really damaged
fuel and you have high activity in containment, you
would normally not run the RNS unless you're in a core
melt scenario or you're in one of those kind of
The first guideline would be turn it on,
but if the activity goes up, turn it off or don't run
it. And that's the same situation as AP600.
In fact, AP600, when you initially start the RNS, it's
pulling water from inside containment to inject it
So when we look at the results of the
COBRA/TRAC we see similar behavior, no core uncovery
relative to AP600. So looking at the COBRA/TRAC
results, they look like we were successful in sizing
ADS and recirculation.
CHAIRMAN WALLIS: How close does it come
to uncovery? It always could be within a micron or
MR. SCHULZ: No. I think feet.
CHAIRMAN WALLIS: More reassuring than
MR. SCHULZ: I don't know -- do we know
CHAIRMAN WALLIS: Maybe we could see that
MR. GAGNON: The behavior is comparable as
I recall in terms of margin.
CHAIRMAN WALLIS: So when we see the
details it will look just like AP600?
MR. SCHULZ: Or better. In summary, for
the passive core cooling system, we've retained the
configuration of the design. We selectively increased
features in terms of the capabilities. We've done
these independent hand calculation margins assessments
which are actually part of the design process to give
us a feeling for how much we've increased the
capacities and where we've done that and we've done
some checks using the SAR codes to look at the
I'd like to now move on to containment.
Mike showed you a little bit of containment and talked
a little bit about what we've done to the pressure
vessel. The shell thickness gets a little bit
thicker, the 1-3/4s is the limit to avoid
post-weld heat treatment which we don't want to do, so
we've gone up to that limit.
The total free volume is actually
increased 20 percent or so. The total volume has not
increased that much, but there's a lot of structures
in there and we've accounted for those. And so most
of the increase that we made goes to free volume,
although we do have bigger steam generators in there.
The design pressure has gone up. In order
to account for that design pressure --
CHAIRMAN WALLIS: How much does it grow
when it gets to 59 psig?
MR. SCHULZ: How much does the --
CHAIRMAN WALLIS: How much does it grow in
terms of inches?
MR. SCHULZ: I don't know the answer to
MR. CUMMINGS: Ed Cummings. On AP600 it
was about an inch and a half. I guess this would be
just a little bit larger than that.
MEMBER SHACK: When I went to my friendly
ASTM handbook, I couldn't see where you'd get a whole
lot more design strength under the SA738. You have to
put a code case together for that.
MR. SCHULZ: Yes. We have to put a code
case together for that. I don't remember. I thought
it was significant, but not -- the other thing that
we've done in the containment cooling area is we made
the water storage tank bigger and this is the maximum
water at the overflow point that we're using as a
reference point here. And we increased it from
540,000 gallons to 800,000 gallons. We also increased
the water flow rate.
Now the initial water flow rate didn't
increase very much. We run 400 or so gallons for
about 3 hours to cover the containment shell, to form
the film quickly, relatively quickly. And since the
containment dome is the same shape in diameter and
it's just a little bit taller, we've maintained that
flow rate, increased it a little bit, but not very
After that three hours, the flow rate
increases more proportional to the power increase. So
it's 70 some percent.
CHAIRMAN WALLIS: Your concrete wall is
thicker, is it?
MR. SCHULZ: It's not thicker. It may
have or probably will have more rebar.
CHAIRMAN WALLIS: More rebar, has
something in it to hold up that water at a higher
MR. SCHULZ: Yes, it's higher and it's a
little bit heavier.
CHAIRMAN WALLIS: Do we get into seismic
considerations, the whole thing shakes?
MR. SCHULZ: We will have to demonstrate
that. I think this is one of the issues with the
staff is how much demonstration we will be doing and
I believe we'll be doing one site, one calculation for
MR. CUMMINGS: This is Ed Cummings. We're
providing hard rock seismic analysis case and we have
done an assessment of the roof structure to show
MR. SCHULZ: We have also again, as we've
done in the other features, taken the SAR analysis
codes and methods which is GOTHIC in this case and
analyzed the containment. One thing that was a little
different than -- at least the SAR reference case, was
using a more realistic large LOCA steam generator
energy input. We had used a very conservative
arbitrary input forcing the generator energy to go
into the reactor coolant system very quickly which
gave us a second peak in pressure that was kind of
unrealistic. So we're using this as what we think is
a more realistic scenario. But otherwise, the codes
and methods and conservatisms in the code is the same.
We looked at two limiting cases, the
double ended large LOCA and a large steam line break.
MEMBER KRESS: Now when you're
transferring heat condensing steam on the wall of your
containment, I recall there was some question about
the model we had in there on how it dealt with the
surface area. It might be covered by liquid and the
part might not be -- did that ever get resolved,
MR. SCHULZ: In terms of the outside of
the containment, the water coverage for AP1000 should
actually be a little better at least around the dome
and the upper part of the shell because the geometry
is the same in terms of the diameter and the shape of
the head and we're in the longer time, anyways, we're
putting more water flow on.
MEMBER KRESS: That was one of the issues.
The one I was recalling though was all on the inside.
MR. SCHULZ: Okay.
MEMBER KRESS: That's a different --
CHAIRMAN WALLIS: Doesn't etching, surface
behavior, if it gets dirty and it doesn't wet so well
or what's better and things --
MR. SCHULZ: That's important on the
outside, not so much on the inside.
CHAIRMAN WALLIS: On the outside.
MR. SCHULZ: The wetting and the spreading
of the film.
CHAIRMAN WALLIS: This thing is going to
rush, isn't it?
MR. SCHULZ: No.
CHAIRMAN WALLIS: Things happen to it.
MR. SCHULZ: Well, it's got a coating on
it that is -- has a safety function and it will be
inspected during the life of the plant and in fact,
we'll be running some tests, running water already
outside periodically which will have a couple of
purposes. It will tend to wash the outside and it
also will demonstrate the fact that the water film
forms. So we can test that part of it, in fact.
CHAIRMAN WALLIS: Do you have any wetting
agent you add to the water to help it spread?
MR. SCHULZ: No, we don't. The coding is
an important factor that we do take credit for and
that's why it has some safety function, btu we don't
add anything to the water.
MEMBER SHACK: But you did up the rate as
well as the total volume of water, as I recall.
MR. SCHULZ: Yes, yes. And I think the
issues with curbage were more in the lower flow rate
regimes, not so much in that initial 400 GPM flow
rate. When we slowed down later in time and there's
where we'll have some more water flow.
MEMBER KRESS: We had some questions about
the GOTHIC assumption of well-mixed flow inside the
containment. And now you've got a slightly worse case
for mixing, maybe, I don't know, because you've got
more heat but how have you dealt with that issue of
whether or not it's well mixed in there?
MR. SCHULZ: I think it's something that
Bill Brown will probably get to, maybe, maybe not.
Would you like to address that now?
MR. BROWN: Bill Brown. If you remember
the last time we went through this, one of the things
that was suggested by the Committee was to do some CFD
and included in this report you'll see a comparison
between AP600 and AP1000. We took a little 2-D slice
through the hull of containment and the results that
we see from it is that it looked similarly mixed to
MEMBER KRESS: Good move.
CHAIRMAN WALLIS: Do we get to talk to you
about that later on today?
MR. BROWN: Yes, if you feel good about
CHAIRMAN WALLIS: Well, the obvious
question is why you take a slice instead of a --
whatever the cylindrical symmetry, a slice really
isn't very typical of a cylinder, but we'll get to
that later on, perhaps when you are standing up there.
MR. BROWN: Yes.
MEMBER KRESS: I thought the slice was
MR. SCHULZ: It is.
MEMBER KRESS: That seems to me like it's
CHAIRMAN WALLIS: Well, we'll talk about
MEMBER KRESS: Okay.
MR. SCHULZ: The large break LOCA
transient looks like this for AP1000. You see the
higher design pressure. The actual margins in terms
of PSI and even percentage are even greater on AP1000
than they are on AP600.
CHAIRMAN WALLIS: This is with
MR. SCHULZ: This is with conservative
assumptions, AP600 methodology margins. The only
difference is in the rate of steam generator energy
input which affects the second peak there.
The other event which we looked at which
is actually limiting is the main steam line break.
And it does with the large steam generators that we
have in AP1000, it is understandable why this is
CHAIRMAN WALLIS: Well, you need some
MR. SCHULZ: Well, I'm not so sure.
Passive containment cooling is not very important
In fact, we've run the same transient
without passive containment cooling and the peak is
barely larger --
CHAIRMAN WALLIS: No, I mean I just wonder
how certain -- do you have conservative assumptions in
MR. SCHULZ: Yes.
CHAIRMAN WALLIS: So the real thing should
be less than that?
MR. SCHULZ: Yes. That is correct.
So our conclusion on the containment is that we expect
margins to increase, to be better on AP1000
. We've increased the capacity of the
CHAIRMAN WALLIS: It's a very funny code,
that one. It goes up linearly and just before it
reaches disaster it stops.
MR. SCHULZ: That's not too surprising.
CHAIRMAN WALLIS: That's what I always
MR. SCHULZ: This is what happens when you
get designers working with the analysis, you figure
out how big you have to make the containment and this
was really limiting.
The other thing is that design pressure is
not a disaster either.
CHAIRMAN WALLIS: No.
MR. SCHULZ: You can go above that --
DR. WALLIS: But you'd expect something which
would have a gentler approach to the maximum or
something instead of going up linearly and coming down
MEMBER KRESS: That's probably just an
artifact of the plotting routine.
At some point up there is where you empty
out the steam generator, I guess.
MR. OFSTUN: This is Rick Ofstun. That's
correct. The time that the steam generator empties or
SVIs are closed at 600 seconds, that's when the break
CHAIRMAN WALLIS: So if something happens
MR. OFSTUN: After the break release stops
then the containment heat sinks continue to soak up
CHAIRMAN WALLIS: So it does turnaround
for a good reason.
MR. OFSTUN: Right.
MR. BOEHNERT: What's the peak pressure
MR. OFSTUN: It looks like around 70 or
70.5 psi. I think we're about 2 or 3 psi from the
MR. SCHULZ: Three and a half or four.
MR. OFSTUN: Okay.
MR. SCHULZ: Again, we didn't see anything
that was really different form AP600 in terms of the
phenomenon involved. The margins look larger. Main
steam line break is the limiting event and the
performance of the PCS is not very important in that
CHAIRMAN WALLIS: When we look at this 2.5
psi between design pressure, this is where
stratification might be important. It may be well
mixed but not that well mixed. It makes a difference
to the pressure.
MEMBER KRESS: Yeah, but in the worse
case, it's well mixed.
CHAIRMAN WALLIS: Well, maybe if that's
the case you can reassure us.
MEMBER KRESS: If you're stratified or not
well mixed, you actually get to a lower pressure.
CHAIRMAN WALLIS: I don't, but the
MEMBER KRESS: The containment.
MR. BROWN: Bill Brown, Dr. Wallis. We
would love it if all the steam could go up and get
with that nice cold water up there and we've actually
taken the worse case and actually assume it would be
CHAIRMAN WALLIS: Are you finished now?
MR. SCHULZ: Yes sir. That was our last
CHAIRMAN WALLIS: We're a little bit
behind in time, I think.
I'm wondering -- we're going to take a
break now, but when we come back probably we'll accept
fairly briefly the reports are going to look about the
MR. SCHULZ: Yes.
CHAIRMAN WALLIS: I noticed that some of
your advisors recommended such phenomena be upgraded
in the PIRT. You might want to mention one or two of
those if they're important, but then we should really
move on to the scaling approach, but we're going to
take a break and I think we should have at least 10
minutes. Let's same come back at 5 after.
(Off the record.)
CHAIRMAN WALLIS: Let's come back in
session. We're looking forward to hear some more.
MR. BROWN: PIRT and scaling assessment.
To give you an idea of the outline here, I want to
briefly go over the PIRT assessment, not the PIRT
itself, there really wasn't much changes. And I want
to spend most of my time in the scaling assessment
which will consist a little bit of trying to identify
the things in which we really assessed relative to
actually did an analysis, I guess.
And a little bit on our approach and then
I want to get into some of the major areas, the ADS-
IRWST transition phase, the ADS phase, the sump
injection phase and go over briefly as to what we did
and what the results were, try to give you a summary
overall for what that meant to the integral effects
test facilities of SPES and OSU, and then move on to
the PSS scaling for containment which will large
address the separate effects test and the containment
mixing and stratification.
The main goal, of course, for our part in
scaling assessment was to try to determine the extent
to which the AP600 experimental test data base was
applicable to AP1000 to support our safety analysis
code validation in accordance with 10 CFR part 52.
So a real simple two-step process we went
through, was the first -- take our AP600 PIRTs as they
were and then have them reviewed by several industry
experts for application to AP1000 and then once we got
the results from that we could then look at the import
of the high rank phenomenon and then use that to
assess these phenomenon relative to AP1000.
MEMBER KRESS: How many of these experts
did you have?
MR. BROWN: Several. I'll show you in a
list here in a second.
MEMBER KRESS: Oh, you've got a list. I'm
MR. BROWN: Yeah, I'm going to give you a
MEMBER KRESS: Okay. The usual suspects.
Who is this Hochreiter person?
MR. BROWN: This guy?
He's a big target back here at one point
with the ACRS. Not a big target at Penn State.
Dr. Bajorek from Kansas State, Dr. Bankoff
from Northwestern, of course, Dr. Hochreiter, Dr.
Larson from INEEL, Dr. Peterson and Mr. Wilson, those
were our primary peer reviewers.
Primarily, Dr. Bankoff and Dr. Peterson
had been involved with our containment PIRTs so they
primarily focused on containment for us and they
looked at the others, and the other four looked at our
other events, our large break LOCAs --
CHAIRMAN WALLIS: I am surprised that
these four academics used industry as an adjective to
describe their expertise.
MR. BROWN: Well. Certainly they worked
MEMBER SHACK: Discipline experts.
CHAIRMAN WALLIS: Because usually
academics are regarded as in other world from
MEMBER KRESS: Independent, right?
CHAIRMAN WALLIS: I'm glad to see that --
well, in a way I'm glad to see at least they're
MR. BROWN: Are you disappointed you're
not on the list?
CHAIRMAN WALLIS: Why do you call them
MR. BROWN: Well, I should say perhaps
academic experts who are certainly familiar with our
industry history issues.
MEMBER KRESS: I thought maybe you'd have
Ivan Katten on there.
MR. BROWN: I did talk to Ivan, but I
didn't get a hold of him quite frankly, early enough
to do that, but when we went through this process,
this was really the starting list and what I sort of
decided was I would send them out to this group and if
I got anything significantly different or got a lot of
comments, then we would continue on with this, but
quite frankly the real result of this was there wasn't
a significant different by most of the reviewers.
And here, gives you an idea of what the
summary of the major changes that they came up with.
A large break LOCA, the core entrainment was increased
a little bit from 6 to 7 from a median to a high. We
addressed this via our BE LOCA methodology.
MEMBER KRESS: Was that because you have
a higher steam flow?
MR. BROWN: Yes, right. Because of the
higher power and the higher steam flow, they expected
additional entrainment than they had up in the upper
plenum area, right.
Small break LOCA, same type thing again.
Same issue, really with increased entrainment and
recommended a high for RWST and sump injection and
I've addressed this via some bottom up scaling on
liquid entrainment inception from the hot leg into
ADS-4 and the scaling report. And then the ADS-4 two-
phase pressure drop was increase the High from IRWST
and sump injection as well. And I've addressed this
from more a top-down perspective during the IRWST and
sump two-phase natural circulation.
In containment, we had no changes
whatsoever. The only issues or comments that came up
there were with respect to comments that you made
similarly earlier, Dr. Kress, with the increased
height of the 25 feet in containment, what would that
do to mixing and later on in our report we have some
CFD analysis to try to address that.
In non-LOCA, there were no important
changes either, so the primary changes were really in
the small break LOCA. So we took these changes in
addition to the things that were already ranked as
high and important from AP-600 and we addressed this
in the scaling assessment.
CHAIRMAN WALLIS: What effect does it
have? Suppose you change a number from 6 to 9 or
something, what difference does it make? What the
procedure for making it actually make some difference?
It's nice to see lots of numbers.
MR. BROWN: I think in the case of the
large break, it really doesn't mean a whole lot
because you're going to really, you're going to end up
doing the analysis anyway, where you're varying each
parameter and going through some uncertainty.
CHAIRMAN WALLIS: I've always been curious
about PIRTs. You have these numbers and medians and
high and all that, that ought to mean that you assign
some weighting factor to sensitivity or uncertainty in
your later analysis where you somehow complete the
loop and come back and see whether you really did a
good job or -- I'm never sure that that's actually
MR. BROWN: I think you could go through
a numerical validation of that after the fact.
CHAIRMAN WALLIS: Otherwise, what's the
exercise for except to put a lot of numbers on the
MR. BROWN: I think it certainly gives you
an idea which I used to make sure that these are
certainly a checklist of items that you should have
included and addressed either in scaling and/or test
facilities, certainly it's a good place to start. I
agree. I don't think it's something -- I think it's
a tool to get started with.
CHAIRMAN WALLIS: You might say, if you
have a 9 you need to have independent assessment from
three facilities and if you have 6 you only need one.
There's got to be some sort of tie in between the
numbers in the PIRT and what you actually do.
MR. BROWN: Right. It sounds like you
could write a paper on that.
CHAIRMAN WALLIS: This PIRT, is it an
empty exercise or does it really --
MR. BROWN: Well, as I said I think it
helps me to focus on what needs to be looked at as far
as scaling. I mean certainly in the areas where if
you initially didn't have a test program before we had
the AP600 test program, I think this was probably much
more valuable, where you said look, I really don't
know how this is going to react. Nobody knows. I
mean the experts here probably don't know how it's
going to react, so we need to do the test.
Yes, now that we've been through the
testing process, quite frankly, these particular items
here are really more -- came out of -- now that I know
what happened in the tests I would have ranked these
higher than I would have initially in AP600. So
really, if I was going back to the AP600 PIRT I would
have increased these a little high as well.
Maybe here if we were using numbers, maybe
this was a 9 in the AP600 and maybe it was a 10 in the
AP1000, but you would have gone back to do it. So I
think when you're initially starting a test program,
I think it's pretty helpful. I think once you have
done a test program, it's probably not quite so
CHAIRMAN WALLIS: It might guide the
staff, if the staff ever gets hold of your codes and
they see that these are 9s, then they might focus on
MEMBER KRESS: It gives you a place to
focus on your sensitivities and things of that nature.
MR. BROWN: I think that helps you as to
where you should spend your effort mostly.
CHAIRMAN WALLIS: There's no check that
you actually did spend your effort. That's the thing
that bothers me. It needs to be a loop, a complete
loop of the PIRT so it leads to some actually
quantitative result in some way.
MR. BROWN: Certainly in the code reports
we do identify, I mean all the PIRT items are -- we
make sure that we have certainly a model and I think
that the scrutiny when looked at the model, the
validation as to how does the prediction of the code
compare to the test facility is much higher,
scrutinized much more heavily when it's a higher
CHAIRMAN WALLIS: If we scrutinized we
would find that correlation.
MR. BROWN: You should. And for the
passive core cooling system then this became the list
to include a couple of items which were increased for
AP1000 as well as those which were already ranked
high. And what I did was I tried to lump these into,
especially for the passive core cooling system, I sort
did top-down versus what I did bottom-up and some of
the things that were mentioned like the ADS two-phase
pressure drop and so on are listed in here. And this
gives you an idea of the type of things that I tried
to do from a system level and quite simply I ended up
with something that's of much more local phenomenon
that was difficult to include top-up such as
entrainment or phase separation and so on. These are
bottom-up. So this kind of gives you a list of what
were the high ranked or most important.
CHAIRMAN WALLIS: Is there something here
about this level swell we were talking about earlier?
MR. BROWN: Level swell in the IRWST?
CHAIRMAN WALLIS: No, in the vessel.
MR. BROWN: In the where?
CHAIRMAN WALLIS: In the vessel.
MR. BROWN: In the vessel. Well, the
closest, I guess, you could look at it as one as I
have a reactor vessel inventory scaling and then also
try to look at the core exit void fraction using the
CHAIRMAN WALLIS: It makes a big
difference now that you carry out into the rest of the
system and a difference in how the actual masses
related to whether or not the two-phase level covers
the core, pretty critical how you model that phase
behavior in the vessel.
MR. BROWN: Yeah, you certainly get into
an area though certainly more important, I think,
certainly the codes right in their answers as opposed
to I'd say scaling where you're not so much after a
best estimate answer, but trying to make a relative
comparison between a facility and a plant.
Okay, this is a list of the phenomena for
the passive core cooling system.
CHAIRMAN WALLIS: How would you scale up
to a vessel the business of level swell in the vessel?
MEMBER KRESS: I don't think what you
measured was collapse level. When you measure it in
the test, I don't think they had a swell level.
MR. BROWN: Right, right. The DP cells --
MEMBER KRESS: The DP cells.
MR. BROWN: -- we had were really
measuring a collapsed liquid levels. That's why we
know it's much better.
CHAIRMAN WALLIS: So we don't have a
measure of these --
MEMBER KRESS: You can do some inferring,
but I don't think you have a direct measure.
MR. BROWN: Yes, I think essentially we
have the -- certainly you have an idea of what that is
based on, the DP cells, but we don't have, again
tensitometers sitting in there in the vessel, looking
at the level.
MEMBER KRESS: What you have is a heat
MR. SCHULZ: This is Terry Schulz. We had
heated rods. If the rods were not adequately covered
and cooled, we would see that in temperatures --
MEMBER KRESS: But generally it's hard to
see with the level swell. It cools the rods pretty
doggone good and it's hard to see it break between
-- where a collapse would be and -- it's hard to find.
MR. BROWN: Unfortunately, in the small
break area we don't get the kind of level swelling
that you certainly would get in the large break on the
initial blow down.
Okay, with that I'd like to move on to the
scaling assessments. There's a lot here.
The scaling assessment really focused on
the high-ranked phenomenon.
CHAIRMAN WALLIS: Will you let us know
when this gets to be priorities?
MR. BROWN: Yes, we're getting there,
we're almost there.
We really tried to focus on the small
break LOCA with respect to core cooling and vessel
inventory and then things like the steam line break
for containment pressure. So that was our focus.
And the assessment, looking at the scope
of the scaling assessment, phenomena that we find in
conventional plants for which there's test data bases
that already exist, we did not scale for AP1000 and
that includes the large break LOCA, blowdown steam
generator circulation phases for the small break LOCA
and non-LOCA with the exception of CMT and passive RHR
which are items that are unique to passive plants.
And things that were low ranked or medium
ranked in the AP600 scaling effort that were already
scaled we did not rescale those.
So our basic approach then was starting
from the AP600 scaling analysis, using that as a basis
for AP1000. We tried to use the insights and lessons
and so on and we did not, as I said, reinvent the week
completely here. Processes that were not important or
minor were not scaled and we certainly tried to use
simplified models to try to highlight these
differences or features in AP1000 such as core power,
volume, ADS vent area and things like that so that we
could see what the real differences were and try to
root those out to be more obvious in looking at the
assessment in AP1000 relative to the AP600.
So we sort of did two types of
assessments, if you will. One, we examined the range
of operating conditions, geometry, those types of
things between AP1000 in each test facility. And
usually in those many cases, the AP600 scaling
analysis was already sufficient. This typically
covered the separate effects test. However, when we
tried to look at AP1000 relative to for example
integral effects test facility, we definitely needed
to be able to supplement this with a scaling analysis,
so that's what you'll see here very shortly as some
examples of what we did in the integral effects tests
to try to assess AP1000 relative to AP600.
So I'm going to give you an idea briefly
as to what's covered. In the integral effects we
looked at an assessment of both SPES, OSU, ROSA, the
ADS test, CMT, PRHR and our DNB tests. And we did an
additional new scaling analysis for both SPES and OSU
for the integral effects test.
In the area of containment, we did the
scaling analysis for the LST, or condensation and
heated flat plate tests, water distribution and water
fill formation tests.
Now we're going to get into the scaling
assessment part which is going to be the priority part
of the meeting.
(Whereupon, at 3:21 p.m., the open session
went into closed session.)
(Off the record.)
(Open session resumed at 4:49 p.m.)
CHAIRMAN WALLIS: Let's come back into
MR. BOEHNERT: We are in open session.
MR. GRESHAM: My name is Jim Gresham and
I'm going to talk about the computer codes used for
AP1000, but I'm not going to get into the details of
the codes. Rather, I'm going to talk about the
approach that we're going to use on the analysis for
And pretty simple, you start with the
codes that were approved for AP600, our starting point
for the assessing the codes for the analysis and used
those as much as possible. Of course, to do that we
have to confirm that they're adequate for performing
the safety analysis for the AP1000 design and address
the concerns that were identified on the AP600
application and make sure that we've reached agreement
on applicability after addressing those and then reach
consensus with the staff and then when we do that,
then we'll use those codes for the FSAR analysis to
complete the safety case.
The advantages of doing it this way is we
step through in an orderly fashion in the review
process which we believe would make it more efficient
for us and for the staff in doing that.
We can identify the major deficiencies in
the codes and address those prior to the final review.
And also, through this we focus on the most important
issues, so with the guidance of the PIRT, the scaling
and test comparisons and then we can really focus our
energies on the important phenomena in evaluating the
CHAIRMAN WALLIS: So how do we resolve
these major deficiencies or find out major
deficiencies? I'm just wondering where the starting
point is. When we had AP600 we started asking
questions and eventually got some code documentation
which had deficiencies in it. I guess they're not the
kind of deficiencies --
MR. GRESHAM: I'm talking about code
deficiencies. There may be --
CHAIRMAN WALLIS: Code. Code is whatever
this magical thing is that is quite different from all
of the --
MR. GRESHAM: I think in a few slides
we'll get to this topic a little more and then maybe
we can -- there are different ways that we may address
CHAIRMAN WALLIS: I think there's a real
question about whether Westinghouse can assess
deficiencies in its own code.
MR. GRESHAM: I believe we can. I don't
believe that will be adequate and we will have to
discuss that with the staff and reach agreement on
that. But I think we're the first ones to address the
deficiencies and then our reviewers will decide how
well we did.
The code, the major codes that are used
for the Chapter 15 analysis, I won't dwell on these,
but the COBRA/TRAC is used for the large break LOCA
and the long term core cooling. NOTRUMP for the small
break LOCA. LOFTRAN is used for non-LOCA transients
and for steam generator tube rupture and there is a
kind of a derivative code, LOFTTR2 which models some
of the operator actions and other phenomena for tube
rupture that was within that umbrella. And then
WGOTHIC for containment integrity.
In terms of identifying the adequacy of
these codes for a performing analysis for AP1000.
First step in the process is to identify the important
phenomena which is done through the PIRT that need to
be addressed and these phenomena have been presented
in the PIRT and the scaling and so that task in this
process for AP1000 is complete. And then identify the
correlations and models used in the code to address
the important phenomena and that really was done under
the AP600 application and again, we are planning to
use the same codes for that. So those have been done.
What remains is to demonstrate that they're adequate.
And that is done through the scaling to demonstrate
that the test data base that we have is adequate for
code validation and then to rely on that. And that
has also been done through the -- is provided in the
scaling report, the determination of the adequacy for
both the integral effects test and separate effects
And then the remaining thing is
demonstrate that the limitations that were identified,
what we just talked about for AP600 are addressed on
the next slide.
The first is just to acknowledge what I've
already said that there were things identified in the
AP600 that there were some concerns about and as
already mentioned today, the approval was restricted
to the AP600 application. So we, I think the burden
is on us to present the case on why the codes can be
used for the higher power plant and that's what we
intend to do.
And again, we have to evaluate those
There are several ways that we think that
this can be done. One would be to make some
modification to the design to increase margin in the
area of the deficiency, to demonstrate more margin.
There may be other test validation that can be done to
demonstrate that the code is adequate to model that
phenomenon. Again, we may just -- maybe additional
evaluation margin may demonstrate that yes, there's a
lot of margin in this area and therefore the code is
adequate to demonstrate safety relative to that
We may do analyses with other codes, for
instance, in the containment analysis example that we
already talked about where we did some CFD
calculations to confirm the mixing in the AP1000 and
the AP600. It would be that kind of analysis or maybe
other systems analysis code to provide independent
confirmation of some portion of the transient or all
And it may be necessary to make some
changes to the code also. Any of those five we may
use or some combination of those five to address
We intend to document this work in the
Code Applicability Report which is targeted to be
complete in April. The key contents of that report
will be a discussion of the importance of the
important AP1000 phenomena referencing back to earlier
reports that we've done, a description of the code
that is being used for AP1000. A lot of it is by
reference to the AP600 work, but we'll also discuss
anything different about it, relative to the work that
was done there.
A discussion of the limits of
acceptability or acceptability for AP600. We intend
to go through the FSER from AP600 and address the
major items on the codes that the staff pointed out
and I believe that's a good systematic way to make
sure that we've addressed all of the items that were
And for each limitation we'll say how that
limitation is addressed and there will be some of
those components that I mentioned on the prior page.
CHAIRMAN WALLIS: So how do we determine
what is a code limitation? Do you look up the SERs,
is that what you do? It was not identified then as
the code limitation?
MR. GRESHAM: Those are -- we're not aware
of any other code limitations that we're addressing,
other than those items that were in the FSER. We were
holding that open for something that happened in
scaling that came out. Nothing did pop out from
either the PIRT or the scaling that we thought had to
be added to that list.
Just quickly on how uncertainties will be
addressed for AP1000. Again, we've talked about all
afternoon is that the phenomena are similar to AP600
and the scaling demonstrates that the validation basis
for the codes is adequate. We intend to deal with the
uncertainties the same way that they were for AP600.
For the large break LOCA we are using our best
estimate methodology and 95th percentile will be
identified as stipulated in the FSER. The reason we
said it that way, I believe it was for the passive RHR
and the CMT -- it was the CMT. The staff said if the
PCT goes higher you may have to do more for
uncertainties in looking at those phenomena. So we
will factor that in. Otherwise, the methodology will
be the same.
For the other codes, we're doing a
bounding analysis as we did for AP600. And we will
ensure that the assessment is conservative, rather
than quantify the numbers in our best estimate.
CHAIRMAN WALLIS: Your AP1000 has higher
MR. GRESHAM: That's correct.
CHAIRMAN WALLIS: It's not clear to me
that you're doing necessarily with the same
uncertainties as you dealt with before with AP600. I
know that the phenomena are the same, but you're
pushing them to --
MR. GRESHAM: The numbers may be
CHAIRMAN WALLIS: For reason of the
envelope. It may be that -- so the way in which
uncertainties work, there isn't quite the same as the
way they did before.
MR. GRESHAM: A good example of that is
the oxidation. Down below 1700, the oxidation is
pretty minor and as you start to increase and
oxidation becomes more important and that is one thing
that will have to be dealt with on the AP1000 that we
didn't on AP600. So your comment is correct.
CHAIRMAN WALLIS: They're not so similar
when you're talking about oxidation. You're actually
going to a much higher degree of oxidation than
MR. GRESHAM: Yes.
CHAIRMAN WALLIS: So it's qualitatively --
you could almost argue it's no longer similar. It's
almost different from an extrapolation, but
MR. GRESHAM: Okay.
CHAIRMAN WALLIS: Different.
MR. GRESHAM: Valid point.
MEMBER KRESS: What exactly does your
sub-bullet up there mean?
MR. GRESHAM: That we will -- the
transients will be analyzed in a way to make sure
we're on the conservative side. For instance, in the
containment where we use test data for the hidden mass
transfer correlations and we do a bounding treatment
of those --
MEMBER KRESS: I understand you use the
conservatisms that are specified in approach. I don't
know what it mean as to say that those bound the
Does that mean to say that if I use those
conservatisms and I will be -- have a value that's
close to the 95 percentile if I did a real
uncertainty? What does it mean to say bound the
uncertainties? I don't understand the statement.
MR. GRESHAM: Actually, we haven't defined
the 95 and won't define the 95th percentile
MEMBER KRESS: I understand that. Nobody
has is the problem.
MR. GRESHAM: That's right.
MEMBER KRESS: That's why I always have a
problem with this. I know those uncertainties are
there, I just don't know how much margin they provide
you with respect to the uncertainties.
MR. GRESHAM: I'm using the small break
LOCA example we talked about to comparison to the test
data on a realistic basis and show that we can model
the phenomena and then we'll add the appendix K
uncertainties on top of that to ensure it's a
conservative assessment. That's the idea on that.
MEMBER KRESS: We'll look at that
difference and say there's some sort of margin there.
MR. GRESHAM: Yes. And that margin covers
the uncertainty and I know when I say that you can't
MEMBER KRESS: It doesn't add any meaning.
MR. GRESHAM: The magnitude.
MEMBER KRESS: Bounding the uncertainties
has no meaning to me.
MR. GRESHAM: Perhaps it's just better to
say we're doing --
MEMBER SHACK: If the analysis is
conservative then you've bounded the uncertainties.
It's hard to demonstrate that analysis is actually
MEMBER KRESS: If you believe it's
conservative you believe it's conservative enough to
have a confidence level in your calculation that's
acceptable, but they're saying something than bounding
uncertainty. I'm just having trouble with semantics
and actually it means to do such a calculation.
CHAIRMAN WALLIS: What do you mean by
that? If you have a break, you assume that there was
no flashing and it simply came out as pure water, the
maximum flow rate you could possibly have is all, is
that what you mean by a conservative analysis,
bounding something? Look at some extreme assumption
which takes you right to the end of what's imaginable
and you use that?
MR. GRESHAM: No, I'm not saying that.
There's a bound on how we're going to be.
That's the idea.
CHAIRMAN WALLIS: Dr. Kress is right then.
MR. GRESHAM: Yes, Dr. Kress has a very
CHAIRMAN WALLIS: I think also when you
talk about realistic, I don't know what the criterion
is for realism. Just that you look at some data and
the curve isn't too far away from them? Once you
start to quantify these things in statements like
realistic and uncertainties, conservative, you've got
to be quite careful in your definitions so that we're
all speaking the same language and we can agree on
criteria for evaluation.
That's all going to be cleared up.
MR. GRESHAM: You're right. There are
different ways to do comparisons and you have to all
agree it's a rational approach. I agree.
CHAIRMAN WALLIS: Sometimes it becomes
MR. GRESHAM: Yes. And then the final
step in my process was to reach consensus and this is
-- again, we're starting from codes that were approved
for AP600 and we certainly want to stand on the
foundation of that effort and not repeat anything that
we don't need to, recognizing that there are still
issues that we need to reach agreement on.
We're providing the reports. We've talked
about them several times, to the staff for their
review, to help to make our case and we will be having
discussions with them.
CHAIRMAN WALLIS: You say you're supply
reports. Are you supply codes?
MR. GRESHAM: We need to go through these
steps before making the decision on that.
If we reach agreement that there have
been, yeah, the codes are also applicable to AP1000
and you've covered all the phenomena and the
validation is okay and we don't have to change the
code, then it's not clear that we need to provide the
code to the staff or the staff needs to exercise those
codes in their review.
MEMBER KRESS: Part of that decision may
be based on exercising the codes.
CHAIRMAN WALLIS: It's all quite possible
for people who exercise a code to do things to get
what they want to get, like this business of the
homogeneous assumption. Until you really dig into an
issue you're running yourself, you don't know how it
is you manage to get this answer. One might feel a
little queasy if the answer only comes from
manipulations performed by Westinghouse. Dials and
things can be changed in a way which always
MR. GRESHAM: I guess my answer to you is
that we need to work with the staff to decide where
they need to have the codes and where they don't. And
CHAIRMAN WALLIS: Work with them or you
need to give whatever they ask for?
MR. GRESHAM: We won't do that without
CHAIRMAN WALLIS: Is there a negotiation
MEMBER SHACK: Persuasion.
CHAIRMAN WALLIS: Yes. Persuasion.
MR. CORLETTI: This is Mike Corletti.
Starting with the basis of the approval of AP600,
there was quite an extensive review of the codes. I'd
really like to start there. We don't want to go back
to 1992 before we submit -- before when we sent in all
the code documentation. We're really looking to start
where we left off building on that. The best way we
see that is what did we learn from that 8-year
certification review, what were the hard issues with
each of the codes and the way we resolved each of
those issues. How does that apply to AP1000 and can
we address that?
I think that's -- that is our approach.
CHAIRMAN WALLIS: But a question I think
someone raised before is whether the staff is capable
of knowing what the issues are with the code unless
they actually exercise it themselves because the real
issues of a code can be so hidden that it's hard to
take what they are.
MR. CORLETTI: I certainly think there's
an onus on us to provide them with the documentation
necessary for them to make the determination that the
methods that we employ, that the methodologies that we
employ are sound and acceptable and I think that onus
is on us to give them that sort of documentation.
Whether that specifically requires them to exercise
the codes, it's not clear to us at this time. I think
maybe perhaps as we go through this code applicability
report, we understand the phenomena. We understand
how the test data applied to AP1000. At that time,
we'll have a clear understanding of where such --
where that would be beneficial and where it would not
be. At this time it certainly is not clear, the first
thing you don't do is jump into exercising the codes.
I think we think that would be the last thing we would
MR. GRESHAM: On my third bullet, we're
asking the staff to agree on the acceptability of
these codes for analyzing the AP1000 transients. Now
we fully expect that they're going to need some good
information to be able to make that decision and we
believe much of that is in the documentation already
mentioned and if there are changes to the codes there
are other significant things it may involve their
running the codes, but we need to be looking at this
-- we need to get further in the process before
knowing the answer to that question on each of the
MEMBER KRESS: What is it -- is it a
difficult thing to transfer a code over to NRC? Does
it cause you heartburn for some reason?
MR. GRESHAM: It will add a lot more time
to the review process, that's one concern we have.
CHAIRMAN WALLIS: Our contention might be
that it would simplify everything and make it a lot
MEMBER KRESS: Save time.
CHAIRMAN WALLIS: Save time. You simply
are open and say here is it. We think it's robust and
you won't find any problems with it. Here it is. But
if one has to dig. If one has to say well, I'm a
little suspicious about this and Westinghouse has to
go away on some things and has come back, and then you
say well, maybe this other thing is something we have
to worry about and then Westinghouse has to go away
and answer that, that might be a very inefficient way
of answering concerns. But you don't really know some
concerns. They may get revealed as you begin that
becomes something that is much quicker to resolve by
having the code. This to and fro, taking months every
time a question gets asked before you get an answer.
MR. GRESHAM: We certainly don't want
CHAIRMAN WALLIS: I think the ACRS has
been saying we really support the idea of the staff
wanting code and we think that's the most efficient in
the long run, the most efficient way there is to
assess a code.
MR. CORLETTI: This is Mike Corletti
again. The other thing that we have seen that was
very successful in us resolving issues on the AP600
was the staff's independent calculations of their
independent codes. And those tended to both our
analysis and their analysis demonstrated the real
large margins of the AP600. We're not in the
situation where we have the PCT of 2175 and it's
really critical do we get -- are we at 28, 25 degree
margin to cut this or not. Kind of where we are with
AP600 for most of the small break LOCA no core
uncovery. We really are far away from a lot of the
limits that historically we've been concerned about.
CHAIRMAN WALLIS: I think that's also in
the ACRS letters you'll find that we also support the
idea of the staff having its independent code.
It really helps with the public confidence
that someone else is running something independently
and gets the same answer. It's something --
MR. GRESHAM: That's right.
MR. BOEHNERT: I was just going to comment
on what Mike said. My recollection is a little
different from Mike's and that is that there was a lot
of problems with the codes, particularly NOTRUMP.
There was a lot of back and forth and a lot of
questions from both the staff and the ACRS and the
code and I think it would have been much easier,
quicker, if the staff had the code to resolve some of
those issues. It turned out to be a very difficult
MR. CORLETTI: And I think at the end all
of our codes predicted the same thing, no core
uncovery and we weren't anywhere near any regulatory
limit, so really, were we talking about the importance
of that code in measuring plant safety?
MR. GRESHAM: We do appreciate your
feedback on that and we did listen.
Just to summarize, building on AP600 codes
that were approved for AP600, we'll confirm the
adequacy for AP1000 and address the concerns and
there's a number of ways that that could be done. And
culminating and confirming they're acceptable and
getting some NRC agreement with that and proceed with
the safety analysis. That's all I have.
I'll turn it back over to Mike for a brief
MR. CORLETTI: That really does conclude
our presentation today. I think it's worthwhile to
summarize. Again, we've been working on this for
quite a while, but I think this is certainly the
beginning for you all and going on, I expect at least
one more of these after the staff has had a chance to
review our submittals. Again, the summary of our
proposed approach, we really focused most of our
efforts so far on how are the plants different and how
are they the same for the important phenomena and
really asses how well does the test data support an
AP1000 Design Certification. How we used it on AP600
provided us the data to validate our codes. We're
hoping that you'll agree that the test data base is
sufficient, that it will provide an adequate data base
for code validation, then we can concentrate on the
validation of those codes and the issues related to
those and how we actually would apply --
CHAIRMAN WALLIS: Why would it be any
different from what we did for AP600 if the data that
you're going to use are the same data and it's the
same code, then why should there be anything different
from what you did before? Is there something
different about AP1000 that would make things
MR. CORLETTI: We're certainly starting
there. We think it does deserve a review to make sure
that there wasn't, like some of these residual issues
that have been raised --
CHAIRMAN WALLIS: Well, what would you do,
you're running the code to say compare with a ROSA
test. You did that for AP600.
MR. CORLETTI: Yes, we did.
CHAIRMAN WALLIS: And you get some data
points. Now what's going to be different about what
you do now? You've got the same tests and the codes,
MR. CORLETTI: That is generally, we had
to wait until we confirmed that the test was
CHAIRMAN WALLIS: Then you don't need to
do any more.
MR. CORLETTI: I think we owe it to the
staff to go through some of the conditions, especially
on like what was mentioned on the oxidation model. We
have to commit to that kind of thing. If there is
anything else that maybe was the way we resolve
certain issues or certain REIs and are they still --
the way we resolve that still applicable? I think it
deserves that it's really going to be the content of
our Code Applicability Report.
But that generally is our approach and
that is the approach that we're --
CHAIRMAN WALLIS: Do you expect some new
comparisons between codes and data or just use the old
MR. CORLETTI: I think we would only use
new comparisons if we had new test data or if there
was test data that was new that was suggested that we
use. There were additional tests that we didn't look
at before because it was not available to us. There
was the ROSA test.
CHAIRMAN WALLIS: You claim you can use
the AP600 test program for AP1000.
MR. CORLETTI: Yes.
CHAIRMAN WALLIS: The AP600 is already
being used to confirm that you can use the codes and
you're going to use the same codes, what is left to
do, except run the code?
MR. CORLETTI: We'll have to hear from the
staff. I agree with your approach.
MR. CUMMINGS: I might comment. This is
Ed Cummings. In a few cases, the staff accepted our
codes. Our use of the codes on the AP600, because of
their assessment of the plant safety rather than their
love of the code. And you have to revisit those
places where they accepted the use of the code because
of the clear safety of the plant, to make sure you
still have the same condition of acceptability.
MR. CORLETTI: Yeah, the core uncovery
issue with some of the models that were employed in
NOTRUMP and I'm not going to go any further than that,
but I know there were some that we didn't have to
evaluate because we didn't have core uncovery and it
may -- if we had core uncovery or if we would get into
that situation, maybe that would be something that
we'd have to address.
MR. BROWN: Bill Brown. Dr. Wallis, a
good example, to come back to what Paul Boehnert said
was a NOTRUMP, for example, there was not a momentum
flux model, okay? This was something that was
identified by the ACRS. It's in the FSAR. If for
some reason in AP1000 it was deemed that this model
needs to be improved for acceptance of the code, then
we would need to go back to just check that model and
specifically validate that.
CHAIRMAN WALLIS: How would the staff know
that? They'd have to really run the code with and
without the momentum flux and find out if it mattered.
They'd have to do it.
In other words, they wouldn't know if it's
an important issue or not, would they?
MR. CORLETTI: I think there's other ways
of doing that. There's independent type of
evaluation, a phenomena with either a different
independent code that says is this important or not.
I guess it would be fully best not to get into too
many what ifs until we really submit our report, but
that's the nature of the kind of assessment we're
CHAIRMAN WALLIS: So what's our role in
all of this, ACRS, we're observing this.
MR. CORLETTI: I guess we would be
interested in feedback on our overall approach. Are
we doing the right thing in regards to our approach
with looking at the test data, the kind of scaling
approach that Bill has outlined, looking at the PIRT,
looking at the scaling and then how we plan on our
application of the codes.
MEMBER KRESS: I think we have a legal
responsibility to sign off on the certification
MR. CORLETTI: That would be the later
phase. I think in this pre-certification --
MEMBER KRESS: Not now, but then.
CHAIRMAN WALLIS: So you're going to make
a presentation to the full Committee?
MR. CORLETTI: In April, there's a
presentation of the full Committee. I think we only
have two hours there.
CHAIRMAN WALLIS: They're liable to ask
questions, so you have to be pretty brief.
MR. CORLETTI: And it will be on the whole
-- there's a couple of issues we really didn't speak
of today, so we'll give you a good -- an overview of
the Phase 2 process and where we are in that process.
We won't probably go into as many of Bill's scaling
equations, but we'll probably give a higher level --
CHAIRMAN WALLIS: You need a matrix or
something showing that the numbers come out all right.
MR. CORLETTI: Some of the members aren't
probably familiar enough with AP600 either and we'll
have to at least tell them how big of a test program
that we did do. I mean they don't realize we did a
$40 million test program on the AP600.
CHAIRMAN WALLIS: Let me ask you a
question, a lot of it, you say, has to be resolved
with the staff. So I wonder how we can give the
Commission advice until we see how your discussion
with the staff works out. We haven't really seen
that. We've just seen something that you've presented
and you say now we have to discuss with the staff.
Should we be giving the Commission advice
until we find out how the staff responds to that?
MR. CUMMINGS: I don't think we can
comment on that.
MR. CORLETTI: We were interested in what
you had to say about our approach so far, that's
probably -- and maybe the staff is also. I don't want
to speak for the staff.
CHAIRMAN WALLIS: Well, then I think we've
seen some scaling analysis or that's most of what we
saw. We might be able to respond to that. The
question about whether or not the staff should accept
these codes without further requirements, I'm not sure
we're in a position to reach any conclusion about that
MR. CORLETTI: Without having seen our
code applicability report. I guess some of this
starting with where we left off on AP600, addressing
the major issues from that and that's sort of an
MR. CUMMINGS: This is Ed Cummings.
Within this Phase 2, however, you will have the code
CHAIRMAN WALLIS: Right.
MR. CUMMINGS: Applicability report and at
the end I think we'd like the NRC to address our
request which is can we do AP1000 without incremental
tests and with the existing codes as modified by our
mutual agreement. That's where we'd like to end up
with in Phase 2. It says nothing, by the way about
what the acceptability of the safety analysis is.
That's a Design Certification.
CHAIRMAN WALLIS: We might agree that the
AP1000 phenomena is similar to AP600. I think we
might be able to agree to that. We might be able to
agree that you've given some demonstrations of
Whether or not the scaling represents
adequate validation basis for codes, I wouldn't be
sure, myself, until I found out what I needed to know
in order to get these adequate validations. So until
you actually start doing some things with the codes,
I'm not quite sure what is an adequate validation
basis. It's a carte blanche that says because you've
demonstrated some scaling you've got an adequate
validation basis for code. It may be a little hard
for us to give you.
MR. CORLETTI: That is based on that data
base was sufficient to validate codes for AP600.
CHAIRMAN WALLIS: We don't know what the
questions are for AP1000. There may be different
questions that come up for AP1000. It's not clear
MR. BROWN: I think one example of that
which worked out well was -- maybe I shouldn't say
well, one example was when we got down to the end to
focus on the ADS to IRWST injection phase and we had
done, obviously -- we had already done a large bulk of
scaling, but for example, this is where that momentum
flux issue had come up and in fact, I had gone back
and Mike Young from Westinghouse and we had looked at
scaling in more detail to come up with the largely
hated level penalty approach in NOTRUMP. But again,
we went back to scaling for that purpose. So that's
probably a good example of something that you're
talking about, Dr. Wallis, that we did in AP600. It
could come up later.
CHAIRMAN WALLIS: Are you expecting a
letter from the Committee or are you expecting to just
inform the Committee and wait until we meet again?
MR. CORLETTI: For today's meeting?
CHAIRMAN WALLIS: No, for the full
Committee. Do you expect the Committee to write a
letter based on what you told them or would you --
MEMBER KRESS: For the April meeting, you
CHAIRMAN WALLIS: View the meeting as
being more informative to say this is -- now you can't
dispute about where we are.
MR. CORLETTI: We didn't have expectations
of a letter, nor did we have expectations of a letter
for this meeting.
CHAIRMAN WALLIS: I think it's a little
difficult again to write a letter without some
substantial input from the staff. In other words,
we'd be short-circuiting them. We don't know what
their questions may be. They may have concerns we
don't know about.
MR. CORLETTI: We agree.
CHAIRMAN WALLIS: Other Members have
points you want to raise before you hear form the
MR. CORLETTI: That ends our presentation.
CHAIRMAN WALLIS: I'd like to thank you
for a pretty clear and professional presentation and
being very willing to respond to our questions.
MR. CORLETTI: Thank you very much. It's
been our pleasure.
CHAIRMAN WALLIS: Can we now hear from the
MR. WILSON: This is Jerry Wilson with
NRR. I don't have a formal presentation. I just want
to say that staff hasn't officially started its review
yet. We're waiting for the remaining submittal from
Westinghouse. At that time we're going to do an
acceptance review to determine if there's sufficient
information to start an efficient review at this time.
If there is, then we're going to establish a review
schedule, review the information, prepare
recommendations on responses to the report questions
that Westinghouse has asked and we're going to send a
report to the Commission telling the Commission how we
plan to answer those questions.
Now we -- I anticipate that the Commission
is going to hear from the ACRS on that so we'll be
prepared to come and brief the ACRS on our response to
the questions that Westinghouse has asked us.
MEMBER SHACK: Will this response be in
the form of an SER, for example, on these reports?
MR. WILSON: I wouldn't call it an SER,
but it's some sort of a NUREG report and it's kind of
like a traditional SER. We're going to have a report
and it will be transmitted via a SECY paper.
CHAIRMAN WALLIS: What about the issue of
exercising the codes themselves? Do you have a
position on that?
MR. CARUSSO: This is Ralph Carusso. I
just reiterate the point that Jerry made earlier that
we have sent the letter to Westinghouse asking for
these codes and we're prepared to run them. And if
they've not made the decision yet to provide us with
the codes then that's something we'll have to talk to
CHAIRMAN WALLIS: It seems to me we're
fairly early in the process. You haven't started a
MR. WILSON: That's correct.
CHAIRMAN WALLIS: It's premature for us to
reach any conclusions at this time.
MR. WILSON: Right.
CHAIRMAN WALLIS: That's it from the
MR. WILSON: Yes, it is.
CHAIRMAN WALLIS: Do you have anything
more to say about any --
MR. WILSON: I don't anticipate that we
CHAIRMAN WALLIS: Do my colleagues have
questions to raise?
MEMBER KRESS: No questions at this time.
CHAIRMAN WALLIS: Could we meet perhaps to
discuss this before we go home, go to dinner?
MEMBER KRESS: Sure.
CHAIRMAN WALLIS: Compare our thoughts and
MEMBER KRESS: Yes.
CHAIRMAN WALLIS: Is there any reason why
I shouldn't declare the meeting closed?
MEMBER KRESS: I think it will be a good
CHAIRMAN WALLIS: I'll do so then.
(Whereupon, at 5:33 p.m., the open meeting
was concluded and the closed meeting commenced.)
Page Last Reviewed/Updated Tuesday, June 09, 2020