490th Meeting - March 7, 2002
Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title: Advisory Committee on Reactor Safeguards
490th Meeting - OPEN SESSION
Docket Number: (not applicable)
Location: Rockville, Maryland
Date: Thursday, March 7, 2002
Work Order No.: NRC-272 Pages 1-38/50-97/118-271
NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers
1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005
(202) 234-4433. UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
+ + + + +
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS (ACRS)
490TH MEETING
+ + + + +
THURSDAY,
MARCH 7, 2002
+ + + + +
ROCKVILLE, MARYLAND
+ + + + +
The committee met at the Nuclear
Regulatory Commission, Two White Flint North,
Room T2B3, 11545 Rockville Pike, at 8:30 a.m.,
George E. Apostolakis, Chairman, presiding.
COMMITTEE MEMBERS PRESENT:
GEORGE E. APOSTOLAKIS Chairman
MARIO V. BONACA Vice Chairman
F. PETER FORD Member
THOMAS S. KRESS Member
DANA A. POWERS Member
WILLIAM J. SHACK Member
JOHN D. SIEBER Member
ACRS STAFF PRESENT:
MAGGALEAN W. WESTON
PAUL A. BOEHNERT
SAM DURAISWAMY
SHER BAHADUR
CAROL A. HARRIS
JOHN T. LARKINS
MICHAEL T. MARKLEY
I N D E X
AGENDA ITEM PAGE
Opening Remarks by the ACRS Chairman . . . . . . . 4
Clinton Nuclear Power Station Core Power . . . . 5
Uprate
By Bill Bohlke . . . . . . . . . . . . . . . 7
By Bill Specer . . . . . . . . . . . . . . .11
By Eric Schweitzer . . . . . . . . . . . . .32
By John Zwolinsky. . . . . . . . . . . . . .78
By Ed Throm. . . . . . . . . . . . . . . . .93
By Bob Pettis. . . . . . . . . . . . . . . .96
Proposed NEI 00-04, Option 2 Implementation
Guideline for Risk-Information for Special
Treatment Requirements of 10 CFR Part 50
By Tony Pietrangelo. . . . . . . . . . . . 170
Arkansas Nuclear One, Unit 2 Core Power
Uprate
By John D. Sieber. . . . . . . . . . . . . 180
By Craig Anderson. . . . . . . . . . . . . 181
By Bryan Daiber. . . . . . . . . . . . . . 185
By Dale James. . . . . . . . . . . . . . . 208
Adjourn. . . . . . . . . . . . . . . . . . . . . 271
. P-R-O-C-E-E-D-I-N-G-S
(8:33 a.m.)
CHAIRMAN APOSTOLAKIS: The meeting will
now come to order.
This is the first day of the 490th meeting
of the Advisory Committee on Reactor Safeguards.
During today's meeting, the committee will consider
the following: Clinton Nuclear Power Station Unit One
Core Power Uprate; Proposed NEI 00-04, "Option 2
Implementation Guideline," for Risk-Informing the
Special Treatment Requirements of 10 CFR Part 50;
Arkansas Nuclear One, Unit 2 Core Power Uprate; and
Proposed ACRS Reports.
Portions of the meeting may be closed to
discuss GE Nuclear Energy and Westinghouse proprietary
information. This meeting is being conducted in
accordance with the provisions of the Federal Advisory
Committee Act. Dr. John T. Larkins is the designated
federal official for the initial portion of the
meeting.
We have received no written comments or
requests for time to make oral statements from members
of the public regarding today's sessions.
A transcript of portions of the meeting is
being kept, and it is requested that the speakers use
one of the microphones, identify themselves, and speak
with sufficient clarity and volume so that they can be
readily heard.
The first item on the agenda is the
Clinton Nuclear Power Station Core Power Uprate. Dr.
Powers is the cognizant member. Please.
MEMBER POWERS: A fact. Thank you, Mr.
Chairman.
We're going to discuss the Clinton power
uprate with both the applicant and the staff. There
is going to be episodic interruptions in the meeting
in order to close it to handle proprietary data, and
I'll beg the Chairman's indulgence for any extension
of the schedule that occurs because of that.
The Clinton power uprate is for BWR6.
We've certainly heard power uprates before, but this
is the first BWR6 we'll hear about. The uprate is
significant. It's overall 20 percent. It's taking
place, however, in two steps -- a seven percent, a 13
percent. It also involves a change in the fuel.
There was a subcommittee meeting dealing
with this subject, and some draft positions have been
taken -- developed by that subcommittee. What the
subcommittee found was that the licensee is, of
course, pursuing this power uprate under what has come
to be called the ELTR1 and ELTR2 methodologies that
the staff have approved.
But, in fact, this is a constant power --
a constant pressure power uprate, and the details of
that methodology are still being reviewed by the
staff. As a consequence, the applicant will take
certain exceptions to the ELTR1 and ELTR2
methodologies, and I encourage the committee to pay
close attention to these exceptions.
At least one of the exceptions is the
familiar large transient testing that we've discussed
before. I'm disappointed Mr. Rosen is not here to
hear the discussion on that particular exception, but
there are other exceptions having to do with the
analyses. And I, again, suggest the committee pay
close attention to it.
The use of a constant pressure power
uprate converts the problem of power uprate from one
that's primarily from a hydraulic issue to one that's
much more neutronic flavor. There are, however, some
thermal hydraulic issues that you have to deal with,
even for a constant power -- constant pressure power
uprate, because you've got to have increased flow
someplace in this system.
And, of course, that flow takes -- those
increases in flow take place in the feedwater and the
steam flow, and that raises some issues of flow-
assisted corrosion in some of the piping systems.
And, indeed, we have issues of flow-assisted corrosion
in this particular unit, and I encourage the committee
to pay close attention to those particular issues.
The applicant and the staff, of course,
think they have this issue well under control through
a combination of modeling and monitoring. There is a
history in the nuclear industry of these methods not
working, with some substantial consequences. So it's
worth paying attention to that.
With that introduction, I will turn to Mr.
Bill Bohlke from the applicant to begin the discussion
of their proposed extended power uprate for the
Clinton Power Station Unit Number 1.
MR. BOHLKE: Thank you. Good morning, Mr.
Chairman, and members of the committee. I'm Bill
Bohlke, Senior Vice President of Nuclear Services for
Exelon Generation.
I just thought I'd spend a minute or so
giving you the background on AmerGen, which is a
company that you may not be particularly familiar
with.
MEMBER POWERS: True.
MR. BOHLKE: AmerGen is co-owned by
British Energy and Exelon. When it was recently
formed, it was co-owned by British Energy and PECO.
But with the Comed and PECO merger, Exelon assumed the
original PECO share. So that's the ownership, and
AmerGen is, in fact, the licensee.
Operationally, Clinton is part of the
Midwest Regional Operating Group, just as Oyster Creek
and TMI are parts of the Mid-Atlantic Regional
Operating Group. What that means is we share a Chief
Nuclear Officer, who last week became Jack Scolds who
succeeded Oliver Kingsley who is now our head of
generation.
And specifically, in the Midwest, the
executive direction and corporate oversight for the
Clinton station is executed by the Midwest ROG out of
Warrenville, Illinois.
The staffing for Clinton, similar to the
staffing for the other two AmerGen units, is a
combination of Exelon employees and AmerGen employees.
Those AmerGen employees have various heritages
depending upon the utility from which they came. In
fact, the station leadership at Clinton currently
consists of a site vice president, plant manager, site
engineering director, site operations director, and
training manager, all of whom are Exelon Nuclear
employees.
So we also use Exelon policies, Exelon
programs and processes, down to a level where we want
station or unit individuality as opposed to common or
standardized processes, so that many of the
organizational structure and management approaches for
Clinton are the Exelon approaches. And the technical
approaches, including the technical approaches
embodied in this request for power increase, is
derived from the Exelon approach.
Specifically, this is a fourth boiling
water reactor station in Illinois that we've subjected
to this. LaSalle was the first one, and you reviewed
that in either late '99 or early 2000. And then, of
course, last fall you heard the presentation on the
Dresden and Quad Cities power uprates.
Dresden 2, in fact, has been uprated and
is operating at 912 megawatts, which is its generator
limit. That startup and testing went extremely
smoothly. Quad Cities 2 has just completed its outage
and is at about 40 percent power this morning going --
undergoing its testing. And so far so good on that
one also. So Clinton will be the fourth in a series
of that.
To try to achieve continuity, the project
manager for the Clinton power uprate was, in fact,
project manager for the LaSalle power uprate. Some of
the technical people are the same. You probably
recognize some of them. So what we do is we allow
ourselves to benefit from the lessons learned and move
it on down the line, so that every project has a
benefit of its predecessor.
And so what we'll see is, when we do the
startup testing for Clinton, it's subject to the
granting of the power uprate license. We'll have
startup testing personnel who have worked at Dresden
or Quad or LaSalle previously, so that we'll have that
lessons learned. We think that's a real strength of
the program.
So that's the extent of my introductory
remarks. I did want to set the stage for that, and
now let me introduce Dale Spencer, who is the Project
Manager for the Clinton extended power uprate.
Thank you.
MEMBER POWERS: Mr. Bohlke, I appreciate
your giving us that introduction to this company. We
see the name all the time, but we really don't know
too much about it.
MR. BOHLKE: You're quite welcome.
MR. SPENCER: Thank you, Bill. Good
morning.
Dale Spencer, Exelon Nuclear, Project
Manager for the Clinton Unit 1 extended power uprate.
Over the next hour, our experts will be providing a
summary of the EPU project, including the
modifications, the analyses performed, and our plans
for implementation. Presentation material has been
chosen based on the agenda that been provided to us by
the ACRS.
As we discussed previously, portions of
our material are proprietary to the General Electric
Company, and we'll ask that a portion of the meeting
be closed. We have grouped the information that's
proprietary together, so we can minimize
interruptions.
MEMBER POWERS: If you will just indicate
to me when you need to close it --
MR. SPENCER: Yes, sir, we will.
MEMBER POWERS: -- we will go through
whatever machinations we have to.
MR. SPENCER: Yes. Yes, sir, we will.
As an introduction, I want to first spend
a few minutes and provide a summary of the overall EPU
project, and then I'll follow by an overview of the
modifications and analyses that we have performed.
We're requesting a license for a 20
percent increase in reactor power. We use the GE
standard EPU process as the guide for our analyses and
the schedule. These GE processes, as you know, have
been used for a number of extended and stretch power
uprates in the industry, both domestically and abroad.
We'll be performing modifications to the
plant to facilitate power ascension, and I'll cover
these in more detail in a couple of slides. And these
modifications will be installed between now and early
2004.
Of these modifications, we'll show that
we're making relatively few changes to the operation
of safety systems. Our plans are to implement the
power ascension in two steps. The first step will be
-- take place when we start up this May after our
refueling outage.
MEMBER POWERS: Let me ask a question.
You make a point that you're making relatively few
changes to the safety system. Am I supposed to derive
comfort from that?
MR. SPENCER: Yes.
MEMBER POWERS: Why?
MR. SPENCER: Essentially, our analyses
have shown that the modifications to the plant and the
limits to the plant post uprate will be on the BOP
side. Our changes are essentially, as I'll get into
in the next slide, the nuclear instrumentation that
we're going into. Other plants have gotten into
modifications in other areas, and with the BWR6 we
have found that this is not the need. And this is a
plus.
MEMBER POWERS: I mean, what you're
essentially saying is that your safety systems have
enough margin to handle the additional 20 percent.
MR. SPENCER: Absolutely.
MEMBER POWERS: Okay. But, clearly,
you're reducing the margins you have in those systems.
MR. SPENCER: Yes, absolutely.
MEMBER POWERS: And somehow that's
acceptable.
MR. SPENCER: Yes, it is.
We talked about our first step for our
license, for our power ascension in May of this year.
And the second step of our power ascension will take
place after our ninth outage, and that's scheduled for
early 2004.
On the next slide is a simple graph of the
power-to-flow map at EPU conditions. For clarity, in
the upper right-hand corner, the gold area, is the EPU
operating region. Simply, as we stated in the
subcommittee, we're increasing power along the
previously licensed MELLLA flow control line.
Other plants that have licensed the
extended power uprate have licensed the MELLLA as part
of their EPU process. In the case of Clinton, this
has already been licensed, so we are not changing any
of the flow control line in our power uprate.
MEMBER KRESS: The axis is 100 percent of
what? The core flow is for what -- percent of what?
I mean, core power -- core flow. Is that 100 percent
of what?
MEMBER POWERS: It's both. I mean, the
question applies to both.
MEMBER KRESS: Yes. What are the units on
your --
MR. SPENCER: The axis on the power is the
100 percent of uprated reactor power, in the top of
the graph right here, the 3473.
MEMBER KRESS: Okay. So that's the full
new uprated power.
MR. SPENCER: Yes, sir.
MEMBER KRESS: What's the one on the
bottom?
MR. SPENCER: The one on the bottom is the
core flow. The core flow is not changing. The core
flow is based on the capability of the recirc system.
So we will need to --
MEMBER KRESS: So when you go up to 110
percent almost there, what does that mean?
MR. SPENCER: I'm sorry. Which --
MEMBER KRESS: Well, at the --
MR. SPENCER: Are you looking right in
here?
MEMBER KRESS: No. Looking at the yellow
part.
MEMBER SHACK: The X axis.
MEMBER KRESS: The X axis, and looking
there. That's like 109 percent or something.
MR. SPENCER: Oh. This is the ICF, the
increased core flow region. This is previously
licensed on --
MEMBER KRESS: This is the previously
licensed core flow.
MR. SPENCER: Yes, sir.
MEMBER KRESS: There's a maximum core flow
in your license?
MR. SPENCER: Yes, sir.
MEMBER KRESS: I see.
MR. SPENCER: This was our license as we
have it right now, and it's in the same X axis, if you
can see on the graph.
MEMBER KRESS: The dotted line is --
MR. SPENCER: Yes.
MEMBER KRESS: It goes all the way up to
108 percent?
MR. SPENCER: In core flow, that's
correct.
MEMBER KRESS: Okay.
MR. SPENCER: You know, that's actually
107, I believe.
MEMBER KRESS: Is there any reason why you
can't use that little triangle up at the top?
MR. SPENCER: That's basically the
capability of the recirc system.
MEMBER KRESS: Okay. You would have to
change out your jet pumps to --
MR. SPENCER: That would be a pretty
significant change. Yes, sir, that's correct.
On the next slide, I just have a brief
summary of the change in plant conditions graphically.
Briefly, the increase in steam flow is accomplished by
replacement of the high pressure turbine, and, thus,
no changes in the reactor steam dome pressure is
needed. And we discussed this at the opening of the
meeting.
I'd like to spend just a few minutes going
over some of the modifications we'll be performing.
As we stated in our uprate safety analysis report, no
safety-related hardware changes will be required to
implement the EPU at Clinton.
Upon issuance of the revised operating
license, we're going to perform changes to nuclear
instrumentation which will allow us to increase our
output. These set point changes include the APRM, the
flow bias, both the SCRAM and the rod block, the main
steam line group 1 isolation, the control and stop
valve and recirc pump trip bypasses, and the low power
and high power set points on the control rod block
pattern controller.
Proceeding to the modifications we'll be
performing on the BOP side of the plant, as I talked
previously, we're going to be implementing our power
ascension in two steps. During our upcoming refueling
outage, we'll be replacing the high pressure turbine.
We'll be replacing the main power transformers, as
well as associated changes to the isolated phase bus
duct configuration and cooling.
The main generator hydrogen coolers will
be replaced, and we'll increase the hydrogen pressure
in the generator from the current 60 to 75 pounds.
The exciter anode transformer will be replaced, and
we'll be upgrading five supports associated with the
feedwater system, all of which will allow us to
achieve the additional 80 plus megawatts for the next
operating cycle.
MEMBER POWERS: When you make changes in
your hydrogen system, changes in transformers, how do
you affect the risk of fire-initiated accidents in
your plant?
MR. SPENCER: The fire-initiated accidents
were analyzed, and we are going to be discussing some
of the risk from all of the risk factors a little bit
later in the presentation. I believe Bill Burchill is
going to get into that at some later time. Can we
discuss that then, or would you like to --
MEMBER POWERS: That would be fine.
MR. SPENCER: Okay. And that is part of
our presentation material at a later time.
Proceeding on, to ensure we get the full
potential from our uprate, we'll be performing
additional modifications to -- and I call them BOP
efficiency improvements in the future. These are
targeted to be installed either online or during the
ninth refueling outage to facilitate future power
increases. And since these are a little bit down the
road, these modifications are in the scoping stage,
and I'm just going to provide a conceptual overview
right now.
Improvements will be made to the condenser
to perform at a higher efficiency. Improvements will
be made to allow condensate polisher stop rate and
balanced flow configuration at the higher condensate
flows we expected. Moisture separator reheat Chevrons
will be replaced to improve the MSR, and that goes
forth to the plant efficiency.
Changes will be made to the breakers,
conductors, relay schemes associated with the
switchyard to allow the increased megawatts electric
and MVA output of the plant. Improvements to the
exciter plan, which will allow the plant to run at the
full capability of the generator. And we do foresee
future improvements in the cooling capability of the
bus duct cooling.
MEMBER FORD: Can you just elaborate on
the --
CHAIRMAN APOSTOLAKIS: Microphone.
MEMBER FORD: Could you elaborate on the
main condenser improvements? What are they, and why
are they being made?
MR. SPENCER: Okay. These are changes
that we're going to make in our ninth refueling
outage, which is currently scheduled for early 2004.
And I'll preface it with the fact that we're doing
conceptual studies, and this is not finalized at this
time.
The most -- I'll say the most front
runners we have right now are changes in online
cleaning system and making sure that we're using the
condenser to its full capability, not having any air
entrained in the condenser. Essentially, making sure
we run it at its highest efficiency.
MEMBER FORD: With the increased steam
flow, are you not expecting vibration problems in the
condenser --
MR. SPENCER: We have --
MEMBER FORD: -- with the current design?
MR. SPENCER: We have performed analyses
of the condenser tubes. We are -- obviously, we are
putting more steam flow in. We have analyzed this to
be acceptable.
MEMBER FORD: Was there any basis for
saying that? I mean, is it based on analysis, or
other plants' experience? I guess these are GE
turbines, and there's plenty of other GE turbines with
the same design out there.
MR. SPENCER: Sure. And --
MEMBER FORD: Are there any with the same
flow rate, the increased flow rates, to draw on?
MR. SPENCER: For our analysis, we used a
specialty vendor who does this kind of work in several
locations, and every condenser is just a little bit
different. There is a mix between analytical
techniques and actual industry experience that he
factors into his work. We also do routinely monitor
the performance and perform inspections on equipment,
even down to the condenser stage in our plant. And
that's an ongoing type evaluation.
We'll continue to do these inspections
even post uprate and continually monitor the
performance of all of our plant equipment.
MEMBER FORD: Okay. So these improvements
aren't necessarily related to increased steam flow,
the EPU. It's just -- you just want to increase the
efficiency. It's got nothing at all to do with --
it's not driven by the fact that you've got increased
steam flow.
MR. SPENCER: At the current efficiency of
the condenser, it's not -- we're not going to be able
to get a whole lot extra out of the condenser, unless
we do something to it. So it is a little bit of both.
MEMBER FORD: Okay. Thank you.
MR. SPENCER: So I'd like to change the
focus just a little bit here, and I want to
concentrate on some of the analyses and evaluations
that we've performed in support of EPU. Listed on the
slide are the specific subjects for which we have
prepared presentation material and our experts will be
talking.
As I stated previously, we have chosen the
subjects based on the agenda provided us to the ACRS.
So at this time, I'd like to introduce Fran Bolger of
General Electric, who will discuss the core and fuels
analyses.
MR. BOLGER: Morning. I'd like to discuss
some of the details of the core fuel analysis that
have been performed. As part of the power uprate,
there was an equilibrium core analysis, which did
demonstrate a full extended power uprate power, that
the core was able to provide the desired energy and
have adequate thermal margins.
I'd like to discuss some of the details of
the actual core design which was performed for
Cycle 9. Cycle 9 is the first step in the two-step
process that was previously described.
Next slide.
To the left, there is a -- this is a
picture of the core design for Cycle 9. What you see
are the color -- these shaded bundles are the fresh
core, the fresh bundles in the core. Up here on the
top you see the locations. And the I and J location,
we'll be talking a little bit about those.
Looking at the core design map, you'll
notice a value in the center of each square, and
that's the bundle exposure, megawatt days per shore
ton. The zero indicates that it's a fresh bundle.
The value here on the bottom, which you
can't see very well, that correlates to a bundle type
used in the simulator. This core was analyzed with
the PANACEA 3D simulator, and that relates to this
value here called IAT down here on the bottom.
If you look on the bottom, you'll see the
bundles that are loaded in this core. The top bundle
is a two cycle depleted bundle, which is a GE10 type
fuel, which is eight by eight design. The next two
bundles are one cycle depleted bundles, which are GE14
type. And the last two are the fresh fuel, which are
also GE14 type.
If you look, you'll -- if you look on this
bundle name, you'll see these numbers here. These
indicate what the bundle average enrichment is for the
bundle. There is 268 fresh bundles being loaded,
which is a fairly large bag size. These values here
are the batch average exposures for the fresh bundles
at beginning of cycle.
Over here on the right is a batch average
radial peaking. These values are actually rounded two
points past the decimal.
What I'd like to do now is talk a little
bit about -- a question?
VICE CHAIRMAN BONACA: Just G14, what is
it, a 10 by 10, 11 by 11?
MR. BOLGER: G14 is a 10 by 10 design.
VICE CHAIRMAN BONACA: 10 by 10.
MR. BOLGER: I'd like to talk now about
some of the results -- the key results in the core
design, cycle design analysis. Over here on the right
you see this column here, which is the cycle exposure,
and this is measured in megawatt days per shore ton.
And you see the core is designed with various steps
through the cycle, and each one of these steps has a
different control rod pattern.
For example, you see here this is the
control rod pattern at the beginning of cycle. You
see the rod positions shown here on the map. The red
boxes are actually the controlled cell locations.
The next column is the critical Eiger
value. When a core is designed, a target critical
Eiger value is developed through the cycle, and this
target is developed based on previous cycle experience
with that plant. It's based on other plants with a
similar fuel design and size, and also based on the
fuel design characteristics.
The next column is a -- is for this
depletion, the core flow as a function of exposure.
If you look at the core flow, you'll notice that some
of the core flow values are below what was on that top
corner of the power flow map previously shown. The
minimum core flow at full EPU is 99 percent, but this
is lower than that because this depletion is actually
about 90 percent of full EPU power.
The next column is the ratio of the
operating limit, minimum critical power ratio -- I'll
call it MCPR -- to the calculated MCPR. When a core
is designed, you try to achieve sufficient MCPR margin
so that the core will operate when it is actually
monitored in the plant. You design the core typically
with about seven percent MCPR margin.
In this case, this core -- the maximum is
about .9, and so there is a little bit of margin
relative to what typically would be the target of
about .93. So there is actually a little bit
additional margin, and this core could probably
operate at a little bit higher power.
These values in parentheses over on the
right of the MCPR margin is the location of the
limiting bundle, and those values correspond to these
locations on this I and J location over here on the
core map. And you see the limiting location does move
around in the core.
The next column is the ratio of the
calculated peak rod LHGR relative over the LHGR limit.
And in this case these locations are the I and J
location as described over here, but also this right-
most is the axial location. The core is designed with
25 axial nodes. So, for example, you see over here
node 4 is toward the bottom corner of the core.
MEMBER POWERS: You said that the core was
designed with 25 axial nodes. Do you really mean it
was analyzed with 25 axial nodes?
MR. BOLGER: Yes, that's correct. The
core was analyzed with 25 axial nodes.
The right -- this column here is the ratio
of the average planar linear heat generation rate to
the average planar linear heat generation rate limit.
And this is where the LOCA limits are factored in.
The right-most column is the core average
axial power shape peak value, and what you see here is
the value for the core average peak and the node. For
example, node 10 is toward the center of the core.
You'll notice that the core peak for most of the time
in the cycle is toward the bottom of the core. And as
you get down toward the end of the cycle, the power
shade moves up to the top.
The BWR will naturally try to peak to the
bottom because of the voids -- the voids in the core.
The core -- this core, as shown, has -- provides the
desired energy and has adequate MCPR and LHGR margin.
Next slide, please.
What I'm showing here is -- this is the
same core design as you saw on the previous slide at
beginning of cycle. This is just to show what you
would get if you depleted the same core with the same
control rod patterns at a lower reactor power -- in
this case, about a seven percent lower reactor power.
You'll see, if you compare the two pages,
that the critical Eiger value is slightly higher,
because it's at a lower void fraction. The thermal
limits are lower because -- obviously, because the
core is at a lower core thermal power.
And you'll also see that the power shape
has shifted up somewhat, because the core is at a
higher void fraction. That allows the power shape to
move somewhat.
If this were an actual design in this
case, there is more than adequate MCPR and LHGR
margin. So you would -- the designer would try to
take advantage of that additional margin. And we'd
try to reduce bundles. We would try to move some of
the bundles towards the center to try and improve the
efficiency of the core. The designer might try to
simplify the operating rod patterns to give more
flexibility to the site.
So the designer will actually try to
target the same margins to limits at the power level
that it is being designed to.
Next slide, please.
MEMBER POWERS: Might I ask you one
question on this? Your axial power peak moves fairly
continuously through this. But there's a
discontinuity in the node where you have your axial
peak power. That occurs around an exposure of 11,500.
Why does that discontinuity occur?
MR. BOLGER: Right here you see the power
shape moving -- moves -- starts moving up to the top.
You'll also -- I can't tell you exactly, but you'll
notice that the actual peak is -- the value of the
peak is not very different. It could be that the
power shapes are -- have a double peak or a fairly
flat distribution, and just a very small variation in
the power shape will shift it up to the top.
MEMBER POWERS: So this is more a
numerical thing than it is --
MR. BOLGER: Yes.
MEMBER POWERS: -- a real discontinuity in
the core performance.
MR. BOLGER: Yes.
Next slide.
In summary, the equilibrium core design
that was analyzed for the EPU, and also the Cycle 9
design, has adequate margin.
Any questions? I'd like to -- the next
presenter is Eric Schweitzer from AmerGen, who will
present the containment analysis.
MEMBER POWERS: Maybe before you depart,
I'll just ask you one question. The fuel exposures
that you've shown in these two analyses are relatively
modest. If I ask you about Cycles 10, 11, and 12,
what kinds of fuel exposures do you anticipate taking
fuel to?
MR. BOLGER: You know, I don't -- I can
just answer you generally.
MEMBER POWERS: Yes, a general answer is
fine.
MR. BOLGER: You know, the fact that the
batch size is fairly large means that the two cycle
depleted bundles are -- it's not a full batch, and
it's -- you wouldn't expect it to go to a three cycle
depleted bundle. So the fuel will be depleted -- will
be discharged after its second cycle. And the two
cycle depleted bundles are primarily on the periphery.
So from a batch average standpoint, the
batch has not a significantly different batch
discharge than you would have if it were depleted at
the current rate of power, just because the batch size
would be lower and it would be possible to have a
larger percentage of the batch loaded in the internal
part of the core on the -- on its third cycle.
MEMBER POWERS: Can you give me a number?
MR. BOLGER: I don't have the value with
me.
MEMBER POWERS: It sounds like you're
going to discharge something a little over 30.
MR. BOLGER: I have a picture of the end
of cycle exposure map. It may be around slide 30 or
so of the backups. Actually, you're interested more
on the -- even the further on cycles.
MEMBER POWERS: I'm really interested in
the --
MR. BOHLKE: Dr. Powers, let me see if I
can -- if I recall correctly -- this is Bill Bohlke
from Exelon. We won't have any burnups over 50,000
megawatt days per ton.
MEMBER POWERS: That's what it sounded
like. Thank you. That's all the more precision I
needed.
VICE CHAIRMAN BONACA: Just have a
question regarding the cycle length. What is the
cycle length of Cycle 9?
MR. BOLGER: The cycle length for Cycle 9
is a 21-month cycle.
VICE CHAIRMAN BONACA: And the following
cycles, are they planning the same cycle length or --
MR. BOLGER: Maybe Exelon would like to
discuss that.
MR. SPENCER: Future cycle lengths.
MR. SCHWEITZER: Your question was, what
is the future cycle lengths?
VICE CHAIRMAN BONACA: Yes. Are you going
to stay at a 21-month cycle or --
MR. SCHWEITZER: Exelon does plan to
transition to 24-month cycles, and there will be a
future license amendment submittal for that.
VICE CHAIRMAN BONACA: Any idea now how
that would change some of this neutronics here? What
you just showed us?
MR. BOLGER: With -- go back to the
previous slide. There is a -- for a 24-month cycle,
there is a little margin for a higher enrichment.
These bundles have not been designed to the maximum
enrichment capability. So they can probably go to
about the 415 level, and that will provide some
additional energy for a 24-month cycle.
There is a benefit in having higher
enrichment bundles in preceding cycles. They will
carry more of the load as they get down into further
cycles. So as you have higher enrichment
transitioning, that will help you to create such a
core design. And there is still a little bit of room
to add a little more fuel, so it'll be challenging but
the capability exists to do that.
VICE CHAIRMAN BONACA: Thank you.
MR. SCHWEITZER: My name is Eric
Schweitzer from AmerGen, and I'd like to present the
Clinton MARK III containment analysis.
To evaluate the containment for extended
power uprate, we followed the established method for
containment analysis in ELTR1. The limiting events
that were analyzed were the main steam line break, the
recirculation suction line break, and the alternate
shutdown cooling.
The next slide shows a summary of the
results. This table shows the drywell and containment
pressures and temperatures and the suppression pool
temperature following the analyzed events. The first
column of values on the left are the original analysis
in the Clinton updated safety analysis report.
The second column of values are the
comparison benchmark cases, which used the EPU methods
with the original licensed power. The third column of
values are the EPU results, and the last column shows
the design basis.
Comparing the first and second columns
shows the effect of the change in methodology. And
comparing the second and third columns shows the
effect of EPU, which is relatively minor with no
vessel pressure change. Comparing the third and
fourth columns shows the margins to the limits.
I'd like to point out that all remain
below the design limit with the exception of the
drywell temperature. This value is above the design
temperature of 330 degrees for less than .5 seconds.
This has been evaluated as acceptable, because there
is insufficient time to heat up the structure.
In conclusion, these results show
acceptable performance for the containment in EPU.
MEMBER FORD: Could I ask a question?
This is more for clarification. This seems fine,
assuming that a containment maintains its original
integrity. Now, I'm out of my depth here, but that's
a big assumption, isn't it? That the containment
maintains its original design integrity. Could you
not have degradation of that integrity?
MR. SCHWEITZER: The containment is tested
on a periodic basis, leak rate tested, and so it's
maintained.
MEMBER FORD: So corrosion of rebar, for
instance, that would be detected?
MR. SCHWEITZER: The leakage would
definitely be detected, if that would cause any
leakage, but the strength of the materials would not
be expected to be changed outside its design margins.
MEMBER FORD: I know this is a topic that
can be -- probably go into the -- a revised version of
GALL. But you're taking as right that the monitoring
programs you have regarding the containment integrity
are adequate.
MR. BOHLKE: Dr. Kress, let me -- excuse
me. Dr. Ford, let me answer that. First of all, it's
a steel containment.
MEMBER FORD: It's a steel containment.
MR. BOHLKE: And it's accessible. It's go
ta shield going around it, so it's accessible for
inspections and it does have the periodic leak rate
test --
MEMBER FORD: So it's rather like Oyster
Creek.
MR. BOHLKE: -- of the penetrations and
the shell as a whole. So it's pretty robust, and it's
pretty inspectable. We don't -- we have been doing
containment ISI inspections on other plants in the
fleet, and the extent of corrosion that we found after
as much as 30 years of operation at Dresden and Quad
Cities, for example, is pretty minimal. But, in fact,
there is an inspection program as part of our
requirements.
MEMBER FORD: The reason why I bring up
the question is I never hear this topic mentioned, and
yet I remember when I was employed by General Electric
that we were concerned about corrosion at Oyster Creek
of the containment.
MR. BOHLKE: That's right. Exactly.
MEMBER FORD: And I have never heard
anything along these lines mentioned since, and this
is why I just bring the question up. As I say, it's
more for my information. Are we kind of opening up a
potential there for --
MR. BOHLKE: No. We think we're okay
because we have a program which specifically focuses
on that.
MEMBER FORD: Fine. Okay. Good.
MEMBER POWERS: I'm a little surprised
that the applicant didn't bring to your attention that
in establishing these limits that they take a certain
amount of corrosion and degradation into account. I
mean, there is a margin built into them for those
reasons.
MEMBER FORD: Yes. A concern -- my
concern is that whenever you look at these corrosion
allowances they are almost going to be picked out of
the air.
MEMBER POWERS: They are, and they are
minuscule compared to what you saw at Oyster Creek.
MEMBER FORD: Correct. Correct.
MR. BYAM: Good morning. I'm Tim Byam
with AmerGen. The next part of our presentation will
address the exceptions that we've taken to the
requirements specified in the extended power uprate
licensing topical reports. That is, ELTR1 and ELTR2.
This portion of our presentation does
contain General Electric company proprietary
information, and we, therefore, ask that the meeting
be closed at this point.
MEMBER POWERS: Okay. There will be a
little interruption while we go through our steps
here.
MR. STROMQUIST: This is Eric Stromquist
from General Electric. I'd kindly ask that Mr.
Wilson, Mr. Huff, and Mr. Moss leave.
MEMBER POWERS: You will switch to --
CHAIRMAN APOSTOLAKIS: You have to speak
in the microphone.
MR. STROMQUIST: I'm sorry.
MEMBER POWERS: I don't know for this
step, but we need to -- no, I don't think this is --
we should be switching at this point.
MR. STROMQUIST: This is Eric Stromquist
with General Electric. All persons are acceptable in
the room now.
MEMBER POWERS: Thank you.
(Whereupon, the proceedings went
immediately into Closed Session.)
. CHAIRMAN APOSTOLAKIS: Okay. Proceed.
MR. BYAM: Kent Scott will now continue
with our presentation on the anticipated transient
without SCRAM event response.
MR. SCOTT: Thanks, Tim.
Now I would like to discuss the response
of the plant to an ATWS at uprated conditions. First,
we found our response as operators to an ATWS remains
unchanged. For example, when we lower reactor water
level to reduce subcooling and trip the reactor
recirculation pumps, we find the plant operating at
the same power-to-flow conditions as pre-EPU.
This is due to the fact that the plant is
currently licensed and operating under the maximum
extended operating domain analysis. This is the
MELLLA analysis with increased core flow that Dale
spoke of earlier.
Since we already operate at these extended
load lines, the plant reacts the same, simply moving
down the existing load line on recirculation pump
trips. Also, the symptoms we must observe as
operators to detect an ATWS remains unchanged.
And, finally, our actions to mitigate an
ATWS remain unchanged for controlling reactor power,
reactor level, and reactor pressure.
MEMBER KRESS: But do you have to change
how far you lower the water level into the core?
MR. SCOTT: No. We haven't changed -- we
still lower level to the same values. We train on the
same bands that we lowered the level to reduce
subcooling. So that did not change at all.
VICE CHAIRMAN BONACA: The time for you to
take action, however, has been reduced, right?
MR. SCOTT: And the analysis that Bill
Burchill is going to talk about a little bit later
about probabilistic risk assessment talks about those
times. Those times are well within the capabilities
of the operators to perform. The sequence of the
actions we take are the same. The required times do
reduce, but they are well within the capabilities of
the operators. We're trained to do that, and I fully
expect everybody to be --
VICE CHAIRMAN BONACA: Could you tell me
what the times are? I mean, just for information.
MR. SCOTT: And Bill may be able to help
me out a little bit with those times. I could tell
you on the order, but I'd rather Bill tell you some
particulars with that. Bill?
MR. BURCHILL: This is Bill Burchill with
Exelon. The realistic analysis indicates that with --
I think it's with one slick pump the time is reduced
from nine minutes to six minutes, and with two it's
from 12 to nine minutes. And those times, as you
recognize, are well beyond the licensing calculation,
which assumes the times are on the order of a couple
of minutes.
MR. SCOTT: Right. And my experience with
initiating standby liquid control in an ATWS situation
-- the times that Bill is talking about are an
eternity for operators to get those actions performed.
MR. BURCHILL: Yes. This is Bill Burchill
again. Again, the operator, of course, operates off
of symptoms. You know, the response is specifically
to the symptom, and, you know, the time is probably
not in their mind at the moment when they're doing
that. They're reacting to a symptom and taking
action.
VICE CHAIRMAN BONACA: Yes. But, I mean,
time available is important to determine whether they
will take the action within a certain time.
Now, you mentioned something about two
minutes. That's the design basis analysis rather than
the -- so these values you gave us, nine minutes
versus six, are a best estimate?
MR. SCOTT: Those are the probabilistic
risk assessment values, realistic analyses.
MR. BURCHILL: Right. Right. Dr. Bonaca,
those are based on map runs specifically to look at
the time available.
VICE CHAIRMAN BONACA: Okay. But your
ATWS analysis that you have docketed with the NRC has
different values.
MR. BURCHILL: Those are the licensing
analysis. You're correct.
VICE CHAIRMAN BONACA: And two minutes are
the reduced times for this design, or the previous --
MR. BURCHILL: Two minutes are the design.
VICE CHAIRMAN BONACA: And before it used
to be three? Two? Two. So --
MR. BURCHILL: Yes, it's the same.
VICE CHAIRMAN BONACA: Why would it be the
same?
MR. BURCHILL: I'm sorry. I didn't
understand.
VICE CHAIRMAN BONACA: I said, why would
it be the same time?
MR. BURCHILL: Because it's already a
bounding time. It's already well within what we
consider as a realistic evaluation of the time
available.
VICE CHAIRMAN BONACA: I mean, as a number
that comes out of an analysis, which is a bounding
analysis, I would have expected to see a change in
that time, too, with respect --
MR. PAPPONE: Yes. This is Dan Pappone,
GE. The two minutes is actually an input assumption
into the analysis. And, again, that's based on --
that's based on the knowledge that the operators are
going to be working off of the symptoms, performing
the same actions based on the same symptoms that are
occurring actually a little faster when you get to
power uprate, when you're looking at power levels
going up and water levels coming down.
When you're getting into those ATWS
situations, the symptoms are going a little bit
faster. So the operator is going to go through his
motions in the same time period. The two minutes is
what we're assuming in the analysis. It's not a
number coming out -- it's not a number calculated from
the analysis results.
VICE CHAIRMAN BONACA: Okay. thank you.
MR. SCOTT: Okay. One thing we have done
is to raise the minimum allowable standby liquid --
CHAIRMAN APOSTOLAKIS: Let me understand
this. The two minutes are used in the analysis. What
analysis is this?
MR. SCOTT: Dan, can you --
MR. PAPPONE: This is Dan Pappone again.
We perform a safety analysis to confirm that the peak
vessel pressure, the peak suppression pool
temperatures, are going to be acceptable. And in
performing that analysis, we have to make certain
assumptions, say, on operator reactions, because we're
using --
CHAIRMAN APOSTOLAKIS: So the shorter you
assume the action is, the more optimistic you are,
aren't you?
VICE CHAIRMAN BONACA: That's right.
CHAIRMAN APOSTOLAKIS: So if you are
saying in the PRA that the actual number will be
around six minutes, and you assume two, then the
deterministic analysis is optimistic.
MR. PAPPONE: No. The --
CHAIRMAN APOSTOLAKIS: No? Am I missing
something?
MEMBER POWERS: It's the time available.
MR. PAPPONE: The PRA analysis is looking
at the maximum time available for the operator to
perform those actions as part of a success criteria.
He has a period of six minutes or nine minutes to
perform that action.
CHAIRMAN APOSTOLAKIS: Okay. And in --
MR. PAPPONE: And if he completes that
action within that time, then the event is successful.
If he fails --
CHAIRMAN APOSTOLAKIS: Right.
MR. PAPPONE: -- if he fails to complete
that action in that six minutes, then that's
considered failure.
CHAIRMAN APOSTOLAKIS: Right.
MR. PAPPONE: It's just a simple success
criteria on the PRA side.
CHAIRMAN APOSTOLAKIS: Yes. Well, I mean,
if I do the deterministic analysis --
MR. PAPPONE: The deterministic
analysis --
CHAIRMAN APOSTOLAKIS: -- it's six minutes
down to two minutes.
MR. PAPPONE: No, no. The --
VICE CHAIRMAN BONACA: The deterministic
was a previous analysis they did for licensing --
MEMBER SHACK: No. But George's point is
if they used six minutes in the deterministic
analysis, they wouldn't have liked the answer.
MR. PAPPONE: Absolutely.
CHAIRMAN APOSTOLAKIS: And, therefore,
would you fill in the blanks there?
MR. PAPPONE: The deterministic analysis
has certain levels of conservatisms in the code. So
that's going to -- and methods that we're using. So
that's going to push the answer above what we'd like
to see. In the PRA analysis --
CHAIRMAN APOSTOLAKIS: But we can't
separate the PRA analysis from everything else. I
mean, it's not a different world.
MR. PAPPONE: But it's a different set of
-- it's a different set of modeling assumptions that
are used in the calculation.
CHAIRMAN APOSTOLAKIS: For the same
system.
MR. PAPPONE: Right.
CHAIRMAN APOSTOLAKIS: Yes. So, you know,
which ones do we go by? I mean, would the calculated
temperatures and pressures change significantly if you
assumed a realistic six-minute response time?
MR. SCOTT: In my experience, as an
operator and watching the crews train, and being a
part of the crews in training, is that the PRA
analysis versus the two-minute analysis, it doesn't
matter to me. You know, my actions are the same.
I'm going to step through, and I'm going
to do those actions in the same amount of time. I can
get those actions done in two minutes whether we have
a power -- whether I'm at 50 percent power or 100
percent power or at 120 percent power. It doesn't
matter.
I think the key is that as long as we can
say that, yes, the times change from nine to six
minutes, and two minutes is still bounding, I'm
comfortable as an operator in being able to take those
actions to protect the plant.
CHAIRMAN APOSTOLAKIS: Right. And you are
speaking now in PRA space, the response of the
operators.
MR. BURCHILL: Well, this is Bill Burchill
again. I want to clarify one thing, Dr. Apostolakis.
The PRA doesn't calculate that it will take the
operator six minutes.
CHAIRMAN APOSTOLAKIS: I understand that.
I understand that. It's the deterministic analysis
that bothers me.
VICE CHAIRMAN BONACA: Yes.
CHAIRMAN APOSTOLAKIS: I mean, the --
MEMBER SHACK: Well, he's using a more
conservative analysis. You know, it's sort of a
bounding analysis versus a best estimate. So that
when he does the bounding analysis --
VICE CHAIRMAN BONACA: Yes. But two
things bother me there. One is that for the
deterministic analysis there must have been a best
case that was analyzed some time in the past for this
plant that said that, based on conservative estimates
of what it takes to reach those points, it takes two
minutes. Okay?
And now you are saying you feel
comfortable with two minutes, and I don't. I mean, at
some point I will become uncomfortable.
MEMBER SHACK: Well, I think the answer is
he takes the action in two minutes.
VICE CHAIRMAN BONACA: I understand.
MEMBER SHACK: And, you know, he gets 1440
in one case and 1477 in the other. So as he keeps the
time fixed and he ups the power, the temperature does
go up, which is what you would expect.
VICE CHAIRMAN BONACA: Yes. No, but I was
saying that now I would have expected that if you now
go up in power you will have a change in that time.
MEMBER SHACK: He kept that time fixed,
presumably because the regulator accepted the two
minutes, and so he lives with the two minutes and sees
what happens.
MR. CARUSO: Dr. Bonaca, this is Ralph
Caruso from the staff. I've been informed that I
believe that the two minutes was a number that came
about as part of the original ATWS rulemaking.
VICE CHAIRMAN BONACA: That's right.
MR. CARUSO: That was an input assumption
that was established at that time as a reasonable
amount of time for an operator to respond to these.
So that's an input to these assumptions.
I believe also there's a comparison going
on here between a deterministic calculation using one
particular GE code and the PRA calculations which are
done using the MAP code, which is an entirely
different code. So you get -- unfortunately, you get
different numbers when you use different codes.
VICE CHAIRMAN BONACA: I understand. But
the fact is, you know, it's important we understand
how going up in power -- okay, what effect it's going
to have on operator reaction.
MR. CARUSO: Yes.
VICE CHAIRMAN BONACA: And time is always
an effect on that. There may be confidence on the
part of the operator that he can perform the action,
but at some point the confidence will be decreased,
just because time is an issue for him to detect, to
respond, and to take action.
So that's why we're pursuing this kind of
questioning, and it's confusing to hear an assumed
number of two minutes for the ATWS and, you know,
there -- I would have to look at the analysis to
understand why it's done in a particular way, because
you cannot get an input from that. And, therefore, we
have to depend on the MAP analysis to get the sense of
time dependency. That's the whole issue.
CHAIRMAN APOSTOLAKIS: So we have to live
with the two minutes, then. This is something
that's --
VICE CHAIRMAN BONACA: I guess so.
CHAIRMAN APOSTOLAKIS: -- NRC given.
MEMBER KRESS: Well, the two minutes --
MEMBER SHACK: NRC accepted at least.
MEMBER KRESS: The two minutes in the
original rule must have come out of observations on
simulators and saying, "Well, they've always" --
MEMBER POWERS: Oh, I wouldn't think so.
I bet -- I bet the original analysis came out a wide-
ass guess with a bunch of conservative --
MEMBER KRESS: Well, it's one of --
MEMBER POWERS: -- considerations.
MEMBER KRESS: It could be one or the
other.
MEMBER POWERS: It could be one or the
other.
MR. CARUSO: I am informed by people who
have some knowledge of this that this arose out of the
recirculation pump trip time. And there are a number
of different things that go into this 120-second value
that's used. We can identify this for you in more
detail if you wish, where it came from, but I think
the issue here really is, for the deterministic
analysis, they use a value of 120 seconds, then come
up with a certain result which is acceptable.
In PRA space, they've determined that they
might be able to go even longer, might be able to go
six or nine minutes.
VICE CHAIRMAN BONACA: Okay.
MR. CARUSO: Okay? But as the operators
here are saying, they feel comfortable that they would
recognize these symptoms and respond to them very
quickly. So there is a little bit of a disconnect
here, but it's -- I think it's explainable and
understandable. It's more of an artifact of the way
that the different methods calculate these parameters.
CHAIRMAN APOSTOLAKIS: So if they respond
in six minutes instead of two, in real life now, the
peak clad temperature will be higher than this, won't
it?
MR. BURCHILL: This is Bill Burchill
again. No. In fact, if they respond in the six or
nine minutes, depending upon the number of pumps, they
will meet all of the success criteria in the PRA
analysis. It's likely you would not meet the
licensing limit, but you would be using an apples and
oranges comparison because --
VICE CHAIRMAN BONACA: The 2200 degrees?
Is that the licensing --
MR. BURCHILL: Right. But that also has
restrictions on the -- you know, the various inputs to
the calculation, the heat transfer correlations, and
all of that stuff that we're -- you know,
traditionally imposed on the design basis analysis,
which would not be true in a realistic analysis that's
used for the PRA.
CHAIRMAN APOSTOLAKIS: So where does that
leave us? I don't understand that. Now, are you
saying --
MEMBER POWERS: Wherever it leaves us,
let's move on, so that I cannot destroy your schedule.
MEMBER KRESS: It leaves us with the
thought that the idea of using best estimate codes
with 95 percent confidence was a pretty good idea,
because you can understand what the number is.
VICE CHAIRMAN BONACA: Well, I think it
would have been interesting I guess -- and I really
don't need -- I accept the fact that two minutes in
the design basis analysis are acceptable for both
conditions. I would have liked to know at what point
I could see those two minutes -- what number does it
become before it becomes unacceptable? You know, to
understand what -- how the margins --
MEMBER POWERS: It depends on -- it
depends critically on whether you're taking core
damage as your criteria for acceptability or 2200
degrees Fahrenheit as your criteria for acceptability.
VICE CHAIRMAN BONACA: I mean, whatever
was the licensing value.
MEMBER POWERS: I mean, it seems to me
that had I known when I did the analysis for peak clad
temperature that my maximum temperature was going to
be 1440, I would have said, "Well, instead of putting
in two minutes for that criteria, I'll put in three
and a half, because I've got more room and I like my
operators. I don't want to put too much torque on
them." And, indeed, they would have probably found
that they met the 2200.
If instead they said your criteria is core
damage, they might well have been able to put in seven
minutes.
VICE CHAIRMAN BONACA: Well, I understand
that, but that was --
MEMBER POWERS: Or 15 minutes maybe.
MR. CARUSO: Dr. Powers, I mean, actually,
you're focusing on peak clad temperature here. For
ATWS events, the more limiting parameter is the pool
temperature.
MEMBER POWERS: And if I bring that up, I
protract a discussion that's already gone on too long.
Okay?
MR. CARUSO: Oh, okay. I just wanted to
make that point. Everyone is focusing on peak clad
temperature, but in an ATWS really the limit that
you're going to hit first is the suppression pool
temperature. That's more important than the peak clad
temperature, because you've got water going through
the core. So you -- it's -- you're going to keep --
MEMBER POWERS: They were trying to
understand where the time comes, and I was --
MR. CARUSO: Okay.
MEMBER POWERS: -- trying to point it out
to them. And even if you had taken the suppression
pool temperature in PRA space, that's not too
important. What is overpressurization of the drywell
becomes important.
MR. CARUSO: Right.
MEMBER POWERS: And the times even go
longer than that.
MR. CARUSO: Right.
MEMBER POWERS: Can we go ahead?
MR. SCOTT: Certainly. Okay. One thing
that we have done is to raise the minimum allowable
standby liquid control boron concentration with --
MEMBER POWERS: The points that you might
want to make is, where do you inject your boron into
this core?
MR. SCOTT: The boron goes in through the
high pressure core spray sparger, which goes right
onto the core, so it's a core --
MEMBER POWERS: On top of the core.
MR. SCOTT: That's correct.
MEMBER POWERS: And not on the bottom.
And so now you're not relying on raising and lowering
the water to mix the boron.
MR. SCOTT: That's correct.
MEMBER POWERS: That's an important
feature of this plant.
MR. SCOTT: Thank you.
So we are raising the minimum allowable
standby liquid control boron concentration to ensure
the rate of negative reactivity addition remains
acceptable after the power uprate. And we have
included the table here. I'm showing some of the
major parameters pre- and post EPU, along with the
associated design limits.
And just to conclude, that these values
show and support the acceptability of maintaining the
existing operator response to an ATWS after
implementing the power uprate.
So now I'd like to introduce Harold
Crockett from Exelon Nuclear, who will discuss plant
response to flow accelerated corrosion.
MR. CROCKETT: Thank you, Kent.
I'm Harold Crockett, and I'm the flow
accelerated corrosion program manager for AmerGen and
Exelon, and I'd like to talk with you a few minutes
about our program and what we have done.
The Clinton station has a program that is
consistent with the industry recognized EPRI
recommendations for flow accelerated corrosion, and
what we have done is we have updated our analysis with
the new design conditions.
MEMBER POWERS: And the spelling of
CHECKWORKS.
MR. CROCKETT: Yes. And as noted, we use
the EPRI program CHECKWORKS. It is a predictive
analysis. And because our analysis is largely cycle-
dependent, dissolved oxygen temperatures flow, we want
to look at each line that is modeled.
And so what we did, we saw the results,
and in our particular station here, the scavenging
steam line had the most significant increase. And we
wanted to cite this example, because the numbers are
a little bit high in the world of fact.
Normally, there are generally small
numbers -- wear rates are maybe five mils per year and
you get a 15 percent increase, so you're up to a
whopping six mils per year. This one was a little bit
higher. We went and focused on this particular line
and looked at the actual measured wear and compared it
with the previous predicted wear. Actual measured
wear was about 20 mils per year, and the old predicted
methodology gave us 38 mils per year.
And what -- the goal is to merge the
predicted with the measured and get a refined
correction factor on this --
MEMBER POWERS: Well, somehow 52 mils per
year doesn't give me much more comfort than 70 mils
per year.
MR. CROCKETT: That's correct. And as
we'll note further down, this particular line we will
be visiting for replacement. We have found that
proactive replacement is a good policy for us if, at
the same station we've seen wear, or at our other
stations we've seen wear.
By the time you put up the scaffolding and
remove the insulation, you've spent so much effort
that a lot of times it's easiest just to go ahead and
upgrade it with chrome-olly and stainless. And yes,
sir, you're exactly right. We have done that, and we
continue to do that.
We've learned a lot in the past decade
about which lines are wearing. And Clinton being a
younger station, they're just getting to the point
where they're doing some of these replacements.
MEMBER FORD: Maybe it's a moot point if
you're going to replace the carbon steel steam line
with chrome-olly. But could you comment on the
qualification of CHECKWORKS for wet steam? Given the
fact that there are different corrosion mechanisms or
different corrosion criteria between wet steam and
water. So how well has CHECKWORKS been qualified by
observation versus prediction, with steam versus
water?
MR. CROCKETT: Yes, sir. The check family
predictions -- the CHECKWORKS, it is set up for a
single phase and two phase, and steam quality is an
input. And we continue to refine the code. There has
not been a dramatic amount of changes in the past
eight years. There's been some small refinements.
EPRI sponsors meetings twice a year, and
we are very active in those meetings, which is the
domestic utilities as well as the international
utilities. We have strong support from around the
world at these meetings. And when we're seeing high
wear we go visit those very areas. So it's --
everybody is pretty much talking to each other.
MEMBER FORD: Would the fact that there
are different mechanisms involved, between those two
environments, is it fair --
MR. CROCKETT: Yes, sir.
MEMBER FORD: -- is it fair to use
CHECKWORKS from one -- from water and then just do a
flip to steam?
MR. CROCKETT: Right. As I mentioned
earlier, steam quality is an input. And what we're
really talking about is not a mechanical attack. It's
a corrosive attack. It's a dissolution of the oxide
layer -- washes away, dissolves the next oxide layer,
and repeats itself.
And it is different, but the code has been
consistently substantiated where it has been used.
And there have been some industry events, and the code
was not properly used. I think at the last
subcommittee meeting --
MEMBER FORD: Okay.
MR. CROCKETT: -- they talked a little bit
about Fort Calhoun's rupture. And when they did go
back and look at the code and properly analyze it, it
did have wear rates that were exactly or very
consistently with the --
MEMBER FORD: But a difference between
less than 20 and 38, between measured and calculated,
that's not unusual. Is that unusual or not?
MR. CROCKETT: That is not unusual to have
a predicted off by that much. And that's why we
continue to select, inspect, and evaluate, and feed it
back into the process. It's not unusual for it to be
off by that much.
MEMBER FORD: And is a discrepancy always
the same way?
MR. CROCKETT: No, sir. It can be --
MEMBER FORD: Plus or minus.
MR. CROCKETT: That is correct.
MEMBER FORD: Oh, okay.
MR. CROCKETT: It could be more than
measured in a prediction or less.
MEMBER FORD: Okay.
MR. CROCKETT: As I mentioned earlier, the
model will continue to be calibrated with post-EPU
conditions. And these changes were anticipated. We
talk about power uprates at our conferences, so the
code is consistently applied.
The schedule replacements, we will
continue to inspect this particular line, both trains
of it. So if we get up to data we receive from this
particular outage that will start up next month, we
may elect to proactively replace this line even before
that time. But it is an ongoing process.
And the programmatic controls are in place
to ensure that inspections continue, and the extent of
the condition is assessed. So if we find wear, we'll
measure upstream and downstream and that -- that
analysis.
MEMBER POWERS: Let me just ask one point
of fact. This line that's corroding at 70 mils per
year in your analysis, what's the wall thickness on
it?
MR. CROCKETT: This is a half-inch wall
thickness.
MEMBER POWERS: Half-inch. And I can't
resist just pointing out that programmatic controls
were in place at Fort Calhoun and Surry.
MR. CROCKETT: Well, Surry certainly was
the birthplace of the modern codes. And Fort Calhoun
-- I was asked to be on a team that assessed that
particular station. And prior to their rupture, they
had not been active with the industry meetings. And
their analysis was partial I guess would be the way to
address that.
MEMBER POWERS: A generous way to put it.
MR. CROCKETT: Yes.
MEMBER SHACK: What are your wear rates in
your feedwater lines?
MR. CROCKETT: Feedwater lines -- BWRs
typically, because of the dissolved oxygen, are not
high. Some of the PWRs have had some feedwater
replacements --
MEMBER SHACK: It's a big difference. But
just -- I mean, is it a couple mils a year?
MR. CROCKETT: On the order of five to 10
mils per year. And that's a much thicker pipe
typically. That's an inch and a half or more.
Yes, sir.
MEMBER FORD: I have a question, not
related to flow assisted accelerated corrosion.
Fluent use vibration -- I recognize that fatigue is
probably another problem in the upper head. However,
there have been stress corrosion problems of core
spray lines and dryers, and this was brought up at the
Quad Cities and Dresden applications.
We raised the question about whether
there's a loose parts problem, and we were assured
that it was not a problem. Has this been revisited
for this particular station -- Clinton?
MR. MOSER: Yes. Dr. Ford, Keith Moser,
Exelon, Reactor and Internals Program Manager.
Yes. We did exactly the same thing we did
for Dresden and Quad. We went back, component by
component, looked at all of the different problems.
And as you suggested, flow-induced vibration was one
of the issues we looked at. You know, for this plant,
we didn't have to put in any mods. Everything worked
out fine. The dryer we looked at before. We're going
to be looking at it right after this outage.
Again, we don't think there will be any
issues. But we do have programs in place to look at
it, and we will be looking --
MEMBER FORD: My concern was entirely the
fact -- I'm not concerned about exceeding -- that it's
going to cause fatigue. I'm more concerned about
exacerbating cracking, stress corrosion cracking, by
the fact that you're superimposing a vibrational load,
which has been increased because of the EPU.
Therefore, if you're going to inspect once
every outage, which is appropriate, is that good
enough?
MR. MOSER: For which component in
particular?
MEMBER FORD: Well, I was thinking of core
spray lines, steam dryers, the brackets holding the
steam dryers to the pressure vessel. They have all
undergone stress corrosion cracking at one time or
other.
MR. MOSER: You know, for the core spray
lines, flow-induced vibration really isn't a big
problem in those lines. The steam dryer, yes, we have
concern, and we have some industry experience on that
-- Peach Bottom, some other overseas plants. But
based on the fact that the loose parts issue got such
a big dryer, big component, even if it would crack
there's nowhere for it to go, and it's not a safety
concern.
MEMBER FORD: Except down.
MR. MOSER: Yes. But that would go right
on top of the separator, correct?
MEMBER FORD: That's correct.
MR. MOSER: And so, in a sense, you don't
have anything that can really cause you concern as far
as a safety perspective. And so from that
perspective, yes, we are absolutely sure that we want
a cycle of looking at the dryers, making sure there
isn't any gross degradation -- is the right thing to
do.
For the brackets themselves on the RPV
wall, yes, we looked at them before. There has been
some industry experiences. We are going to look at
them after the outage -- after the EPU conditions and
make sure that we have everything modeled correctly.
MEMBER FORD: Okay.
MR. MOSER: Does that answer your
question?
MEMBER FORD: Yes. Kind of.
MR. MOSER: Okay.
MEMBER FORD: Yes.
MR. CROCKETT: In conclusion, the EPU
changes are acceptable to the FAC program.
I'd like to turn the --
MEMBER POWERS: Well, at this point, I'm
going to intercede. We've exceeded the allotted time
for this. This is not a risk-informed submission. I
believe those interested in the risk significance of
this submission can read the viewgraphs, and I propose
that we move right to the closing. And any points you
want to make about the implementation you can make
there.
MR. SIMPKIN: I am Terry Simpkin. I'm the
Manager of Licensing for Exelon Nuclear. First of
all, I'd like to thank the staff for their rigorous
review and I'd like to thank this Committee for its
consideration of our request to uprate the power level
at the Clinton Power Station.
We have completed extensive analyses,
using accepted methodology. We have identified no
significant impacts on plant response or system
integrity. Our request involves minimum changes in
plant risk and we believe that plant operation is
acceptable at the extended power uprate conditions.
Subject to any questions from the Committee, this
concludes our presentation.
MEMBER POWERS: Do Members have any other
questions they would like to pose on this to the
licensees about this?
Well, thank you very much, gentlemen, and
I'm sorry to eliminate a couple of sections of your
presentation, but I think the visual aids were very
clear and made your essential points there.
I'd like now to call on Mr. Zwolinsky to
make a presentation for the Staff and their review of
this application.
(Pause.)
Mr. Zwolinsky, I understand that in the
course of teh Staff's presentation we'll have to
interrupt the meeting for a protection of proprietary
interests?
MR. ZWOLINSKY: This is my understanding.
MEMBER POWERS: You'll let me know when
that has to take place?
MR. ZWOLINSKY: Yes sir.
MEMBER POWERS: Please.
MR. ZWOLINSKY: Good morning. For those
of you that don't know me, my name is John Zwolinsky.
I'm the Director for the Division of Licensing Project
Management. Staff is here to present its review of
the 20 percent power uprate request for the Clinton
Plant.
I'd like to take a minute to acknowledge
several of our management team that are in attendance
today that are clearly supportive of our staff and are
here to represent that support, beginning with Suzanne
Black, our Deputy Director for the Division of System
Safety and Analysis; along with her, we have Gary
Holahan, the Division Director of the Division; John
Hannon, our Plant Systems Branch Chief; Ted Quay, our
Equipment and Performance and Human Factors Branch
Chief; Singh Bajwa, our Project Director responsible
for power uprates. We also have a number of our
Section Chiefs, our first line supervisors responsible
for assuring a high quality product: Dale Thatcher in
the Equipment Performance Branch; Corney Holden,
Electrical and Instrumentation and Control Systems
Branch; Matt Mitchell from our Materials Branch; Ralph
Caruso, of course, from Reactor Systems, Kamal Manoly
from our Mechanical Branch; Brian Thomas from our
Plant Systems Branch; Louise Lund from our Materials
Branch.
I go through that only to articulate the
sense of importance that we place on assuring that top
notch products are generated and thatwould be in
response to the Committee and any concerns or
questions that may arise.
The Staff made a presentation on this
review to the Subcommittee on thermohydraulic
phenomena on February 14. The Clinton power uprate is
similar to the Duane Arnold, Dresden and Quad Cities
power uprates which were reviewed by the ACRS late
last year. Clinton's application does deviate from
teh approved ELTR1 and 2 methodologies for GE BWRs and
extended power uprates in four areas. These areas are
and we did go through this with teh Subcommittee,
transient analysis, LOCA analysis, stability and large
transient testing. The Staff will discuss these areas
today.
The Staff has conducted thorough reviews
of the Clinton power uprate with the focus being on
safety. The reviews were conducted consistent with
the existing practices which includes the lessons
learned from Maine Yankee. As indicated in the draft
safety evaluation, many areas affected by the power
uprates have been reviewed and evaluated and results
were transmitted in that draft safety evaluation
report. We have additional work to perform in
cleaning the safety evaluation up.
With that, I'd like to get on with the
presetatnions. Our lead project manaber for this
particular facility, Clinton, is John Hopkins. John
will walk up through the presentations as we go
forward and as I said earlier to Dr. Powers, our staff
is available to answer any questions associated with
the presentation or beyond.
MEMBER POWERS: Let me ask you one
question. You said this was similar to the ones we've
looked at before including Quad Cities, Dresden. It
strikes me, in fact, that this is simpler than those.
Clinton just seems like a much easier
power uprate than those other plants have. Is that
your kind of sense or not?
MR. ZWOLINSKY: The amount of time that we
spent, staff time in reviewing especially Quad Cities
and Dresden was quite large compared to the other
applications. We did not spend as much time reviewing
this application. That would be a metric.
MEMBER POWERS: That may or may not be a
metric, but I mean the general amount of changes, the
effort that they have to go to and the changes -- I
mean, I point to just the power uprate is much easier
in this plant than --
MR. ZWOLINSKY: Yes sir. I think as a
general comment, I think we can agree with that.
MEMBER POWERS: Good.
MR. ZWOLINSKY: John.
MR. HOPKINS: Good morning, I'm John
Hopkins, NRR Senior Project Manager assigned to
Clinton. I'll go quickly over the overview. To start
with, I'll be starting and then our next presenter
will be Plant Systems area and then we'll have --
discuss large transient testing and then in the end,
we'll discuss reactor systems and those exceptions.
We will need to close the session for when we discuss
those, even though the handouts are all
nonproprietary. We really have to close it to discuss
it.
To start with, as has been previously
stated, this is a BWR6 Mark III. After the 20 percent
uprate is completed, Clinton will still be just the
third largest BWR6 as far as megawatt thermal power is
considered. Perry will be slightly larger and Grand
Fulf will be about 400 megawatt thermal larger.
The licensee went through many balance of
plant mods to accomplish this uprate. GE14 fuel is
being used and they'll have about a two-thirds core
after they start up from this upcoming refueling
outage which is projected to start April 2nd.
MEMBER POWERS: Maybe you better be clear
by what you mean by two thirds core.
MR. HOPKINS: Two thirds GE14.
MEMBER POWERS: You are not loading jsut
two thirds of the core.
MR. HOPKINS: Okay, I'm sorry. I was
trying to go quickly.
(Laughter.)
This application came in in June of last
year and so it's been a fairly quick review. To
respond to you, Dr. Powers, what you asked John, I
think this has been a simpler review. They already
have had GE14. They already have had MELLLA approved
for this plant. There's no recirculation, new recirc
runback system associated with the plant, so I do
think it's been simpler in those regards.
It is on 18-month cycles. However, this
next cycle is expected to be run approximately 20 to
21 months and I expect to get an application to go to
24-month cycles during that time.
It is nonrisk informed as previous EPUs,
however, risk was looked at and we did not identify
anything that would argue against the uprate.
AmerGen is the licensee and they have
previous experience in operating applications as the
staff does also.
We have one license condition at this
time. It's on a feedwater nozzle cumulative usage
factor. The licensee is still performing analyses of
this and they expect to submit the analyses to us
fairly soon and so we'd condition the license for the
next operating cycle for us to review these analyses
and then find them acceptable.
MEMBER POWERS: John, just for Members'
information. Your cumulative usage factor refers to
the fatigue issue?
MR. HOPKINS: Yes.
MEMBER POWERS: And it is a thermal
fatigue or vibrational fatigue?
MR. HOPKINS: My understanding is it's
thermal fatigue.
They list the four exceptions there. We
will discuss each of the four exceptions during this
presentation. Again, teh first three will be
discussed at the last presenter and that will be
closed.
Right now, unless there are any other
questions, I'm going to briefly discuss flow
acceleration corrosion. This is a question that came
out of the Subcommittee meeting and the qeustion I
received was basically when NRC inspectors look at
flow acceleration corrosion, what understanding do
they have of it and what resources can they tape to
help them?
MEMBER POWERS: I think more so than that
particular question is do the people doing the
inspection of programs at the plant understand from
you in looking at this power uprate quest that there
are certain critical copmonents including teh
scavenger line where flow acceleration corrosion could
be high and the licensee is relying very heavily on
programmatic issues, constraints to assure this
doesn't get out of hand.
MR. HOPKINS: My answer to that is we are,
the staff is developing a power uprate inspection
procedure at this time. It's out to the Regions for
comment. We expect to finalize it in a few months.
One issue that's being considered to be
included in there, specifically FAC. Now all of our
inspections are based on risk importance and mainly
from a nuclear safety perspective. So I don't --
MEMBER POWERS: It seems to me one of the
problems you're going to run into is that we're
talking about flow acceleration corrosion in a line
that probably doesn't rank very high on a risk
analyses, but when you break these lines, they
typically have some pretty substantial conseqeunces,
nevertheless.
So you worry about using risk where you're
talking about damage to the public in tehse kinds of
context.
MR. HOPKINS: I undrestand that. I think
we'd have to get back to you as we develop our uprate
inspection procedure to fully respond.
MR. ZWOLINSKY: I think your comment is a
fair comment. Yesterday, at the Regulatory
Information Conference, Jack Robe was our Division
Director for Reactor Safety and Region 3 was
presenting the inspection program that was conducted
at Quad Cities, Dresden and Duane Arnold and he did
not get to that level of specifics, but they did
implement a specific inspection program, targeted to
power uprate, seeking key vulnerability that they felt
had been identified not just in the application, but
in the safety evaluation.
MEMBER POWERS: I think that's what needs
to -- there needs just to be some communication here.
MR. ZWOLINSKY: Okay, the left hand and
right hand were clearly communicating and I think that
was one of the major points he was making.
MEMBER POWERS: Very good.
MR. ZWOLINSKY: But as John alluded to, we
are developing the temporary instruction for more
uniformed inspection across the country.
MEMBER POWERS: If you happen to have the
slides from his presentation, I'd enjoy seeing them.
MR. ZWOLINSKY: We can forard those to
you.
MEMBER POWERS: Okay.
MR. HOPKINS: Okay, at this time I'd like
to introduce Richard Lobel who is from our Plant
Systems Area. And he will talk about another question
from the Subcommittee on spent fuel pool temperature
distribution and briefly discuss contributory
containment analyses that we performed on Clinton.
MR. LOBEL: Good morning. I was giving a
presentation on the plant system's areas of review for
the Subcommittee and a question came up about the
temperature distribution, the water and the spent fuel
pool and I said that I believe there had been studies
done on that. Let me just go over it briefly.
The heated water in the spent fuel pool is
collected around the periphery of the pool and
circulated through heat exchangers and then discharged
at the bottom of the pool to enhance circulation. The
power uprate doesn't change the design aspects of the
spent fuel pool, cooling, the circulation mixing
patterns and the operation of the spent fuel pool.
Based on staff experience, it's not power uprate, but
spent fuel pool reracking that results in the greatest
increases in spent fuel pool temperatures and we have
reviewed many rerack applications. As part of that,
the staff reviews the thermal hydraulic analyses
including the maximum water temperatures with and
without water circulation, without forced circulation.
In one rerack review that was done a while ago, the
staff performed extensive two and three-dimensional
calculations of the water distribution in the spent
fuel pool, compared the calculations with the
licensee's calculations and concluded that the
license's calculations were conservative.
As we discussed with the Subcommittee, the
spent fuel pool water and the fuel temperature
increases aren't a concern for the Clinton power
uprate. Their analyses show that they're below the
spent fuel pool limits.
I hope that answered the question.
MEMBER KRESS: Those limits are set based
on the concrete --
MR. LOBEL: Right. The limits are really
separate from the question of the temperature
distribution. There's a limit because of the material
that's used in the purification system and there's a
limit of 150 degrees on the concrete. And Clinton was
well below both of those limits.
MEMBER KRESS: Was the temperature
distribution significantly different from what it is
normally?
MR. LOBEL: Well, it depends on the
loading pattern and the density of the loading pattern
and that's why the reracking really has more of an
effect than the power uprate. Typically, in a rerack
you're moving the fuel closer together and higher
energy density into the same amount of water.
MR. ZWOLINSKY: Dr. Kress, the biggest
issue we've identified is length of time that the
licensee retains the fuel in the core and it's initial
configuration from shutdown. The longer it cools in
the reactor vessel and then transfers over to the
pool, the smaller effect it has on the pool's
temperature.
MR. LOBEL: Also at the Subcommittee
meeting I mentioned that we were doing confirmatory
analysis for the containment calculations and at that
time I wasn't sure whether we'd be done in time for
this meeting, but it turns out that the calculations
are completed and if you'd like, we can talk about
that a little.
MEMBER KRESS: Yes, I think we would like
to.
MR. LOBEL: Well, let me introduce Edward
Throm from Plant Systems Branch who did the
calculations and he can discuss it.
MEMBER POWERS: Well, maybe you better get
the speaker on with a mobile microphone.
(Pause.)
MR. THROM: Am I on? Okay, real quick, my
name is Ed Throm. I'm with the Plant Systems Branch
and what we attempted to do in a very short time was
do some confirmatory calculations for the Clinton
extended power uprate.
What we wanted to do was look at the
contained two which is the staff containment code and
compare it to the M3CPT and SUPERHEX results that GE
typically calculates.
We started off with an existing Grand Gulf
Mark III deck and modified it to look like Clinton.
This modification is dry well/wet well. Volumes,
initial conditions to reprsent th eplant. We used the
mass and energy releases provided by the licensee
directly and this leads to a little bit of a
discrepancy and one of the results of that I will show
you.
This particular time we've done three
calculations, two short term for the recirculation
line break and then the steam line break and we also
did the recirc line long-term cooling temperature
response calculation. We couldn't do the shutdown
calculation in a short time because that would have
required additional model changes that I didn't have
time to do. And basically, by looking at the
qualitative comparison for the studies we've done, we
believe that our conclusion that the licensee's
analyses are acceptable for the extended power uprate
is a true statement.
The model is fairly simple. It's
basically got three models of dry well with the
annulus region that connects to the wet well and for
this particular design it has the three vent paths.
I'm just going to put up two results
because of time. This is the short term recirculation
line break. This is a comparison of the contained
results to the M3CPT. As you can see, M3CPT is
calculating a little higher pressure than the staff's
calculation and overall the dry well temperature
response is very consistent with the licensee's
calculation.
And the long term break, this is one of
the areas where you learn as an analysis that you may
have missed a piece of information that was important
if you were trying to do it, an audit calculation and
what's happening here is by using the licensee's
provided mass and energies, we got a very coarse set
of data and the data tends to show more of a steam
release, high energy release over the initial portion
of the transient. That's why we're predicting this
higher initial pressure response and a slightly higher
temperature in the suppression pool.
This is very evident from that little
spike there, where we have three points, one at a
liquid, one at a steam, and one at a liquid. So we've
just done a linear interpolation so that's what the
offset is doing for these types of calculations. But
again, qualitatively, we don't see anything between
the two codes that suggest that the analysis method
that's been approved and accepted by GE, by the staff
or GE plants is any different for Clinton than it was
for the Duane Arnold which basically the staff did a
similar evaluation for Duane Arnold. That's pretty
much what I have to present.
MEMBER KRESS: Thank you.
MR. ZWOLINSKY: Thank you, Ed.
MR. THROM: Okay, sure.
MEMBER POWERS: Thank you.
MR. THROM: Sure.
MR. ZWOLINSKY: Okay, our next presenter
will be Bob Pettis and he'll talk about the exception
from ELTR1 and 2 to not perform large transient
testing.
MR. PETTIS: Good morning. My name is Bob
Pettis and I'm with the Quality and Maintenance
Section within the Division of Inspection Program
Management.
Our review of the Clinton application
focused on the testing section of the application with
some specific attention to the exception for not
performing a large transient test. The Clinton EPU
tst program follows ELTR1 which is basically
delineated in Appendix L2.
As discussed previous by the Applicant,
Clinton will perform a limited subset of original
start up tests to demonstrate capability of the plant
systems to perform as designed through the EPU power
extension. Routine measurements are taken for reactor
and system pressures, flows and vibrations, up through
EPU conditions and main steam and feedwater systems
will be monitored for vibration.
The exception to ELTR1 is in the area of
the mainsteam valve closure and generator load reject
tests. This exception was also previously approved
for the Dresden and Quad Cities EPUs.
The staff felt that the exception to the
ELTR1 is acceptable for several reasons. First
reason, GE had stated the constant reactor dome
pressure simplifies the analyses and plant changes to
achieve EPU conditions.
Text spec surveillance testing will
confirm the performance capability of the compnents
challenged by large transients. Another point is that
CPS is not installing any new safety-related systems,
features or significant additional components as a
result of achieving EPU. There are some balance of
plant modifications that were discussed previously.
An analysis was also performed by the
licensee in coordination with GE that reviewed some of
the compnents that would be challenged by large
transient testing. Some of those components included
MSIVs, safety relief valves, turbine stop valves and
turbine control and bypass valves. They also reviewed
main steam line geometry and control rod insertion
times.
The CPS test program also will monitor
important plant parameters during power ascention,
operating preassures and flows, temperatures,
vibration and closure times for MSIV turbine stop adn
control valves. Operating history and experience at
other BWRs has also been recognized. There's been a
slide that was presented yesterday that was more
extensive than this, but the KKL plants or the KKL
plant is operating at 116 percent; the KKM at 117;
Monticello at 106; and Hatch around 113. And also,
the Dresden, a plant as well.
MEMBER POWERS: Have any of these plants
performed the equivalent of a large transient test?
MR. PETTIS: To our knowledge, the only
plant that has was the KKL plant in Sweden and from
what we have reviewed, at least in our section, those
results appear to correlate very well with the
analytical models that would ahve predicted the
resonse to the tarnsients.
MEMBER POWERS: What I'm struggling with
is what you mean by operating history and experience
at other uprated BWRs. If say one has performed the
test, that's not a whole lot of experience to judge
them, is there?
MR. PETTIS: Well, the way that should be
looked at is KKL did perform the large transients and
there is information that correlates well for the KKL
plant.
MEMBER POWERS: Okay, so if I were to
rewrite the line, it would say we have one plant
that's done this test and it seemed to match the code
predictions and so we'll live with code predictions?
MR. PETTIS: The expectation level, I
think, has been achieved with KKL. With respect to
the domestic plants, the only experience tehre that
we're trying to demonstrate is the fact that they have
undergone EPU conditions. They are operating at EPU
or near EPU conditions with no anomalies.
MEMBER SHACK: They had a transient at
Hatch, right?
MR. PETTIS: Yes, that was in 1999. It
should also be noted that Clinton is not making any
modifications to the reactor recirc runback system
which was an area of concern for one of the ACRS
Members previously for Dresden and Quad Cities.
Large transietn testing is also not needed
for code validation. I believe that's probably the
ODEN code, but I'll let our reactor systems folks
discuss that.
And also, the incident that did or the
event that happened at Hatch in 1999 where they
experienced a load reject from 98 percent power and
KKL had a turbine trip at 113 and a low generator
reject at 104, and both of those events followed again
code predictions.
Our conclusion is taht conducting large
transient tests would not provide significant new
information regarding transient modeling and component
performance and that the Clinton EPU test program is
acceptable.
Thank you.
MR. ZWOLINSKY: Good job. Okay, I would
ask that the session be closed now because we'll be
doing our reactor systems.
(Whereupon, the proceedings went
immediately into Closed Session.)
. MR. HOPKINS: Again, this concludes our
presentation. The Staff finds that 20 percent power
uprate for Clinton can be accepted and approved. We'd
request a letter from the licensee discussing our
presentation.
MEMBER POWERS: You discuss, you request
a letter from the licensee discussing yoru
presentation? I'm sure they'd be happy to critique
you.
(Laughter.)
MR. HOPKINS: They probably will.
CHAIRMAN APOSTOLAKIS: Or disagree with
you.
MR. HOPKINS: You got me. A letter from
the Committee. Thank you.
MR. ZWOLINSKY: I do thank the Committee
for your time and while we may have rushed through a
few of our presentations, we were trying to push a
number of our Staff before the Committee to give an
indication of some of the areas that we focused on.
With that, I feel the staff has completed
their presentations.
MEMBER POWERS: Thank you, John. I'm
going to pass on a comment from teh Subcommittee and
that is that the Subcommittee found that this safety
evlauation report was among the most readable in the
power uprates that they had. I know that's been an
area of concern for you and the Subcommittee detected
real progress in improving the readability of those
reports.
MR. ZWOLINSKY: Thank you.
MEMBER POWERS: Mr. Chairman, I think at
thi spoint I think we're done with this session.
CHAIRMAN APOSTOLAKIS: Thank you, Dr.
Powers. I also thank the representatives from the
licensee and the Staff for their presentations and we
will recess until 10 minutes past 11.
(Off the record.)
CHAIRMAN APOSTOLAKIS: Okay, we are back
in session. The next agenda item is the proposed NEI-
00-04 report, Option 2 Implementation Guideline for
Risk-informing the Special Treatment Requirements of
10 CFR Part 50.
We sent to the Staff and NEI a set of
questions last January and we had an opportunity at
the Subcommittee meeting in February, February 22nd to
discuss these questions and the rsponses from NEI.
The Staff has forwarded the questions to NEI. So --
well, I must also point out as it will be pointed out
later, this Committee has also written two reports to
the Commission, one dated October 12, 1999 and the
other February 11, 2000 on importance measures and
related matters.
So today, we will -- the Staff has
requested a letter. Last time, at least, at the
Subcommittee meeting, you said that you would like to
see a letter with the Committee's views. I assume you
still would like to have a letter from us. You can
change your midn, if you wish.
MR. REED: Yes, as you'll see in the
slides, currently, we're asking for a letter. If
there are big issues, show stopper issues that the
Committee has that we really need to address in order
to go forward and get your concurrence on the proposal
--
CHAIRMAN APOSTOLAKIS: Okay, so why don't
we start then and see how well -- and then we'll
discuss the issue.
MR. GILLESPIE: Well, knowing that the
Committee was really going to focus today on
categorization process which is kind of a cornerstone
of the rule, let me kind of give you just in a
nutshell the status of kind of where we stand.
Frank Gillespie, NRR. The rule has been
delayed. So the nature of the letter and the nature
of this meeting might be slightly different, so let me
kind of give you the context of that. Two things
we're still wrestling with that within the staff. One
is having a categorization document, guidance document
and working with NEI which right now is kind of a work
in progress. We sent 17 pages and comments to them
and they're digesting them and so it's going to be
another iteration.
And the importance of the categorization
document going with the rule and being substantially
finished or at least let's say at 80 and 85 percent,
so the people can understand what the cost of doing
this rule is, having a categorization process and
committees and other things. So that was considered
an important element. Our original schedule for
April, even my optimistic view, I said well, let's
send up what they have with our 17 pages of comments
and that wasn't going to be the right thing to do. It
wasn't going to be -- we would not have worked out
back adn forth with all the stakeholders the right
kind of document, so we've delayed the rule from April
to July.
Now, if we can beat July, our criteria
really is less to date and is more having both the
rule and a guidance document that goes to set. And
taht's our real criteria. And the guidance document
has to be close enough that we have a high possibility
of general acceptance that it's rational.
So the guidance document is still kind of
a work in progress. Besides the guidance document and
it might be a subset and I know you're not here to
discuss this, but in my mind they're not unrelated, is
the question under special treatment, if you notice
the last draft words, the only real special treatment
that still is left in for RISC-3 components is
50.55(a). And what the Staff is still grappling with
is how to achieve -- how do we write achieving what
50.55(a) is trying to achieve, even for low risk
components. And do you have to continue to have
50.55(a) apply or is there a way to deal with that
wthin a more performance-oriented approach, for
example, within cateogroization. And I"ll give you
one example and then turn it over to Tim of -- this is
kind of an on-going discussion in the staff so I can't
even say ther'es a position on it. There are several
positions.
It's not clear to me right now that our
current rule and the current guidance says explicitly
that you have to consider known degradation mechanisms
as part of your consideration in what you're going to
do relative to a RISC-3 component. Yet, does not
50.55(a) try to continue a certain level of assurance
for a compnent, even if you're ina risk-informed space
which fundamentally is saying we're maintaining our
input reliability or the input to our decision process
is being sustained.
And so what kind of what we're grappling
with is the only way to say to do that for some
mechanical components and pressure things and passive
things is by dictating 50.55(a) or is there another
way of dealing with it in waht the various committees
would be considering within the guidance and within
categorization.
I've taken my best shot at waht we're
grappling with and we've got the staff here. If you
want to discuss that later as a point, Tom Scarboro
and John Fayer are here, Gareth and Mike Cheof, so
we're still grappling with that one point. And it's
an important point and we haven't come up with the
exact way to deal with it and make that decision yet.
There is a meeting next week where we're
trying, going to try to make a definitive decision and
the Staffs are kind of working on different points of
view and how do you approach that problem. Not a
problem, but how do you say what you want to say and
get what you want to get in th emost performance-
oriented risk-informed way?
So, that's in my mind those two decisions
are not mutually exclusive. And the other comment we
did get back to NEI was that these committees, the IDP
Committee, we haven't necessarily articualted how they
should establish criteria or what the higher level
criteria they should be considering is. And that's a
piece that we need some time to go back and forth on
to discuss because clearly all the components are not
considered in the PRA and the mathematical model and
in particular, how do you deal with the passive
components. Again, not disconnected from 50.55(a) and
is that in or out, particularly for passive pressure
kind of things.
So that's waht we're wrestling with and
with that, I'm going to turn it over to Tim. I talked
to George, yesterday, and I've asked Tim to go through
the presentation, the formal presentation as quickly
as possible, so taht maybe we can get your advice on
these questions we're grappling with and we got the
staff here that maybe we can have some interactive
discussion on it. Because we don't know the right
answer.
CHAIRMAN APOSTOLAKIS: Good idea.
MR. GILLESPIE: Thanks, George.
MR. REED: This is Tim Reed from NRR and
I have with me also over at the side table Mike Cheof
from Gareth Parry to assist here today. The focus
here really is focusing on the categorization pieces,
Option 2 and looking at what are the remaining key
issues that we need to resolve so that ultimately we
can get this Committee's endorsement going forward.
So that's what I'll be trying to focus this discussion
towards.
And this won't look -- this should look
very familiar to most of you from the Subcommittee.
This tries to give a high level overall status of
where we stand on categorization. We recently sent
our third round of comments to NEI back in early
February and those comments, as Frank mentioned, like
14 pages or whatever, reprsent what the Staff belives
are the key issues that need to be resolved for us to
reach agreement on the categorization guideline NEI-
00-04. They reflect both the Palo Verde activity
feedback that we've had to date. We've observed three
pilots. In effect, next week is the last pilot, Palo
Verde, and we'll be observingt that one two. And they
also reflect the staff's review of draft revision B of
NEI-00-4. As you're aware, that document is goig to
be revised and it's a work in progress and ultimately
will become, I believe, draft revision C.
There are some major issues and I've just
hit a few here and there are several other issues in
this 14 pages that, in fact, I agree with some of the
comments that ACRS brought up in the Subcommittee, but
hitting some of the bigger ones here, the issue of
long-term containment integrity and how you consider
that within the element of defense-in-depth and how
the IDP considers taht. That's an issue, that's a
comment, you'll see in our comments back to NEI. It's
been an on-going issue.
The element of the IDP, the IDP guidance
and whether that's sufficiently structured, I think it
probably needs a little bit more structure there.
That's the nature of our comments. I think the
Committee has that. In fact, I think NEI would agree
to that to some extent. That's a feedback coming back
to the pilot activities also. And then the whole
issue of the PRA quality, the use of the PRA review
process, how we roll tht in and how the staff develops
review guidance to judge whether, in fact, a PRA has
sufficinet to support the categorization process. So
that's a very key issue here.
All these issues, I believe, from a
technical standpoint, I think the staff belives, can
be resolved. So there's nothing here that can't
technically be resolved. Taht doesn't mean that we
have signfiicant work, but nothing appears to be a
technical roadblock at thi spoint.
Of course, we have to come back to this
Committee and get -- the Committee needs to see more
of a final product, obviously. And so we need to come
back for the proposed rule concurrent stage and as
Frank mentioned, our schedule at this point is to try
to get this to the Commission by the end of July. And
so that puts us actually in a very, very tight,
difficutl schedule to try to get this Committee, get
a letter from this Committee to support that schedule
to get this whole package to the Commission by the end
of July.
And I don't know if NEI is going to speak
today or not, but NEI is in a very tight situation
too. They have our comments. They'll hvae the IDP
next week. They have to take the feedback from that.
Roll it back into NEI-00-04, work through their side
to get agreement and then send it to us and then we
need to look at that and with the draft reg. guide,
get it to the Committee and all taht before July. So
you can see there's an awful lot that has to happen
here in the schedule.
So that kind of gives you the high view,
the important pieces of categorization and wehre they
fit into the proposed rule schedule.
Now these are some of the high points and
I'm not going to success that this is the Committee's
views. This is what the Staff heard, waht we think
are important and I'm sure the Committee would think
there are other issues that were discussed that are
more important to indiviudal members. But I'm just
going to hit a few of the high oints here. But an
overriding theme, I think, we heard numerous times
during the Subcommittee meeting was that sevearl
members mentioned or expressed concerns with the
underlying basis that supports the NEI-00-04 document
and whether or not that basis is really there, or the
document or hte studies there -- is there something
you could point to that says yeah, we all agree. I
think most people, in fact, agreed, taht they thought
it was conservative, but that's more of a subjective
judgmetn and I think it was the Committee's view that
we probably ened to have more in place as far as
something to point to, in fact, that demonstrates more
clearly, in fact, that a lot of these assumptions that
we're making are, in fact, robust and lead to a robust
categorization.
These were some of the isseus that we took
away, the staff took away, you see listed there, what
should be the -- what kind of failure rates and what
should be the increase in failure rates. AS you're
well aware, the numbers are being bantered about
between 3 and 5 increase in factor of failure rates
for the sensitivities. When you roll this up into the
CDF and LERF sensitivities, as you're well aware,
South Texas uses a 10. What hsould be the
sensitivity? Wht should the nujber be? Should it be
a distribution? There's lots of discussion there.
The fuseell-Vesely and the risk
acheivement worth screening criteria or guideline
values, whta they should be? I think there was even
mention of distributions and how you might want to
handle that.
There was consideration of rolling up and
addressing uncertainties in the whole model, whether
that should be done or not. And the whole issue of
common cost failure was addressed pretty extensively
also. Adn also in combination with RAW, as a matter
of fact, at the Subcommittee.
And as I mentioned, the Committee's --
I'll let the Committee speak for itself, but I think
some of the Committee would like to see some of this
more documented out there for everybody so everybody
could -- and I believe the ACRS mentioned another
important comment. I think it agrees with the staff
that perhaps the 00-4 guidance, the NEI-00-04 guidance
should have a little more structure, a little more
guidance to the IDP. I think we think that too. I
think it would help to make more effective, efficient,
repeatable decisions from IDPs, wherever they occur.
I think we all agree with that. I think even Mike
even agrees with that to some extgent also.
CHAIRMAN APOSTOLAKIS: After the expert
panel makes its determination in whatever way, we're
going to work on the process, but let's say at teh
very end, they ahve categorized now system structures
and components, how do you know that they're right?
You're putting all your trust in the process or are
there other mechanisms at the end for you to gain some
confidence that yes, what they have done is, in ract,
reasonable?
Remember now, we're talking about
thousands of system structures and components. IN
other words, are we approving a process and then
telling the licensee go ahead and impleent it and as
long as you follow the process, we're happy.
MR. REED: I'll answer the easy part. It
is a process approval. To the extent of the
confidence in the process and whether the validity of
that process is maintianed over time, I'll answer
another easy piece of it. That's a function, I think,
of monitoring and bringing information, operational
experience back into the process and ensuring that
you're assumptions, the cateogrization assumtoins are
maintained valid over time.
I'll let Gareth and Mike, if you want to
say something intelligent than that.
MR. CHEOF: Basically, I think after the
IDP categorizes all the SSCs, the PRA is supposed to
qualify the change in risk from all the SSCs that are
put in RISC-3. Adn this change in risk is supposed to
be shown to be small, according to the guidelines we
provide in Reg. Guide 1174, but I don't think that's
the question you're asking. I think the question is
--
CHAIRMAN APOSTOLAKIS: That's part of the
process.
MR. CHEOF: That's right, that's part of
the process.
CHAIRMAN APOSTOLAKIS: The process has
been completed and they go ahead an they categorize
things.
How do yo uknow that what they are doing
is reasonable?
MR. CHEOF: I think --
CHAIRMAN APOSTOLAKIS: Not that they are
trying maliciously to not implement it, but we're
talking about thousands of SSC's here and peoplea re
people. They make mistakes and so on.
MR. PARRY: I think Tim was right. In a
part of this follow-up is the monitoring of the SSCs
and we do have an update requirment for the risk
analysis and the process itself that takes into
account operating experience. I think the one problem
we see is that in the monitoring for the RISC-3 SSCs
might also decrease which might not provide you with
enough feedback that you might need. So it's
something we need to work on.
CHAIRMAN APOSTOLAKIS: Operating
experience in these things is probably admitted by
you, I would say, because a lot of these systems and
components, I mean we are really intersted in their
performance during an accident, right? So I mean you
would probably be concerned about the proper
categorization of particular things and you would not
wait until something happens to see whether things
work. Are you planning to have sort of an audit on a
random basis or some other basis to get that feeling
that things are being implemented the way they're
supposed to and the results are reasonsable?
I've heard -- Frank told me the otehr day
that the MSIVs wer according to some people were
miscategorized at teh South Texas projec. I'd like to
know a littl emore about it, why they feel that way.
That's a major component. I mean it's something that
we can look at.
I don't know, Gareth, if you don't have an
answer now, that's fine, but that's osmething that
concerns some Members of this Committee. Waht is
happening? Are we turning over the responsibilities
now to the licensees?
MR. PARRY: There's one other progarm we
have on place, the reactor oversight program that I
think also has a role in finding degradation for
components.
CHAIRMAN APOSTOLAKIS: Good.
MR. PARRY: It obviously is not foolproof,
but I think it is -- if there is a finding, then it
has to pushed through the SDP and that can -- that may
reveal another --
CHAIRMAN APOSTOLAKIS: The actual
oversight process it not really looking at the
categorization.
MR. PARRY: No, no, it's not looking at
the catgegorization. It would be looking at
conditions that might arise because, for example, a
lack of treatment in certain areas.
MR. GILLESPIE: This is hard to answer,
but the answer is we haven't figured it out yet and I
think I want to be careful. The overight program is
us overseeing a licensee carrying out its
responsibilities.
CHAIRMAN APOSTOLAKIS: Right.
MR. GILLESPIE: And we cannot build in a
dependency on our actions for the safety of the
facility. And one of the things that we're gropiong
with just a little bit this morning and in an earlier
meeting was how do we deal with the eact question
you're asking. And have we actually, have we
necessarliy given either the right commetns in the
area of reinforcing -- you're attesting is the quality
of the decision. What's our confidence when we made
the decision that RISC-3 is really RISC-3? I don't
have an ansswer for you, but it was a question that we
had on the table and we were kicking it around this
morning.
The other question is how do you know that
the input criteria tha tyou made you rdecision are
continuing to be sustained?
For example, if you did a sensitivity
study and you did varythings by a factor of 4, how do
you know that all those RISC-3 components were still
within that envelope? And then we got into a
discussion of the word degradation mechanisms need to
be considered and right now it's not clear that within
our guidance and clearly not within our rule that the
word considered known degradation mechanisms up front
is actually any of our rule or the guidance. And I'm
going to -- Tom Scarboro has got an example he gives
of - and I've got to give him credit, but I got it
third hand, so if I don't say it right, Tom, jump in,
is if you have valves, a lot of samll valves that have
grease on the stems and they are in a steam tunnel and
you're going to get hardening of the grease, that's an
environmental condition that needs to be considered in
all aspects of what we now have in the rule where
we've got monitoring and all those different
paragraphs. And yet we haven't necessarily written in
the rules the idea of considering environmetnal
conditions. So we're grappling with that right now.
And I'm going to suggest that that's my connection to
50.55(a) which is a mechanism right now within our
requirements that tries to grapple with assuring that
continued reliability. So what we're trying to do is
get to the roots of how do we say that. We don't have
it righ tnow.
CHAIRMAN APOSTOLAKIS: It's something that
woudl be --
MR. GILLESPIE: WE've got it on the table.
I'm not sure qyite how to do it, and do it without
superimposing just a bevy of QA requirements on the
PRA and decision process also. I mean in my mind I've
got to keep, we've got to keep this as a staff in
perspective. We are delaing with low risk components
to the greatest degree possible. We hope we have a
credible process, but how do you check that your
process was carried out the way you thought it hsould
be.
MEMBER SHACK: But in yoru particular
example, that's an active component. Wouldn't that be
covered under the maintenance rule?
MR. GILLESPIE: The miantenance rule is
one of the exemptions within 50.69. So that's a
special treatment rule that this RISC-3 copmonent
would be exempted from. So it's consideration in
advance of that to meet the other aspects of the rule,
good enough, or do you need a requirement that says go
inspect it every eyar and we're caught between how do
we get at the essence of the deterministic go and
inspect it every year and the right onctext of kind of
a risk-informed, performance-based and we're wrestling
with it. Again, I don't know the answer.
CHAIRMAN APOSTOLAKIS: I undrsatnd.
MR. GILLESPIE: I think we've got the
question.
MR. REED: Saying it a little bit
different way, when you look at whether you feel
comfortable with this process going forward as an NRC,
a regulator, you look well, what, whwere were the
requirements reduced, so that focuses you down on
RISC-3. That's where we removed ruqirements.
Everything else, we're keeping requirments and putting
more on. So I look down at Box 3 first and say what
could go wrong there. Well, what could go wrong there
is obviously they could degrade over time. YOu could
lose either functionality or you could go outside the
bounds of your sensitivity analysis. How do you fix
that? Well, then you look at what's the feedback
mechanism. So you can see how our logic works to try
to get to the exact sisue you just brought up and
maybe we can do it in a performance based manner
that's consistent with the principles of Option 2. I
don't know. We're wrestling with it.
MEMBER BONACA: On the issue that you
raised regarding how do you know, okay, one comfort we
got from SDP was that they claimed that for each
component that was in a certain category there was a
full document description of how they got to that
particular based. So therefore, one could envision
that you could have an audit to understand how it was
applied and there was -- I didn't understand that this
would be a requirement under the NEI document.
Traceability wasn't clear there to me.
MR. GILLESPIE: Yes, and that's part of
what I said. The guidance document is a work in
progress, and in fact our thoughts are actually
evolving even right now as people are putting these
issues, like the valve in the steam tunnel, or there's
RISC-3 and there's RISC-3. Which 3 is the spectrum,
and you draw a threshold. But as you get closer to
the top end of that threshold, the sensitivity study
actually takes on more importance as to whether you're
within the threshold or not.
How do you that in rules space and in
guidance space without overburdening a system? It's
a compromise to some degree. We're wresting with
that, and that's why I think we're going to actually
have some more interaction with NEI on it, because we
want them wrestling with us with some of these same
questions, and we might not have articulated them the
same way in the letter we sent them. Because as we're
talking to you and other people, our thought processes
are saying how can we get along with this? How can we
kind of -- we're focusing in on these specific kind of
questions that we might not have -- our thinking might
not have been completely clear even three months ago.
CHAIRMAN APOSTOLAKIS: Anything else on
this topic? The second bullet there, "Some ACRS
Subcommittee members would like studies to perform."
I think it's important to make it clear what kind of
studies and what kind of analysis we're talking about
here. I think there are two categories. One is
genetic type of studies. And, again, we're not
talking about multi-year kind of things. I mean
experienced people can do these things in a short
period of time. But generic kinds of studies that can
support various approximations or assumptions that are
being made routinely, and NEI 004 articulates those,
what is being done. I don't expect that if one does
these studies, the basic approach will be really upset
too much.
But it seems to me that we ought to be
doing things that we understood very well. For
example, this issue of uncertainty in importance
measures. Frankly, I don't think it's going to be a
big issue, but I would hate to have the only evidence
that I have come from a professor and a graduate
student somewhere who did something six years ago.
Can we make sure that we understand that this is not
an issue? And how long will it take to do that? I
don't think it's going to be a long study, and this is
kind of generic. It's not something that every
licensee will have to do later.
Now, speaking of what the licensees will
have to do, I don't understand why -- I mean, I agree
with you that the sub-bullet there, parametric as well
as model uncertainties, the real issue is really model
uncertainty here, not the parametric. And yet we're
mixing the two. We know -- I mean if there is one
thing we know now is how to handle parametric
uncertainty. And it's easy to do, to propagate. But
still the document ask that this be done. It plays
with sensitivity studies that it's not clear now which
part of the sensitivity study addresses the model
uncertainty, which part addresses the range of the
parameters of lambda and so on.
Some of this stuff, it seems to me, can
become much cleaner and more convincing. And, again,
I don't -- I think it's going to be of great value to
the independent panel. It really will be. In fact,
in our letter of three years ago, we said that one
should do studies of this type, and then the insights
that will be gained -- let's see how we put it. Now,
at that time we were looking at Appendix T, but "The
guidance to be provided in the proposed Appendix T for
the Expert Panel should include insights gained from
the implementation of Recommendation 4 above, which
was really doing all these studies that I just
mentioned."
And that I see as an essential part to
making sure that the whole process is on solid ground.
In other words, just because somebody's an expert on
plant systems I'm not sure he's really qualified to
use fussell vesely and RAW without any other
information to categorize things. I mean, we need a
combination of types of expertise here to come up with
a reasonable product for the same reason that I
wouldn't trust a guy who understands RAW and fussell
vesely and never been to a plant to do this
categorization. Should you have some --
MEMBER POWERS: But I mean isn't that
fairly a hypothetical thing? I mean who is going to
be involved in a categorization process that only has
fussell vesely and RAW data only?
CHAIRMAN APOSTOLAKIS: We should take PRA
guy who has never really been to a plant? I mean, I
don't think would be a proper member, but that's not
my point here. My point is that when you're
presenting the results from the PRA using RAW and
fussell vesely, to what extent should you train the
Expert Panel, or educate them, as to what these
measures mean, limitations perhaps, and so on. It's
like the expert opinion thing that was done for the --
MEMBER POWERS: Well, I think that's the
point is that it's the limitations on these measures.
I mean nobody's going to use them with no other
information. You simply can't. I mean, it's just not
physically possible to close your mind to other
information.
CHAIRMAN APOSTOLAKIS: No, no. That's not
what I meant. But I mean --
MEMBER POWERS: No. I think your point
that the limitations of these things are not well
appreciated.
CHAIRMAN APOSTOLAKIS: Right.
MEMBER POWERS: And they are severely
limited, and it's the one of time variation that's the
principal limitation, to my mind.
MR. PARRY: George, I'm not sure that the
members of the IDP necessarily to be looking at the
RAW and fussell vesely values themselves. I think
that they would be looking at the overall results of
the categorization that would have been performed PRA
analysts that would have taken into account all these
uncertainties about fussell vesely and RAW. I mean
it's not clear to me that they need -- that the IDP
needs to actually understand what a RAW is.
CHAIRMAN APOSTOLAKIS: If I were them, I
would like to understand.
MR. PARRY: But that's not the only thing
that goes into the categorization that they're going
to be presented with. There's a whole slew of --
CHAIRMAN APOSTOLAKIS: But it's a major
input, though. It's a major input.
MR. PARRY: It's an input. I'm not sure
it's that major of an input. It's the starting point
for the categorization.
MEMBER POWERS: Well, I mean even if it's
just that, even if it's just the starting point for
the categorization, then I think it's important to
understand the limitations on that starting point.
MR. PARRY: Yes. And I think the process
recognizes the limitations on the starting point and
compensates for it by requesting some other additional
studies. In particular, it also requests the
valuation of delta CDF and delta LERF.
MEMBER POWERS: Well, I think the concern
is that the choice of those augmenting studies that
you speak of here, the additional information,
requires some substantial understanding of what the
limitations of the fussell vesely and RAW numbers are.
MR. PARRY: And shouldn't that be a part
of the process NE 00-04 that recognizes those
limitations and designs the process to compensate for
them?
MEMBER POWERS: Yes.
MR. PARRY: And that's what it tries to
do.
CHAIRMAN APOSTOLAKIS: But the point is
that the Expert Panel also should become aware of
these at some level anyway. You don't want them to
become expert but at some level.
MR. PARRY: I think that might involve a
quite considerable amount of PRA training if go
through all that.
CHAIRMAN APOSTOLAKIS: Well, I don't know
about that. You know, you can -- I'm not sure that's
the case.
MR. CHEOF: I guess let me add something.
I guess in the current NEI 00-04 guidance and in the
IDP that we have observed so far the IDP members have
been pre-trained in a one-day PRA training as to the
results they are getting and what they mean, and I
guess, like you all say, the limitations of the PRA.
I am not sure that the training includes things like
the uncertainties in terms of parameters and models.
And perhaps the models might get -- the modeling
uncertainty might get mentioned in the fact that these
initiators are not modeled and we will account for
these initiatives this other way using this flow
chart, for example, in NEI-00-04. But there is
training for the IDP members, for all IDP members, and
that training does include importance measures.
CHAIRMAN APOSTOLAKIS: I don't remember
that, but if there is, fine.
Now, you guys agree with the NEI, at least
the version we saw, that it's good enough to use so-
called point values and then do sensitivity analysis?
MR. PARRY: In general, yes.
CHAIRMAN APOSTOLAKIS: Even when you
calculate delta CDF, which is a very small number?
MR. PARRY: Well, going back to --
CHAIRMAN APOSTOLAKIS: And why do it that
way? What do you say? Is it a big deal to do it
right?
MR. PARRY: Actually, for some people it
may be. But then that's just a technical issue on the
codes.
CHAIRMAN APOSTOLAKIS: For some people.
MR. PARRY: It depends on the
quantification code you have, George. Some of them
are not set up to do the proper state-of-knowledge
correlation on uncertainties, and you have to do that.
CHAIRMAN APOSTOLAKIS: Maybe they
shouldn't then be following Option 2.
MR. PARRY: No, no, no, no. Because what
it also -- we address this issue also in Reg Guide
1.174. We had the same issue. And what we said there
was you should do it by propagating uncertainties to
get the correct mean, but you could also -- if you
could demonstrate, by reviewing the cutsets, that in
fact this impact of the state-of-knowledge correlation
was not significant, then that would be another way of
demonstrating that you got close enough to the mean
value. Because that's the only thing that changes the
mean value from using input mean value as a point
estimate in all the calculations.
CHAIRMAN APOSTOLAKIS: No, no. It's not
the only thing.
MR. PARRY: I believe it is.
CHAIRMAN APOSTOLAKIS: When you propagate
point values, don't you need the values?
MR. PARRY: No.
CHAIRMAN APOSTOLAKIS: Yes.
MR. PARRY: No. Not with cutsets, you
don't. Only if you have correlated variables.
CHAIRMAN APOSTOLAKIS: The early PRAs were
done that way. I did it by hand, and you have to use
the variance too. The expressions for the mean
involve the variance.
MR. PARRY: If you are multiplying
together basic events that depend on the same
parameter for their probabilities, yes, you have to
propagate the variance, but otherwise the mean
translates. If they're totally independent variables,
it doesn't matter if you add them or multiply them,
it's the mean value.
CHAIRMAN APOSTOLAKIS: Anyway, we don't
need to debate that now, but I don't think you're
right. I think to propagate the nth moment, you need
the N plus one moments.
MR. PARRY: You don't need the nth moment
if --
CHAIRMAN APOSTOLAKIS: Well, you need the
first one; therefore, you need the second too. But
the point is that even though the -- by and large,
you're right, the number will be close enough.
Wouldn't the sensitivity studies be more
meaningful if you had such a baseline analysis to do
them, rather than playing with the point values and,
say, "I multiplied by two and I will go now to the
95th percentile." What's the basis for all this?
MR. PARRY: I think you'll find in the
latest version of NEI-00-04 that you were looking at,
certainly as far as the independent failure rates,
that taking them to the 95th percentile and the 5th
percentile's been taken out. The parameters that are
varied in that way are things like common cause
failure parameters and human error probabilities.
And the reason I think they're put in
there that way is because we know that those are
pretty uncertain values, and what we want to do by
doing those sensitivity studies is to make sure that
either -- the importance of certain components has not
been either inflated by using pessimistic common cause
failures value or deflated by using very optimistic
values. It's like a safety net, if you like. And I
think that's the reason for it being there, and I
think it is informative to do that.
CHAIRMAN APOSTOLAKIS: And how will you
know that the point values that they will use will be
the mean values?
MR. PARRY: Because they've been declared
to be the mean values. And I think you'll find that
that's a lot of the way things were done in the past
and it's historically. But those point values are
probably chosen from generic -- for a start, if they
were point values and purely point values, they would
probably have been got from generic estimates. And I
think most people, in choosing the values, would have
chosen the mean value of any generic estimate. And if
it's a --
CHAIRMAN APOSTOLAKIS: So, in essence,
what you're arguing is that we should forget about the
additional uncertainty analysis, because it doesn't
really matter. I mean that's what you're saying.
MR. PARRY: That's not quite what I'm
saying.
CHAIRMAN APOSTOLAKIS: Well, when does it
matter then?
MR. PARRY: I'm saying that you don't
necessarily need -- that you can survive without doing
it.
CHAIRMAN APOSTOLAKIS: The document was
very clear that you don't need it. It didn't say
necessarily.
MR. PARRY: We all said it's preferable to
do it.
CHAIRMAN APOSTOLAKIS: But you should have
some benefits when you do it.
MR. PARRY: If there is indeed a major
benefit to be obtained from it. And that's perhaps
one of your studies.
CHAIRMAN APOSTOLAKIS: Well, I mean one of
the goals of the Commission is to inspire public
confidence.
MR. PARRY: Yes.
CHAIRMAN APOSTOLAKIS: Also doing things
right has to have some value.
MEMBER POWERS: Well, I don't know that
doing it right, but calling something a mean value
that isn't a mean value does not sound very confidence
inspiring.
CHAIRMAN APOSTOLAKIS: That's right.
MR. PARRY: But there is -- I think for
very many of the parameters it's actually quite
difficult to get the generic distributions. And I can
give you an example. A long time ago, I was involved
in an exercise to generate a database, a generic
database for ASEP. The way it went was that everybody
voted on what value they should use for the
parameters, and they said, "Okay, what uncertainty
should we put on this? Put down a factor of three or
should we put ten?" And that was a vote as well. And
that's how some of these old, generic databases were
generated.
Now, it turns out, actually, as data is
being collected that they were not all that bad, and
as more and more people have done data collection on
their own plants and updated the distributions using
the Baysian methods, they haven't found that the mean
values, in general, have strayed too far from those
originals.
CHAIRMAN APOSTOLAKIS: My experience has
been different. In some plants, their operating
experience did indeed change the mean values.
MR. PARRY: There will be some specific
cases, yes.
CHAIRMAN APOSTOLAKIS: But were the IPEEEs
done that way? I mean did they use --
MR. PARRY: They're all over the map.
CHAIRMAN APOSTOLAKIS: -- mean values that
--
MR. PARRY: They're all over the map.
CHAIRMAN APOSTOLAKIS: They're all over
the map.
MR. PARRY: Yes.
CHAIRMAN APOSTOLAKIS: So how do we know
that these will not be all over the map as well?
MR. PARRY: I think one of the things that
we're proposing to put in our view guidance with the
staff is that indeed the parameter value should be
compared to well-documented generic databases to see
if they are significantly different.
CHAIRMAN APOSTOLAKIS: But it seems that
the reason why you're arguing that way is because some
people might have difficulty doing it rigorously. So
why don't you then say, "This is the rigorous way of
doing it, and if you don't do it that way, you do it
another way, you also have to do something else, so
there is a penalty"? And that's fine.
MR. PARRY: But they do have to do
something else.
CHAIRMAN APOSTOLAKIS: The version we saw
did not ask for uncertainty propagation at all. It
just said --
MR. PARRY: Right.
CHAIRMAN APOSTOLAKIS: -- these are the
point values.
MR. PARRY: But in calculating the delta
CDF.
CHAIRMAN APOSTOLAKIS: It didn't say
anything there either.
MR. PARRY: Maybe it didn't, but -- well,
I can't defend the NEI document in that regard, but --
CHAIRMAN APOSTOLAKIS: I mean when we
calculate the difference of two very small numbers,
are we arguing that uncertainty doesn't count? I
mean, boy, that's really -- when you calculate delta
CDF and delta LERF, has anybody demonstrated that if
you work with mean values, you get a reasonable
difference when you're talking about ten to the minus
six and seven? I don't know. Because as we know, the
uncertainty increases, right?
MR. PARRY: Again, I think you can -- by
reviewing the cutsets that drive those deltas, you can
see whether there's likely to be a difference. That's
the extra work you have to do to show it.
CHAIRMAN APOSTOLAKIS: One of the papers
that was published in '68 by Allen Cornell, that was
one of the first papers that showed that probablistic
methods can indeed give you counter intuitive results,
had this example in it. You have the difference of
the stress minus the strength, you have some
uncertainty in each, okay? And the variance of the
difference is in fact the sum of the variances, which
is kind of counter intuitive. It increases the
uncertainty. The difference of the means -- I mean
the mean of the difference is the difference of the
means, and everybody says, "Yes, big deal. I knew
that." But the uncertainty increases by how much, and
the variance is the sum of the variances.
So now we are calculating these delta CDFs
and delta LERFs that are such small numbers, and this
morning we saw ten to the minus seven, and we are
completely ignoring these subtleties, if you wish, of
the methods. I mean somebody has to demonstrate that
it doesn't matter, if it indeed it doesn't. I don't
know.
So even if we accept the premise that
point values, mean values -- the declared means values
really don't matter when you do the baseline
calculation using fussell vesely and RAW, when you go
to the delta CDF they still don't matter, the
uncertainties? I mean when you're now calculating
differences of very, very small numbers? I don't
know. I'm not saying it does, but can someone show me
some evidence that it doesn't matter?
And I think it's the same thing as we were
talking about the availability or unavailability of
PRA. If you have a PRA, your life should be easier
using fussell vesely and RAW and all that stuff. If
you don't, then your categorization process should be
much more conservative, right?
MR. PARRY: Yes.
CHAIRMAN APOSTOLAKIS: I mean is that
evident in NEI 00-04? I don't know.
MR. CHEOF: We think it is. We have made
that comment before that, you know, if a licensee has
a PRA for an external event, that they can be less
conservative, and if they were to categorize using a
method that's not PRA quantified, they have to be a
lot more conservative.
CHAIRMAN APOSTOLAKIS: Yes. We can agree
that that's the way it should be. But does the
document do something that guarantees that this will
happen?
MR. CHEOF: I think there's at least one
statement in there that says that. I'm not sure if
the process itself --
CHAIRMAN APOSTOLAKIS: Oh, okay.
MR. CHEOF: I mean they do have flow
charts in there and how they can treat the other
external events. And the staff has looked at those
flow charts, and I think we are working with NEI as to
how we can ensure that those flow charts will indeed
give you more conservative results than if you had a
PRA.
MEMBER KRESS: Let me ask what might sound
like a strange question. This process of
recategorization of SSCs is -- the view, seems to me,
we've got to already have a categorization. That's
the reason we end up with four categories. And that
the process is going to be applied to plants that have
already categorized and we're just going to
recategorize. Can the process be applied to a brand
new plant that comes in and says, "I haven't
categorized yet."
MR. REED: Yes.
MEMBER KRESS: Do they have to go through
the old process first and categorize and then --
MR. REED: Yes.
MEMBER KRESS: So is the old --
MR. REED: I think they would have to.
MEMBER KRESS: So the old process would be
part of the overall process.
MR. REED: Obviously, this hasn't happened
yet, and so it's going to be a little bit of a
speculation on my part, but if somebody was to, today,
decide to apply for a new license and follow the
current Part 50 and then try to adopt 50.69 in the
process, I think what they would first have to do is
basically go through like the old design basis,
safety-related world --
MEMBER KRESS: They have to go through the
whole design basis.
MR. REED: Do that on paper now.
MEMBER KRESS: Yes.
MR. REED: And then basically do a
categorization, and then take that whole safety-
related and non-safety-related world, translate it
into the four boxes, now all on paper, come in
basically with that submittal, and they would procure
from the get-go Box 3 to be RISC-3. So they would --
right from the start the whole plant would be procured
that way. That's a big difference versus current
facilities.
MEMBER KRESS: Yes. That's my impression
of what would have to be done.
MR. REED: That's correct.
MR. GILLESPIE: Yes. On the other side,
if it's a certified design, the certification is, in
and of itself, a rulemaking. And so someone who has
a certified design could apply the concepts of Option
2 likely within the context of the certification. So
Tim described someone who would be coming in applying
for a license under the current Part 50 as if they
were 20 years ago. Yet we have a different process
which might actually allow a little more freedom and
innovation. Because a rule can offset a rule.
MEMBER KRESS: Yes.
MR. GILLESPIE: And the certification
itself is a rule.
MEMBER KRESS: I understand.
MR. GILLESPIE: I think that's how we'd
end up getting around this without rewriting all of
Part 50 again.
MEMBER KRESS: Yes.
CHAIRMAN APOSTOLAKIS: I can give you an
example from this morning's presentation of abuse of
PRA models, and only if you really have seen a lot you
appreciate it. We were told that the time to respond
to something was reduced from nine minutes to six
minutes. And there was a table that said, "and the
core damage frequency increases by less than one
percent." Now, Gareth, you don't believe that, do
you? You know that it can't and on the face of it is
a misleading statement. Are there any models -- any
reliability models that can really tell the difference
between a nine-minute response and a six-minute
response, and everybody agrees that, yes, that's true?
I mean there are ideas, there are judgments, there is
this, there is that, and yet it was presented this is
what it is.
Now, the application was not risk-informed
so it didn't matter, but you see that my point is that
somebody who knows will look at this thing and say,
"What's going on here? This is really nonsense." But
it's not essential to the decision, so you let it go.
MR. REED: I think I'll just add a
comment; hopefully it's constructive. But I think
some of the issues you bring up are why in Option 2
space why we're risking forming only assurance
requirements and maintaining the design basis down in
Box 3, albeit with less assurance.
It almost -- I'm not saying I know what
you think, but sometimes I get the feeling that we're
trying to justify significant technical changes here,
and I think you'd have to know a lot better some of
these uncertainties if you were trying to make
technical changes to the facility. At least that's my
own personal opinion; perhaps you don't. But in
Option 2 space, in think we get some comfort from the
fact that we're going to be maintaining the design
basis functions.
CHAIRMAN APOSTOLAKIS: The problem is that
when we start using PRA in real decisions, we seem to
be going backwards, and we seem to treat methods and
models in a cavalier way. You know, you want the
delta CDF involving time-dependent human errors? I'll
give you one. Okay. And everybody says it's one
percent. I mean when in fact the answer is there are
ten different models out there, and you can get any
answer you want. And the truth, in my mind, is that
you can't really quantify such a difference, I mean,
with any degree of confidence at this level.
I mean you know that it's a good thing.
It's actually a bad thing in this case because
available time was decreased by a little bit. But you
can't really put a number. But if you keep doing it
that way and you never really raise the flag and so
on, eventually it will become practice, and that's
bad.
MR. PARRY: Well, I think though, George,
in that particular example, I think it's incumbent
upon the reviewer to figure out what model's been used
and whether if they used alternate models if they'd
get a different answer. And that's part of --
CHAIRMAN APOSTOLAKIS: And why isn't that
applicable here? If you use a different model, you
may get a different answer.
MR. PARRY: I think that there are -- I
mean at least in one of our comments, one of the
things we said was that you should identify the
assumptions that drive the changes, and they should be
candidates for performing sensitivity studies --
CHAIRMAN APOSTOLAKIS: That's right.
MR. PARRY: -- to see whether they impact
the categorization.
CHAIRMAN APOSTOLAKIS: I guess we're
talking about two different things now. You're
referring to comments you have submitted to NEI, which
I am not aware of.
MR. PARRY: Right.
CHAIRMAN APOSTOLAKIS: And I'm referring
to NEI 00-04, the document I have viewed.
MR. PARRY: I know. And one of the things
in 00-04 is if you look at the sensitivity studies
that are specified, they do have a category there
which are those that I think that are revealed by the
facts and observations from the peer review process.
So that gets part of the way to where we want to be,
but I don't think it gets all the way there.
CHAIRMAN APOSTOLAKIS: Anyway, shall we go
on?
MR. REED: Sure.
CHAIRMAN APOSTOLAKIS: Unless there are
other --
MR. REED: Anymore questions on this
slide? Okay.
So continuing with two more key points
then, one of which has been made several times already
today, NEI 00-04 is an interim product, it's in a
state of flux. It's certainly going to change. NEI's
going to update it and roll back in the feedback
they've gotten from pilot activities and also address
our comments. And it's understandable and of course
appropriate that ACRS would reserve its final judgment
until they have a more final product. So that's all
this slide simply reflects.
MR. GILLESPIE: I think, George, part of
the discussion this morning has highlighted why it's
a work in progress. Fourteen pages of comments, and
I forget how long NEI 00-04 was, but I think it was
only -- Tony, help me out, about 30 pages long?
MR. ULSYS: Categorization piece --
MR. GILLESPIE: Categorization piece.
MR. ULSYS: -- was 17 pages long.
MR. GILLESPIE: Okay. So we've submitted
comments that are in excess of approaching 15 percent
of the total document. And we need to see how it now
comes out of this next step in the process and have
another iteration. I'm not promising that everything
you've said will be considered, but I think the idea
of assuring the applicability of the study to what
we're using it for, the need to do that is a concept
that we will be trying to embody in it.
I don't know how to do it, and there's
been discussion that's over my head on how to do it,
but I think we understand the comment and have some
sympathy for it. Now, we have to figure out is how to
articulate it where it's consistent with what we think
are low-risk components also.
CHAIRMAN APOSTOLAKIS: The way I see it,
Frank, is you have to make approximations, you have
to. I mean that's the way life is. If you do, it
seems to me, somewhere there you have to demonstrate
that it's an approximation, rather than saying, "Well,
we've done it many times, and it came out that way."
The other thing is I think it would be
useful to say, "Look, if you do it this way, in a
rigorous way, this is the benefit you get. If you do
it in a less rigorous way, you should be a little bit
more conservative, and this will guarantee that that's
case. And typical example here is when you have a PRA
or you don't. A guy who has a PRA for fires, for
example, should be able to get more benefit out of it
than somebody who did five, right?
MR. GILLESPIE: Yes.
CHAIRMAN APOSTOLAKIS: Or the seismic
Heathcliff, whatever they call it, or rigorous PRA.
And then when you go to the places where there is no
PRA at all, then you make sure that you have a
conservative approach. So this kind of phased
approach, I think, would go a long way towards
convincing me, at least, that this is a good, solid
approach.
MR. REED: I think everybody agrees with
that concept, and I believe we are trying to assure
that's in this guidance document.
MR. GILLESPIE: Yes. In fact, I'll say it
a little different way. What the staff would like to
do is give people who have gone that extra mile to do
an external event PRA in some detail or shut down PRA
in some detail or fire analysis, we'd like them to be
able to maximize the benefit they get from their
investment.
CHAIRMAN APOSTOLAKIS: Exactly.
MR. GILLESPIE: So in principle, we're in
total agreement. And right now, though, we're trying
to endorse what -- I could generalize, this is a
generalization -- a one-size-fits-all kind of guidance
document. And what you're saying is when you read the
guidance document, you didn't see the spectrum that
would actually encourage people to do the right thing
because of the benefit from it.
CHAIRMAN APOSTOLAKIS: Exactly.
MR. GILLESPIE: So I think I got -- I know
where you're coming from and we're in total agreement,
but the staff being in agreement doesn't mean the
industry who's writing the industry guidance is
necessarily writing it with that same concept in mind.
We can give them the comment and they're here. If
they heard the comment. But how they embrace that
comment, we're going to be reviewing for the purpose
involved, whatever they submit. But I think it's a
valid comment. And, indeed, in disk space and risk-
informed tech spec space, we would also, in a risk-
management sense like to give people the maximum
payback for what they've invested and what should be
a better decision tool.
How do we get there? We haven't figured
that out yet. You know, it's difficult. As a
regulator you can't dictate that to somebody, but we
would hope the industry would kind of grasp that
concept and maybe figure out how to factor it into
their document also.
A mutual gain on that, because there is a
spectrum of facilities out there with a spectrum of
tools available. And how do you give that guy who's
put a big investment in the maximum return, and that's
really your question, versus writing a guidance
document to the median level? It's a fair comment.
Tony, you've got the comment.
MR. PIETRANGELO: We'll wait our turn.
MR. GILLESPIE: Okay.
CHAIRMAN APOSTOLAKIS: Well, are there any
more questions for the staff?
MR. REED: I guess I'd just like to say on
this last slide, it kind of rolls everything up, I
would simply mention that, probably said already, I
think we've said everything here, but if the Committee
has major issues, then we'd like to have a letter, and
we certainly appreciate the Committee's input, and
it's obviously a great deal of expertise in the PRA on
this Committee, so appreciate that.
MR. GILLESPIE: And I hate to say it, this
being March and our rule due in July means we probably
need a letter in June. So it gives us about 30 days
to get you something for potentially an April or May
meeting.
MEMBER POWERS: We're not that slow.
MR. GILLESPIE: No, but we may -- just in
counting back, if we're trying to get something to the
Commission and we need a letter by July, I'm not sure
that we might not have to have something to you so
that you can read and review it for a June meeting.
Which may mean, I don't know, a May Subcommittee
meeting.
So, anyway, we're going to be back again.
I'm anticipating, George, that we'll probably have
another Subcommittee meeting and full Committee
meeting.
CHAIRMAN APOSTOLAKIS: Yes. That would be
good. When you say that the staff requests a letter,
you mean now, this letter you are --
MR. REED: Yes. The letter I'm asking for
right now if there are major issues that -- I'd like
to be able to --
MR. GILLESPIE: Yes. We're going to have
to come back again for a second letter. So the
context of this letter is we're really having a
dialogue now and we're looking for suggestions, advice
and anything you want to give us. We'll be back again
for another letter.
CHAIRMAN APOSTOLAKIS: The two letters
that I mentioned that we have already written they
still stand.
MR. GILLESPIE: Yes.
CHAIRMAN APOSTOLAKIS: Okay?
MEMBER SHACK: Just coming back to your
degradation problem for a minute, isn't part of that
an artificial thing where you insist on -- or at least
the guidance sort of gives all components the same
increase in failure rate. If you really went through
and you said, you know, "A valve and a steam tunnel,
if I don't have the special treatment, it's failure
rate may go up X a valve sitting out in a benign
containment environment."
MR. GILLESPIE: Well, exactly, and that's
why --
MEMBER SHACK: And it seems to me that
that might be a sensible place to begin to attack that
kind of a problem. I think we're going to have to go
to sensitivity analyses, but I'm not sure that -- you
know, the broad X approach that's been chosen at the
moment really, I think, penalizes the industry to a
certain extent.
MR. GILLESPIE: And that's something we
just really started coming to grips with, because it's
kind of an overriding consideration for all those
paragraphs we've got in the rule for monitoring and
conditioning and all those different things. So it's
kind of an -- there's a sentence in there someplace
that we might not have right now that says, "Consider
this when you're doing these things."
We're not saying hide a monitor, but
clearly monitoring of these components, whether it's
needed or not in that decision, the intent is to
maintain the credibility of the decision process that
you made when you made your decision on what you
needed to do or the decision that it was RISC-3. How
do we sustain the credibility of the decision process,
both in the beginning and as an ongoing basis? We're
wrestling with that right now. Because you would
penalize people. If you make a blanket statement,
it's like, well, we're going back to the old way of
doing things. So we want to leave some flexibility in
there.
MEMBER SHACK: Okay.
MR. GILLESPIE: Thank you. I think that's
it for the staff.
CHAIRMAN APOSTOLAKIS: Thank you very
much. Tony? Welcome.
MR. PIETRANGELO: Good afternoon. I'm
joined by Adrian Haymar and Biff Bradley from NEI. We
wanted to -- we were primarily going to talk about the
categorization guides today, and I know you had a
Subcommittee meeting where you got a pretty extensive
presentation on that guidance. We're not going to
redo that today, obviously.
What we did want to talk about, though,
and I think Tim set it up quite well, this is a work
in progress. We have your comments, we have the
staff's comments. We think we have a comprehensive
guidance document for this categorization. Before we
began the pilot categorization effort that began last
fall, I think we had a critical time last July with
the staff to make sure there were no show stoppers in
the guidance that would preclude the pilots from
trying to demonstrate the usefulness of NEI 00-04. We
got to that point. I'm sure there's additional
comments that can be incorporated.
I think Tim captured the issues we're
working on quite well, in terms of addressing the late
containment failure and the IDP guidance. I think the
peer review piece is a separate issue. Quite frankly,
that applies to all the PRA applications. We have an
entirely separate effort dealing with that. We met
with the Office of Research last month on the guidance
to endorse both the ASME standard as well as our peer
review guidance, so that's going in parallel with
this. Certainly, this is one of the most important
applications that exercises all elements of the PRA.
One thing I do want to point out where we
are in kind of our stage of development here, and this
issue has come up and we've talked about the ASME
standard and PRA quality, in general, this is an
evolutionary process. I think if we get too hung up
with the level of precision about where we are today,
I mean that stops progress. That's why we've been, I
think, pretty adamant from the outset on here that
there's a need for NRC staff review of any Option 2
application.
We didn't buy the Appendix T concept of
you raise your right hand and swear that you meet all
the stuff that's detailed in Appendix T and then the
staff doesn't have to review it. We're not in the
stage of development with PRA and the comfort level
yet to do that. And we were glad to see the staff
pull that out of 50.69, because, again, that's a nice
goal to shoot for maybe in five or ten years when we
do have added confidence in the studies, but we're not
there yet.
So if Gareth or Mike or any other NRC
staff reviewer has a particular question about what
the licensee did in the Option 2 application, they can
ask the question in the process. We still intend to
define a template, just like we did -- we're using
risk-informed ISI as a model for this, a template of
what the licensee would submit as part of an Option 2
application.
So I think the process -- that's another
part of the process. We've tried to follow 1.174.
That has worked quite well. I think some of what I
gathered from the discussion this morning, George, is
that we're starting reopen some of the things that
were discussed when 1.174 was developed. And I think
in the context of an application, that's not the time
to do it. It should be done independently of the
application.
We think sensitivity studies were one of
the things that 1.174 said you could do to address
areas where you have uncertainties. And that's what
we're doing. So we're trying to follow the guidance
that's out there that's been successful, and not try
to reinvent that, with the Option 2 guidance.
Your Committee spent a lot of time on
1.174. All those issues were debated quite
thoroughly. And to reopen that as we go through this
process again, I'm not sure is the most --
CHAIRMAN APOSTOLAKIS: Well, and I would
have to be convinced that I'm reopening issues. I'll
go and read 1.174 again, but lets go on.
MR. PIETRANGELO: Okay. The other thing
we wanted to briefly chat about today is the
treatment, and Frank teed up some of those issues. I
mean to us we've been talking about treatment I think
going back to graded QA now for almost ten years. And
it continues to bewilder us that the focus of the
reviews are on the low safety significant SSCs.
That's not what risk-informed regulations are all
about. That's what we see all the hand wringing about
just about in every application we get into. And it's
like each application, before we get to the end of it,
it seems like the entire regulatory process has to be
dumped into this one application, and we forget about
all the other things that are at play out there.
Besides the revised oversight process,
there's requirements in the rule that the licensee has
to meet, all right? And they're subject to inspection
and enforcement. And if they don't do what they said
they were going to do, then that's willful non-
compliance, and the staff knows very well how to deal
with willful non-compliances.
So I mean we have to remember there's
whole other regulatory construct around Option 2 that
doesn't go away. And we just started hearing about
these known degradation -- failure mechanisms and
degradations. Well, we know what those are. They
don't go away when we go to Option 2. If a valve or
any other piece of equipment goes into the
preventative maintenance program, you don't forget
about all the other stuff that happened to it. All
that operational experience is still there.
And so NRC can audit the implementation of
Option 2. They can audit the performance monitoring
that goes on with Option 2. Most of it's already
captured in the maintenance rule.
And the other thing, you know, the
industry has experience with categorization, even
before the maintenance rule with some of the MOVs and
a graded QA application. So we've been doing
categorization for a long time. Let's demystify
what's going on here. I mean this is not going to
hinge on some number in the PRA as the one thing going
up or down. I mean to even suggest that that's
happening through this is absurd, okay? The Expert
Panel there's guidance in our document now that talks
about the requirements for the Expert Panel, that's
similar to the folks that are around the table here,
okay?
So, again, to get tied up in the level of
detail and the PRA calculations and all that, we want
to do a good technical job, I'm not here to say
otherwise, but unless it has an impact on the result,
then we can spend an awful lot of time in the noise
level of this and not get to the fundamental purpose
of it, which is we've been categorizing SSCs for a
long time, we have a very comprehensive process we're
trying to develop and get the staff to endorse, and
your endorsement, and give the best guidance we can to
the licensees who wish to go into Option 2.
And the short answer to your question that
you posed at the end, if they don't have a sound
technical basis from which to defend moving the thing
down from RISC-1 to RISC-3, they're not going to do
it, because they're going to be subject to the NRC's
review, and if you don't have an argument, you can't
just wave your hands at it and say, "Now, it's RISC-
3." There's an extensive process to go through, both
PRA and other studies done, as well as the Expert
Panel review, and that will be subject to the staff's
review also. We have to put faith in that process.
And even afterwards in implementation, if
there's new information that comes to the table,
there's mechanisms to treat that, just like there is
with any other regulation. And how do you know we're
implementing a(4) from day to day, how do you know
we're doing 50.59 right from day to day? So that
whole rest of the regulatory process is there to bring
those up again within this context. I mean we've all
been in this business a long time, and to wring our
hands over that kind of stuff at this point just seems
to me to be a waste of time.
I'm being very candid with you this
morning. We've been working on this for a long time,
progress has been slow. We've got the last pilot on
categorization next week. We'll be rolling that into
the next draft of our guidance. We're way ahead of
where we are normally on a rulemaking with regard to
the development of guidance. In most cases, the
guidance document isn't even developed until the final
rule is done.
(Laughter.)
We're way ahead of the game on this one.
It's even been piloted already to a certain extent.
So remember where we are in the context there. That's
all I'm asking.
And we've been listening to the questions
back there and we could point out places in the
guidance to try to address your questions, but I'll
ask Adrian and Biff if they want to add anything to
that.
MEMBER POWERS: In thinking about things
to add, I guess this philosophical issue that George
brings up on rigor in the use of PRA, I'd be
interested in your comments on that. You've already
addressed it somewhat. But I guess what I'm
interested in is the desirability of having a nice
rigorous treatment with respect to PRA, understanding
that the PRAs that we have today bear faint
resemblance to the PRAs we'll have in ten years or 20
years.
MR. PIETRANGELO: Right. We want them to
be rigorous, we want it to be a repeatable
application, we want to have some stability in the
process. We're all for rigor, okay? But if you don't
have the study for a particular scope element of the
PRA, then you've got to use another means. You're not
even going to be an Option 2 potential applicant --
MEMBER POWERS: Because you're too far
away, yes.
MR. PIETRANGELO: -- right, if you're too
far away from that. I mean no one's going to submit
themselves to the gauntlet of review here if they
don't think they've got a good technical basis from
which to do the categorization.
MEMBER POWERS: But doesn't everybody
think that because of the IPEEEs?
MR. PIETRANGELO: No. I don't think
everybody thinks that. Not to the extent we're
talking about here in Option 2. I think for purposes
of the maintenance rule when we were trying to
establish the level of monitoring that was done, I
think the answer to your question was yes. For this
purpose, I think this is much more rigorous, and I
think we're finding out from the pilots, not only more
rigorous but more costly to do and resource-intensive.
So unless you're really serious about it, you're not
going to do it.
CHAIRMAN APOSTOLAKIS: Anything else?
Thank you.
MR. PIETRANGELO: Thank you, George.
CHAIRMAN APOSTOLAKIS: I always appreciate
Tony's way of -- elliptical way of making a point. I
would like to thank the staff as well. And we will
reconvene at 1:25.
(Whereupon, the foregoing matter went off
the record at 12:27 p.m. and went back on
the record at 1:27 p.m.)
. A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N
1:27 p.m.
CHAIRMAN APOSTOLAKIS: We're back in
session. The next item is the Arkansas Nuclear One,
Unit 2 Core Power Uprate. The ACRS cognizant member
is Mr. Sieber. Jack, it's yours.
MEMBER SIEBER Thank you, Mr. Chairman.
The application and the SER that we're going to
discuss this afternoon differs from the previous power
uprates that we've had in that this is the first
pressurized water reactor that has applied for an
uprate in power greater than our cutoff limit, which
has been five percent. And so this will be the first
PWR that we have undertaken to examine. On the other
hand, the staff has done previous uprates of lesser
increases in power in the past.
Interestingly enough, the Arkansas One,
Unit 2 is a combustion engineering plant, and in the
process of deciding what guidance the applicant would
use in order to make sure that they have covered all
the aspects that are recommended or necessary to do a
power uprate, they ended up going to a Westinghouse
document, which was published in 1983, and it's WCAP
10263, and that was the basis for the applicant's
process of coming up with the analysis and studies
necessary to do the power uprates. And on the other
hand, in 1997, the staff did a uprate SER for the
Farley Plant, and the Farley uprate was also based on
the WCAP that I discussed and mentioned.
And so there is a sort of de facto
template out there for the staff to write its SER and
applicants to do the analysis for a power uprate for
a PWR. Even though the Plant's combustion
engineering, there are plenty of similarities between
the combustion engineering plants and Westinghouse
plants so that these documents are generally
applicable.
What I'd like to do now is to introduce
Mr. Craig Anderson from Entergy, and he is Vice
President of Operations, and he will guide us through
Entergy's presentation on the power uprate.
MR. ANDERSON: Okay, sir. Thank you very
much.
MEMBER SIEBER Sure.
MR. ANDERSON: Again, I'm Craig Anderson.
I'm the Site Vice President for Entergy at Arkansas
Nuclear One. We've got several other presenters here
I would like to introduce, and if you all would raise
your hand in the back. Bryan Daiber is the Senior
Staff Engineer that will present a lot of the
technical information today. Rich Swanson is the
Senior Reactor Operator that we brought. I think the
operational aspects of a power uprate are very
important, and we felt like we needed to discuss
those. Dale James, the Manager of Engineering
Programs and Joe Kowalewski, the Director of
Engineering. We also have other folks here, members
of our staff, that might address questions that might
come up. And also Westinghouse folks here to address
any questions that they might help us with.
One of the thoughts might be, well, you're
combustion engineering in NSSS, you've got
Westinghouse here. If you recall, Westinghouse
acquired combustion engineering a few years back, and
these folks were previously on the combustion
engineering staff, so we've got good technical
expertise here to try and address those questions.
Over the next hour, we will discuss the
results of years of work to both analyze our Plant
and, where necessary, to install modifications to
support a safe power uprate of the Unit. You'll see
that we were careful to maintain the operating and
design margins, and just as importantly to us, to
minimize new challenges to the operators. We
certainly don't want a power uprated plant that's not
reliable or that presents difficulties to the
operators.
Where we weren't comfortable with our
margins we modified our Plant, and we will talk about
several of those modifications during our
presentation. We, of course, used accepted
methodologies and we've, in all cases, demonstrated
compliance with regulatory and safety limits.
A little bit about the project before I
turn it over to Bryan. Our goal is a seven and a half
percent uprate. That's where we've completed our
analysis to support that uprate. And it essentially
was a balance on the financials between the investment
that you make in the Plant and the return you get from
the investment, and, of course, that without adversely
impacting the available design and operating margins.
I think one of the things that's important
to point out, the majority of the modifications that
were needed to support power uprate have already been
made. They were installed during the last refueling
outage in the fall of 2000. And we've operated this
operating cycle, which ends next month, with those
modifications, and the modifications have performed
quite well.
The most significant modification was
steam generator replacement. Steam generator
replacement, while it was not driven primarily by
power uprate, it was driven by the degradation of
alloy 600 tubing. We took advantage of the need to
replace the steam generators and increased the heat
transfer area, both to support power uprate and also
give us some more margin.
The rest of the mods, most of them, in the
balance of Plant will be installed in the spring, and
we will complete all the necessary work for the power
uprate, including the start-up testing following the
outage to support the power uprate.
So we believe we're prepared for the
uprate. We've done a thorough review of the equipment
and our analysis and our procedures, and that's been
completed. We have been and are continuing to train
our operators on the uprated Plant to make sure that
they are ready, and we believe our people and
equipment are ready.
And let me turn it over to Bryan Daiber
who will go through the technical portion of our
evaluation. Bryan?
MR. DAIBER: Make sure the microphone's
working here. I'm Bryan Daiber. I'm the Safety
Analysis Lead. I was the Safety Analysis Lead for
both the RSG and the power uprate projects. I will be
going through several of the presentations today. The
first one I'm going to go over are the plant
modifications for considerations of power uprate.
For ANO2, we've been considering power
uprate for the past four cycles. We were obviously
considering steam generator replacement due to
degradation of the alloy 600 tubes in those. In
preparation for that, we were trying to move the
copper from the secondary side system, so we replaced
the condensers and other major components to do that.
And in doing those replacements, we kept power uprate
in mind in the design of all those components we've
replaced over the past four cycles.
So we have replaced many of the
components, many of the major modifications have
already been implemented, and we've operated with
those for at least one cycle on most of the major
components. And as I mentioned, we did keep that in
mind, the power uprate was considered for those
modifications.
On this slide, we list many of the
modifications, balance of plant, and other
modifications needed to support power uprate
conditions. Many of the major modifications, like I
said, have already been installed and have
accommodated for power uprate conditions. Rather than
go over these balance of plant type modifications, I'd
rather focus in on three key areas, the highlighted
ones in blue here: the replacement steam generators,
the containment uprate considerations and the fuel
core design considerations that we implemented for
power uprate itself.
The first key modification, the steam
generators. We did replace steam generators last
outage. There were degradation concerns with the
alloy 600 tubes. When we replaced these steam
generators, we replaced them with steam generators
that were specifically designed for the power uprate
condition. In light of that, when we designed these
generators, we did increase the tube sheet diameter by
four inches to accommodate greater number of tubes in
the steam generator. We also increased the surface
area and the number of tubes in the generator by going
from three-quarter-inch diameter tubes to eleven-
sixteenths diameter tubes. The net effect of these
changes allowed us to gain 25 percent surface area on
the tubing material in the new steam generators.
The result of that also resulted in an
increase in the primary side volume. Now this is key.
The increase in primary side volume did cause a
challenge to the containment building pressure. As a
result of that, we did have to look at the building
pressure considerations. The volume, we essentially
went from 1,600 cubic feet to 1,839 cubic feet per
steam generator. That increase in volume obviously
resulted in an increase in mass of energy available to
blow down to containment for our LOCA analysis.
In comparing that to the effects of power
uprate, for power uprate considerations on the
containment analysis, power uprate results in a
slightly higher increase in Tav, and we're also
proposing to increase Tcold by two degrees. Both of
these effects essentially increase the energy content
in the RCS available for the blowdown, but it does
decrease the mass available. So really the net effect
is the increase in volume had a much bigger impact on
the containment pressure considerations than the power
uprate considerations.
The other thing on the new steam
generators, the secondary side volume also went up as
a result of the change. To offset that change in
secondary side volume, we didn't modify the steam
generators. The new steam generators have an integral
flow restricting nozzle in them. This integral flow
restricting nozzle, in combination with high
containment pressure actuation signal, the containment
spray actuation signal, was sent to isolate main steam
and main feed. The combination of those two
modifications essentially reduced the peak building
pressures associated with the higher power steam line
break considerations. As a result, the hot zero power
steam line break is actually the most limiting.
MEMBER SIEBER Did you not increase the
sprayed area of containment also?
MR. DAIBER: No, we did not -- the sprayed
area of containment by the containment spray system
stayed the same.
MEMBER SIEBER All right.
MR. DAIBER: So we have designed a new
steam generator to accommodate power uprate
conditions.
The second key design consideration that
we made to accommodate power uprate was, as we
mentioned, we did uprate the containment design
pressure. We went from a design of 54 pounds to 59
pounds. We accommodate this increase in design
pressure by recognizing the fact that the Unit 2
containment is very similar to the Unit 1 containment,
although not identical. And the Unit 1 containment is
already designed to 59 pounds.
There was a detailed finite element
analysis done to verify the structural capabilities of
the containment. That detailed finite element
analysis did credit additional tendons. We didn't
install additional tendons, but there were -- as part
of uprate, but during original construction,
additional tendons were put into the containment to
account for surveillance considerations and
construction considerations that weren't originally
credited in the original structural analysis. We did
credit those in this analysis to accommodate the
increase in design pressure.
Not only did we verify that the structure
itself was capable of operating at the higher design
pressure, we also verified the equipment inside
containment was also able to accommodate the higher
design pressure. The containment was tested at 68
pounds to verify its capabilities. All of this work
was done obviously as part of the replacement steam
generator project and has already been approved by
License Amendment 225.
The third key change or consideration with
respect to power uprate deals with the fuel core
design itself. At ANO2, we are currently using a
Gadolinia integral burnable poison. We are going to
switch that integral burnable poison from Gadolinia to
Erbia. Now, back in Cycle 13, Cycle 16 being the next
core design and it's our operated core design, but
back in Cycle 13, we started replacing the poison
schims before c-schims with integral burnable poisons,
and Gadolinia was the burnable poison of choice at the
time. By replacing those schims with these integral
burnable poisons, we have effectively gained almost
four percent additional pins available in assemblies
for additional fuel considerations.
The Gadolinia burnable poison is a much
more potent poison, and the typical assembly will have
about eight weight percent Gadolinia versus about two
weight percent for Erbia. The Gadolinia within an
assembly that has Gadolinia pins, they'll be somewhere
between four and eight pins, or four and 12 pins per
assembly. Whereas with Erbia, we'll have somewhere
between 30 to 100 Erbia pins per assembly.
The Erbia is current approved methodology.
There are many plants within the CE fleet already
using the Erbia core designs. There's essentially
over 159,000 Erbia pins already in operation, 64,000
of which have already been discharged. As I
mentioned, the Erbia is a more dilute poison, it
allows us to have better power control and gives us
better peaking control within the assembly itself.
And that helps us out just during normal operation
conditions and also as a result of any transient
conditions that would occur. It also allows us to
have a better control over the moderate temperature
coefficient.
MEMBER POWERS: What was the attraction of
selecting Erbia as the poison?
MR. DAIBER: Again, within the assembly
itself, the Erbia allows for a much more equal power
distribution within the assembly.
MEMBER POWERS: Well, I understand that.
That's based on the number of pins that you put in
there.
MR. DAIBER: Pins and the amount of the
poison. Erbia's more at two percent, whereas
Gadolinia's more at six to eight percent.
MEMBER POWERS: But you could have just as
well have put two percent Gadolinia and put more pins
in and done the same thing, couldn't you have?
MR. DAIBER: Jeff?
MR. BROIDA: Use a microphone and identify
yourself, please.
MR. DAIBER: I'll let Jeff Brown from
Westinghouse speak to that question.
MR. BROWN: Jeff Brown, from Westinghouse.
Another major difference is the cross-section of
Erbia. It has about 200 barns cross-section for Erbia
compared to, I believe, about 10,000 barns for
Gadolinia. Gadolinia, on a per atom basis, is much
more stronger thing, and it depresses the local power
distribution almost like you had a small control rod
there. Whereas Erbia, you know -- so even in a two
weight percent concentration, the Gad would have a
similar effect as it does not.
MEMBER POWERS: Oh, okay. So you're just
avoiding the high cross-section --
MR. BROWN: Yes.
MEMBER POWERS: Sure, I understand.
MR. DAIBER: I'd like to make two major
points here with these comparisons of the core
designs. The most -- the first issue, which we've
already talked about is that going to the Erbia
burnable poison allows us to do the flatter power
control within the assembly itself. Also, the number
of assemblies we're putting in for Cycle 16 it's a
larger reload. Eighty new fresh assemblies are going
into the Cycle 16 core design. By doing that, we also
control the peaking factors, the radial peaking factor
is going down by over seven and a half percent for the
uprated core designs. That gives us a flatter power
distribution within the whole core itself.
The other important point I'm trying to
make on this slide is the energy content. The energy
content for Cycle 16 is actually bounded by the energy
content that we've already implemented under our
current power rating conditions in Cycle 14. The
Cycle 14 length was 557 EFPD. For Cycle 16, the EFPD
is 485. When converted to a comparable power of
2,815, it's more like 521 EFPD. So it's actually
lower energy content. That also can be noticed by the
cycle burn-up value. The cycle burn-up for Cycle 14
was 19,770 megawatt days per ton, whereas for Cycle
16, it's 18,825 megawatt days per ton.
I'd like to make two clarifications from
the Subcommittee presentation. There was a question
asked at the Subcommittee about the fuel zoning. For
Gadolinia fuel assemblies, we typically have three
zones of U235 considerations. I believe we may have
mentioned only one in the Subcommittee presentations.
With the Erbia, there are essentially two zones of
zoning in the Erbia designs.
The other question that came up at the
Subcommittee meeting was with respect to the cycle
burn-ups, and we've just discussed the cycle burn-ups
for the different core designs. Yes?
MEMBER BONACA: I had a question, but I
couldn't find the answer here. You must have changed
your Delta T, T hot to cold.
MR. DAIBER: Yes. Our RCS flow stays the
same, so the T hot goes up.
MEMBER BONACA: Yes. Have you changed
your pressurizer program?
MR. DAIBER: The pressurizer --
MEMBER BONACA: Program.
MR. DAIBER: Yes. The pressurizer level
control system was reviewed and verified and updated
as necessary to account for the Tav increase.
MEMBER BONACA: Because some of the early
CE five percent power increases didn't, and they used
to lose pressurizer level when they were SCRAMing. So
you did look at that.
MR. DAIBER: Yes, we did look at that.
With that, I'd like to move on to the next
agenda item, which deals with compliance with
regulatory requirements, and in particular they deal
with the Plant margins. We did review the ANO2 design
to make sure it could accommodate the power uprate
conditions. We did obviously look at all the balance
of plants, and we made modifications as necessary on
the balance of plant to ensure that it could
accommodate power uprate within its design
considerations.
We also looked at the NSSS, the nuclear
steam supply system, which is a CE-manufactured
product, and we verified that all the design
components there could also withstand the
considerations on the power uprate conditions. We
also looked at the control systems, the pressurized
level control, feedwater control system and all the
control systems and made sure that they also could
accommodate power uprate. And as we discussed, steam
generators containment and the fuel design were also
considered, along with all the safety systems.
In the review of any of these systems, in
any place where we felt margin was not being
maintained as a result of the power uprated
conditions, modifications were implemented or will be
implemented in the next outage to ensure that all the
components could operate satisfactorily at uprated
conditions. And for the control systems, appropriate
set point changes have been made for those systems.
MEMBER SIEBER I'd like to go back to RSC
flow.
MR. DAIBER: Yes.
MEMBER SIEBER During the Subcommittee
meeting, it was stated that the replacement of steam
generators had a lower DP --
MR. DAIBER: Yes.
MEMBER SIEBER -- on the primary side than
the original ones.
MR. DAIBER: That is correct.
MEMBER SIEBER That would increase RCS
flow instead of having it stay the same, right?
MR. DAIBER: That is correct.
MEMBER SIEBER And that's why your Delta
T change was not as much as you would ordinarily
calculate from a seven and a half power increase?
MR. DAIBER: There are several things that
went on. Obviously, with the old steam generators we
had plugged those quite a bit, and flow -- the delta
P had gone up, and flow had gone down. When we
installed the new steam generators, we designed those
to essentially restore the delta P of the steam
generator essentially comparable to the unplugged
original steam generators. So flow went back up as a
result of that, but it was more due to the plugging --
the removal of the plugging restrictions.
MEMBER SIEBER Now, it would seem right
now with the new steam generators that you have a --
you take into account the fact that I690 has a cooler
heat transfer coefficient, so that takes away surface?
MR. DAIBER: That's true.
MEMBER SIEBER Looks like you have a plug
of about ten percent. Is that correct?
MR. DAIBER: When we did all of our work,
we did it with a ten percent consideration, so all of
the efforts that we undertook, we assumed a ten
percent plugging consideration.
MEMBER SIEBER And so that -- if I add the
eight percent and the ten percent and the seven and a
half percent increase in power, that accounts for all
the additional surface that's in there.
MR. DAIBER: Essentially, yes.
MEMBER SIEBER Okay. Thank you. Now, I
have one other question, going back to containment.
What was the containment test pressure?
MR. DAIBER: Sixty-eight psig.
MEMBER SIEBER Sixty-eight.
MR. DAIBER: Yes, 68.
MEMBER SIEBER That's 110 percent of the
design. Okay. Thank you.
MR. DAIBER: Hundred and fifteen percent.
Right, 115 percent.
One other point I'd like to make is that
when analyzed the safety analysis and control system
considerations, we are a CE NSSS Plant, and we
utilized many of the CE Westinghouse methodologies for
performing the safety analysis, core design
considerations. These methodologies that we utilized,
that Westinghouse utilizes are the same methods used
by other CE plants of higher power rating than where
ANO2 is projected to go.
As we've discussed, we did install new
steam generators. Those steam generators were
specifically designed to accommodate power uprate to
ensure adequate margin was accommodated in those steam
generators. Containment was uprated from 54 to 59
pounds. We installed integral flow restricting nozzle
in the CSAS actuation to accommodate the secondary
side inventory associated with that. We did have to
modify the containment cooling fans. The horsepower
requirements in the cooling fans went up above the
motor rating at the 59 pound consideration. To
accommodate that, we reduced the pitch, we lowered the
flow a little bit, brought the horsepower requirements
back down within design considerations. To offset
that effect, out tech specs only required us to have
one containment cooling fan per train. We upped that
to two containment cooling fans per train to offset
that.
MEMBER KRESS: How did you increase your
design pressure?
MR. DAIBER: The design pressure on the
containment, again, we went back and we looked at the
-- did a finite element analysis on the structural
containment design itself and verified that the
structure could maintain additional pressure
associated with that.
MEMBER KRESS: So you reanalyzed it.
MR. DAIBER: Yes.
MEMBER KRESS: Okay.
MR. DAIBER: We did a reanalysis.
MEMBER KRESS: And you needed that for
five pounds?
MR. DAIBER: Not all of the five pounds.
The 59 pounds came more from the Unit 1 design
consideration.
MR. ADAMS: Let me address that. My name
is Doyle Adams. I was on the containment uprate
project itself, also the steam generator replacement,
and then was also the responsible engineer for the
mods and things that was done to the containment and
the repair and the testing of the containment when it
came out.
The way we came up with the amount we
could actually go, which is only about five pounds,
it's about ten percent additional capacity -- Unit 1
is very, very similar. It only lacks in tendons in
some areas due to the design time that it was actually
come in. What happened when we went through and
developed a complete reanalysis of the containment
using the BECHTEL BSAP program they have, which is
used for San Onofre. It was designed for concrete
containment.
There was additional tendons put into the
containment, like we said a while ago. There was
three additional tendons in each grouping for the dome
and the hoop and the vertical tendons. There was
three additional ones for surveillance only. They
added an additional three tendons in each group to
take care of construction problems that you might go
into and then loss of wire with the surveillance
processes over the life of the containment itself.
So you have these 18 additional tendons
that were not accounted for in the original analysis,
and that's where this additional capacity came in.
You also have additional -- use very conservative
creep values for the concrete when it was originally
done, and that allowed us to maintain more of our
compression in the concrete due to less loss of creep
in the concrete.
MR. DAIBER: So we have reviewed the Plant
as a whole and verified that the Plant with design
margin considerations can be operated at uprated
conditions.
With that, I'm going to move down to the
fifth agenda item, dealing with ECCS analysis. I'm
going to switch things up here a little bit. And for
the ECCS analysis, emergency core cooling system
analysis, we analyzed the ANO2, large break LOCA and
small break LOCA considerations using 10 CFR 50.46
Appendix K, approved methodologies. We used approved
methodologies, a combustion engineering, all
Westinghouse methodologies, to do that analysis for
power uprate considerations. These are in compliance
with Appendix K and hence have the conservatism
associated with Appendix K built into them. They are
not the best estimate methodologies that are
available.
We did -- in order to accommodate power
uprate, we did move to the 1999 evaluation model for
the large break LOCA considerations. That was
necessary to ensure that under uprated conditions we
did not have to impose any additional operating
restrictions.
So we did use approved methodologies. The
large break LOCA methodology is documented in CENPD-
132, Supplement 4-P, Revision 1. The small break LOCA
methodology, for that we used the same methodology
that we were currently licensed to -- are currently
licensed to, which is referred to as the S2M. It's
documented in CENPD-137, Supplement 2-PA.
In performing these analyses, obviously we
got acceptable results. We stayed within the
acceptance criteria. The peak clad temperature for
the large break LOCA analysis went up from 2,029
degrees Fahrenheit, which was analyzed with the old
methodology, and it went up to 2,124 degrees with the
new methodology. The small break LOCA, the peak clad
temperature went up from 1,905 degrees to 2,090
degrees. We also verified that the maximum clad
oxidation, the maximum core-wide oxidation and
coolable geometry requirements were also maintained.
MEMBER SHACK: Did you do a full spectrum
of breaks or you analyzed your limiting breaks from
your previous?
MR. DAIBER: We did a spectrum of breaks.
We did a spectrum of large break LOCAs and a spectrum
of small break LOCAs.
MEMBER SHACK: How did that spectrum
compare with your previous analyses?
MR. DAIBER: It was effectively the same
spectrum, a very similar spectrum. The break size
changed on the large break LOCA, so the spectrum
changed a little bit to accommodate that, to make sure
we bounded it on each side.
MEMBER BONACA: For the large break LOCA,
you say you used a new, approved methodology?
MR. DAIBER: That's correct.
MEMBER BONACA: Was it specifically for
this change, to support this modification?
MR. DAIBER: The large break LOCA
methodology was developed not for power uprate. It
was developed generically.
MEMBER BONACA: Okay.
MR. DAIBER: But it was implemented, and
it was necessary. The margin gained by going to the
1999 EM was necessary to ensure power uprate
conditions without any additional operational
restrictions.
MEMBER BONACA: Okay. So you were looking
for some margin there, and this new methodology gave
it to you.
MR. DAIBER: That is correct. Again, the
methodology, though, is still in compliance with
Appendix K --
MEMBER BONACA: I understand.
MR. DAIBER: -- considerations.
MEMBER SIEBER It was --
MEMBER KRESS: What were these values for
the ANO2 without the uprate?
MR. DAIBER: The peak clad temperature for
large break LOCA was 2,029.
MEMBER KRESS: Okay.
MR. DAIBER: For the large break. And for
small break LOCA, it was 1,905. I don't have the
other ones readily available.
MEMBER KRESS: Okay.
MEMBER SHACK: And that's with the same
analysis methodology.
MR. DAIBER: The same break LOCA, yes.
The large break, we switched.
MEMBER SIEBER It was my understanding
that the large break LOCA evaluation model used FLECHT
data, or reflood heat transfer coefficients. Is that
correct? And that was one of the factors that gives
you additional margin?
MR. DAIBER: I'll let Joe Cleary from
Westinghouse address that. He can more appropriately
answer that question.
MR. CLEARY: Yes. The large break
evaluation model does use FLECHT-based reflood heat
transfer coefficients, and one of the improvements we
made going from the 1985 EM to the 1999 EM was to
improve the procedure for applying the FLECHT
correlation.
MEMBER SIEBER Could you tell me about how
much margin you think you gained on a Plant like this
in degrees between old and new --
MR. CLEARY: For that change, I believe it
was a little bit less than 100 degrees on that
particular one.
MEMBER SIEBER Okay.
MR. CLEARY: The sample calculations we
showed in the topical gave a range of 64 to 72 degrees
--
MEMBER SIEBER Okay.
MR. CLEARY: -- for a couple of
calculations.
MEMBER SIEBER Okay. But 100 is a good
number?
MR. CLEARY: I would go a little bit less
than 100. Overall, the change from the '85 EM to the
'99 EM resulted in a change of 150 degrees net.
MEMBER SIEBER Okay. Thank you.
MR. CLEARY: Approximate.
MR. DAIBER: So we have performed the LOCA
analysis and verified acceptable results under
operating conditions.
With that, I'm going to jump back up to
Agenda Item Number 4, which are the review issues, and
with this I'm going to switch things on around here a
little bit again too. I'm going to start out with
ATWS considerations.
ANO2 is a CE-designed Plant, and so our
approach to ATWS is different than that that the
boilers and some of the Westinghouse plants have
considered. Boilers and some of the Westinghouse
plants do credit operator action and perform analyses
to ensure compliance with the ATWS considerations.
ANO2, being a CE Plant, for our compliance with 10 CFR
50.62 ATWS requirements, we installed a diverse and
redundant SCRAM system. We also installed a diverse
emergency feedwater actuation system and took credit
for a diverse Turbine Trip system at the Plant.
For power uprate considerations, we
verified that these systems and their set points and
response times associated with these systems would
still remain valid under uprated conditions to ensure
compliance with the ATWS considerations.
I'm going to move on to the impact of
containment response. We did, obviously, redo the
containment analysis. When we redid that analysis, we
looked at both the steam line break and the LOCA
considerations. The mass and energy that was
generated for that peak building pressure
consideration, they were generated using Westinghouse
CE combustion engineering, Westinghouse methodologies
to generate mass and energy release. That mass and
energy release data is input into the BECHTEL COPATTA
code, which is our containment peak building pressure
analysis code, to get the new peak building pressure
considerations. When we did all this, we did it as
part of the RSG project, and we did it to account and
cover power uprated conditions, and it's all been
approved as part of License Amendment 225 already.
For the LOCA, we did look at cold leg, hot
leg -- cold leg discharge, cold leg suction, hot leg
break considerations. We did look at various single
failures to come up with the limiting LOCA peak
building pressure considerations, and the loss of an
EDG was a limiting single failure. For the steam line
break, we looked at a range of power levels and a
range of single failures associated with this.
And as I mentioned before, we installed
integral flow restricting nozzles in the CSAS
actuation signal to isolate main steam and main feed,
such as the hot zero power steam line break is now the
most limiting break with the single failure of a
spray. The new peak building pressures associated
with the LOCA was 57.6 psig, and with the hot zero
power steam line break, it's 57.4 psig.
As part of compliance with Appendix K
methodologies for peak clad temperature
considerations, we also do a minimum containment
pressure analysis, and that peak pressure was 27 psig,
but that's for Appendix K compliance considerations,
just to show the relative margin between peak building
pressure and minimum building pressure for LOCA
considerations.
With that, I'd like to turn it over to
Dale James for alloy 600 considerations.
MR. JAMES: Thank you, Bryan. Good
afternoon. My name is Dale James. I'm the Manager of
Engineering Programs and Components at Arkansas
Nuclear One. I will be discussing the impact of the
power uprate on our alloy 600 nozzles in the RCS and
on the secondary components due to the flow
accelerated corrosion.
As Bryan mentioned, the power uprate was
made possible by the replacement of the ANO2 original
steam generators with new generators made with alloy
690 tubing, but also with a heat transfer area of
approximately 25 percent greater than our original
steam generators.
By increasing the heat transfer area by
this magnitude, we were able to accommodate the power
uprate with only a marginal increase of the T hot to
609 degrees. Historically, our T hot has run between
600 and 607. Under the power uprated condition, T
cold will be approximately 551, which is actually a
reduction in the T cold from our original cycles of
operation by about two degrees. The pressurizer
conditions will remain unchanged. Temperatures and
pressures there will be consistent with the power
uprated conditions.
Therefore, for the uprate, we evaluated
the effects of the increase in temperature on the
reactor vessel head nozzles and the hot link nozzles.
The increase in T hot for the reactor vessel head
nozzle has been evaluated using the same methodology
as the industry has used to evaluate the conditions
identified in NRC Bulletin 2001-01. That was dealing
with the Oconee 3 circumferential cracking issues.
The methodology is founded -- or is based
upon EPRI Material Reliability Program documents 44
and 48. And this process ranks components based upon
their potential for a primary water stress corrosion
and cracking of the reactor vessel head nozzles. And
that ranking is based upon a plant's operating time,
adjusted for the difference in reactor vessel head
operating temperature using an activation energy
model.
Considering the increase in T hot at ANO2,
the ranking time was decreased for the power uprated
condition from 17.1 EFPY to 14.2 EFPY. With this
reduction, ANO2 remains in what I've characterized as
a moderate category. That is a range of five to 30
EFPY that the bulletin established for reaching a
condition similar to that at Oconee 3.
For this category of plant, the bulletin
recommended that the licensee perform an effective
visual examination of the reactor vessel head nozzles
during the upcoming refueling outage. Due to
constraints that we have with respect to our
insulation design on ANO2, we are unable to perform a
100 percent visual examination of the reactor vessel
head. Therefore, during our upcoming refueling
outage, we will be performing a 100 percent UT
examination from below the head.
With respect to the hot leg nozzles --
MEMBER SHACK: When is that outage, this
spring?
MR. JAMES: This spring. It begins this
April.
For the hot leg nozzles, we will be
continuing to perform a 100 percent bare metal
examination at each of our refueling outages to detect
any signs of leakage. To date, we have replaced nine
of the 19 hot leg nozzles, and those replacements are
performed with alloy 690 material. All the nozzles
below the water line in midloop have been replaced to
date. As I mentioned, we will continue to perform
those examinations in the future to detect any
leakages of any additional nozzles.
MEMBER SHACK: I asked this question
before, and I can't remember the answer. Your surge
line, is that stainless, so do you have 182 butters
anywhere?
MR. JAMES: Yes. Because they're all
shop-welded safe ends, then connected to the stainless
nozzles.
MEMBER SHACK: But that's just for the
pressurizer.
MR. JAMES: Yes. Now, we have other
stainless components. Our reactor coolant pump
casings are stainless also in the cold legs. Those
also have shop-welded safe ends on them and butters at
the shop. Okay?
With respect to FAC, the impact of power
uprate on secondary components were evaluated
utilizing the EPRI CHECKWORKS Program. A parametric
study was performed assuming a maximum steaming rate
under the power uprated conditions. The Check rate
model predicted minimal impact on FAC wear rates.
This prediction is consistent with those that other
utilities have evaluated under power uprate conditions
and is also consistent with measured values following
uprated conditions.
Following uprate, we will continue to
monitor those areas that are most susceptible as a
result of the power uprate condition, and if we see
any deviations from what the model predicted, we'll
factor that back into our modeling for any future
repair and replacement decisions.
MEMBER SIEBER Could you give me an
estimate, from a percentage standpoint, about how much
increase CHECKWORKS predicted for FAC?
MR. JAMES: Yes. What we did was looked
at some of the more susceptible components as were
identified as a result of the power uprated
conditions. What we saw there is probably an average
increase in wear rate of about five mils per year.
That's added on to what we would consider a relatively
low wear rate right now. So we were not anticipating
any major modifications or any major changes in our
wear rate.
MEMBER SIEBER Okay. Thank you.
MEMBER SHACK: Have you done chrome-olly
replacements?
MR. JAMES: Yes. All of our replacement
is done with two and a quarter chrome-olly, which
essentially eliminates FAC wear.
MEMBER SHACK: But how much of your
secondary piping now is chrome-olly or you just do it
as you go?
MR. JAMES: Well, we do it as we go, but
we take a very proactive approach to that. We're
replacing probably on the order $300,000 to $400,000
worth of piping each refueling outage. So we're not
waiting until a system wears to a point where we're on
threat of losing a component.
Okay. In conclusion, our evaluation shows
power uprate will only have minimal impact on both our
alloy 600 nozzles and our FAC wear rate, although we
will continue to evaluate and monitor those systems to
ensure our predictions are consistent.
I'm going to turn it over now to Rich
Swanson in our operation organization.
MR. SWANSON: Hi. I'm Rich Swanson. I'm
a senior reactor operator on Unit 2. I'm the ops lead
for power uprate, and I was also a member of the Steam
Generator Replacement Team.
Training has already started on our new
plant. Simulated changes have been made, and we have
two training cycles that are concentrating on power
uprate. Each crew will be evaluated on an uprated
plant prior to outage. And I'd like to point out, the
changes we're doing for power uprate have much less
impact than those we did last cycle for steam
generator replacement.
Changes to controls and displays have been
minimal or none. We've made no physical modifications
to control stations due to power uprate, and there's
no change in the format or the Safety Parameter
Display System.
We have made about 75 procedure changes
for power uprate, and that includes emergency abnormal
and normal operating procedures. There's been no
change to the type and scope of procedure, and we
haven't had to write any new procedures for power
uprate. As far as emergency operating procedures,
once again, there's no change to type and nature of
actions, and we have added no new actions.
Operations is heavily involved in the
development and implementation of Power Ascension
Testing. We have test teams designated to perform all
the testing coming up out of outage. They'll be
working with the test group. And these are
experienced teams. The operations leads on these test
teams are also involved in the steam generator
replacement testing.
This slide shows our power ascension
profile for coming up out of our next outage. The
first four points are standard for coming up out of
any outage. You have turbine over speed testing and
three points for physics testing. And we'll stop at
90 percent power, which is approximately 98 percent of
our current power level. And they'll be performing
walkdowns, vibration checks, control system checks,
parameter verifications. And we'll make sure
everything is where we predicted it to be before we
increase power.
You see those hold points? About 24 to
48 hours at each hold point. From there we'll go up
in 2.5 percent increments and repeat all the testing.
I'd like to turn it over to Joe
Kowalewski, who will talk more about our Start-Up Test
Program.
MR. KOWALEWSKI: Joe Kowalewski. I'm the
director of engineering, and going to review the
Start-Up Testing Program that we've got outlined.
Our Start-Up Test Program is in compliance
with Test Spec 6.9.1, which requires that we review
against our original start-up testing program as
documented in the Safety Analysis Report. Original
testing for the plant was in compliance with
Reg Guide 1.6.8.
We've gone through the Safety Analysis
Report, reviewed approximately 150 tests specified in
that report. We've also looked at the scope of all
the modifications that were done both for the
replacement steam generator as well as the power
uprate.
We've used industry experience to look at
our Start-Up Test Program. We looked at recent CE
System 80 plants that have started up and reviewed
their test programs. We reviewed the start-up test
programs associated with other steam generator and
replacements in power uprates. And then after we
completed the development of our test program, had an
assessment done with industry expertise, both
combustion engineering in Westinghouse and start-up
leads from other plants that validated our test
program.
We have done extensive start-up testing
for the steam generator replacement already. Much of
that is credited for the power uprate. That includes
post-modification testing associated with each of the
modifications that was performed in the plant;
performance of the components as well as the control
systems; the containment testing for the uprate of
containment, which was the Structural Integrity Test
as well as the Code Test there. And steam generator
performance testing-- both components effects on the
plant as well as performance of the generator itself.
Additional testing we intend to do, we
will as part of our shut-down into 2R15 do a
25 percent load rejection to further benchmark our
integrated control system response. We have tested
each of the control systems, and this will give us
additional data to see if there's any final
adjustments we need to make before we go up further in
power.
And we will be doing the routine
pre-criticality low power physics and power range
testing to validate the core design. So we'll do the
power range testing both at our 90 percent, and then
again when we reach 100 percent in our operating
conditions.
As Rich talked about, we have a overall
work plan for control of the power extension coming
out of the outage. We'll stop at 90 percent, take
extensive data, baseline the plant there, and then go
in 2.5 percent increment above that. As we take the
data both on the primary and secondary side, we'll be
looking and comparing it to our heat balance as well
as all of our design predictions. And it will be
reviewed by a test working group made up of senior ANO
plant management, including the operations manager,
systems engineering manager, design manager, and our
on-site Review Committee chair.
We'll be verifying our heat balance at
each of those points and collecting a wide variety of
key parameters, both on the primary and secondary
side. We'll be doing biological shield surveys at
each point and piping vibration testing both inside
and outside of containment. And inside containment
we'll be using hand-held instrumentation on the feed
water and steam lines.
Once we get up at our operation condition,
we will be doing a moisture carryover test as well as
performance testing of the steam generator. A
question that came up in the subcommittee meeting was
relative to steam quality and effect on the turbine.
Right now, at our current conditions, in the
replacement steam generator we're seeing .02 percent
on one steam generator and .013 on the other, with an
acceptance criteria of .1. That's compared to a steam
quality of .2 percent -- approximately
.2 percent -- for the old steam generator. So the
steam quality is actually an improvement over what we
had before. And we don't expect any negative effects
ont the turbine.
MEMBER SIEBER Do you offhand know what
the turbine rating is for inlet quality?
MR. KOWALEWSKI: The rating.
MEMBER SIEBER A lot of times they're
something like 1 percent. And so, below 1 percent,
that sort of tells you how much margin you have.
Do you know what it is?
MR. KOWALEWSKI: I don't know offhand what
the turbine rating is.
MEMBER SIEBER Who's the turbine
manufacturer?
MR. KOWALEWSKI: It's a GE turbine.
MEMBER SIEBER Okay.
MR. KOWALEWSKI: Vince, do you know that
MR. BOND: I've heard the term 1 percent.
I'm Vince Bond, start-up testing group
supervisor. I've heard 1 percent before from various
design people. I don't know that for a fact myself,
but 1 percent is the term that I've heard.
MEMBER SIEBER I guess it's not very
important. But it looks like you have a lot of
margin.
MR. KOWALEWSKI: Okay. The acceptance
criteria for the test is 1 percent.
MEMBER SIEBER Thank you.
MR. KOWALEWSKI: .1 percent. I'm sorry.
The plant will be verified form --
MR. WILSON: Excuse me. I'm Roger Wilson
with Entergy.
On moisture carryover, the design of the
RSG gave us a lot more volume for feedwater control.
The original steam generators had a conical section
that went into a cylindrical sectional. Now it's
strictly in a conical section. So we've done a lot of
looking at high-level trip, and we have a lot more
margin for that than we had with the original steam
generators. And, of course, the turbines being more
efficient, they're designed for going deeper into the
two-phase dome. And they're probably designed for
that.
MR. KOWALEWSKI: Our test program will
verify that we're performing in accordance with the
design parameters, and we'll document that in our test
report within 90 days of the plant start-up.
Now I'd like to return it to Bryan Daiber,
who's going to talk about the impact of power uprate
on point risk.
MR. DAIBER: I'm Bryan Daiber, again.
For the power uprate considerations, not
only did we look at the safety analysis
considerations, but we also looked at the potential
risk impacts associated with power uprate. And we did
this effectively in several forms. We did quantify
the effects of power uprate on the core damage
frequency and the large early release frequency
considerations.
We also in more of a qualitative fashion,
we addressed the effects of power uprate on the
external events-- seismic, fire vulnerabilities,
tornadoes, winds, failures, transportation accidents
at nearby facilities and awful long shut-down risk
considerations.
For looking at the core damage frequency,
the Level 1 considerations and the impacts of power
uprate on those, we reviewed the initiating event
frequencies, we reviewed the success criteria,
component failure rates, system fault trees, and
operator responses associated with the Level 1 CDF
considerations.
We reviewed all of these and implemented
the effects of power uprate in all of these areas.
The area that was most impacted by power uprate were
the operator responses.
For the operator response considerations,
we did review the operator responses credited in the
Level 1, core damage frequency considerations. To
quantify the impacts of power uprate on those we ran
a CENTS analysis for various sequences.
The CENTS code is a Westinghouse code used
to do the Chapter 15, Thermal Hydraulic Analysis.
When doing that analyses, we ran that code to
determine the time to core uncovery. And we did a
comparable run both at current power rating and at
uprated conditions to determine the different times
associated with each. We then took those times, and
we put them into the human reliability analysis to
come up with a human error rate.
We took those human error rates along with
all the other changes that were necessary with respect
to the success criteria, initiating band frequencies,
and the fault tree considerations. We put those into
a power uprated model. We quantified both the
pre-power uprate model and quantified the post-power
uprate mode, came up with a delta CDF. The delta CDF
was 2.7E-6, which was essentially a 16 percent
increase. This falls within Region 2 or small change
as defined by Reg Guide 1.174.
In a similar manner --
MEMBER KRESS: Is that the same number,
that pre, that you had in your IPE?
MR. DAIBER: No, it is not. Over the
years, we have updated the model several times, and
this value is different than the IPE value.
MEMBER KRESS: Okay.
MR. DAIBER: In a similar fashion, then,
we accounted for the effects of power uprate, and came
up with a change in the large earlier release
frequency, the LERF. The delta LERF was 9.3E-8, which
is a 24 percent increase associated with power uprate.
This fell within Region 3, which was a very small
change from Reg Guide 1.174.
MEMBER KRESS: That LERF is almost two
orders of magnitude lower than your CDF.
MR. DAIBER: Yes.
MEMBER KRESS: Is ANO2 a large dry?
MR. DAIBER: Yeah, it's typical for a
large dry EWR.
MEMBER KRESS: So that's why you get that
kind of --
MR. DAIBER: That is correct.
As I mentioned, we also looked at the
external event considerations-- fire, seismic
considerations, shut-down risk considerations. And
when we did those assessments, we looked to see if
there was anything unique about power uprate. And
doing those assessments, we determined there were no
unique or significant insights to be gained as
associated with the power uprate impacts on the plant.
So in summary, we've looked at the plant
from --
MEMBER POWERS: Are you changing any
electrical equipment at the plant?
MR. DAIBER: Major electrical equipment,
no.
MEMBER POWERS: No transformers are
changed?
MR. DAIBER: No. The transformers
themselves --
MEMBER POWERS: No relay being changed.
That doesn't affect your fire?
MR. DAIBER: No, not in a fire-initiating
event frequency consideration.
MEMBER POWERS: How can it not?
MR. DAIBER: I'm sorry?
MEMBER POWERS: How can it not?
MR. DAIBER: Affect the fire frequency?
MEMBER POWERS: Sure.
MR. DAIBER: Mike, are you aware of the
basis for the combustible loading considerations with
respect to fire?
MR. LLOYD: My name is Mike Lloyd. I'm
the ANO lead engineer in PSA area.
We did a separate fire analysis, and I
don't -- that part of the analysis was done by our
fire protection folks. They did a fire loading. And
the loading itself considered those aspects of the
plant. I don't believe that the increase loading,
however, was explicitly considered. But there are
large, I guess -- degree of conservatism in the fire
analysis that we did perform. We used the five
methodologies, an EPRI method.
MEMBER SIEBER I guess we're talking
specifically about the main unit transformer.
MR. DAIBER: Yes.
MEMBER SIEBER That's located outside?
MR. LLOYD: Right. That is correct.
MEMBER SIEBER Is that away from the
buildings.
MR. LLOYD: Yes.
MEMBER SIEBER Twenty or 30 feet?
MR. LLOYD: Yes.
MEMBER SIEBER Do you have a dike around
it?
MR. LLOYD: I'm not -- yes, there's a dike
around it.
MEMBER SIEBER And it has water
suppression? Automatic water suppression?
MR. LLOYD: Yes.
MEMBER SIEBER Okay. Thank you.
MEMBER POWERS: Have we ever had
transformer fires at nuclear power plants?
MEMBER SIEBER Pardon?
MEMBER POWERS: Have we ever had
transformer fires at nuclear power plants?
MEMBER SIEBER Yeah, I had two of them.
MEMBER POWERS: I mean, it just seems
remarkable to me that we can do an analysis that says
we've increased the power running through the
transformer, and we didn't change the fire frequency.
MEMBER SIEBER Well, the initiation
frequency should change because the linings are
hotter. The potential fault currents, as long as the
breakers continue to be interrupted, don't explode.
That's usually not an issue. But the transformers are
located anywhere from 20 to 50 feet from the nearest
building. And I haven't seen -- even with major
fires, I haven't seen it spread to the buildings.
They seem to get trapped in the diked area.
MR. DAIBER: I would venture to say that
that was one of the screen zones, below the 1 x 7-6.
MR. LLOYD: But the impact of the fire,
because of the exterior location of these
transformers, would cause a loss of off-site power,
yes. And we did evaluate the loss off-site power in
our analysis. But that would be, I believe, the major
effect of such a fire in that exterior to the plant
location.
In addition, the location is very distant
from the safety-related equipment. It's in the aux
building, which is quite distant from the location of
the transformers.
MEMBER POWERS: Well, saying that the only
effect of the fire and the transformer is to increase
the frequency of loss of off-site power is not what I
would call heart-warming. That's usually a fairly
significant accident.
MR. LLOYD: We evaluated that. And I
believe that roughly it represents about 5 percent of
the CDF. It's not a major single contributor. And
BWRs, typically this loss off-site power represents a
much, much larger fraction of their risk.
MEMBER POWERS: It tends to be plants,
specifically.
MEMBER SIEBER Typically, in a PWR, you
have two buses fed from the system and two fed from
the main unit. If you blow the transformer, then you
lose two of the four, and then they automatically
cross-connect.
Is that the way your plant is --
MR. LLOYD: Our unit has two divisions.
Should we lose off-site power, what would happen is we
would use on-site diesels, one emergency diesel
powering each of the emergency buses. And in addition
to that, quite distant from the transformers we have
another unit that's a station blackout unit which was
installed for the station blackout rule. And it is a
stand-alone island of power basically dependent on
only itself for all intents of purposes. It has its
own DC system for starting. It can be started from
the control room. It has its own air cooling system.
It's totally independent of service water. So it sits
there ready to be used from the control room.
MEMBER SIEBER Okay.
MEMBER KRESS: Are there two units on that
site?
MR. DAIBER: Yes.
MEMBER KRESS: About the same power level?
MR. DAIBER: Unit 1 is slightly lower.
Thermal is 2856.
MEMBER KRESS: Is it the same kind of
reactor and containment?
MEMBER SIEBER No.
MR. DAIBER: No. It's 25 --
MEMBER SIEBER It's a BMW.
MR. DAIBER: BMW.
MEMBER KRESS: It's a BMW.
MR. DAIBER: With that, we've looked at
the plant both from a design capability standpoint to
make sure all the components could operate properly
within the design criteria for the plant. We also
looked at it from a risk perspective and verified from
a risk perspective the plant and their operating
conditions were acceptable.
With that, I'd like to turn it over to
Craig Anderson for concluding remarks.
CHAIRMAN APOSTOLAKIS: Is the staff going
to make a presentation, Jack?
MEMBER SIEBER Yes.
MR. ANDERSON: All right. Let me start
off by thanking this committee for your time this
afternoon. I'd just like to close by saying, our
focus has been throughout this project -- as it should
be -- in keeping the plant safe and reliable. And
through analysis, through modifications, through
training, we believe that we've done that.
Our plant, and our equipment, and our
people are ready for the power uprate. And if there's
not any additional questions, that concludes our
portion of the presentation.
MEMBER SIEBER Okay. Well, I thank you
and your staff and Entergy for putting together a good
presentation that we can understand. And it appears
to me like you have done a lot of work to get this
unit ready to run at a higher --
MR. ANDERSON: Yes, sir.
MEMBER SIEBER Thank you very much.
MR. ANDERSON: Thank you.
MEMBER SIEBER What we'd like to do now is
have the NRC staff come forward. And as they get set
up for their portion of the presentation, which will
discuss the Safety Evaluation Report, I would like to
introduce to you someone we haven't seen for several
hours, which is John Zwolinsky, who seems to show up
for every operation.
MEMBER POWERS: He just can't stay away.
We're so kind to him that --
MEMBER SIEBER So when you folks are all
set, you can begin.
MR. ZWOLINSKY: Give us just a couple
minutes. Thank you.
MEMBER SIEBER All right. No problem.
MR. ZWOLINSKY: Can I get started?
MEMBER SIEBER Yes.
MR. ZWOLINSKY: Great.
Good afternoon. To those of you that
don't recall who I am, I'm John Zwolinsky, director of
Division of Licensing Project Management in NRR.
Joining me today are our management team and
first-line supervisors that have overseen the review
of the Arkansas power uprate. I'd like to take a
minute to identify those folks. They're here in
support of our staff. And, certainly, we have a large
number of staff here to answer any of the questions
that may go beyond the agenda.
I'd first like to recognize Ms. Suzanne
Black, our deputy director for the Division of Safety
and Systems Analysis; Richard Barrett, our branch
chief in the PRA Branch; Stu Richards, our project
director for PD4.
We have a number of our section chiefs,
our first-line supervisors. Bob Graham out of PD4.
Frank Akstulewicz of Reactor Systems Branch will be
making a presentation; Ralph Gruso of Reactor Systems;
Kamal Manoly of Mechanical Branch; Brian Thomas from
Plant Systems; Matt Mitchell from our Materials
Branch; Corney Holden from our Electrical and
Instrumentation Control Systems Branch; Louise Lund
from our Materials Group; Mark Rubin from our PRA
Group, at the table.
I feel it's important to ask the staff to
join me for meetings such as this. We place high
emphasis on bringing these to closure. And as you
know, the Commission has placed a high degree of
importance on power uprates in general, and I
appreciate our staff being here with me.
The staff is here to present its review of
the 7.5 percent power uprate for the Arkansas
Nuclear 1 Unit 2 plant. The staff made a presentation
on this review to the subcommittee on thermal
hydraulic phenomena on February 13, 2002.
The ANO2 uprate is the largest extended
power uprate for PWR we have reviewed to date. The
staff has conducted a thorough review of the Arkansas
plant, focusing on safety. Reviews were conducted
consistent with existing practices, which include the
license arm from Main Yankee.
We used the Farley power uprate as a
template for this particular review; that is, all the
sections of Farley dictated the sections that we would
review here. Scope and depth were driven to some
extent by our standard review plan for various
sections. We'll talk about that in greater detail
throughout the presentations.
All areas affected by the power uprate
were reviewed by the staff. The staff has critically
examined the methodologies in their application for
these power uprate requests, and concluded that the
analytical codes and methodologies used for licensing
analysis are acceptable for these applications.
Without further ado, I'd like to turn it
over to Tom Alexion. Tom is our project manager for
this plant and has shepherded this entire project from
beginning to end.
Go ahead and get going, Tom.
MR. ALEXION: Thank you, John.
Good afternoon. I'm Tom Alexion. And I'm
the NRC project manager assigned to Arkansas.
By way of background, the 7.5 percent
power uprate application by Entergy represents the
largest PWR uprate to date, as you heard earlier. The
highest PWR power uprate previously approved was
5 percent.
Some background into the CE designed PWR.
The architect engineer and constructor were BECHTEL.
The full power license was issued on September 1,
1978. And the current license maximum reactor core
power level is 2815 megawatts thermal. And it to has
a large dry containment.
The steam generator was replaced in the
fall of 2000. Some of the differences between the old
and new steam generators are shown in this slide. The
licensee designed the replacement steam generators to
accommodate the increase in reactor power.
I would also like to note that when we're
doing the power uprate application, the NRR staff
relied upon analysis previously done at the uprated
power in support of steam generator replacement. And
this is in the fall of 2000.
The NRR staff used the following power
uprate as a guide for the scope and depth of its
review. To further review guidance, the standard
review plan is utilized. The staff have used their
licensee's application of codes and methodologies to
ensure that they are used within the appropriate
restrictions and limitations, and to ensure they're
appropriate at the higher power level. During the
course of the review, the staff issued many requests
for additional information. The licensees responded
to all of them.
For the containment, the staff had a
contractor perform independent calculations of the
pre-containment pressures and temperatures following
a postulated LOCA and main steamline break. In the
area of vessel materials, the staff performed
independent calculations of the pressurized thermal
shock reference temperature and end-of-life upper
shelf energy for each reactor pressure vessel
material, and performed independent calculations on
the susceptibility to vessel-head penetration
cracking.
In the area of dose assessment, the staff
performed independent calculations of the atmospheric
dispersion for the exclusionary boundary and low
population zone, and the dose assessments for the
LOCA, steam generator tube rupture, CEA ejection, and
fuel-handling accidents.
In the area of risk assessment, the staff
audited the licensees risk evaluation for power
uprate, which included manipulating various parameter
in the human reliability analysis spreadsheet, and did
an independent calculation to gain a perspective of
the seismic risk.
The principal areas of review are the NSSS
and accident analyses, evaluation of structure,
systems and components, BOP systems, human factors,
radiological analyses, and risk assessment.
But for today, the order of presentation
is as shown. We plan to present these four areas.
And we're also going to show -- we have some examples
where the staff focused this review.
When they were issued the draft safety
evaluation, the only open items were in the
radiological assessment area. But these items have
been resolved. So, therefore, the NRR staff has no
open items.
And with that, those are my opening
remarks. We can move to reactor systems, unless there
are any questions.
MR. ZWOLINSKY: Frank Akstulewicz is our
section chief responsible for this area. Chu Li Yang
is senior staff reviewer.
MR. AKSTULEWICZ: Thank you. My name is
Frank Akstulewicz. I'm the section chief in the PWR
section of reactor systems. And to my right is Chu li
Yang, who is the lead reviewer for this particular
power uprate.
What I'd like to do is jump to Slide 2.
Slide 2 identifies in general terms the areas of
review that we focused on, and I'd like to make a few
remarks about each of the bullets.
The first bullet specifically looked at
the design operating characteristics and requirements
for reactor coolant system ECCS and shut-down systems.
And as you've heard today, there were very few
modifications, if any, other than the steam
generators, to these systems in order to support the
power uprate. So our effort here principally was
examining what the operating requirements were,
verifying that the analyses supported those operating
requirements, and confirming that the analysis was
done using acceptable methods.
The second area, fuel system design.
Again, this particular power uprate does not use a new
or different fuel type. It's the standard CE 16 x 16
array. The only thing different here is the poison.
In this particular case, you've heard the licensee
that particular effort. Thermohydraulically, it's no
different than anything else that's used in other CE
plants of higher power level. And analytically, we've
done a number of accident evaluations using this fuel
and have found no problems.
The last area, the LOCA and transient
area, again, principally here we look at the specific
initial conditions and assumptions used to assess the
accidents. We look to make sure that the codes that
are being used to assess those accidents are
appropriate for application, and whatever the terms
and restrictions are in those codes have been
satisfactorily either resolved or complied with. Then
we look at the results to make sure that from our
experience and familiarity with how these transients
should occur, whether the results are anomalous or
not. And depending on the outcome there, we either
pursue further information from the licensees or we
verify that it satisfies the acceptance criteria
that's established for that particular transient, and
approve the safety evaluation.
The two examples on the bottom will be
more specific as two the actual nature of the review
and some example -- maybe to the level of detail that
we went into some of the assessments.
So with that, Chu will take over.
MR. YANG: My name is Chu Li Yang. I'm
the reviewer for the Arkansas Unit 2 Power Uprate.
And I'm going to discuss the staff review of feedwater
line break and LSS performed by the licensee to
support its power uprate.
As a part of power uprate, the licensee
revised it's feedwater line break and LSS methodology.
This slide presents some of the changes to the
methodology and initial conditions and assumptions to
perform it's power uprate reanalysis. However, the
principal changes to the methodology
involves -- proposed the use of low-level water trips
at point associated with affected steam generator in
their new analysis. And the previous low-level
analysis -- low level trip of in tact steam generators
was used instead of the affected steam generators.
And this methodology calculates the
limiting feedwater break size by concurrent high
pressurizer, pressure trip and the low steam
generator -- what level trip affects the steam
generator in the new analysis.
The change of the methodology slight
resulted in a reduction in margin. And also, the
reanalysis assumed uprated power level. The changes
also in the areas of initial conditions and
assumptions, such as a high initial pressurizer
larger, in mill steam safety of tolerance and early
mill steam isolation. And those conservative
assumptions were added to provide safety margin in the
new analysis.
For review of changes in methodology, we
accept everything of the changes and will be discussed
next.
The acceptability of the revised
methodology used in the new feedwater line break
analysis is reviewed in the following steps.
First we consult INC staff regarding
accuracy of the low-level trip set point on affected
steam generator in the new analysis. INC staff
concludes that the instrumentation of certainty
calculations were acceptable for this application.
Also, we look at the documentation for the NOTRUMP
computer code, and find that the code was initially
reviewed with ability to evaluate the steam generator
water level behavior and stability and damping
predictions during feedwater line break dynamic
conditions. And the staff concluded, NOTRUMP is
acceptable for those specific areas of calculation.
And inputs found in NOTRUMP computer code from the
simulator steam generator -- and provide input to
system transient code for primary system response
simulation.
Finally, the approach used to taking
credit for low-level trip affected steam generator is
currently used in Westinghouse plants. And this
approach has been approved in WCAP-9230 for
Westinghouse steam generators. And based on those
facts, we conclude that the approach of new level
water trip in affected steam generator to giving
credit for feedwater line break is acceptable for ANO
Unit 2.
We would like to discuss the impact of the
revised methodology.
The first bullet of this slide lists the
major impacts found in the revised methodology. In
the new analysis, the licensee indicated that the
limited break size is slightly reduced from
approximately .17 square feet to approximately
.15 square feet. And the reactor trip react early
during the transient. In the steam generator, the
water inventory is increased at the time of the
reactor trip. And those changes result in slightly a
reduction in margin. But the calculated peak
transient primary and secondary pressures are slightly
reduced. It's reduced to 2647 PSIA, and the previous
analysis was 50 points higher.
The result of this analysis met all
acceptance criteria specified in the SRP. And the
peak primary and secondary pressures remained below
10 percent of the line pressure. The pressure lines
would not go solid, and the DMB is not a concern for
this event.
Next, we'd like to discuss the staff
review of control or the withdrawal from subcritical
conditions. This event is classified as AOO. And the
staff acceptance criteria for this event does not
allow fuel damage. The safety limit in existing text
specs is the peak linear hit rate limit, less than
21 kilowatts per foot.
The licensee's power uprate analysis shows
the transient linear hit rate approximately
40 kilowatts per feet, which is about the existing
text specs. However, the impact of this high linear
hit rate limit is very limited by a short transient
duration of less than two seconds. And the calculated
peak fuel center line temperature is approximately
2800 degrees. And the results of the analysis shows
all SRP acceptance criteria are satisfied with respect
to fuel performance, fuel pressure and fuel center
line temperature. And the licensee has revised the
text specs to remove linear hit rate limit and include
a fuel center line temperature limit value adjusted
for fuel burn, which is roughly --
MEMBER POWERS: Would you say that again,
please?
MR. YANG: The text specs have been
revised. And the existing text specs, the linear hit
rate limit is eliminated, and instead, the fuel center
line temperature is defined as a limit for this event.
It's consistent with SRP.
MEMBER POWERS: How does it change with
burn up?
MR. YANG: It's adjusted for fuel burn up.
MR. AKSTULEWICZ: The fuel
adjustment -- the temperature is actually decreased
with burn up. It declines -- I think -- well, there's
a proprietary restriction on actual number for CE.
But I can say that for Westinghouse plants, it's
approximately 50 degrees per 10,000 megawatt days per
ton.
MEMBER POWERS: Why do we think that's an
adequate thing?
MR. AKSTULEWICZ: It defines the point at
which the fuel centerline actually will begin to melt
for this particular -- for these kind of power
insertion events. It's a calculated value that's part
of the design basis.
MEMBER POWERS: How do they calculate the
melting point of the fuel?
MR. AKSTULEWICZ: It looks at the rate of
reactivity insertion and the energy deposition within
the fuel itself, and then does a temperature
calculation, looks at the heat up of the fuel.
MEMBER POWERS: Yeah. But when does it
melt? I mean, how do we know when it melts?
MR. AKSTULEWICZ: The -- it's assumed to
melt when it reaches a certain temperature. And that
temperature is based on experimental data that the
fuel vendors have.
MEMBER BONACA: This is a bank withdrawal,
right? Not a single rod.
MR. AKSTULEWICZ: No, this is single rod
withdrawal.
MEMBER BONACA: A single rod.
MEMBER SIEBER You mean a single rod out
of its bank configuration, right?
MR. AKSTULEWICZ: Yes. This is a single
rod being looped.
MEMBER SIEBER Otherwise, you can't
go -- a single rod by itself won't do anything.
MR. AKSTULEWICZ: That's correct.
MR. ALEXION: Okay. If there's no further
questions, we'll move on to the plant systems branch
with Dave Cullison and Rich Lobel.
MR. CULLISON: Good afternoon. I'm Dave
Cullison from Plant Systems Branch. With me is Rich
Lobel also from the Plant Systems Branch. I perform
the majority of the reviews of the power uprate. Rich
did the containment reviews that were done as part of
the replacement steam generator and containment uprate
of the project.
My two slides I want to discuss just show
the SRP sections we used in the performance, we used
as guidance for completeness and accuracy. We
determined in all our reviews that there's no
significant impact on the system operations through
the power uprate. And this is just the continuation
slide.
Rich is going to discuss the independent
confirmatory analysis done for the containment
response to the power uprate.
MR. LOBEL: Richard Lobel from Plant
Systems Branch. As part of the replacement steam
generator review, we contracted with Los Alamos
National Laboratory to do a
calculation -- confirmatory analysis of the
calculations done by the licensee for the peak
temperature and pressure for both a LOCA and a
steamline break. They used the MELCOR code to do the
calculation. But it was a designed basis calculation,
so it didn't really exercise most of the models in
MELCOR.
The analysis looked at, like I say, both
the LOCA and the steamline break, and in general
agreed with the licensee's analysis. The one area
where there was a large discrepancy between the
analysis was in the case of the steamline break, the
licensee calculated a much more conservative
temperature than we did. And after discussing it with
the licensee, the licensee suggested that it might be
an assumption they made for containment spray. They
assumed a very low efficiency, very low heat transfer
from the atmosphere to the spray. And MELCOR used
pretty much a physical model of the spray. We went
back and adjusted the spray model and got fairly good
agreement with the licensee's calculations.
When I talked to the subcommittee, I said
that the report on this was available in ADAMS, and
everybody laughed. So let me just say now that it's
in ADAMS.
That's all I have, unless there's any
questions.
MR. ALEXION: Okay. We'll move on to the
Materials and Chemical Engineering Branch, and Barry
Elliott will be the presenter.
MEMBER SIEBER Before we get to that, I'd
like to ask what the ultimate heat sync is at -- is it
a lake or river or --
MR. CULLISON: They have two. They have
a pond, which is their -- and they also have the
Dardinel Reservoir.
MEMBER SIEBER Okay.
MR. CULLISON: The one with -- that Rich
reviewed, ultimately heat sync evaluations as far as
steam generator replacement in the pond.
MEMBER SIEBER Okay. Thank you.
MR. ELLIOTT: I'm Barry Elliott with
Materials and Chemical Engineering Branch. This slide
shows all the areas within our branch that we review.
The first six items I'm not going to go over today.
I'm going to go over the last three, which I think are
the most significant, which is the reactor vessel
integrity, steam generator tube integrity, and the
Alloy 600 Program.
Before I go on, are there any questions
about the first six items? No.
The Alloy 600 Program is intended to take
the primary water stress corrosion cracking of
Alloy 600 and Alloy 182 wells and the reactor pool and
piping, the pressurizer and vessel head penetrations.
Cracking in vessel head penetrations were the subject
in NRC Bulletin 2000 and '01. PWRs were ranked by
their MRP, according to the operating time and
temperature, and effective full power years required
for the plant to reach the effective time and
temperature corresponding to the Oconee 2 event, where
they had crackings -- circumferential cracking in
their Alloy 600 head penetrations.
Plants with high susceptibility to primary
water stress corrosion cracking are those which are
predicted to have a ranking of less than five
effective full power years from the Oconee 3
condition. Plants with a moderate susceptibility to
primary water stress corrosion cracking are those
which are predicted to have a ranking of more than
five effective full power years and less than 30 full
power years from the Oconee 3 condition. Depending on
which ranking you are determines which inspection
program you're involved in.
In the case ANO2, before the uprate, they
were in the moderate category, and after the uprate,
they're still in that category. The uprate increases
the T-hot temperature from 604 to 609. Increase in
T-hot will not substantially increase primary water
stress corrosion initiation and growth rate; however,
it does affect the ranking somewhat.
Potential for primary water stress
corrosion cracking developing in Alloy 600 nozzles
will not be significantly affected by the power
uprate, and, therefore, there is no change in the
Alloy 600 and the vessel head penetration inspection
program as a result of the power uprate.
MEMBER POWERS: Somehow this uprate will
increase T-hot form 604 to 609, and then say that
won't increase the primary water stress corrosion
cracking initiation and growth rate didn't strike me
as quite what you mean here.
Don't you mean that, though they have a
T-hot going from 604 to 609, that's not what the
temperatures of the head --
MR. ELLIOTT: There's two issues here.
There's a head issue, and then there's a piping issue.
MEMBER SIEBER They're different.
MR. ELLIOTT: The intent of that lip was
the piping and pressurizer issue.
MEMBER POWERS: Oh, okay.
MR. ELLIOTT: The head is -- it has a
lower temperature than the head -- than the piping and
the pressurizer.
MEMBER FORD: I've got no quarrel at all
with what you put down there, except that it is, as
Dana intimated, fairly qualitative. And although
you're quite right, it still remains in the moderate
range, that temperature time, erraneous type metric
that is being used is pretty rough. The times are
still fairly short in absolute terms -- 5, 15
years -- compared with license-renewal time schedules.
During your thought on this -- during your
analyses of this -- was there any quantification along
these lines?
MR. ELLIOTT: Well, the quantification
is -- the purpose of the Bulletin 2000 and '01 is to
determine the inspections that are going to be
occurring at the next refueling outage.
MEMBER FORD: Correct.
MR. ELLIOTT: So the whole point of it is,
is to get how susceptible your plant was to decipher
cracking. If you were very susceptible, then you had
to do some more inspection. And it was just -- based
upon the models that were developed as part of the
MRP, you were put into different categories. And that
was the intent, to determine what kind of inspection
is required at the next refueling outage. The model
itself was developed from data of material crack
growth.
MEMBER FORD: Yeah. But pretty well,
every plant which is in the first category has, in
fact, shown cracking.
MR. ELLIOTT: Right.
MEMBER FORD: So can we expect cracking on
this plant within the next three years?
MR. ELLIOTT: Well, according to our
model, it won't be.
MEMBER SHACK: Well, we have cracking at
Millstone, right, at 14 years.
MEMBER FORD: Right.
MR. ELLIOTT: It could. I mean -- when
they do the inspection --
MEMBER SHACK: But they're going to do the
inspection.
MR. ELLIOTT: -- we'll find out. They're
going to do a volumetric inspection, which should be
able to detect these cracks.
MEMBER FORD: As part of a process, I'm
asking, this plant will crack. And it will crack
before the end of its life.
During your reasoning, does that come at
all into your arguments?
MR. ELLIOTT: You mean that the plant will
eventually crack?
MEMBER FORD: Yeah. I mean, is it a thing
that comes into your --
MR. ELLIOTT: I think the issue is -- as
you say, it will eventually crack --
MEMBER FORD: Sure.
MR. ELLIOTT: -- and it will crack before
the end of 40 years probably. But that becomes -- as
long as we have an inspection program that is capable
of detecting the cracks before they become critical
and affects the integrity of the reactor coolant
system, that's all we're looking for. We're looking
to make sure that's there.
MEMBER FORD: But that important fact is
not set out there. And that's reassuring, your saying
that. Again, for the public confidence aspect, it's
useful to have that enunciated.
MR. ELLIOTT: Well, we're relying on an
inspection program to detecting these cracks before
they become critical.
MEMBER FORD: Right.
MR. ELLIOTT: And that's what the model is
intended to do, to lay out what we suspect to be the
worst plants, and that they need more inspection than
the less susceptible.
However, this plant, even though they're
in the moderate category, is still doing a volumetric,
which is very good.
The next slide deals with reactor vessel
integrity. Just a quick background for you who are
not knowledgeable.
10 CFR 50 establishes a Scharpey
upper-shelf screening criteria. And 10 CFR 50.61
establishes RTpts screening criteria for pressurized
thermal shock.
The licensee has made evaluations of the
upper-shelf energy and the RTpts values, and they're
done in accordance to Reg Guide 1.99, Rev 2. For this
plant, the materials have a low rate of brittlement.
The upper-shelf energy is predicted to drop even with
a power uprate of only 60 foot pounds. And the RTpts
value is around 120 only, which is 150 degrees below
the screening criteria in the pts rule.
The staff reviewed these calculations. I
want to point out also, we also reviewed in the
previous slide the Alloy 600. We did our own
susceptibility calculations.
We did the calculations here for the
upper-shelf energy and the RTpts values. In addition,
Appendix G requires pressure temperature limits. And
from those pressure temperature limits, low
temperature over-pressure set points are determined.
These limits in set points were provided in a separate
application and are being reviewed by the staff to
ensure that they meet all regulatory requirements.
Based on these analyses, the reactor vessel meets all
regulatory requirements.
As far as steam generator integrity, the
Alloy 690 tubes are more resistant to stress corrosion
cracking than the Alloy 600 tubes. Degradation of
tubes resulting from the deposition of copper was
eliminated by removing copper from the secondary side.
We've done analysis of vibration or frequency
responses of antivibration bars, minimized wear. Reg
Guide 1.121 analyses were performed to ensure
structural integrity. And based upon this analysis
and the changes in the system, there is no change in
the tube inspection program required at this time.
That completes my presentation today.
MR. ZWOLINSKY: Thank you, Barry.
MEMBER POWERS: When you say there's no
need to change the tube inspection program, you mean
that there's no need to increase it, right? That's
all you looked at.
MR. ELLIOTT: I --
MEMBER POWERS: You need to look at the
possibility --
MR. ELLIOTT: The inspection program is a
text spec item, and is a certain program they have to
follow. This will not change that.
MS. LUND: Right.
MEMBER POWERS: You didn't look at the
possibility that they could increase or decrease their
inspection.
MS. LUND: It wasn't considered under just
the power uprate situation. We're evaluating that
separately under the NEI-97-06. And we're still -- as
you know, we're still evaluating that.
MEMBER POWERS: Okay.
MR. BOEHNERT: Can you identify yourself
for the record, please?
MS. LUND: Oh, I'm sorry. It's Louise
Lund of Component Integrity and Chemical Engineering
section.
MR. BOEHNERT: Thank you.
MR. ZWOLINSKY: If I might play off your
interest in the bulletin that Barry alluded to. We
continue to receive information from licensees
conducting inspections. They are finding cracks. And
our challenge going forward are our next steps. And
this program matures over the next couple of
years -- you're probably aware, many licensees have
committed to head replacements. And the concept or
thought of inspecting at every cycle seems to be not
the best answer. So we still have our challenges
before us. But as we go forward through the spring
outages, it may be appropriate to come back to the
committee and give you a status report essentially one
year later, so to speak, with the fall outages having
taken place.
MEMBER POWERS: An issue I'd like to know
more about is what is the risk importance described in
the vessel head.
MR. ZWOLINSKY: The vessel head or the
independent CRDMs?
MEMBER POWERS: Either one or both.
MR. ZWOLINSKY: Part of the basis of the
bulletin when it was developed was the lost of one of
the CRDMs.
MEMBER POWERS: Yeah, I understand the
bulletin. I guess I'm asking the probablists in this,
if I tried to guide a risk importance parameter for
the vessel head -- who are the CRDMs housings from a
PRA -- what number would I get?
MR. BARRETT: This is Richard Barrett.
I'm with NRR staff.
We have been looking at this question of
the risk significance of the CRDM cracking issue. And
clearly there are two important questions. One is,
for any given situation, for any given head at any
given time, what's the probability that it would
result in a LOCA. And I think we're talking about a
medium LOCA. And then the second question is what is
the conditional probability that that LOCA would then
lead to a core damage accident, and then possibly only
to a LERF, large early release.
The second part of the equation is the
easier part. You can look that up in most PRAs, and
it's of the order of conditional probabilities of 1 in
1,000. The first part, however, is much more
difficult to assess. And it has to do with your
perceptions as to the initial conditions of the head,
of a particular CRDM, in terms of the probability that
a crack exists, the size of the crack, the depth of
the crack, and then the crack growth rate. And the
type of analysis that is required is not that
different from the kind of analyses that we've been
talking about in the context of 97-06 for the steam
generator tubes.
In the fall of this year, we went through
a lot of what I'll call qualitative analysis in trying
to resolve -- make our regulatory decisions with
regard to the operation of the high susceptibility
plants, and proposals that were made by various
licensees as to the schedule for when they wanted to
shut down.
But I think as we go forward, we need to
get a better handle on this. And we're working with
our Office of Research who are developing and refining
models for crack initiation, crack growth, and how
that relates to the probability of a catastrophic
failure. It's not an easy question to answer, but
we're working on it.
MEMBER POWERS: Good.
MR. ZWOLINSKY: Okay.
MR. HARRISON: Good afternoon. I'm Donny
Harrison. I was the lead for the PRA part of the
review.
We can just move to the second slide.
This slide just identifies the -- I think you've heard
this before a number of different times, primarily
with BWRs, but we look at the internal events,
external events, shut-down operations. And we look do
a look at their PRA quality. We do that to see if
there's any insights and just to confirm that there's
no new vulnerabilities being created as part of a
power uprate.
MEMBER POWERS: Let me ask you what the
significance of looking at the IPEs and the IEEEs for
this plant is. The previous speaker told us that he
modified his plant, and PRA all over the place. So
why would you bother to look at the IPE?
MR. HARRISON: Often times you'll see in
IPE and IPEEE either a statement -- an example of that
would be the seismic area for Arkansas. They do a
seismic margins analysis. In the process of doing
that analysis they make assumptions that they've fixed
things. We come back, and I take a look at that, and
I then send a request for additional information to
the licensee and say, did you fix it?
The thing we found out at Dresden was, in
one area, no. That's worth knowing. For Arkansas,
the answer was yes. Everything we took credit for
that we used in that analysis we've now fixed, and we
fixed it the way we said we were going to fix it. And
it meets the assumptions of the IPEEE, so it gives you
really some confidence that the IPEEE now actually
reflects the plant that's there.
MEMBER POWERS: But here we know that the
IPE does not reflect the plant that's there.
MR. HARRISON: Right. The IPE even still
may say during a technical evaluation, the staff found
weaknesses in initiating event frequencies. I think
one of the comments that was made on Clinton was that
it was a new plant, and they didn't have a whole lot
of plant-specific data. So you can look and see what
has the plant done in response to what the IPE or
IPEEE found.
In a way, it's kind of a way to check to
make sure that plants are improving their analysis and
not just using the same old analysis and staying with
it, not changing.
CHAIRMAN APOSTOLAKIS: You said the magic
words, improving the analysis. I think -- I mean, you
are not responsible for people's models and so on.
But, unfortunately, your silence may be misunderstood.
I see here in Section 8 a fairly detailed
analysis of the operator actions that affect it. And
that's the safety evaluation.
MR. HARRISON: Right.
CHAIRMAN APOSTOLAKIS: And the licensee
says that they used three EPRI reports to come up with
human error probabilities. And you have a table here
where, for example, for failure to reenergize such and
such and such from SD2, the pre-power uprate available
time was 42 minutes, and the HEP was .19, and the
post-power uprate available time was 39
minutes -- three minutes down -- and the HEP was
2.9 x 10-1. And then you go on and have a very nice
discussion of how you really wanted to make sure that
there were no other operator actions that were left
out and not evaluated, and I think that's very good.
The thing that bothers me, though, is that I don't
think there is a model anywhere in the world that can
tell the difference between 42 minutes and 39 minutes
and produce a number like 2.9 x 10-1.
Now you are very carefully here saying,
the staff finds, based on the information provided by
the licensee and the staff site review, that the
licensee's human reliability analysis application is
consistent with their identified methodologies-- a
beautiful statement. It says nothing. Right?
But then you go on and say, and that the
assumed increases in the HEP values for the identified
operator actions reasonably reflect the reductions in
the times available for the operators to perform the
necessary actions.
Now, I don't know how you've gotten it.
I suspect you're right. But you didn't get it from
the EPRI methodologies. Now, a minor reduction in
time tells me that the performance of the operators
are expected to be more or less the same as it was
before. But to say in a table that the number went
from 1.9 x 10-1 to 2.9 x 10-1, I mean, is an illusion.
MR. HARRISON: Right.
CHAIRMAN APOSTOLAKIS: And I would expect
you to say that this methodology -- I mean, find nice
words -- that these methodologies are not widely
acceptable; they have not been approved by the NRC.
You know, something to that effect. Because, frankly,
they are not widely acceptable. That's why this
agency has spent a lot of money trying to develop
ATHENA. That's why the French are spending a lot of
money developing MERMOS, the Fins are spending a lot
of money developing something else. If EPRI had done
it, we wouldn't be doing this.
So I think your silence on this may be
misconstrued by other people. Now, I realize it is
not your job to evaluate human reliability models, but
you should not accept uncritically results such as
this one.
Now your sentence here is really
beautiful, but I would expect it to say something more
than that. The fact that something is consistent with
some methodology, the numbers, I mean, what does that
tell me? Not much. Although, the ultimate
conclusion -- I mean, this is the day where the
conclusions seem to be reasonable, but the models that
led to them are terrible. Not terrible. Not
terrible. You know, they're still in evolution.
I think your conclusion is okay, that the
times probably are not affected that much, and the
human error probabilities are probably the same as
they were before. But to go ahead and produce a delta
CDF of 3 x 10-6, I just don't believe that. If the
major input of this calculation is these human error
probabilities, I don't believe it.
Now, is it much larger than that? I don't
believe that either. Should you deny their request?
Based on this, no. I'm not saying that either. Okay?
And what perplexes me is that this is not a
risk-informed application. So whatever you're
presenting here really does nothing, does it?
But I just can't let it go. This is a
difficult situation here. I don't think the licensee
should be penalized for this. But, you know --
MR. HARRISON: I'm glad you bring it up.
Because just as an analyst, I get concerned when we
focus too much on the numbers, and we don't sit back
and say what did the plant learn from all this. If
we're just focused on did the number go from .1 to
.2 --
CHAIRMAN APOSTOLAKIS: Well, if you had
written it that way, I would be much happier. Because
I really appreciate the difficulty that you're in.
Your job is not to evaluate at-risk models or whoever
models. But if somebody says, I used these models,
and here are my numbers, and you say nothing, then, I
mean, we have a problem there.
MEMBER KRESS: But 1.174 says you have to
come up with a number.
CHAIRMAN APOSTOLAKIS: Well, I'm sorry.
But that's not the number.
MEMBER KRESS: How would you have come up
with a number is my question.
CHAIRMAN APOSTOLAKIS: I couldn't. I
mean, if you don't have a model, why should you come
up with a number no matter what? You just don't have
it. Maybe you can give a bounding value, change the
attitude completely and say, look, I don't have model,
but I don't think that such and such and such. But to
say I use this model because --
MEMBER SHACK: But isn't that what the
result is saying, is it didn't change all that much?
CHAIRMAN APOSTOLAKIS: Yes.
MEMBER SHACK: Maybe you don't believe
either number or the notion that it didn't change all
that much, is what you're --
MR. HARRISON: And it becomes a relative
decision, not an absolute.
CHAIRMAN APOSTOLAKIS: But my problem is
that, if this is not in there, the next guy will say,
oh, they used the EPRI methodology; that's pretty
good. The staff didn't say anything.
MEMBER SHACK: But now you know --
CHAIRMAN APOSTOLAKIS: What?
MEMBER SHACK: Now you know why none of
these applications are ever risk --
CHAIRMAN APOSTOLAKIS: Why don't we just
eliminate all the risk references? I don't know what
all this means.
MEMBER BONACA: But it's remarkable. You
go from 122 minutes to 113 minutes, and they have a
distinct difference in number. How you figured that
out, I don't know.
CHAIRMAN APOSTOLAKIS: Yes.
MR. HARRISON: That's just an
analytical -- it's an analytical exercise.
CHAIRMAN APOSTOLAKIS: It's not use of the
concept of model.
MR. HARRISON: All right.
CHAIRMAN APOSTOLAKIS: So I don't know
what to say. On the one hand it doesn't matter; on
the other hand, you know, it's a document of the
agency.
MEMBER KRESS: Well, when you have a
LERF --
CHAIRMAN APOSTOLAKIS: Tell me. I mean,
why is this agency spending all this money developing
ATHENA if one can pick up the EPRI reports and do
this? Why? Because there's a different group?
MR. HARRISON: We already took care of
ATHENA.
CHAIRMAN APOSTOLAKIS: I'm completely
confused now. I mean, we spent more than a million
dollars.
MR. HARRISON: We already took care of
ATHENA.
CHAIRMAN APOSTOLAKIS: Huh?
MR. HARRISON: We already took care of
ATHENA.
CHAIRMAN APOSTOLAKIS: Because of this.
Anyway, you understand where I'm coming
from. I mean, I'm not criticizing you, because that's
not your job. Well, maybe a little bit I am. Better
words.
I mean, I thought this was brilliant.
"The licensee's human reliability analysis application
is consistent with the identified methodologies."
Brilliant.
MR. HARRISON: And that's about all I can
say.
CHAIRMAN APOSTOLAKIS: It sounds good, and
it says nothing.
MR. HARRISON: Okay, enough.
MEMBER SIEBER Yes, why don't we move on.
MR. HARRISON: Okay.
The bottom line, though, to answer your
question, is as our review, the only -- if you want to
say the only value is, is it's a negative review of
looking for is there an issue out there that's going
to come up on some plant down the road -- it didn't
happen here -- that puts us into an adequate
protection question.
CHAIRMAN APOSTOLAKIS: And I think this is
a very good point.
MR. HARRISON: And at that point -- and I
would say, if a plant like Turkey Point came in that
did the five methodology too and got a real high
number, and they've got a high IPE value -- I don't
know what their PRA number is now -- we'd want to look
at that.
CHAIRMAN APOSTOLAKIS: Actually, the part
that you did where you really questioned whether there
were additional operator actions that the licensee did
not address and so on, that was really nice. That was
really nice. I think you did a good job there. It's
the quantification that bothers me.
MEMBER KRESS: Well, in terms of
quantification, the fact that the LERF, whether you
believe the bottom-line number or not, is around 10-7,
tells me that you've got a pretty good plant here.
CHAIRMAN APOSTOLAKIS: I agree with that
too. Because even if you increase it by a factor of
100 --
MEMBER KRESS: That's right.
CHAIRMAN APOSTOLAKIS: But I would much
rather see something like that than saying, is EPRI
such and such.
MEMBER KRESS: So I didn't pay much
attention --
CHAIRMAN APOSTOLAKIS: Well, I am. I am
paying attention.
MR. HARRISON: No, I appreciate the input.
Because, again, like I said, one of the concerns I
have as an analyst is an overfocus on trying to get
the precise number and worrying about did the CDF go
up by 1 percent, when the ultimate answer is adequate
protection, and am I up at 10-3.
With that, actually, I really won't bother
to go on.
CHAIRMAN APOSTOLAKIS: I'm sorry that I
had to say all these things.
MR. HARRISON: Oh, that's okay.
CHAIRMAN APOSTOLAKIS: It wasn't you.
MR. ALEXION: That concludes this staff's
technical presentation. I just have one last slide
I'd like to show. And that is our conclusion. We
felt we've done a thorough review, they're extensive.
We spent a lot of time on RAIs, a lot of information's
been communicated. We don't have any open items. We
feel the application meets applicable regulations.
And the NRR staff recommends approval of the power
uprate application.
MR. ZWOLINSKY: And I'd like to also take
just a minute to thank the committee for this
opportunity to present our review of the Arkansas
extended power uprate.
As you've heard, the vast number of
sections and the review areas that the staff has
addressed and the independent analysis performed is to
me quite impressive. And I trust the committee finds
it the same way. I'd certainly recommend approval for
this particular power uprate. Thank you so very much.
MEMBER SIEBER Thank you and members of
your staff. I read the SER more than once, and I
found that it was pretty well organized, which I think
in part is because of the existence of the Farley SER
in the work that have been done by the applicant. And
it was pretty easy to read. And I thought that it was
important to tell us about confirmatory calculations
and the analysis that you did so that we can
appreciate that the SER is not a rubber stamp; that
it's actually an independent analysis and confirmatory
calculations. And to us that's important. It allows
us to be able to see what the basis is when you say
that this plant is satisfactory or the requested
amendment is satisfactory.
So if there are no questions from the
members at this time, Mr. Chairman, I give it back to
you.
CHAIRMAN APOSTOLAKIS: Thank you very
much, Jack.
We'll recess until 3:40.
(Whereupon, the foregoing matter went off
the record at 3:27 p.m.)
Page Last Reviewed/Updated Monday, July 18, 2016