Thermal-Hydraulic Phenomena - September 26, 2001
Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title: Advisory Committee on Reactor Safeguards
Thermal-Hydraulic Phenomena Subcommittee
Duane Arnold Energy Center Power Uprate
Request
Docket Number: (not applicable)
Location: Rockville, Maryland
Date: Wednesday, September 26, 2001
Work Order No.: NRC-033 Pages 1-177
NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers
1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005
(202) 234-4433 UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
+ + + + +
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
THERMAL-HYDRAULIC PHENOMENA SUBCOMMITTEE MEETING
DUANE ARNOLD ENERGY CENTER POWER UPRATE REQUEST
+ + + + +
WEDNESDAY
SEPTEMBER 26, 2001
+ + + + +
ROCKVILLE, MARYLAND
+ + + + +
The ACRS Thermal Phenomena Subcommittee
met at the Nuclear Regulatory Commission, Two White
Flint North, Room T2B3, 11545 Rockville Pike, at 1:00
p.m., Dr. Graham Wallis, Chairman,
presiding.
COMMITTEE MEMBERS PRESENT:
DR. GRAHAM WALLIS, Chairman
DR. F. PETER FORD, Member
DR. THOMAS S. KRESS, Member
DR. DANA POWERS, Cognizant ACRS Member
DR. STEPHEN ROSEN, Member
DR. WILLIAM SHACK, Member
DR. VIRGIL SCHROCK, ACRS Consultant ACRS STAFF PRESENT:
PAUL A. BOEHNERT, ACRS Staff Engineer
I-N-D-E-X
AGENDA ITEM PAGE
Introduction . . . . . . . . . . . . . . . . . . . 4
Duane Arnold Power Uprate Presentation . . . . . . 7
Concluding Remarks . . . . . . . . . . . . . . . 176
P-R-O-C-E-E-D-I-N-G-S
(1:00 p.m.)
CHAIRMAN WALLIS: The meeting will come to
order. This is A meeting of the ACRS Subcommittee on
Thermal-Hydraulic Phenomena. I am Graham Wallis,
Chairman of the Subcommittee.
Dana Powers will be the ACRS Cognizant
Member for this meeting. Other ACRS Members in
attendance are Peter Ford, Thomas Kress, Stephen
Rosen, and William Shack. The ACRS Consultant in
attendance is Virgil Schrock.
The purpose of this meeting is for the
subcommittee to review the license amendment request
of the Nuclear Management Company fora core power
uprate for the Duane Arnold Energy Center.
The subcommittee will gather information,
and analyze relevant issues and facts, and formulate
the proposed positions and actions as appropriate for
deliberation by the full committee. Mr. Paul Boehnert
is the Cognizant ACRS Staff Engineer for this meeting.
The rules for participation in today's
meeting have been announced as part of the notice of
this meeting previously published in the Federal
Register on September 19th, 2001.
Portions of this meeting may be closed to
the public as necessary to discuss information
considered proprietary to General Electric Nuclear
Energy. Please let us know if and when that is the
case.
The transcript of this meeting is being
kept, and the open portions of this transcript will be
made available as stated in the Federal Register
notice. It is requested that speakers first identify
themselves, and speak with sufficient clarity and
volume so that they can be readily heard.
We have received no written comments or
requests for time to make oral statements from members
of the public. I have a brief opening comment.
The ACRS, before this meeting, received
stacks of paper which amounted to over a foot in
height. We obviously don't have time to read and
digest every word.
So I think that it is very important that
the speakers focus on what issues the ACRS needs to
consider and what information we are going to need to
reach decisions on those issues. And I believe that
Dr. Ford has a statement to make.
DR. FORD: Yes. I am a GE retiree, and
therefore I have a conflict of interest.
CHAIRMAN WALLIS: Now I would like to ask
my colleague, Dana Powers, to take over my job for a
while and to run the meeting.
DR. POWERS: Thank you Professor Wallis.
We are going to be looking at one of the first of the
major power updates that we seem to have coming along
the pike here. This is a truism that the boiling
water reactors in this country typically operate at
powers that are less than what they were originally
conceived of operating at.
And in part that was because of a historic
-- a long time ago many ACRS' before this current
version of it had particular concerns about DWR
stability at the higher power.
What we are going to try to cover is a
huge amount of material. Professor Wallis' is over a
foot, and he must have only gotten half of it if he
only had a foot.
CHAIRMAN WALLIS: Over a foot. I was
being conservative.
DR. POWERS: And the plan of attack is
that we are going to listen to the applicant this
afternoon, and then tomorrow we are going to listen to
the staff tell us why we should have believed
everything that was told to us from the applicant.
And so I am going to turn now to Ron
McGee, the power uprate project manager, to start the
presentation, and you will introduce the additional
speakers as the need arises.
And remain cognizant that should we have
to deal with proprietary material, that creates a huge
disruption. So you have to let us know beforehand.
MR. MCGEE: Good morning then. My name is
Ron McGee, of the Nuclear Management Company at the
Duane Arnold Energy Center. I would like to thank the
committee for taking the time to review our submittal
and for meeting with us today.
We recognize the importance of power
updates as part of the solution to meeting the
country's future energy needs, but foremost we must
ensure that public safety is not jeopardized.
We believe that through our engineering
evaluations and the staff's review process the DAEC
application for a power uprate has shown an adequate
amount of operational design and safety margin for the
various facets of the project.
Today, we have been asked to present the
following topics. I will be presenting the plant
changes and modifications, and then we will talk
quickly about the regulatory compliance, the analysis
performed as part of the project, and then we have
been asked to discuss margins, which I will get to
here in a minute.
Then we go through the operator training
that we have applied. Then we will have discussions
on thermal hydraulic stability, the ATWS response, and
ATWS instability fuel response, material degradation
issues, the containment analysis, the effects of power
uprate on the steam separator and dryer, ECCS analysis
as part of the project.
And then the last presentation is the PRA
analysis, and then we will have concluding remarks and
a wrap up of any open issues that come up.
CHAIRMAN WALLIS: Did you rehearse this so
that it can be over in two hours?
MR. MCGEE: Two hours, no. We have --
CHAIRMAN WALLIS: You are supposed to
allow two hours for our questions.
MR. MCGEE: We have accounted for four
hours, including questions. We believe that the
presentation material, including questions, should be
concluded within four hours.
DR. POWERS: It is going to be so clear
that that we will have no questions whatsoever.
MR. MCGEE: The first presentation will be
where we go over the power uprate modifications that
we performed as part of this project. The safety
related modifications -- and I will point out that
these are the only safety related modifications that
were necessary to accommodate the power uprate.
These were installed in our recent outage,
and we installed new APRM cards, installed higher
range main steam line flow implementation, and we
through a previous amendment, we have increased our
required boron concentration for a standby liquid
system.
The balanced plant modifications. We
installed higher capacity transformer; coolers --
improved cooling capacity on our hydrogen coolers for
our main generator.
A major modification was that we replaced
the high pressure turbine, and the feed water level
control for the feed water heater system had to be
modified to accommodate the higher capacity. As part
of the ELTR, we have installed flow induced vibration
monitoring.
CHAIRMAN WALLIS: What does Phase One
mean?
MR. MCGEE: As part of our uprate, we
intend to go up from 1658 megawatt thermals, our
current license power level, and we intend to operate
at 1790 megawatt thermal. And then following a future
outage, we plan to ascend to the rest of the license.
CHAIRMAN WALLIS: So you are applying for
the whole thing?
MR. MCGEE: That's correct. We are
applying for the license to 1912 megawatts thermal.
But a balance of plant modifications will only
accommodate operation up to 1790 for this interim
period.
MR. BOEHNERT: What are the percentages,
Ron? Do you know roughly?
MR. MCGEE: Approximately halfway each
time; 8-1/2 percent right now, and then another 8-1/2
percent on top of that. Feed and condensate pump
breaker, protective relaying set point, condensate the
demineralizer capacity as part of the feed water flow
stream.
And the main condenser tubes because of
the increased steam flow, and added structural
support.
CHAIRMAN WALLIS: I actually have to agree
on the language. I noticed in this staff review that
they are talking about 120 percent increase?
MR. MCGEE: The 120 percent that is from
original license --
CHAIRMAN WALLIS: No, no, no.
MR. MCGEE: Oh, increase.
CHAIRMAN WALLIS: A twenty percent
increase.
MR. MCGEE: Yes.
CHAIRMAN WALLIS: The staff was talking
about 107 percent, and it is really mind-boggling.
MR. MCGEE: That would be Unit 2.
DR. KRESS: You went by one of our slides
a little too fast. You talked about whether one of
the mods was a MELLLA APRM card, and I know what the
MELLLA is and all of that, but his this card an
automatic controller to make sure that you go along
the MELLA line? What is the card?
MR. MCGEE: The card actually monitors
your flow by SCRAM set points, and supplies the trip
function into your RPS system, reactor protection
system.
DR. KRESS: Okay. That's what I thought,
but I wanted to make sure.
MR. MCGEE: That's correct. The next
slide. Continuing with our balanced plant
modification; isophase bus temperature monitoring for
the electrical load increase; and monitor the
temperature.
And the main steam line relief value
snubber was one support that we needed to increase.
One of our feed water heaters was going to have a
significant increase in its load carrying capacity,
whereby bypassing the flow to the main condenser.
And control room indications and alarms
have been modified to accommodate the previous
modifications that you have seen here.
Phase Two, when we go from 1790 up to
1912, preliminarily, we have identified feed water
system capacity, and we will need to increase the
system capacity from about 8.1 million pound mass to
something just greater than 8.75 million pound mass
per hour.
Feed water heaters. Their load bearing
capacity will need to increase, and so we are
anticipating the need to increase various feed water
heaters. And then our isophase bus to carry the
increased electrical loading.
MR. ROSEN: You said the increased feed
water, and the slide says replacement. Are you going
to replace the heaters?
MR. MCGEE: We do plan to replace feed
water heaters, certain ones.
MR. ROSEN: But not all of them
MR. MCGEE: But not all of them, that's
correct. Some will be at the upper -- increased at
the EPU power level, and will be within their design
to carry the amount of increased loading, but the
three, four, and five heaters -- we have six heaters,
and the 3, 4, and 5s right now looks like they will be
marginal. So we will be looking at a wholesale change
outs of those.
MR. ROSEN: Those are the low pressure
heaters?
MR. MCGEE: Those are high pressure
heaters.
MR. ROSEN: And the low pressures are all
right, but the high pressure heaters need to be
changed?
MR. MCGEE: The ones and twos are the
lowest pressure, yes, that's correct. Those are okay.
MR. ROSEN: So you are saying that you are
looking at it. It is curious language. You are
looking at replacing them.
MR. MCGEE: Yes. We are planning to
replace all of the heaters. We are looking at designs
and depending on how and which ones you replace, that
will determine the need to replace others. Next
slide.
The next topic was regulatory compliance.
Our application was a deterministic application, and
it is not a risk-informed application. It was
performed in accordance with previously approved ELTRs
1 and 2.
The process for the application and the
studies includes a feasibility study, which was
conducted in late 1999. Engineering evaluations
throughout the year 2000, et cetera. The licensing
reports, which you have seen most of, I believe, if
you have gotten a foot of paper.
The hardware modifications that we have
just reviewed, and then post-approval, and we have
testing to perform, and we have performed preliminary
testing up to our current license power uprate.
CHAIRMAN WALLIS: You said this was not a
risk-control, and yet one of the major consequences
here is the operator reaction time during ATWS. One
of the major concerns in that seems to be resolved on
a risk basis rather than some sort of compliance.
MR. MCGEE: We do have a presentation that
will include discussion of that topic if you would
like to wait for that.
CHAIRMAN WALLIS: Okay. You will address
that at that time?
MR. MCGEE: We will.
MR. MCGEE: The analysis performed. The
general topics were the reactor operating conditions,
accidents and transients, the radiological
consequences of a power uprate, component system
capacity, including NSS and BOP; instrumentation and
controls; the environmental impact of the power
uprate.
And then a review of the station programs.
For instance, PSA, environment qualifications, station
blackout, et cetera.
The generic topic of margins. We were
asked to discuss that. And what we have done is
included a discussion in the rest of the presentations
today to address the impact on margins on the specific
topics.
And then if the committee would like to
follow on with questions during those times, I would
propose that is how we address those. And next will
be Steve Kottenstette.
MR. KOTTENSTETTE: Hi. I am going to be
talking about operator training. My name is Steve
Kottenstette, and I am an operations shift manager on
loan to the power uprate project.
We started out using classroom training.
We took the material from the power uprate project and
we discussed with the operators how the procedures are
going to change, and how the tech specs will change,
and the testing that we will do as far as the power
uprate probe.
And then we moved into the simulator,
where we took what we believed would be the best guess
on how the plant will operate, and use a simulator,
and through the various operational transients that we
would see a trip over recirc pump, a trip of the feed
pump, turbine trips, reactor SCRAMS.
And then we also went into some of the
accident scenarios, where we went through an ATWS and
showed the operators the benefits of injecting standby
liquid control early on in the scenario, and then
showed what would happen if we didn't inject standby
liquid control, or had a failure to inject.
We did a turbine trip and SCRAM scenario,
and then we also did an MSIV closure, and did show the
operators how much it did change. And for the most
part, there was very little change as far as our
actions and how the plant responded once a plant was
shut down.
DR. SCHROCK: Could you embellish a little
on what you mean by your best guess as to how the
plant is going to perform?
MR. KOTTENSTETTE: We took the model or
the design information of the plant modifications, and
the change out of the conset pumps and the feed pumps,
and the reactor model, and we basically gave it our
best guess on how it should respond.
And then we also benchmarked it against
the accident analysis that we got from GE, as far as
this is how the plant should respond to a turbine
trip, or an MSIV closure.
And we looked to see how the simulator
responded, and it pretty much matched up to what we
saw or the information that we got from the analysis.
CHAIRMAN WALLIS: Do you have any sort of
feedback for how well the operators responded to this?
I mean, you trained them and you would tell them these
various things, and then it is supposed to change
their performance in some way or their reaction at the
time, or whatever.
Do you have a measure of how well they did
after training?
MR. KOTTENSTETTE: During the training, it
was obviously observed during the training that the
operators responded per our procedures, and as far as
containing the scenario and responding to it, there
was no marked decrease in operator performance.
CHAIRMAN WALLIS: So you don't try to
measure the probability of them doing the right thing?
This is somehow only an analytical assessment?
MR. KOTTENSTETTE: As far as how the
operators responded, or --
CHAIRMAN WALLIS: There are numbers that
are given to us in the paperwork about the probability
of human error during an ATWS, and the numbers have
increased.
And I just wondered if there was any
measure from this training to show whether or not the
operators were under more pressure, and made more
mistakes or whatever with the power uprate than you
would get from the simulator experience.
MR. KOTTENSTETTE: We didn't see any
increased errors.
CHAIRMAN WALLIS: But did you look for
any?
MR. MCGEE: This is Ron McGee once again.
The operators during their training scenarios have
critical tasks that they have to perform in their
dynamic scenarios, and the operators on the power
uprate tasks all successfully performed the critical
tasks.
There were no operator failures or
remediations necessary during the power uprate phase
of the test of the classroom or simulator training.
MR. ROSEN: I think we will hear later
that the time for action, taking the required actions,
is shortened by this uprate, but that what you are
saying is that the operators were able to take the
necessary actions in spite of the shortened times
allowed.
MR. KOTTENSTETTE: Yes, because there is
a very simple process of actually initiating standby
liquid control.
MR. ROSEN: It is all right there in front
of them, the keys and the mode switch?
MR. KOTTENSTETTE: The keys and the mode
switch, and take it out of the mode switch and put it
in the switch for the controls for the pumps.
MR. MCGEE: And we specifically monitored
for that, for the operator taking those appropriate
actions on standby liquid. And all of the currently
licensed crews were able to perform that action
satisfactorily.
MR. ROSEN: And I think we will hear more
about that in the PSA discussion I assume?
MR. KOTTENSTETTE: Yes, you will.
MR. MCGEE: Yes, that's correct.
CHAIRMAN WALLIS: I'm sorry, but you have
your schedule. When are we going to get into other
questions that are not on your plans, such as the
stresses in the components and a question that we
might have about something that is not in your
outline? Do we leave that to the end?
MR. MCGEE: That is an option.
DR. SHACK: It seems to me that we ought
to bring them up when they are appropriate, Graham.
CHAIRMAN WALLIS: Well, if they never
raise the issue, we are going to have to bring in up
sometime.
MR. KOTTENSTETTE: There is a section on
material issues, and that may be the appropriate time
to do that.
CHAIRMAN WALLIS: Well, I had a very basic
question, which is that this is going to be a constant
pressure power uprate?
MR. KOTTENSTETTE: That is correct.
CHAIRMAN WALLIS: And I just wondered why
the stresses went up in things like the main closure
flange, the vessel and the head if there were no
changes in pressure? And they go up by 10 percent or
more than 10 percent, and why is that?
MR. MCGEE: Gary is looking. We have
several people looking.
CHAIRMAN WALLIS: Well, maybe we can after
the break come back to this question. That is a very,
very basic question that I had to raise at some time.
So, think about it.
DR. POWERS: It had to be a flow. It
can't be anything else.
MR. BROWNING: Carrying on, next we are
going to talk about the thermal-hydraulic stability,
and my name is Tony Browning, and I am from Duane
Arnold.
We will have General Electric and Jason
Post here to my right giving part of the presentation;
and Mr. Kottenstette said he will get back up and talk
about the impact on the operations of the power plant.
So we will kind of go through a little bit
of quick background and the calculational methodology.
Then I will get back up and discuss the analytical
results. Then we will get into some of the issues
that the committee has raised about the Solomon
monitoring system.
And then the operational aspects as I said
by Mr. Kottenstette, and then we will have a quick
conclusion. Our purpose here is to demonstrate to the
committee that we have adequate operational and safety
margin at the EPU conditions. So with that, I will
turn it over to Jason.
MR. POST: Good afternoon. My name is
Jason Post, and I am with GE. Stability solutions.
The general design criteria is that GE C12 does
neither prevent or reliably and readily detect and
suppress reactor instabilities.
Duane Arnold has stability option 1-D,
which does both. So they have features that both
prevent and detect, and suppress. Their prevention
feature is in an exclusion zone in the power flow map.
It is down in the low flow and high power
corner of the power flow map. It is defined with the
frequency -- the main model, and it has a very
conservative decay ration margin, .8.
So, of course, we wouldn't expect an
oscillation to grow and continue to grow until the
decay ratio was 1.0 or higher. But they have a .8
decay ratio, and so there is margin built in right
there.
And they have a buffer zone outside of the
exclusion region. Of course, as an exclusion region,
you cannot operate in that region at all. If you
enter that region, you have to immediately exit.
The buffer zone is five percent of power
and flow outside of that region, and so that would be
even a more conservative, and even a more lower decay
ratio value for that.
And you can go into that if you are sure
that you have a low decay ratio, and that is from the
SOLOMON code. It is an on-line monitor based on the
ODYSY code, and that is how you ensure that you
maintain a low decay ratio in that region.
CHAIRMAN WALLIS: This exclusion zone is
based on theory isn't it at this stage? You have not
built these cores with flux and higher power, and so
we don't yet know when oscillation is actually going
to occur with these cores do we?
MR. POST: Well, we have had instabilities
in cores, and so we have a pretty good idea of where
they occur, and the most recent one was at the
Columbia Power Plant back in about 1995, I believe.
And so we have seen them, and we have a
pretty good idea. We benchmarked those cases with our
models, and so we have a pretty good idea.
DR. KRESS: But those aren't at the power
densities that you are talking about, at the power
levels and the flows. They are at the original
values, right? The question is does that instability
region expand or change?
MR. POST: Right, and we do expand it
based upon the same models.
DR. KRESS: You use the models to expand
it?
MR. POST: Right. The key factor is
really the highest license rod line. So the change to
EPU is not as significant as the change from ELLLA to
MELLLA.
In fact, we already have plants that are
operating with MELLLA. So, it is not a significant
extension of the methodology. It is consistent with
what we have already operated plants at.
CHAIRMAN WALLIS: And presumably when you
start up this plant, you use some sort of a warning
that the are exclusions that we have calculated, and
if you get close to it, you had better be observant.
MR. POST: Exactly, and Steve -- and later
in here we have a start-up map to show how they go
outside the region and discuss how the SOLOMON code is
used as they come up outside the region. So that is
a specific part of this in a couple of more slides.
CHAIRMAN WALLIS: So you might have to
modify SOLOMON based on what you actually observe, or
is that independent?
MR. POST: The modification is simply in
terms of the inputs to the code to make sure it knows
what the operating conditions are. The code itself
doesn't have to be modified.
So that is the prevention features. The
detect and suppress features. It is important that
the oscillation for Duane Arnold is proven to be only
a core-wide mode, and so the entire core is going up
and down at the same time; as opposed to a harmonic,
where you get a side to side.
And if you had a side by side, then the
average power tends to be flattened out, and your
APRM, which is your average power range monitor, gives
a relatively flat response.
But if it is a core-wide mode, then the
oscillation is easily picked up by the APRMs. So that
existing hardware is where we demonstrate that that
existing hardware does provide adequate protection of
the safety limit.
MR. ROSEN: And you are sure that the
Duane Arnold core will respond in a core-wide mode
because of its tight neuronic coupling?
MR. POST: That's correct, and that is
part of the demonstration; that the core-wide remains
the predominant oscillation mode.
MR. ROSEN: Will you say more about why it
is tightly neutronically coupled?
MR. POST: We will say a little bit more,
and we will make sure and see if you have any more
questions about that.
MR. ROSEN: Okay.
MR. POST: Next slide. So the prevention
methodology is the ODYSY code. When the stability
solutions were initially developed, we used the FABLE
code, which was another frequency to the main model,
and ODYSY is just a much better code.
It was initially applied for another
stability solution, the enhanced Option 1-A solution,
which is a prevention solution. And so we extended it
to Option 1-D, and the SER for that was just issued in
April.
So the Duane Arnold extended power uprate
is the first application, and Duane Arnold is
operating in Cycle 18 right now for their current
license power, with stability regions based on the
ODYSY code.
It is important to note that when we
licensed ODYSY that we replicated the results with
FABLE, and the way that we did that was by adding an
additional .15 margin in the decay ratio criteria. So
while we say there is a .8 criteria, really .65 is
what is being calculated by ODYSY as our limit for the
exclusion region.
And then we add .15 just as an adder to
get the .8 that we are using on the stability
criterion now. And as I said before, ODYSY is also
the basis for SOLOMON. Next slide.
So this is the stability criteria map, and
in the lower right-hand side is where you have a high
channel decay ratio, and a low core decay ratio.
That is the type of condition where you
can get a channel flow instability, and the fuel is
specifically designed to avoid or to ensure that that
cannot happen.
And then in the upper left-hand side, that
is where you have a high core decay ratio and a low
channel decay ratio, and that is where you get a core-
wide mode instability.
And then where the cupus is taken out in
the upper right-hand corner, that's where you have
relatively high core and channel decay ratio, and
that's when the higher harmonics can cause a regional
mode instability.
So we use the 0.56 as the stability
criteria and as the dividing line between when a core-
wide mode and a regional mode instability can occur.
And this stability criteria has been around for a long
time.
It goes back to the time when the LaSalle
instability happened or before, and it has been
supported by various tests and events to show that
that is the difference between when core-wide and
channel instabilities occur.
So we don't do a separation type
calculation, for example. We don't go into all those
other arguments. We just use that as long as the
channel decay ratio is below .56 when the core decay
ration exceeds .8, that that is a basis for
demonstrating that core-wide is the predominant mode.
And I will turn it back over to Tony.
MR. BROWNING: And this gets into some of
the things that we have talked about, and we will walk
through them fairly quickly. The development of the
exclusion zone power to flow map was a critical piece
of this.
Jason just discussed the confirmation that
we were still having core-wide oscillations only, and
our maximum channel decay ratio was a value of .36,
which you can see is well below the .56 acceptance
criteria that you just saw on the stability maps.
So we are still predominantly core-wide.
We also integrated the new flow bias trips from going
from ELLLA to MELLLA, and you will see that has the
biggest impact on the results. Not so much the power
uprate itself, but it is more driven by the change to
MELLLA.
And then we go through the confirmation
that the flow bias SCRAM at the MELLLA level will
protect the safety limit, minimum power critical power
ratio in the fuel, and that is a critical part of the
analysis.
And then we will go through a little
comparison of pre-EPU and EPU results so that you can
see that change. Okay. Steve. Here is the power to
flow map that we have been talking about, and as Jason
mentioned, the exclusion zone is the area where we are
not allowed to operate in steady state, and it is in
the high power low flow portion of the power flow map.
And one of the other things that I would
like to point out --
CHAIRMAN WALLIS: That zone has boundaries
that go all the way around it. It is above the red
line?
MR. BROWNING: You are not allowed to
operate in this region. You have to state on this
side of the line. The exclusion zone is inside --
CHAIRMAN WALLIS: Up there?
MR. BROWNING: In here. And the other
thing that we would like to point out is we keeping
talking about the change from ELLLA to MELLLA. The
black line is the ELLLA current load line.
And then you see the impact of MELLLA.
And while it is fairly dramatic at the top end at
rated power, down here in the stability region, you
will notice that it is not that dramatic, and that
really explains later why you are not seeing a huge
change when we go from ELLLA to MELLLA, or going into
power uprate.
And that's because we are talking about
this area down here, and you can see that the impact
is not that big. Next slide. And what we do is we go
through the APR flow by flux trip, and like we said,
going from ELLLA to MELLLA, we raise all the trip
points up consistent with that.
And then the impact, of course, is that by
doing that we have moved slightly further away for the
SCRAM. So when we get into the oscillation
calculations, and look at the hot bundle oscillation
magnitude at the H BOP, it takes just a little bit
longer for the automatic SCRAM to terminate the
oscillation, because we have moved that much further
up away from that corner of the power flow map with
the automatic trip.
And that is the predominant thing that we
see. Looking at the impact, and I know that this
slide is a little busy, but we are trying to show that
all the changes we made aren't that dramatic.
When you look at where the current power
level rated exclusions and buffer zones are, and how
they just shift slightly with the uprate. You will
notice that they are almost anchored on the natural
circulation line, because almost nothing changes
there, and slightly greater sub-cooling has a slight
effect down here.
But the biggest impact is shifting up from
the ELLLA to MELLLA point. You just take it and you
drag it over.
CHAIRMAN WALLIS: Well, I think you ought
to explain some things to me. I mean, if I am
starting up a plant, and I have no flow and no power.
I am at zero. How do I get here?
MR. BROWNING: We are going to show you
that.
CHAIRMAN WALLIS: Because it looks to me
as if you can't get there without -- you know, I don't
see any reason that you are allowed to operate in it.
MR. BROWNING: Steve will show you that
shortly. We have a power flow map that actually shows
that.
CHAIRMAN WALLIS: But that lower code, the
one you call natural circulation.
MR. BROWNING: Yes.
CHAIRMAN WALLIS: You have to be to the
right of that, or the left of it, or what?
MR. BROWNING: We are going to be to the
right of it. We are going to come up this way.
CHAIRMAN WALLIS: Then you have got to be
below that other black line haven't you?
MR. BROWNING: Yes, we have to clear the
feed water protection line here on the recirc pumps.
CHAIRMAN WALLIS: And you have to be below
that?
MR. BROWNING: Yes. We have to be above
it.
MR. MCGEE: That is at minimum.
MR. BROWNING: We have to clear this line
before we can increase recirculation flow. So we pull
rods and heat up the plant on minimum pump speed, and
clear this interlock, and then we can increase in flow
and go in this direction.
And then continue to pull control rods out
and go up on this side of the exclusion zone. And we
will show you a power flow map where we have an
example of that.
DR. KRESS: And then when you get up on
the MELLLA line, that is strictly a flow power. You
don't have to pull the rods out there anymore?
MR. BROWNING: Right. You get as high as
you can on the load line, and then you just cruise on
up with the recirculation flow. That's correct.
Now we are going to talk about SOLOMON a
little bit because we had some inquiry from the
committee about the SOLOMON software, and the
stability monitor. As Jason has already said, it is
the ODYSY model.
It has been integrated into the plant core
physics monitoring software. It is an integral part
of it. It does not run separately. It runs with it,
and it takes its input from it.
Its purpose? It is a backup for what are
called power shaped controls, because back in the
original days of this stability, one of the things
that was of a concern was that the power shapes that
were modeled in the bundles, how did we know that we
were going to stay in that operating environment, and
be bounded by what was assumed in those analysis?
So we have what are called backup power
shaped controls. So other options use things like
boiling boundary. For the Option 1-D plants like us,
we use the SOLOMON software and the buffer zone as our
backup power shaped control to maintain that margin.
And what it does is that it merely allows
us to sustain operation in the buffer zone and the
power flow operating maps. So that little band that
we were talking about adding on of the 5 percent, when
SOLOMON is available, the operators are allowed to
transgress through that area and go through it.
If the SOLOMON software is not available,
it becomes an extended exclusion zone, and they are
not allowed to operate there.
MR. ROSEN: Is the ODYSY code actually
running in the background, or is it SOLOMON looking up
results, and pre-store the results of ODYSY?
MR. BROWNING: It is actually running, but
it is not real time. It takes a while to do the
calculation. So it takes its input from the core
physics program, and then runs its time domain
calculation, and comes out with the decay rations.
MR. MCGEE: Frequency domain.
MR. BROWNING: I'm sorry, frequency domain
calculations, and comes up with a decay ratio, and
this displayed to the operators.
MR. ROSEN: So it is possible for the
operators, if they are moving very quickly, to outrun
ODYSY and SOLOMON, or how do you prevent that?
MR. KOTTENSTETTE: You mean as far as --
you know, if I change the power real fast, then --
MR. ROSEN: If ODYSY sees you changing
power, it goes back and tries to calculate the new
outputs, but you have changed again before it ever
catches up with you.
MR. MCGEE: Are you talking about a
predictive capability?
MR. ROSEN: yes.
MR. BROWNING: That is one of the things
here. It is a predictive capability. You can look
ahead and the reactor operators do that on a rod
sequence exchange, or a start up. They will have
planned the sequence that they are going through in
the start up process.
And they will have done predictive SOLOMON
cases ahead of time, and tried to map out exactly
where they are in stability space as part of their
normal package that they bring up to the control room
for the operators to use during those operational
scenarios.
And it is because of this, because it is
not real time, and exactly what we were talking about.
It can't keep up. It just can't do the calculation
fast enough.
And the other thing that we need to
understand is that it is not monitoring the in-core
neutron detectors in a time domain. It is not
actually looking for an oscillation. It is doing a
predictive calculation in the frequency domain, using
the inputs from the physics, just like you would run
it to do it for a reload.
CHAIRMAN WALLIS: But you need this rather
than just having a code which is permanent because of
changes in the burn up or something? Why do you need
to have any calculation at all if you have already
calculated it once?
MR. POST: Can I answer that?
MR. BROWNING: Sure.
MR. POST: Jason Post again. When we
first proposed Option 1-D to the staff, they wanted an
extra measure of protection to make sure that you were
maintaining your stability condition with a loaded K
ratio and with the core-wide mode as the predominant
mode. So it was added as an extra feature at the
staff's request.
CHAIRMAN WALLIS: And so it follows the
burn up changes and changes in power distribution or
something?
MR. POST: All that is built into it from
the input.
DR. POWERS: From the physics.
MR. ROSEN: So the operators know how to
start up and avoid the instability region, and they
also have SOLOMON cases which they have run along the
line of their intended path to full power.
MR. POST: Right.
MR. ROSEN: And have basically checked out
to make sure that they are stable using SOLOMON, which
is really running ODYSY, or taking ODYSY results.
MR. POST: That is correct. And that is
a great lead in to Steve here, who is next.
CHAIRMAN WALLIS: All right.
MR. KOTTENSTETTE: I am Steve Kottenstette
again.
DR. SCHROCK: Could I ask one more
question. Where does the thing typically begin
steaming in this start up period?
MR. KOTTENSTETTE: Usually at one percent
power. Usually we get to the point of adding heat is
usually right around 4 to 5 percent power.
DR. SCHROCK: So, 4 to 5 percent. Okay.
MR. KOTTENSTETTE: Okay. As far as the
use of SOLOMON, again it is always running, and once
a day we get a printout when we are up and running.
But since it is always looking at where we are at on
the power to flow map, it is automatic as far as when
it sees that the plant has gotten into either the
exclusion or the buffer zone.
It automatically calculates a case for us
and prints it out for us. So like if we have an
operational transient where we lost a recirc pump,
which is the most probable cause for us to go into the
exclusion zone, it will sit there after a time delay
and do the calculation, and print out for us where we
are at as far as the stability plot.
As far as how we would monitor for thermal
stability, we use our APRMs as our primary means to
either detect it and to suppress it either when we see
it initially, or if it sees it before we can actually
take action.
As far as a plant start up, you can see
here in the pink line there that that is our typical
plant start up, as we come up in power and maneuver
around the exclusion and buffer zones.
So once we get up in power and get the
generator on line, we are now 3-D, and can actually
operate and provide that input to SOLOMON. And you
will see that we raise power and lift the control rods
enough to get above the interlock for the recirc
pumps, where we can increase flow now.
And then once we get up to a point where
we can now pull rods again to get it close to our
rated low line, and then after that it is just to go
up in power with recirc flow.
And then as we get up close to our target
rod pattern, there will be minor rod adjustments. And
that is where you see all that squiggle up here at the
top, as we are making adjustments to account for zanon
and other poisons that are burning then.
CHAIRMAN WALLIS: And you said that
SOLOMON is run once a day. Is the output from SOLOMON
simply to move around these orange lines isn't it, and
they change a little bit from day to day; isn't that
what it really does?
MR. ROSEN: You said that you run SOLOMON
once a day when you are at a steady state?
MR. KOTTENSTETTE: Right. Or at the other
end of the power flow.
MR. ROSEN: But when you are going up, how
often do you run the SOLOMON when you are making this
maneuver?
MR. KOTTENSTETTE: We normally don't run
software or ask for a SOLOMON case during a start up.
MR. ROSEN: Because you have several of
them already done ahead of time, and you have looked
at that as part of your operating plan?
MR. KOTTENSTETTE: Right.
MR. ROSEN: So you know where you are
going to be as long as you go up?
MR. KOTTENSTETTE: We know that we are not
going to be close to the buffer or the exclusion area.
MR. ROSEN: You got a little close this
time didn't you?
MR. BROWNING: Well, this illustrates a
good point. This was an actual start up of the plant.
This is actual plant data from this past January. So
this is at the current power levels and you can see
that we stopped it at the current power level.
So we were just trying to highlight here
that it is feasible to get around the exclusion and
buffer zones, and yes, we got close here, but
obviously when the operators get to the uprated
condition, they will just move a little further over
and just shift it a little bit to the right.
MR. ROSEN: What was the cause in this
particular case of the flow dropping off from 26
million pounds per hour to 25?
MR. KOTTENSTETTE: As you go up in power,
you get increased resistance to the core, and so flow
is going to actually die down or reduce.
MR. BROWNING: Right. We increase the
steaming as you are pulling rods and going up.
MR. ROSEN: So with a constant recirc pump
speed, you are getting lower -- well, we can see the
flow dropping off about a million pounds per hour?
MR. KOTTENSTETTE: Correct.
MR. BROWNING: That's correct. And so the
main point and the emphasis was that this region down
here, where it appears that we could be constrained,
we were just trying to show that the operators do
actually have operating margin room in this area to
get around this low end.
Because maneuvering out here is not the
issue. It is clearing the interlock and then skirting
around. And like we said, in conclusion, we tried to
demonstrate that the methodology that we use builds in
margin, and that the calculational methodology of the
ODYSY code, we have accounted for that.
And that the acceptance criteria that
Jason talked about adding the extra .15 on to the
decay ratios so that it maps back to the old FABLE
code, and so that's how we build in the margin.
And then in our case, plant specifically,
we have seen no impact on the safety margins. We have
got lots of margin to the decay issue.
CHAIRMAN WALLIS: And how would you define
a safety margin?
MR. BROWNING: Well, the safety limit
minimum critical power ratio, and that the fact that
the APR and flow by scramble protect that,and protect
the fuel, and we have demonstrated that in the
analysis.
CHAIRMAN WALLIS: So safety margin is a
measure obtained by comparing some number with some
other number?
MR. BROWNING: Yes.
CHAIRMAN WALLIS: And one is lower by some
amount and the safety is the difference between the
numbers or something?
MR. BROWNING: Well, what we do is we look
at several scenarios, and do the calculation to show
the change in the critical power ratio is for those
particular transients.
And then we compare that to the safety
limit and show that we have the margin that is
required to demonstrate the safety margin is met.
CHAIRMAN WALLIS: Is there something in
the law which says what the safety margin has to be?
MR. BROWNING: It is built into the safety
limit MCPR, and the value that we use has got margin
built into it.
CHAIRMAN WALLIS: So it is clear what is
meant?
MR. BROWNING: Right.
DR. SCHROCK: What is the duration of this
start up process typically?
MR. KOTTENSTETTE: The typical start up
process, from initial start up to 100 percent power,
it normally takes about two days to get all the way
there.
DR. SCHROCK: So it is very slow?
MR. KOTTENSTETTE: Yes, it is.
MR. ROSEN: Is it very slow during the
time that you are going through the door, through that
window? How long does it take to get from -- if you
will put the slide back up with the maneuvering.
Let's say to go from 13 million pounds per
hour, which is the natural circulation, to 20 million
pounds per hour? How long does it take you to do
that?
MR. KOTTENSTETTE: That should take
probably about 15 minutes, because when we increase
power with recirc, we pretty much go -- our normal
rate is like 2 to 3 megawatts of electric. So, with
five percent power, that is going to take about 15 to
20 minutes of adjusting recirc flow.
MR. ROSEN: So in terms of the critical
operational period, you are going to go through all
those critical maneuvers and be watching the critical
parameters.
And it's not like if you have to watch
that for two days. You are in the critical region for
about 15 or 20 minutes, and from then on you have got
a lot more margin.
MR. KOTTENSTETTE: That's right.
CHAIRMAN WALLIS: So one shift does it.
It's not as if you are in a critical reason for a
shift change or anything like that.
MR. KOTTENSTETTE: No.
MR. ROSEN: In fact, that is a good
question, Graham. When you start up do you change
shifts at any point during this period?
MR. KOTTENSTETTE: It depends on where we
start up on the shift.
CHAIRMAN WALLIS: Well, if it is two days,
you better.
MR. KOTTENSTETTE: Well, yes. But we
pretty much hold points throughout the start up and
get to the point where you get up to the rated
pressure, and from there to get to the point where we
can roll the generator. And then from there --
MR. BROWNING: And there are prerequisite
tests that are required --
CHAIRMAN WALLIS: Are you on 8 hour shifts
or 12 hour shifts?
MR. KOTTENSTETTE: We are on 12 hour
shifts.
MR. ROSEN: And I guess the operative
question is that one shift actually takes you up from
the natural circulation line up into the 30 million
pounds per hour or something like that?
MR. KOTTENSTETTE: Yes.
MR. BROWNING: These guys look ahead and
try and target those windows to make sure that they
don't have a shift turnover right in the middle of
some critical task in the middle here.
And our conclusion is that we have shown
that the operation at the extended power uprate with
respect to the thermo hydraulic stability has been
acceptable. Next slide.
I get to continue on, and again with
Jason, and this time we are going to talk about
anticipated transients without SCRAM. We are going to
go through and give you a little bit of background on
how Duane Arnold complies with the ATWS rule.
And then Jason again is going to talk
about methodology, and how we went through and did the
calculations, and then I will get back up again and
talk about the analytical results and the conclusions.
Again, to demonstrate that we have
considered the operational and safety margins from the
ATWS perspective at the EPU conditions. First, the
system that everybody is most familiar with when we
talk about ATWS, and that is the standby liquid
control system.
For Duane Arnold, we have gone to the two-
pump operation, where the single switch in the control
room starts both pumps simultaneously. And they are
required to pump a minimum of 26.2 gallons a minute
each, and to get the equivalency requirement, we use
naturally enriched boron.
And we use a minimum concentration of 11.8
weight solution of sodium pentaborate. That means the
rule requirement for the 86 GPM equivalency that was
spelled out in the rule.
And the other thing to understand is with
our design being a BWR-4, we do inject the boron
solution below the core through the sparger, okay.
We have installed the alternate rod
insertion system, and one of the things that you will
hear about in the conservatism and the way we do the
calculation is that we take no credit for that in the
analysis, which is the system that pneumatically
bleeds off the air from the control rods, and allows
them to go in as the back up.
We have the recirculation pump trip
system, that when we detect conditions that would be
indicative of an ATWS of a high pressure in the water
level, the recirc pumps will trip off and run back to
flow.
And that we have adopted the BWR Owner's
Group emergency procedure guidelines for dealing with
the ATWS, which include lowering the water level and
taking those actions.
The ATWS rule established pretty much
hardware requirements, and then behind it, we go back
and we look at and demonstrate that we comply with the
analytical basis that that rule was predicated on.
So the things that we look at are the peak
pressure below the ASME service level of 1500 psi for
the events. We demonstrate that the peak cladding
temperature remains within the 50.46 requirements of
2200 degrees fahrenheit.
We look at the local oxidation fraction,
and make sure that it stays below the 17 percent
requirement of 50.46. We look at the suppressible
temperature and ensure that it remains below the plant
design limit of 281 degrees fahrenheit.
And we also look at the containment
pressure to make sure that it stays below the plant
design limit of 62 pounds. And then we go back and we
benchmark to not the current power level, but the
original license power level, which for us would be
50.93 megawatts, to demonstrate that the impact of the
EPU is acceptable.
So we go all the way back to the original
plan and do the comparison. And at this point, I will
turn it over to Jason.
CHAIRMAN WALLIS: How close do you get to
these limits when you do this? Let's say it is within
10 CFR 50.46, are you opening up against, say, 2200
degrees, or are you still a long way from it?
MR. BROWNING: We are a fair ways away.
MR. POST: We are a long way from it and
we are going to show you those specifically as well.
CHAIRMAN WALLIS: But you are closer than
you were before?
MR. BROWNING: Yes, and we will show you
the results later.
MR. POST: So this is Jason Post again.
I have one slide here on methodology. We use the ODYN
code when we did the first generic licensing topical
report on power uprate. That was also the same time
that we also submitted the application to use the ODYN
code to do the ATWS calculations.
And ODYN, of course, has been used for a
number of years for transients, but we had to get the
approval of the various models that we needed for
ATWS, and specifically the boron mixing model.
And the boron mixing model is the key
conservatism that we have in the ODYN analysis, and we
demonstrated that it adequately bounds the best
estimate calculation with the TRACG code. That was
our benchmark that we used.
It is important to note that we do use a
best estimate approach for ATWS. Some of the
conservatisms that we have in there are on the SRV
subpoints. We do use conservative SRV subpoints in
the calculation to compare to the peak reactor
pressure criteria.
And we do use reasonable operator action
times. It is important to note that the SLCS
initiation, which is two minutes after the ATWS
signal, has not changed. So someone made a comment
earlier about how the operator response has decreased.
In fact, we use exactly the same operator
action time that we used for power uprate, or that we
would use for an ATWS analysis at current license
power.
So it does -- you do have a slightly
steeper uprate during those first two minutes, but we
have not changed the operator action time. We use the
same action time.
We use pool cooling and service about 11
minutes, and that is basically 10 minutes of nothing
happening and one minute to align the system is where
that 11 minutes comes from.
And as we have talked about before, this
is supported by the emergency procedure guidelines,
and the emergency procedure guidelines actions are
fully adequate for EPU. There is on change to the
basic actions that are taken in the simulator
training.
CHAIRMAN WALLIS: Now, how is the level
controlled during this ATWS? Is the operator looking
at the level in the vessel?
MR. POST: Yes, he is. We do do a water
level reduction.
CHAIRMAN WALLIS: And isn't there some
action to control that level?
MR. POST: One of the key mitigating
features of an ATWS response is to lower water level
below the feed water spargers so you don't have that
low subcooling.
And actually you reduce clear down to the
top of the active fuel to reduce the power level.
Once you reduce the power level, then you are
mitigating the stream that is going to the suppression
pool for the bounding ATWS event.
So the termination of feed water happens
in about the same time frame as the initiation of
SLCS. They are both what we would call immediate
operator actions in the power control portion of the
guideline.
CHAIRMAN WALLIS: So with the uprated
power, there is somewhat less time to do this?
MR. POST: Again, we make the same
assumption on operator action time. We assume the
time is the same. It does give you a little bit worst
result, which you will see in a few minutes. But we
are using the same time to take both actions or for
both conditions I should say.
DR. POWERS: When you have analyzed these
in the generic sense, or in the specific sense, either
one, do you look at the pressure on the fuel rods
during the water level drop and then the mixing?
MR. POST: It is not an explicit part of
the calculation. Remember that during the water level
drop, we maintain the core covered, and we do have a
peak clad temperature calculation which shows that the
temperature stays quite low.
I don't think the response to the fuel is
any more severe than one of the transient events, the
response for ATWS. The real threat from ATWS is the
containment temperature. I mean, that is the biggest
worry.
DR. POWERS: Here is what I am interested
in, is whether any of the fuel rods having large
stresses put on them or strains?
MR. POST: Not more severe than any other
event in the envelope, in the design envelope.
MR. ROSEN: Well, you have got me a little
confused now frankly. I am reading the staff's safety
evaluation, and it is on page 75, and in that they are
talking about the PSA and the screening that was done,
which identified five operator actions that were
evaluated for their impact on plant risk.
One of those actions is the initiation of
SLCS for turbine trip and main steam isolation valve
closure ATWS events. And in that paragraph, it says
due to the extended power outrate, the early SLCS
initiation timing is reduced from 6 minutes to 4
minutes; while the late SLCS initiation timing is
reduced from 20 minutes to 14 minutes.
Now, just looking at the early, that says
from 6 to 4 minutes; and yet your slide says 2
minutes.
MR. POST: Two minutes after the ATWS
signal is what we use in the analysis, and I am not
certain the basis for what is in the PRA.
MR. BROWNING: In the PRA analysis -- this
is Tony Browning again. In the PRA analysis -- and we
will speak to it later when we get to that
presentation, those are actually acceptance criteria
that are applied in the PSA model.
If the operator performs to that level by
that time, the event tree goes in one direction. If
he is not successful at that juncture, it takes a
different path and goes down through the event
analysis in a different way.
And that is really all that is driving
there. It is driven by the map results, and the
change in time, and the calculation of the human
performance. But really what it is doing is setting
up later in those events what the successful criteria
is, or how much suppression pool cooling is required
to keep the containment within the design.
MR. ROSEN: I understand that, and these
results that are reported here by the staff have an
effect on our probablistic safety analysis, and the
impact of the change in power on the resulting core
damage frequency.
MR. POST: Correct.
MR. ROSEN: And what you are saying here,
I think, and help me to understand this, is that even
though the PSA uses four minutes to draw some
judgments about operator success likelihood, and that
four minutes speaks to some analysis of the
performance shaping factors for the PSA, in the
thermal hydraulic analysis, you initiate SLCS in two
minutes rather than four minutes.
MR. POST: Yes, we do. That's correct.
MR. ROSEN: It's different and I don't
understand why.
MR. BROWNING: This is the basic -- if you
will, deterministic modeling and methodology which --
like we said in the beginning, that is the basis for
our application.
The PSA look was to look for risk
insights, and you are seeing that. You are seeing the
result of us looking with our PSA model for risk
insights, and we have incorporated those.
That's why we went up with the simulator
with the operator training early on and ran through
these scenarios. And Mr. Kottenstette said he has
discussed that earlier, and that we really did not see
any degradation of human performance in the simulator.
That was the take away from this. We saw
the result, and we got the lesson learned, and we went
up to the simulator to see if in reality we were
seeing a challenge to the operators, and if that had
been the case -- and it wasn't, but had that been the
case, and we had seen that, we would have had to make
adjustments.
And that either at operator training or
some other mitigative strategy, if the effect of the
uprate had been that we needed to get standby liquid
in much sooner, and we had seen a degradation of human
performance in the simulator, we would have had to
address that, but we didn't see it.
MR. ROSEN: I am puzzled, and a little
troubled about using two different numbers; one in the
PSA analysis and one here, for when you initiate SLCS.
If you used the four minutes here to be consistent
with what the PSA people do --
MR. POST: Our PSA expert is going to get
up and address that.
MR. BOEHNERT: Come up to the mike and
identify yourself.
MR. POST: But that is not uncommon. I
mean, we have different ways that we look at things in
deterministic space from the way that we look at
things in probablistic space. It is not different.
MR. HOPKINS: This is Brad Hopkins from
Duane Arnold. I am the PRA engineer at Duane Arnold.
I think I can provide a little clarification.
In the PRA, we allow containment pressure
and temperature to go much higher than the design
values before we assume failure. So in our thermal
hydraulic analysis, we are able to live with later
standby liquid control injection before we would
exceed our criteria.
The criteria is different because in the
PRA our containment failure occurs at much higher
pressures than the design pressure. That is, we would
not expect containment failure until about 120 psi.
Whereas, in the licensing basis, 54 psi. So that
would account for some of those differences.
MR. POST: So you could allow a higher or
a longer action time and still meet your criteria?
MR. HOPKINS: Yes. We are looking at or
trying to look at realistic evaluations and not
putting in the conservatisms that are applied for the
licensing based evaluation.
MR. ROSEN: That clarifies it, but I would
point out that you have the differences there.
MR. BROWNING: And as we get into the
event, specifically with the results, and we were
asked to look at the acceptance criteria and how we
compare those, and do a comparison of pre-EPU to EPU
results.
And then a recent topical issue, we are
going to look at our Evaluation of Information Notice
2001-013, which was the inadequate SLCS relief value
margin issue. This is a comparison of Pre-EPU to EPU,
and if you look at the --
MR. BOEHNERT: Excuse me, but could you go
back to the slide. Can you highlight that relief
value margin issue, please?
MR. BROWNING: We will get to that.
MR. BOEHNERT: Fine.
MR. BROWNING: And first off we will look
at the EPU results. You will see that the peak
reactor vessel pressure, the acceptance criteria is
1500 pounds, and we are at 1343, and we are below that
limit.
We are going to look at peak fuel cladding
temperature against the 2200 limit, and as you can
see, we are at 1380 degrees fahrenheit. So we are
quite low there.
And the peak suppression pool temperature
limit, the design limit is 281 degrees, and we are at
215.6 for the EPU; and the peak containment pressure,
the design on that is 652 psi, and we are at 18.3. So
as you can see here, we have lots of margin.
And looking at the impact of the EPU,
again, reminding everyone that this comparison goes
all the way back to the original rated thermal power
of 1593 psig, you are seeing the impact of not only
the full 20 percent increase, but you also are seeing
the impact of reactor pressure change, and a ELLLA to
MELLLA change as well.
Because at our previous uprate, when we
did this stretch of 5 percent in 1985, that was when
it raised reactor pressure. So you are seeing all
those effects rolled up into this.
So the reactor pressure as expected goes
up, and the suppression pool temperature and the decay
heat goes up, and it takes the containment pressure
with it.
The interesting result is the fuel
cladding temperature. You see a slight reduction, and
that is because of the flattening of the radial power
in the core, and where the peak bundle isn't working
any harder to bring up the average.
So we redistribute the flow, and the net
result of that is that the peak bundle gets a little
bit more flow because the average bundles are getting
a little bit less because of the increased pressure
dropped from their steam production.
So the peak bundle gets a little bit more
cooling, and so the FCT comes down.
CHAIRMAN WALLIS: Do you actually have the
acceptance criteria on this? You told us what there
were, but --
MR. BROWNING: We have a back up slide
with that if you would like to see it.
CHAIRMAN WALLIS: But you told us what
they were.
MR. BROWNING: Here is the background on
the information that was on the SLCS margins, and the
concern is that in an ATWS event, and the loss of off-
site power is the specific one that was addressed, the
concern is that you have high reactor pressure at the
time of standby liquid and ejection.
And you have reduced margins to the relief
valve setpoint, and one of the things is that you have
an operating margin that is required between the peak
system pressure at the relief value next to the pump,
and the nominal relief valve setpoint.
You have a required delta that you are
required to maintain there. So what happens is that
you are trying to account for uncertainties, a set
point drift in the relief valve and other things, and
also because these are positive pressure pumps.
And they are very dynamic, and you get big
pressure pulses as it ejects, and so you are trying to
absorb all that with this margin. And the concern is
that if the reactor pressure is too high, it can eat
into this margin, and you have the potential to
interrupt the standby liquid ejection, and actually
the circulate the boron in a loop around the pump, but
not actually inject it into the core.
The results for Duane Arnold is that we
have grater than a hundred psi operating margin to the
nominal valve set point, and we saw no interruption in
the SLCS ejection.
In conclusion, we talked about the
methodology and how we use that to capture margin,
which we also go back and do at the benchmark to look
at the impact of the EPU on the plant, and to look at
the margin that would go there.
And then again in the plant specific
results, we satisfied all the acceptance criteria, and
so we saw no impact from the safety margin. If you
have adequate margin for the acceptance criteria, we
have operational margin sustained by that.
And then we have an acceptable comparison
to the benchmark case, and so we didn't see a huge
change there. So from that we can conclude that the
operation of the EPU from the ATWS perspective is
adequate.
DR. SCHROCK: What was this best estimate
of the --
MR. POST: That was approved at the time
of the ATWS rule and the first time that we started
doing ATWS analysis. It is because of the low
probability, and also because it was not part of the
original design basis. It was an added analysis, low
probability.
DR. SCHROCK: To put it in the context of
best estimate analysis, can you put a date on that for
me? Was it the early '80s?
MR. POST: I don't remember when the ATWS
rule came out, but --
DR. POWERS: Around '85 or '85.
DR. SCHROCK: And ODYN was its basis at
that time?
MR. POST: No, at that time we were using
a READY code, and as I said earlier, we didn't -- ODYN
had been used for the transient calculations, but we
had not qualified the boron mixing model until the
time that we did the generic submittal on power uprate
in the mid-1990s. That's when we started using ODYN
for ATWS calculations.
DR. SCHROCK: And in the original ATWS
problem, you didn't present it as a best estimate
calculations?
MR. POST: Well, I am sure in the original
analysis that was done with READY, and the ATWS rule
compliance, I am sure that those were done as best
estimate calculations. Yes, I'm sure that they were.
DR. SCHROCK: I don't remember it getting
reviewed in that time frame, but it must have been.
MR. BOEHNERT: I remember his argument,
and I remember that they did a lot of that as he said.
I believe it was ODYN, but I don't really remember.
MR. POST: There were a couple of very
large topical reports that GE wrote on the ATWS
response for various events, and trying to determine
what the limiting events were and that needed to be
analyzed, and what the assumptions for the analysis
should be.
And I know that those were presented to
the NRC, and whether they were actually presented to
the ACRS, I am not certain.
DR. SCHROCK: Well, I am just curious to
know a little more about what is in the ODYN one.
MR. POST: All right. This is ATWS
instability, and we talked about the instability
prevention to ensure that you prevent an instability,
and if it does occur, you do get an automatic SCRAM to
show the reactor down and terminate the oscillation.
But one of the concerns previously was
what happens if that SCRAM fails, and so the
oscillation continues to grow and it is not
terminated, and how bad does it get.
And so I am going to talk a little bit
about the background, and the methodology, and what
happens to an ATWS instability if you don't have
mitigation, and why the application for Duane Arnold
is acceptable.
And again we are trying to demonstrate
that the existing ATWS instability analysis, which was
done for a high power density plant with MELLLA is
adequate for the Duane Arnold extended power uprate.
So there were two topical reports written
and that were both reviewed simultaneously by the NRC,
and one SER was written on both reports. The first
one is the NEDO-32047, which is the ATWS rule report.
And the purpose of this report was to
determine if fuel rod failures are unlikely from a
worst case instability event with the SCRAM failure.
And the result of the evaluation was that
this had no mitigating operator actions of any kind,
and so it maintained water level high in the reactor
and so it maximized the power production.
And we found that the power spikes become
very tall and narrow. It is almost like a reactivity
excursion type of event, in terms of what the fuel
experiences. It becomes -- so the peak energy
deposition, and we found it is within the fuel design
limits as you would get for reactivity excursions.
But the power becomes more severe as the
core inlet subcooling decreases, and in fact you get
to a point where as the subcooling decreases that you
actually get a power spike that causes an extended dry
out, and where the fuel doesn't re-wet in an
oscillation.
And when you get that extended dry out of
fuel service, then you get a very excessive clad
temperature, to the point where a portion of the fuel
could fail, and so we calculate that the number of
bundles that this could happen on, and the actual
location of the bundles. And it is about a half-a-
percent of the core volume.
DR. POWERS: When you say that it is
within the fuel design limits, you mean that it is
less than --
MR. POST: That's correct.
DR. POWERS: And that is if it is fresh,
but how about if it is burned up a bit?
MR. POST: I think we are at around 70 or
80 calories per gram, and --
DR. POWERS: And can it tolerate that when
you --
MR. POST: I am going to call on Dr. Jens
Anderson. Jens, would you mind helping me with this?
DR. ANDERSON: This is Jens Anderson to
talk about that fuel. When you get these high powered
oscillations, you get those in the fuel bundles from
high radial peaking. High radial peaking will only
occur for fuel with low exposure.
Once you get to the higher exposure, two
things happen. First of all, you don't have as much
activity left in the fuel and so you don't get the
higher radial peaking, and that was actually analyzed
as part of this work that was done in the first
report, the NEDO-332047.
And it shows us that as you go down in
radial peaking, you cannot get these high powered
oscillations. Secondly, this is very -- the other
things that happen is that even if you have high flux
peaks, with lower activity in the fuel, you don't get
the power response.
So I think the short answer is that you
can get the high oscillation for fresh fuel, but for
highly exposed fuel, you cannot have the high power
oscillations.
MR. POST: Again, this is the event and
the results that were analyzed previously, and what we
are demonstrating or discussing is the fact that those
were adequately severe in the analysis that was done
already, and did not get any worse for the MELLA EPU
condition for Duane Arnold.
So we are discussing things that were
already presented, and so it is not a new consequence.
CHAIRMAN WALLIS: It has already been
judged then as an acceptable --
MR. POST: Yes, and again this is the no
mitigation result. I mean, that's why we had EPGs,
and this is intended to demonstrate the worst case
-- no mitigation, maintaining water level high,
letting subcoolant go dry, and how bad does it get,
and that is what that report was intended to show.
And there could be a larger fraction of
the core. The .5 percent may not be a valid number.
It may go up to one percent. I'm not sure exactly b
because of the flattening, and the radial power
distribution, and you have more bundles that are
closer to the limit.
So I would agree that the .5 percent that
was reported in that report was based upon the core
design at that time. So that could get a little bit
worse, and frankly we have not calculated that.
DR. SCHROCK: And the most immediate
consequence is that gaseous fissure products are
released from rods that have failed.
MR. POST: Yes, certainly.
DR. SCHROCK: And it would seem to make
more sense to express it as a fraction of rods failed,
as opposed to fraction of core volume failed.
MR. POST: That was the way that it was
expressed originally. So if I can continue with the
next report, which is the mitigation report, 32614.
What this one did was recognize that that condition is
not acceptable.
You certainly would not want to have your
plan operate there, and get into that condition. So
they looked at what are effective mitigation
strategies.
And the two that are reported are as most
effective, one is to lower the water level to below
the field water spargers. Now, of course, the EPGs
say lower it to -- there is two levels approved by the
NRC.
One is to five feet above top of active
fuel, and the other is to the minimum steam cooling
water level, which is actually below, a collapse level
below the top of active fuel.
But to mitigate the ATWS instability, you
don't have to get it that low. That gives you a
bigger power reduction, but the key thing is the
subcooling, in terms of the instability.
You only have to get it about one or two
feet below the feed water spargers in order to have
the water that is coming in spray into a steam space,
and raise the temperature enough so that you mitigate
the core and let sub-cooling.
So that can be accomplished very quickly
with feed water run back, and that gives a real quick
water level reduction. And it eliminates, completely
eliminates the large power pulses.
Now, you can still have a small
oscillation that continues, but these very large
dramatic power pulses are completely eliminated. The
other feature is the boron injection, which is of
course also specified in the EPGs.
And boron injections is very effective for
the long term shutdown, but it is not quick enough to
prevent the kind of extended dry out that gives the
fuel rod failures by itself.
It does eventually make the oscillations
go away completely, but it doesn't happen -- the delay
time from the time it was initiated, and to the delay
time until it actually gets into the reactor core, and
until it mixes, and until it shuts down enough to
terminate the oscillations, it just does not happen
fast enough. So the water level reduction is the key.
And the methodology we used is TRACG, and
which you have reviewed before. We use multiple
channel groups and it gives you a detailed 3-D
kinetics in the thermal-hydraulic model of the core.
It is a very effective model for doing this.
And then the next chart there is a
benchmark from -- it is a current calculation of the
repeat of the case that was in NEDO-322047, and you
can see here the type of power spikes that were
reported in NEDO-322047 and that go up above a
thousand percent.
And as subcooling continues to decrease
then, you are kind of reaching a maximum of your
subcooling at about 200 seconds, and that is where if
you go to the next chart on the peak clad temperature
--
CHAIRMAN WALLIS: And that has been going
along for quite a long time hasn't it?
MR. POST: Yes. Right. And this is again
where the operator isn't -- you know, this is assuming
that whatever actions the operator has taken to try
and insert control rods have been completely
ineffective and the water level has not been reduced.
DR. POWERS: And is this level for the
fuel cycle?
MR. POST: I don't remember exactly. I'm
sure that it is at the most reactive point in the
cycle. It is probably around the middle of the cycle
is probably when it is done.
And again they are very conservatively
bumping up the radio peaking factor to make sure that
they get it. So the next slide talks about the ATWS
instability with mitigation.
Now, I don't have a chart to show that,
but what happens is that at about 150 seconds, feed
water -- the core in-let subcooling turns around, and
the oscillations start to die back down again, and you
don't get anymore of those huge random power peaks up
to a thousand percent.
CHAIRMAN WALLIS: Well, you have showed us
the bad looking ones, and it would be very good if you
showed us the good looking one as well.
DR. POWERS: And even so, within the first
150 seconds, you are putting some pretty good pops
into that fuel. I mean, even before the 150 seconds.
MR. POST: Well, I didn't have an
electronic version of that available, and so we will
go to the old paper method. But this shows how the
core -- the base case about mitigation is that
subcooling continues to increase and it goes up to
about 60 degrees K, and then that is about here, and
about 200 seconds is where they just keep going along
and this is where the dry out occurs.
And with the feed water reduction, the
water level reduction below the feed water spargers,
you get a very effective turnaround of the subcooling,
and you can see that we don't like these kinds of
oscillations, but they are enough so that they stay
within the capacity of the fuel.
CHAIRMAN WALLIS: And it is counter-
intuitive, and if you make the water colder, you think
it would cool better. But in fact it makes the
oscillations worse.
MR. POST: That is correct. Warmer water
gives you a better response from the hydraulic
instability.
MR. ROSEN: As long as you have raised the
question of counter-intuitive. From an operator
perspective, Steve, a little bit counter-intuitive,
isn't that to lower the water below the feed water
sparges?
MR. POST: As far as auxiliary power?
MR. ROSEN: Are you trained to do that?
MR. POST: We know that we lower power or
reactor water power reduces with it.
DR. KRESS: And it is not so counter-
intuitive for BWRs in other words.
MR. MCGEE: It is not counter-intuitive.
MR. POST: In nearly every training
scenario, they will have some sort of ATWS scenario,
and they are trained as to that.
MR. ROSEN: So you are modeling this in
the simulator, the compliance simulator is that you
are saying?
DR. KRESS: No, this is TRACG.
MR. POST: This is TRACG.
MR. ROSEN: No, I am saying that you are
modeling this event.
MR. POST: If I maintain water level high,
I will still see the high power because I am not
getting the increase in subcooling going on. So I
know that it is going to be a longer scenario for me
because power is going to be higher.
MR. ROSEN: So in your simulator crews are
trained to run feed water back and get the core level
below the sparges.
MR. POST: That's correct. The operator
action is to lower the water level all the way to the
minimum steam cooling level, which is near the top of
the active fuel, and which is well below the feed
water sparges.
So they don't stop when they clear the
feed water sparger. They are running it all the way
down until they get power to clear the APR and down
scale --
MR. ROSEN: Is this one of the critical
actions in the operator training?
MR. KOTTENSTETTE: It is a critical task
for us to lower the level down to a certain point, and
for us it is 87 inches above the top of the active
fuel. At that time, we have a decision to make; is
power now less than five percent power.
If it is, then now I have arranged to
maintain water level from the top of active fuel up to
whatever that water level is that the power is less
than five percent.
DR. KRESS: Would you go back to your
slide on the peak cladding temperature without
mitigation that you had up there. Yes, that one. No,
the next one.
What is happening to the center line fuel
temperature during this process? Do you have an
equivalent curve for the center line fuel temperature
in that slide?
MR. POST: I do not have that available.
It wasn't included in that report I don't believe.
DR. KRESS: But does it oscillate or does
it have a steady rise because of the lack of good
coupling, the thermal coupling between the clad and
the --
MR. POST: Well, I am sure that it is
oscillating on this same kind of frequency as well.
So I am sure that it is not a steady temperature. I
mean, the surface heat transfer coefficient is varying
as the fluid conditions changes at the surface.
DR. KRESS: But your thermal conductivity
and the fuel is not varying very much, and it is a
pretty good heat capacity in those fuels compared to
the clad, and I was mentally thinking that you might
get some oscillations, but you have got a steady rise
in that --
DR. POWERS: The fuel looks like a bunch
of stair steps.
DR. KRESS: Yes, but not little or big
stair steps. But I was trying -- what I am thinking,
Dana, is the total deposited energy in the fuel itself
compared to this limit of how many calories per gram
you get, as opposed to what you get in one
oscillation.
DR. ANDERSON: This is Jens Anderson
again. What you can see in this plot prior to 200
seconds is that you have repeated boiling transitions
and reword, and in that period the heat removal, the
net heat removal from the surface of the fuel rod, is
the same as the energy generation.
So, yes, you get some oscillation in the
center line temperature, and the center line
temperature is higher than the cladding temperature,
and on average it is constant.
DR. KRESS: It's not steadily climbing up
then.
DR. ANDERSON: No, it's not. It doesn't
start climbing up steadily until you fail to leave it,
and then you go up to a higher clad temperature, and
a correspondingly higher center line temperature.
DR. POWERS: I cannot believe that in two
seconds that you thermally communicate with the center
line of a fuel rod.
DR. KRESS: That was my problem.
DR. POWERS: And I would find that
remarkable, especially with a BWR rod.
DR. ANDERSON: No, that's correct, and you
are going to have a significant face shift between the
center line temperatures and the surface, because the
fuel is time constant.
Some are time constant, and the fuel is
typically in the order of 6 seconds, while the period
of the oscillations are more like 2 seconds.
And which tend to give a fair amount of damping in the
temperature response.
CHAIRMAN WALLIS: Were you adding boron to
the water at this time?
DR. KRESS: No, this is no mitigation.
CHAIRMAN WALLIS: No mitigation at all?
So what is the long term prospect?
MR. POST: The long term prospect is --
CHAIRMAN WALLIS: How does it eventually
shut down?
DR. POWERS: That is a special plot that
shuts it down.
MR. POST: Well, that's why we move around
a little, but then this is the effect of boron
mitigation.
CHAIRMAN WALLIS: Eventually you want to
raise the water level eventually.
MR. POST: Well, not until you get the
reactor shut down.
CHAIRMAN WALLIS: You have to get some
boron in there or something.
MR. POST: Yes. When you get the boron in
the oscillations go away completely.
CHAIRMAN WALLIS: That's right.
MR. POST: So this particular plot has
only the boron injections, and so if you did the water
level reduction by here in 150 seconds, you would make
all these power spikes go away.
And then as you would continue to inject
boron, you would make them go away completely. So it
is a combination of the two that allow for getting rid
of those oscillations and --
CHAIRMAN WALLIS: It is this drop in the
level that is just to shut down the neutronics, and so
it is counter-intuitive from the point of your
cooling, but it is what you need to do to shut down
the nuclear reaction?
MR. POST: That's right.
CHAIRMAN WALLIS: Then you need to get
some boron in for the long term.
MR. POST: It mitigates the containment
response dramatically, as well as avoiding this type
of power spikes in the fuel.
DR. POWERS: Graham, not everything is
thermal hydraulics.
CHAIRMAN WALLIS: No, it's not. I think
it is great.
DR. ANDERSON: This is Jens Anderson
again. I would like to point out one thing, is that
the curves that Jason Post has shown is what you have
is when you have an ATWS, the high density plant at
the middle line. It is really not an easy EPU issue.
DR. POWERS: Well, we don't have analyses
for this plant at the high and the low power levels to
see what they do.
MR. POST: And you are right, we do not
have that. Because the MELLLA boundary had previously
been analyzed and the peak bundle power for Duane
Arnold is consistent with what the bases were that
were performed, we have done some GE14 studies to
confirm the GE14, which is the newest fuel design that
they have already loaded, I believe.
And the response for GE14 is similar, and
the ATWS mitigation techniques are still effective,
and so we did not do a Duane Arnold specific TRAC
calculation for this.
So the methodology, it evaluates the
margin and it uses limiting initial conditions, and
limiting peak bundle powers. And there isn't really
a safety margin associated with this. We are past the
safety margins for this particular evaluation.
There is no degradation of the fuel
response for EPU. We already have a pretty severe
response, and so the margins that you had before are
sustained. So from this point of view, EPU is
acceptable for Duane Arnold.
DR. POWERS: Are there any questions on
this portion of the program that we have gone through
so far? Seeing none, and not looking very hard for
any of them, I am going to call for a recess until 10
of.
CHAIRMAN WALLIS: And during the recess,
I would like to respond to the question that I raised
about capacity, because it may be just a
misunderstanding.
(Whereupon, at 2:37 p.m., the meeting was
recessed and was resumed at 2:50 p.m.)
DR. POWERS: Let's go back into session.
We are now going to move on to the non-controversial
topic of the corrosion. I know that there will be no
questions at all, and so we will be able to whip
through this topic with speed and direction, I'm sure.
MR. SEVERSON: I am Russ Severson, and I
am here to discuss our flow accelerated corrosion
program at Duane Arnold, and what the impact will be
from what I expect the impact is from the extended
power uprate.
To quickly explain flow accelerated
corrosion in five seconds, flow accelerated corrosion
leads to wall thinning, and many perimeters, including
water chemistry, material composition, and the
hydrodynamics effects, affect this wear rate.
And, of course, carbon steel piping is a
especially susceptible material. Duane Arnold has had
a long term monitoring program focusing on the
susceptible high-energy carbon-steel piping system.
We include both single and two phase
systems throughout the balance of the plant site, and
at DAEC, we completed a tailored collaboration with
EPRI back in the mid-1990s, which helped us base line
and determine what our modeling was and to evaluate
what should be modeled, and what our inspections
should be.
And we have been progressing with that
base line inspection. All our lines are continuously
operating lines, and are modeled in our EPRI CHECWORKS
program.
The inspections are performed to verify
their model and to monitor the wear specifically. We
typically do 40 to 60 inspections or actually
locations. We inspect locations.
CHAIRMAN WALLIS: To verify a model, do
you actually do enough readings to verify a model?
MR. SEVERSON: To verify this model? Yes.
CHAIRMAN WALLIS: So you actually have a
prediction at the rate at which --
MR. SEVERSON: We have a prediction, yes.
CHAIRMAN WALLIS: And it works?
MR. SEVERSON: Of where it is, yes, and of
our different rates within our continuously operating
lines. And in that prediction, what we had to do was
we went back and evaluated the beginning of the
operation, and decided what our wear rates were
through all 18 at the time, or 15 cycles of where.
And to evaluate what our chemistry was
through all those 15 cycles, and how we operated, and
we have a heat balance, a simplified heat balance
within the program to identify what the hydrodynamics
are.
And adding all of that up, we do these
inspections. Now, we didn't start out doing 40 top 60
locations. That is now after many years of having
this model and verifying, and ensuring that it is
correct.
CHAIRMAN WALLIS: So now you have enough
information that you can safely scale it up to higher
velocities.
MR. SEVERSON: Correct. Originally, I
think we did around like 200 inspections the first
time we put the model together. But since then, we
have been having to do less.
DR. FORD: Is it qualified for the higher
flow rates? By qualified I mean there are data for
the higher flow rates?
MR. SEVERSON: Yes, there is. Within
CHECWORKS, it will let you vary the feet per second
wear rate within your systems. Our plant has by
design fairly low flow rates. And so with the 20
percent increase, you are within the boundaries.
DR. FORD: And are there other data of
what the CHECWORKS flow rate would tell you?
MR. SEVERSON: Well, within their book
that they publish with EPRI, they show graphs of up to
40 inches per flow rates, and I don't know if some
plants have this or not, but I do know that Duane
Arnold is a low wear plant, and that is partly because
we were built with larger pipe diameters than what
they built with some of the later model plants.
DR. POWERS: And one has to recognize that
CHECWORKS has an empirical database that extends well
beyond just the nuclear business.
MR. SEVERSON: That's correct.
CHAIRMAN WALLIS: Could I ask if
Susquehanna is one of the plants with higher flow
rates, the reason being that they have had erosion
problems? I understand that they are going to have a
limited power uprate.
MR. SEVERSON: It is in the model. I have
not had data back from Susquehanna that they wouldn't
have had and that CHECWORKS would not have worked.
And I can't tell you as to what extent
they use CHECWORKS at Susquehanna, and so I can't
speak from that qualification of knowledge. I do know
that within our flow rates there are plants out there
that model lines that will be at these newer flow
rates, 20 percent higher, and they have not seen that
issue.
And I would have to see what the
Susquehanna issues are. I am not sure if they are a
reheat plant, or a second reheat plant like we are,
which makes a huge difference into your wear rates.
CHAIRMAN WALLIS: And the fuel piping, has
that been exposed to --
MR. SEVERSON: That is correct.
CHAIRMAN WALLIS: And those carbon steel
pipings have been exposed to --
MR. SEVERSON: To the feeder water lines,
yes.
CHAIRMAN WALLIS: And the feeder water
lines are checked?
MR. SEVERSON: Yes, and those are the ones
that I tried to provide data on since they are the
ones that can show you the velocity changes since --
and some of our other lines were going to a less -- to
a higher quality line, and less wear from the quality
standpoint of the steam coming in.
CHAIRMAN WALLIS: And versus the
observation?
MR. SEVERSON: Yes.
CHAIRMAN WALLIS: And also for the
platinum covered carbon steel?
MR. SEVERSON: I have not seen with the
platinum covered carbon steel as to -- well, I have
not seen where CHECWORKS significantly differs yet on
wear rates.
Now, one thing about flow accelerated
corrosion, which is that it is a very long term
phenomenon, and I am modeling history back to '75, and
we have had none since '96, and so far the Noble Chem
has not shown a significant difference.
DR. SHACK: But those lines at Noble Chem,
those would be very low flow accelerated corrosion.
I mean, aren't they the ones that just sort of sit
there? The carbon steel lines that actually see Noble
Chem.
MR. SEVERSON: Not in the high flow rates.
Go ahead.
MR. KNECHT: This is Don Knecht from GE.
The feed water -- the carbon steel feed water lines do
not see the Noble Chem injections. It is only the
stainless steel.
MR. SEVERSON: Yes, it should not have
come back that way, because they do it with the
recirc.
CHAIRMAN WALLIS: CHECWORKS predicts a
continuous variation of wear rate versus velocity or
something, or is there a transition, and a critical
velocity? What sort of dependence is it?
MR. SEVERSON: They have an empirical
formula of -- I will throw up a slide here to give you
a feel for what the impact of the velocity is.
CHAIRMAN WALLIS: Is it velocity to some
power or something? So it is a continuous behavior.
It is not a step chain. It is level or something? I
mean, downstream of a connection, it is not --
MR. SEVERSON: There is another one, and
that is true, too. In CHECWORKS, they have a certain
factor formula. Let me throw that up.
DR. SHACK: And that's where you would see
dramatic changes if you suddenly went through some
sort of a transition, but this is kind of a -- you are
in the same flow mode.
CHAIRMAN WALLIS: It is probably the
boundary that matters, and if the high velocity gets
right close to the boundary --
DR. POWERS: You have to understand in the
middle that they really can't calculate anything, and
so they develop this incredible empirical library, and
it is called CHECWORKS.
MR. SEVERSON: And we are constantly doing
testing and we use French data, and what have you.
Here is the formula to give you a feel.
CHAIRMAN WALLIS: The geometry effect is
this fudge factor G.
MR. SEVERSON: And from their experimental
evidence they apply this geometry effect, and what I
just showed you was an effect that they provide. This
is what is in the CHECWORKS model for liquid velocity
changes.
This is by keeping the other issues
constant, and here for the BWR is the oxygen level.
It is a very low oxygen level for what this graph is
showing.
DR. SHACK: What is the normal oxygen
level in your feed water?
MR. SEVERSON: Right now, 30 parts per
million.
DR. SHACK: Now, when you change out your
high pressure turbine, and is chrome moly steel, or
are you stuck with the associated lines?
MR. SEVERSON: A couple of them have
already been changed with chrome moly steel, and a
couple of them will remain carbon steel. The ones
that I have not been seeing with significant wear.
And a couple of them were the old alloy --
the copper based alloy that we as a plant have not
seen significant wear in, and there is a smattering of
different lines throughout that we watch.
DR. KRESS: Why do those curves peak at a
given temperature?
MR. SEVERSON: Why do they change in
temperature?
DR. KRESS: No, why do they peak?
MR. SEVERSON: Why do they come like this
and come back down?
DR. KRESS: Yes. Why do they come back
down?
MR. SEVERSON: Because flow accelerated
corrosion is a temperature dependent phenomena.
DR. KRESS: I know, but I thought it would
have just kept going up.
CHAIRMAN WALLIS: There is no why about
any of this. It is empirical.
MR. SEVERSON: Well, around 300 degrees is
your highest wear rate for flow accelerated corrosion
with everything else said.
DR. KRESS: But my question is why is
this?
DR. POWERS: It is the solubility data
from Oak Ridge.
MR. SEVERSON: He's exactly right.
DR. SHACK: It is solubility.
DR. KRESS: It is dissolving the oxide off
of it.
MR. SEVERSON: Yes.
DR. POWERS: It is solubility for EPRI
304, and goes through a maximum, and that is what
underlies those curves. That was figured out by the
chemists.
Now, the metallurgists came along and they
said that in order to do anything they had to put
fudge factors in because they can't calculate
anything.
CHAIRMAN WALLIS: So it is washing away
the rust.
DR. POWERS: Yes, washing away the rust.
CHAIRMAN WALLIS: Now I understand.
MR. SEVERSON: Now, we already went
through this, and let's go on to the next slide.
CHAIRMAN WALLIS: Well, you have predicted
what the change will be and it is going to be very
small presumably. Is it?
MR. SEVERSON: Yes.
CHAIRMAN WALLIS: What sort of change do
you predict?
MR. SEVERSON: Down here at the end, we
will show you. It is about half to 1-1/2 mills,
depending on where you are within the system because
of temperature, and flow rate because of the size of
the geometry.
So what I did was that I took the highest
flow area in the feed water, and I took the worst
temperature case in the feed water, and did a
parametric study and showed what the differences were.
And this is about a half to one-and-a-
half, where we are seeing about four mill now, and so
we should be seeing about 5-1/2 mill, which with the
piping, we will have about 150 mill margin to code
allowances.
CHAIRMAN WALLIS: In a hundred years?
MR. SEVERSON: No, at 25, I think, or 30.
Let's go to the next slide. So as I would conclude
that we will monitor what the water rate changes are
with the power uprate, and with the increased
velocity.
CHAIRMAN WALLIS: Where does all the water
waste go? Does it actually stays in the solution, and
just deposits somewhere else?
MR. SEVERSON: It ends up in the
condensate polishers.
CHAIRMAN WALLIS: Does it build up in
other parts of the system?
MR. SEVERSON: Yes, you will see it
throughout. And we found direct actual evidence with
our chemistry numbers with iron, and we found actually
a pretty good correlation as to what our wear rates
are compared to the iron is at the end of the feed
water.
DR. POWERS: You don't have any regions
where you have corrosion product build up that is
going to strip off, mechanically strip off?
CHAIRMAN WALLIS: A piece of scale that
will then wrap around?
MR. SEVERSON: I don't believe that we
will have any impingement problems. Is that your
issue?
DR. POWERS: I was just thinking of the
Surry incident, where they stripped some oxide off
mechanically because they jacked the flow rates up.
MR. SEVERSON: I don't see that. We are
not in those flow rate ranges, and I think that's
where these max numbers are. But we are like going
from 16 to 18 feet per second generally.
DR. POWERS: So you are really quite low.
MR. SEVERSON: Yes.
CHAIRMAN WALLIS: Are we talking about
flow reduced vibration in your analyses, too? I would
think that reduced vibration would affect where, too,
because of the boundary areas change when you
oscillate the things.
MR. SEVERSON: Well, I don't think that
this --
CHAIRMAN WALLIS: And then the reduced
vibration would affect that.
MR. SEVERSON: I don't know if we have
seen that, but I don't know if that phenomena really
exists.
MR. ROSEN: When you estimated 25 years of
margin, that was for beyond the 40 year? In other
words, a total of 65 years?
MR. SEVERSON: That is from now. That is
about from now with -- well, the differences that I am
seeing in wear rate, I probably would not change my
designs from when I think we should change by about,
and depending if we went another 60 years, or another
10 years, I would probably have about the same
numbers, whether we had a power uprate or not.
Because the wear rate right now until when
we do a piping change, or decide to do a piping
change, is not that much of an added effect, compared
to what we have had since the beginning.
In actual fact, I think our chemistry
probably in the early days wore us more than what we
are going to wear now with a power uprate.
MR. ROSEN: You are saying, I think, that
if Duane Arnold were to get or to come in for a
license renewal that it would do it at the higher
power level which it is now asking for, and not have
to plan a piping replacement. Am I correct?
MR. SEVERSON: Not in this area. I don't
believe so in feed water, and in some other areas, we
are probably going to be doing pipe replacements
anyway.
But some of the other areas that we were
looking at, like some of the extraction steam lines
are actually going to be improved under a power
uprate, but change them anyway just because of where
we are at.
But overall the majority of the piping
will not be affected by a power uprate, and what we
are going to decide to do, and what we are going to
decide to change out, won't be affected.
I can't answer your question directly
partially this is a continuously monitoring program,
and we have done some pipe replacements, and we will
probably do some more because of varying different
reasons. And some of the reasons that we do pipe
replacements is because we don't want to inspect it
anymore.
We know that if we put in a better piece
of pipe that I can reduce my inspections, and I can
save money that way. But I don't consider that the
EPU will have much effect on the decisions that we
make.
MR. PARK: Good afternoon. My name is
Gary Park, and I am the ISI Program Engineer for the
Duane Arnold Energy Center. I administer all the
inspections that we do on the reactor vessel, and on
our ASME Section 11 components.
The first slide that I would like to talk
a little bit about is about the program that we have
at Duane Arnold. I think we have a pretty aggressive
ISI, IVII, program, and IVII being internal vessel and
internal inspections.
If you will notice for the Class One
components -- and I have only counted back to 1985,
but by the year 2005, we would have done 1,875
inspections just on the Class One systems.
And so the power uprate as far as the
effect on the structural integrity of these
components, we have already got a pretty good base
line inspections for those.
The thing that I need to bring out about
the inspection program is the fact that we find
problems before they actually exist to a failure. We
also utilize in our inspection program the recommended
inspections of the boiling water reactor vessel
internals project.
And I hope that everybody on the panel or
on the committee is familiar with that, because I am
sure that you have been addressing different safety
evaluations from that particular group of utilities
and their recommended inspections.
I think one of the materials that we
should address in this presentation is the stainless
steel materials that we have inside the reactor
vessel.
Again, we perform the recommended
inspections of the BWRVIP, and we follow all of their
documents, and we have a pretty aggressive program in
doing so.
For example, the course route, we have
inspected all the H-1 through H-7 wells twice since
1985, and we have not found any IGSCC, intergranular
stress corrosion cracking, in any of those welds.
So that shows that we have a good base
line prior to power uprate in a particular important
component that the industries have been finding
problems with.
DR. FORD: And on that particular item, it
is true isn't it that most of the VIP disposition
curves, et cetera, have not been obtained, or are not
based on data rather at relevant flow rates?
MR. PARK: The recommendations made from
the VIP is in fact on safety and not based on any
pressures or temperatures. It is just based on if
that component fails, where are the areas that we
should inspect.
DR. FORD: But are the frequency of your
inspections based on disposition curves?
MR. PARK: I am not quite sure I
understand what you mean.
DR. FORD: Well, what are the inspections
based on?
MR. PARK: It is based on material and
your --
DR. FORD: And if you find a crack?
MR. PARK: Then you increase your
frequency, yes.
DR. FORD: And the frequency is dependent
on the degradation rate?
DR. FORD: Sure. Sure. And crack growth
rate would be one of them, yes.
DR. FORD: My point is that most of the
crack growth rates which go into deriving what those
disposition curves are, are being based on data not at
high -- well, do you understand what I am saying?
MR. PARK: Well, I understand what you are
saying. I don't know that I know the answer to that.
DR. FORD: I guess going back to the very
first slide, "Inspection Programs finds problems prior
to failure."
MR. PARK: Right.
DR. FORD: And which assumes that you are
inspecting --
MR. PARK: At a frequency, that is
correct. That is correct.
DR. FORD: And that is the origin of my
words. And it goes on to the next question, and
talking about DAEC performing examination of vessel
internals, and we are particularly interested in
IASCC/IAGSC.
It was mentioned earlier that the profile
has changed.
MR. PARK: Well, I will defer to Tony on
that, but it is more flattened out, but it has changed
some.
DR. FORD: And therefore the pressure at
the core shroud has increased?
MR. PARK: Yes.
CHAIRMAN WALLIS: Do we know how that will
affect cracking at the core shrouds, and at that prior
flux, and therefore fluence, especially if you are
going to extend -- the fluences are all going to
increase at a higher rate?
MR. PARK: Yes, and there is some
threshold and that's when IASCC starts, and I am not
sure where we are at as far as Duane Arnold. I think
we are approaching that.
MR. BROWNING: This is Tony Browning
again. We have already exceeded the VIP threshold for
IASCC in like the top guide area in the upper shroud
area. The other thing you mentioned was the increase
in fluence.
One of the things that we noted when we
did the fluence calculation was that the increase to
the shroud area wasn't as dramatic as you were
expecting, and that was because of the partial rods
from the GE-14 design that we were going to. There is
just less neutrons there. It is not as dramatic as
the uprate itself.
MR. PARK: And then I think the other
important thing to note is that we have done probably
the highest percentage of any inspection that is done
on these particular welds, and we have not found any
cracking at all.
So we have a real good history of water
chemistry, and then as I will address in a later
slide, we have done the mitigation measures to help
support and continue operation of that.
In fact, that is a good lead into the next
slide. Duane Arnold has implemented hydrogen water
chemistry which protects our recirc piping, which is
stainless steel, and we were the lead plant in the
industry in getting a relief from inspection
frequencies based on the results of our hydrogen water
chemistry.
And as we have continued to do
inspections, we continue not to find anything, and so
we believe that HWC is effective in mitigation of
IGSCC in our stainless steel piping, particularly our
recirc piping.
And then in 1996, which has already been
mentioned here on the committee, we were the pilot
plant for the Noble Chem, and we have since injected
Noble Chem another time. So we have injected twice,
which does enhance the effectiveness of HWC in
protecting the reactor internals.
DR. FORD: Which is of importance in
monitoring, and not just crack monitoring, but
environmental monitoring. Remind me, but at Duane
Arnold do you have corrosion potential monitors in the
core?
MR. PARK: We have installed those, and we
do have a caste system that is external that has
reactor fluid in it, reactor water that runs through
it.
DR. FORD: The reason for my question is
not quite the answer to the question that I asked. My
concern is that, yes, you have Noble Chem, and yes, it
will stop cracking in the core, but the question now
is that if you increase the flow rate in the core is
there going to be any additional danger by that one
action of increasing the flow with Noble Chem?
And that to a certain extent is only going
to be answered if you have corrosion potential
monitors in the core.
CHAIRMAN WALLIS: Well, that hasn't
changed, the core flow hasn't changed in the power
uprate. That is only the feed water and the steam
flow that have changed. The core flow stays the same
doesn't it?
MR. BROWNING: But back to your earlier
question, and this is Tony Browning again. We do have
in core monitoring. We replaced one of the LPRMs
streams with the ECP monitors at the time.
MR. PARK: We have done that in the past,
yes. They don't last very long as everybody knows.
MR. BROWNING: Right.
MR. PARK: But we have done it in the
past.
MR. BROWNING: Yes, to demonstrate the
effectiveness of the Noble Chem injection.
MR. PARK: Right.
MR. BROWNING: And as Gary pointed out, we
have the external cracks verification system, the
outer clave with the pre-crack specimens in it to
monitor the effectiveness of water chemistry.
MR. PARK: Before I got to my conclusions,
I think I will turn some time over to Mr. Al Roderick
to answer the stress question that was brought up
earlier if I may, and we have an overhead of that.
MR. RODERICK: I am Al Roderick with Duane
Arnold. The question that was raised earlier was
based on a review of a response to a staff's REI in
the area of stress analysis.
In looking at the main closure flange from
current to EPU, I believe if you do the math of that,
I think it is about a 12-1/2 percent increase that has
been evaluated. What that is a result of is from GE's
methodology in looking at EPUs, is to not redo a
complete code stress analysis for the vessel.
They have in their methodology is the
determination of scaling factors based on changes in
perimeters from the code of record, or the calc of
record, to the EPU conditions. It could be in the
area of pressure, temperature, flow rates,
particularly with flow rates with impact nozzles.
And they determined the stress of the
scaling factors that would be applied. And they don't
do it on an individual component basis. When looking
at the reactor vessel, it is split up into zones if
you go back and look at the original diagrams for
defining operating conditions.
And so wheat they did was to
conservatively apply the maximum scaling factor that
came out of a particular region in the vessel, and as
I pointed out earlier, as you are going back to the
calc of record where the stresses are coming from, and
in radioing up the EPU conditions.
So I don't have the specifics of what all
fed into the 12-1/2 percent, but it is based on a
conservative screening methodology for a good
description, because it is a first cut, and it is
applied to the entire stress intensity.
It is not usually split out in terms of
pressure thermal mechanical loads, et cetera. The
highest ones apply to the total stress intensity to
get a conservative extrapolation or prediction of the
stress, that is then compared to the code allowables.
And because all the code allowables were
met, nothing more detailed or refined was done.
CHAIRMAN WALLIS: Are you saying that the
reason that there is a 12-1/2 percent difference from
current to power uprate is because a different method
is being used?
MR. RODERICK: We are not using a detailed
computer or code calculation. We are using a scaling
--
CHAIRMAN WALLIS: So then the 12-1/2
percent is somewhat illusionary?
MR. RODERICK: It is based on changes in
parameters, and I don't have all the details.
CHAIRMAN WALLIS: I would think the main
closure flange is mostly influenced simply by the
pressure in the vessel isn't it?
MR. RODERICK: Well, as I said earlier, it
is not done on a component specific basis. It is done
for the whole region in the vessel. So a scaling
factor of 12-1/2 percent increase may have come from
a different component in that Region A of the vessel,
and is conservatively being applied to the flange to
evaluate those.
CHAIRMAN WALLIS: Well, it still doesn't
explain why the numbers come up by 12-1/2 percent when
the pressure has hardly changed at all. There is
still some mystery, which maybe you can clear up with
the staff or something.
MR. PARK: Let me try something now.
Instead of doing a full-blown code stress analysis for
a power uprate, GE took the conservative approach to
make sure that all these regions in the vessel would
still meet the code allowable.
CHAIRMAN WALLIS: That's okay. So if you
simply look at the EPU versus code allowable, that is
what you are saying.
MR. PARK: Right.
CHAIRMAN WALLIS: But the problem that I
have is that when I look at the difference between
current and EPU, which should tell me by how much are
you changing things, then that 12-1/2 percent is not
something that I should take seriously?
MR. PARK: And you brought up the point
that the pressure is not changing, and so why do we
see a change there, and all it is there is a
conservative --
CHAIRMAN WALLIS: It is a different method
of calculation.
MR. PARK: It is just a conservative
number being added to see if we still meet code
allowable, as opposed to doing the number crunching on
a full-blown code stress for the component.
CHAIRMAN WALLIS: So the comparison
between current and EPU is different because different
methods are being used. They weren't so conservative
before? Is that what I am gathering?
MR. PARK: Well, I am sure that the
original design was very conservative.
MR. MCGEE: This method was adequate to
demonstrate the margin --
CHAIRMAN WALLIS: You are getting close in
terms of the 80,000 and the 77,364. Presumably the
staff asked this question for some reason, and this
was supposed to answer some question was it? The
question was whether or not the stresses were code
allowable was it?
MR. PARK: It is just to demonstrate that
we are still meeting code allowable designs.
MR. MCGEE: We did have discussions with
the staff and with the particular reviewer on the
method that was utilized.
CHAIRMAN WALLIS: Well, maybe when you
come to the full committee that you can have a better
explanation of why the numbers differ by so much from
current to EPU, because it still seems to me that we
are just saying that if somebody used a different
method -- if you use a different method, then why show
the comparison, and it is a little foggy what the
comparison is really showing us.
MR. RODERICK: The request from the staff
was what did we use to access the acceptability of
stresses in these components, and in the work that had
been done was a conservative scaling up of the current
calculated stresses based on a maximum scaling factor
in the region, and probably in this case came from a
different component.
And then compared to the allowable or the
acceptance criteria. So this was the basis for
demonstrating margin and acceptability at EPU
conditions for these components. And the two pieces
that I was able to look at for the closure flange
itself is in the original analysis, and the original
drawings for the pressure term was using a thousand
PSIG.
And in doing consideration of this area of
the vessel, we are now looking at a 1,025. So just
looking at that ratio itself would be at 2-1/2 percent
increase. So that obviously is not it.
The temperature change is 3 degrees, and
that is just based on a saturation temperature. So
with those two pieces of information, I am very
comfortable that this 12-1/2 percent scaling factor is
from another component that is still part of this
region of the vessel.
CHAIRMAN WALLIS: Well, I don't want to
pursue it anymore. I think when you come back, if you
could identify what that component is, and give us a
clearer explanation of why the numbers are so
different, and the full committee will be satisfied.
DR. POWERS: I don't know whether you are
the correct speaker or not, but who should I ask about
the fatigue usage factors?
MR. PARK: Fatigue usage factors?
DR. POWERS: Right.
MR. PARK: Do you have a question?
DR. POWERS: Well, in looking at your SAR,
I noticed that your fatigue usage factors usually went
down, and it was kind of surprising. And when I read
the text, it said that they had used a less
conservative method of analysis when they calculated
the fatigue usage factor.
And in some cases they produced some
remarkable reductions in the usage factors. For
instance, the hydraulic system return nozzle went from
about .85 down to .57. There is another case where it
went from .97 to .2.
And I just wondered what the less
conservative analysis method was. I mean, what was
entailed. But then I went on and I noticed that your
feed water nozzles were -- that the usage factors
actually went up pretty dramatically.
They went from about .85 -- and this is
end of license times, and so .85 up to .968, and it
doesn't surprise me that the usage factor would go up
on the feed water nozzles.
But it seems like a big jump, even if you
were using a less conservative analysis method.
MR. PARK: So you want us to just address
what that analysis was?
DR. POWERS: I just would like to know
what the differences were in the method of analysis.
MR. PARK: I was not prepared to do that,
but we certainly can write something up. Do you want
us to bring that back before the full committee?
DR. POWERS: You can just tell me one way
or the other, formally or informally.
MR. PARK: Okay. As far as my
conclusions, I think we have pretty much addressed
those during our discussion. We follow the
recommendations of the VIP, which I think is an
industry standard that is going to be developed, and
I believe that the VIP has also come out with a
recommendation for going out and doing self-
assessments to make sure that we are implementing
those products.
We used Noble Chem and HWC, which has been
shown to be an effective mitigation, and that those
effects are going to help in the power uprate. And
then also our vessel internals have been evaluated,
and it is important to note that they still meet the
design criteria with some margin.
CHAIRMAN WALLIS: Now, how did you decide
what is sufficient margin? They meet the criteria,
but --
MR. PARK: Right.
CHAIRMAN WALLIS: And you start getting
into one margin that is sufficient, and that gets
again fuzzy doesn't it?
MR. PARK: Well, they meet the criteria.
They are still under what the design margins are, or
the design is.
CHAIRMAN WALLIS: But you were very
uncertain about your predictions. You presumed that
they have a bigger margin.
MR. PARK: Excuse me?
CHAIRMAN WALLIS: If you meet the
criteria, but you are close, and then you say that we
are uncertain in our predictions, and we had better
back off, then that would be increasing the margin
because of uncertainty wouldn't it?
MR. PARK: Well, there was some
conservatism in developing that criteria, and in
developing what it was.
CHAIRMAN WALLIS: So, criteria with
conservatism.
MR. PARK: Well, yes, that might be a
better way to put it, yes. Is there any other
questions? Thank you.
MR. BROWNING: Quickly. Dr. Powers, we
have the calculation for the hydraulic system return
line, but it is proprietary material. We can show it
to you over the break if you would like to see it.
DR. POWERS: That would be fine.
MR. BROWNING: Great.
MR. HUEBSCH: My name is Steve Huebsch and
I with the Duane Arnold Energy Center, and I am going
to present some information pertaining to the
containment pressure temperature response from the
EPU.
Specifically the areas of interest that
were looked as parameters as part of the analysis were
the drywall pressures, the drywall gas temperatures,
the drywall shell metal temperatures, the wet well
pressures, the suppression pool water temperatures,
and the containment loads.
These parameters were looked at both in
the short term and in the long term. The analysis
looked for both peaks, as well as the specific results
and comparison between the two.
DR. SCHROCK: This containment, it is
BWR4, is the toros containment?
MR. HUEBSCH: Yes, it is. It is the Mark-
1.
DR. SCHROCK: Thank you.
MR. HUEBSCH: One thing that I want to
address and that is probably the most important thing
as far as evaluating the containment structures is
when you look at the analysis and the way the analysis
is done for both the peak drywall pressures and the
temperatures, as well as the
Mark-1 containment analysis for the load stuff, they
start basically with a thermal hydraulic analysis to
develop the loads based on the Mark-1 program, and
testing that was done in accordance with those days.
Once those loads are developed, those
loads are put into the structural calculations, and
those structural calculations then are required to
meet the ASME code requirements for a containment
vessel.
And in all of the cases that we looked at,
we were able to maintain the ASME code allowables as
defined by the original design requirements. The
methodologies that were used for the analysis
CHAIRMAN WALLIS: Are you going to talk
about this 5 percent hydrogen limit? Is that part of
your discussion or somebody else's?
DR. POWERS: I don't know what limit you
are talking about?
CHAIRMAN WALLIS: Well, I was trying to
understand the SAR, the draft SAR, and there is a lot
of stuff about combustible gas control and 5 percent
hydrogen, and it seems to be pretty obscure.
MR. HUEBSCH: That is not directly
associated with this presentation, but we can discuss
it. I guess
DR. KRESS: That is a corrosion production
of hydrogen at 5 percent. It generally is not
important generally.
CHAIRMAN WALLIS: It is not important?
MR. HUEBSCH: It is dealing with post-
accident flammability issues with hydrogen-oxygen
generation, post-LOCA.
DR. KRESS: Their past system was supposed
to be designed to deal with that kind of levels.
CHAIRMAN WALLIS: That's right, and
monitoring --
DR. KRESS: Yes.
MR. HUEBSCH: Monitoring, and then dealing
with it such that we don't end up with a flammability
situation post-accident.
CHAIRMAN WALLIS: Well, maybe we can just
ask the staff to explain that one then if you don't
want to.
DR. POWERS: We will meet with the staff
tomorrow on that. You are not responsible for the
SAR.
MR. MCGEE: We can discuss that, but --
DR. POWERS: Well, we can have the staff
do that tomorrow.
MR. MCGEE: Well, I would be more than
happy and if you want to wait until this is done, then
I can answer any direct questions.
MR. HUEBSCH: The analysis methods that
were used to do the containment analysis, in the short
term cases, to come up with the peak drywall
pressures, and to determine a short term temperature
in both gas, as well as suppressible temperatures, was
the M3CPT model that GE has.
This is the model that was used in the
Mark-1 containment analysis. It was approved for use
at that point for the short term analyses. In the
long term event, which is looking at heat up based on
decay heat changes and things of that nature after 8
hours, 10 hours, out.
CHAIRMAN WALLIS: Why would you expect a
difference with the power uprate? Is it because of
the heat stored in the metal and the fuel?
MR. HUEBSCH: For which case, the short
term?
CHAIRMAN WALLIS: Is it a difference heat
source; is that what it is? Why is there a difference
in the power uprate?
MR. HUEBSCH: In the short term, you see
certain things, and in the case of the Duane Arnold,
we see a little more sub-cooling. So when you have
the break, you have more mass transferred to the
containment structure.
You see some changes in the pressure and
in the longer term, you have a higher decay heat, and
you transfer that heat. So you will see some changes
in this analysis, and the changes were in accordance
with what was expected because of those specific
attributes to the power uprate.
The long term model that was used in
accordance with the ELTR is called SHEX, and that was
done to do not only the DBA-LOCA cases, but the other
longer term analysis -- station blackout, the MPSH
analysis for ECCS, and other methods.
And the SHEX model has been approved only
a case by case basis. It i not generically approved
like the M3CPT model was, but it is in accordance with
the ELTR.
The loads, the specific loads on the
containment structure, the Mark-1 containment loads
were done in accordance with the Mark-1 program. The
new loads as developed by, or as looked at, were
compared back to the original test data, and the
original program to determine whether or not it was
previously bounded by the cases that were analyzed for
the initial program.
The methodologies used were bounding
correlations, and the models are conservative by
nature, and they are benchmarked back to the original
analyses, and they are qualified against the test data
that was done for the Mark-1 stuff.
One specific issue that is important is
the increase in the containment peak pressure, and
this inputs into our local leak rate testing and
various other things that we do to maintain
containment integrity.
DR. SCHROCK: This is on a scaled test
data and is the range of the parameters that are
changed by the power uprate, and is that covered by
that testing range that exists?
MR. HUEBSCH: Yes. The way the original
testing was set up, it was based on things like pool
swell and various things, and loads from the SRVs, and
the blow down model through the vents.
These were analyzed numbers, and then they
were -- and then the specifics of Duane Arnold were
compared to those values that were tested in the low
definition report developed by GE, and then other
analysis.
And, Dan, I don't know if you wanted to
add anything to that or not.
MR. PAPPONE: This is Dan Pappone with GE.
There are two basic test approaches. One was a
generic bounding test configuration that was developed
to bound all Mark-1 containments.
So they ran the one test for all
containments, and what we are doing in the individual
plant applications is that we are comparing either the
original analysis or in this case the power uprate
analysis, to confirm that we are still within that
original test basis.
There are some tests that are done on m
more of a plant unique basis, where the test facility
itself may be -- you know, the geometry there is
fairly fixed, but some of the parameters, the initial
parameters, were set up to bound a specific plant.
And there again we are looking at the
power uprate conditions to confirm that we are still
within or bounded by the actual test.
DR. SCHROCK: And I guess that was the way
that I was thinking of it. Ordinarily, you would want
your tests to cover the range of parameters to which
it is going to be applied.
And here you are extending that range of
parameters in a power uprate program.
MR. PAPPONE: Right, but we are going back
and confirming that once we have extended the plant
specific values to the power uprate conditions, we are
still within the original bounds of the test, those
parameters.
DR. SCHROCK: Okay. Thank you.
MR. HUEBSCH: In this case, it shows that
basically the peak containment pressure analyzed has
gone up 3 pounds.
DR. KRESS: Do you have to do that also
for ATWS events?
MR. PAPPONE: Did we run these cases for
the ATWS events?
DR. KRESS: Yes.
MR. PAPPONE: There was a pressure
temperature analysis that was done.
DR. KRESS: Was it less than this?
MR. PAPPONE: Yes. The 45.7 psi occurs
very quickly in the DBA LOCA event, and it is the peak
pressure that is identified as analyzed per the whole
series of accidents.
DR. KRESS: For the whole series of
accidents. Okay.
MR. HUEBSCH: One of the issues that the
long term SHEX model gets involved in is the use of
containment pressure for an ECCS pump performance. At
the Duane Arnold Energy Center, the plant was
originally licensed with the use of containment
overpressure for the core base systems specifically.
And in the original RHR core spray pump
specifications, and in the containment specifications,
there is actually criteria for how to analyze for the
containment pressure models.
We have stayed within the original license
bases, and the design bases for the containment
analysis that we did today as part of the EPU.
The specific analysis, because when we got
involved with the ECCS strainer issues, there were
some aspects of the PRA that looked at what happens if
you lose your injection capability, as well as lose
your containment.
So those aspects have been looked at for
insights, as far as the use of containment pressure,
and what would happen if you lost it. The other thing
is that when we ran the containment overpressure
analysis that we were consistent with both the branch
technical position that was written for this is how
you should analyze to mitigate -- to minimize your
pressure and maximize your pool temperatures.
As well as the original specifications for
the plant. So we applied those aspects when we ran
the cases, and the analysis also includes things like
containment leakage, and it factors those in so that
you are decaying off your containment pressure as the
event goes on.
What you see here is the results of the
analysis and where after the MPSH calculations were
calculated what are the reliance on containment
pressure is.
And there were two specific issues or
points that were significant. One was at the 10
minute mark, because prior to 10 minutes the pumps
were at run out conditions.
And at the 10 minute mark the operators
restrict the pumps to rated conditions. Although
there is pressure available in accordance with the
analysis, we have no reliance on containment pressure
in the first 10 minutes of the event.
But what we have found at Duane Arnold is
that the reliance on pressure -- and this is in
accordance with the original license -- occurs at peak
pool temperatures.
And the black is the available, and the
others required for original license, we require 3.1
psi for over pressure. And we will be looking at 5.3
psi and EPU conditions --
CHAIRMAN WALLIS: That's because the water
is hotter in the pool?
MR. HUEBSCH: Correct. The water
temperature has gone up, and I believe where we were
analyzed after completion of the ECCS strainer
installations was roughly 202 or 203 degrees
fahrenheit peak pool temperatures, and we are looking
at 209.2 degrees now.
And so a seven degree increase because of
EPU for this specific analysis. One thing at Duane
Arnold specifically is that the pressure is used for
core spray, and you run into a temperature issue.
Core spray requires over pressure roughly
at 180 degrees. So anytime the pool temperature
reaches 180 degrees or above there is some reliance on
over pressure with the current analyses assumptions,
which are very conservative.
For the RHR system, the way we are
configured is that after the events of the LOCA and
divisional failure, you are down to one RHR pump. We
don't require containment over pressure for that one
RHR pump.
If you had two RHR pumps running, there is
a requirement, but that's not our design basis, but we
have analyzed all those cases. In the continual load
section, Dan talked about that a little bit.
The specific loads that were evaluated for
EPU were in line with the original Mark-1 pool swell,
vent thrust, condensation oscillation, considerations
of chugging, and SRV discharge, both the first pop, as
well as the second pop, and the impacts of low, low
set.
And whether there were any changes between
our current configuration and EPU. The only one of
these that had any impacts on the original loads that
were analyzed were the vent thrust section, because we
are seeing a higher dry wall pressurization rate.
And so you have a larger load on the vent
system as the blow down model comes through the vents.
The loads were increased roughly by five percent. It
was a scaling or a linear evaluation rather than a
detailed evaluation as was done in the Mark-1.
CHAIRMAN WALLIS: This is just a momentum
of the fluid coming out of the pipe; is that what it
is?
MR. HUEBSCH: I believe so. Dan, is that
correct?
MR. PAPPONE: This is Dan Pappone. The
basic vent thrust loads are from the momentum of the
flow through there, with the power uprate looking at
a little bit higher -- well, it is a trade off between
a little higher initial break flow due to the
subcooling, and a little bit lower energy coming out
of the flow.
So every pound coming out is a little bit
lower because of the higher subcooling, but we are
getting -- the flow is coming out a little faster.
The next effect of that is a little higher
pressurization rate in the dry well, and that shows up
in the flow through the vents and the thrust loads.
And we run that through the Mark-1
calculational methods to come up with that 5 percent
increase in the load definition.
MR. HUEBSCH: And those values were then
compared to the structural allowables, and we are
still within the allowables for the program. So it
still meets the requirements of the ASME code, and all
the margins are maintained.
Let's go to the conclusions then. One
other area where one of the limits were challenged is
in the station blackout event, at about roughly 3.7
hours into it the temperatures exceed the 281 degree
containment design temperature.
And what was done in that case was the
pressure and temperature requirements were looked at
in comparison to the design requirements. Our
containment design is 56 pounds at 281 degrees
fahrenheit.
In the case of the station blackout event,
it reached 283 -- well, just short of 284 degrees at
the four hour point basically, 3.7 hours out, with 8.7
psi.
So because you are at such a low pressure
and the temperature is only there for a short period
of time before the four hour coping period is over, it
was analyzed as being acceptable.
As we said earlier, the vent thrust loads,
and the dry wall temperature as I just said, and the
station blackout, were the only two events that
challenged the thermal hydraulic analysis that had
previously been done for the plant. So everything
else was bounded.
And the structural analysis of all the
events, including those two, after the loads were
changed or evaluated for the higher considerations,
were still within ASME code. So there were no
challenges to the DAEC containment.
MR. KNECHT: I am Don Knecht from GE, and
I am here to talk about the separators and dryers, and
really a specific aspect of it. As you see here on
the outline, the basic things that we are going to be
focusing here on are the loads, and the separators,
and the dryers, and some of the dryer experience that
we have been having.
There was an RAI asked by the NRC dealing
with the flow induced vibrations, and that's really
the emphasis here. There are some other aspects, but
I am not going to address those now.
First off, a little bit on the impact of
EPU. Obviously there is a steam flow increase, and
both the steam flow increase coming out of the core
affects the separators and turns the excitation forces
which are transmitted to the shroud.
The dryer sees the increase flow pretty
with regards to the power increase, and along with
this is an increased pressure drop across the dryer.
The other issues that I am not going to deal with here
are the moisture content issues and the effect of the
carry under change that goes on with the dryer
performance, to just to try to contain the discussion
a little bit.
Now, on the separator, the excitation
forces that are going on are primarily from the flow
increase, and also the swirling action in the
separator as it is going out.
Those are increased, but Duane Arnold, not
coincidentally, but Duane Arnold was the prototype
unit for the BWR4 in terms of the stresses on the
separator, and were instrumented at the time of start
up.
And they found that the stresses at that
time were only about 15 percent of the allowables.
With the EPU scaling it up for the increased flow and
what not, it shouldn't be more than about 20 percent
of allowables.
So as far as the separator is concerned,
there really is no concern, and there is quite a bit
of margin preserved for that. So I don't really see
any issues with the separators themselves.
Now, on the dryers, the dryers are
designed -- first off, they are a non-safety related
component. It's main function is to keep the moisture
content of the steam below a certain goal.
From a safety standpoint, we don't want
any failure that a dryer such that there would be a
lose part that could go and impact, let's say, an MSID
closure or some other consequence.
So the dryer is designed for the main
steam line break event and it has sufficient margin as
it was originally designed to show that a main steam
line break would not result in any adverse
consequence.
Now, with the EPU, that event does not
change because we are at constant pressure, and the
main steam line break is a choke flow type of
consequence. So there really is no impact on the
loads on the dryer due to that.
So the structural integrity of it should
be maintained. Now, the question in the RAI dealt
with flow induced vibration, and because it is a non-
safety component, it is not something that is analyzed
with codes and what not.
Instead, it is more of a qualitative
evaluation that is done, and because of the flow
increases the load should increase by about 31 percent
was the estimate.
CHAIRMAN WALLIS: This is based on a --
MR. KNECHT: Yes.
CHAIRMAN WALLIS: Now, is that really the
whole story? I mean, don't you get vibration due to
resonances and things which are not just proportional
to momentum?
MR. KNECHT: This is really dealing with
the amplitude of the flow induced vibrations. The
frequency stays the same, because they are based on
the natural frequency.
CHAIRMAN WALLIS: Unless you have some
sort of resonance between some wall shedding or
something and the mechanical behavior. You are way
away from that and maybe you are right.
MR. KNECHT: That is not the concern.
What we have done traditionally on the dryer
performance, or not so much the performance, but the
flow induced vibrations, is that we have looked at
this on a fleet wide basis.
As it turns out, Duane Arnold has not had
any particular problems with their dryers, in terms of
this, but there have been cracks that have gone on in
the dryer drains and some other components.
And so it has been looked at for several
years, and we have a database going back into the mid-
1980s tracking various dryer cracks that have been
found.
So those have been used in a way that
tries to identify areas that we think should be looked
at. The VIP program talks about since the dryer is
going to be removed during outages anyway that a
visual inspection should be done on the dryers, and
that is what has been done in pretty much all plants,
but at Duane Arnold at any rate.
We use the fleet experience to try to
guide those inspections as to what ought to be
inspected, but the cracks that have been seen have
been pretty odd, and they have not been so much of a
problem.
So the areas where we have seen some of
the more dramatic cracks have been in the drain
channels, where we have seen some fairly significant
cracks. But none of these cracks have led to any
concern with the integrity of the components.
And so we use this program as sort of an
operational way of evaluating the integrity of the
components.
The other main point here is -- and
getting on to the next slide, is that once these
cracks are identified, they are readily repairable
because the dryer is available in the pool, and they
are generally repaired, unless they are so small that
another cycle or so would not lead to any real
concern.
The experience that we have had so far is
that there have been two types of cracking. The IGSCC
cracking has been a little bit more than half the
cracks that have been observed. But those are not
really impacted by EPU.
The chemical environment in the steam has
not really been changed by EPU, per se. It is mostly
just a steam environment. So we don't see any impact
of EPU on IGSCC.
Now, the high cycle fatigue is the other
area, and clearly there is an impact there. But again
we have seen no cracking at Duane Arnold, and many
plants have seen cracking, and they have been all
repaired.
MR. ROSEN: Do you have a visual of this
dryer where you can show us where the cracking has
been observed?
MR. KNECHT: Over there I do.
(Brief Pause.)
MR. KNECHT: This is the general area.
This is a brief diagram here of the dryer, and this is
the top of the separators here coming up, and there is
just a little bit of a gap here between the top of the
separators, and these are the typical dryer drains
where the steam will come up through these channels
here, and through the dryer assembly, and then out.
Now, the moisture that comes off of the
dryer collects down here in these troth areas here,
and that leads into -- well, these are the bottom
drains that lead into a troth, and then these are the
drains that go down here and into the separator area,
and combine with the separated moisture that is
removed and then back.
But what doesn't really show on this
diagram is that the cracks that have been seen are in
some of the drain channels that lead from here out and
down.
And subject to the vibrations that get
generated here in the dryer drains, and so it is
transmitted back down through that structure.
MR. ROSEN: You called them channels. But
are they open at the top or are they pipes that are
closed?
MR. KNECHT: The troth is open down in
this general area, and then those troths drain into
some pipes.
CHAIRMAN WALLIS: Well, the things that
shake are the louvers aren't they? Whatever they are,
the things that have the initial impact on --
MR. KNECHT: The drains here?
CHAIRMAN WALLIS: Yes. And those are the
things that shake?
MR. KNECHT: Yes.
MR. ROSEN: So I am still trying to figure
out what cracks.
MR. KNECHT: The drain channels -- and
unfortunately they don't show this, but if you go in
3-dimensionally, there is some --
CHAIRMAN WALLIS: Well, it is a funny
place to crack if the drains are shaking.
MR. KNECHT: That is the forcing drain.
CHAIRMAN WALLIS: It is transmitted down?
DR. FORD: Are those welded to conform
down there?
MR. KNECHT: There are some welds, yes.
DR. FORD: And the cracking, presumably
the stress is associated with those probably?
MR. KNECHT: It could be contributing.
CHAIRMAN WALLIS: And so because the pipe
is further, it is the rigidity of the whole structure?
The pipe is helping to retain --
MR. KNECHT: There are probably some
stresses there. Because they are easily repaired, I
don't think we go into a lot of analysis as to --
MR. ROSEN: Well, you are worrying about
the wrong end of the problem. I mean, I grant that
they are easy to repair, but what I am concerned about
is one of those parts carrying away during operation,
and what would happen then.
But I can't get a good feel for what would
carry away since I don't have a picture of it. Can
you help me with that question? What if the crack
proceeded to where it severed the component?
CHAIRMAN WALLIS: It would just leak
wouldn't it? I mean, it's whole --
MR. ROSEN: I don't care about leakage.
MR. KNECHT: If a part is completely
carried away -- well, first off, we have never seen an
experience where we saw that it was completely covered
away.
If it did, it would become some kind of a
lost part, but I don't think it would go -- this is
down below the dryer assembly, and it would probably
find its way into this area someplace.
Now, there would be an increase in
moisture coming out of the dryer, because you would be
bypassing things and we are not concerned about that.
So it has not really been a concern.
MR. ROSEN: Where would a plate of steel
or an elbow of pipe that came lose there go? Where
could it go?
MR. KNECHT: I suppose that it could find
its way up here, and block part of the drain here.
MR. ROSEN: There is no way that it could
get down below the separators?
MR. KNECHT: No, because steam is going
up.
MR. ROSEN: Yes, but not all the time.
When you shut down --
MR. KNECHT: It could go back through.
MR. ROSEN: Go with me for a minute on
this. You have got a crack, and the crack proceeds to
where the part fails. It is a piece of steel now,
regular shaped.
Now for some reason in their wisdom, the
operators decide to shut the plant down, and now there
is very little steam. Where does the part go?
You said don't worry about it, there is lots of steam.
Well, not all the time.
MR. KNECHT: Wouldn't it just lay down on
top of --
CHAIRMAN WALLIS: The pipe is held at the
other end if it cracks off at the place you indicated,
and it is just held at the other end, and the forcing
function has gone away because it is broken off.
MR. KNECHT: Well, it might come down
between the separators.
CHAIRMAN WALLIS: What if it doesn't come
down at all?
MR. ROSEN: Well, it has broken loose.
CHAIRMAN WALLIS: No, it is only broken on
one end.
MR. KNECHT: Well, it would wind up
somewhere in that region, and probably lay on top of
the separators.
DR. KRESS: I don't think I would put this
one in my PRA.
DR. FORD: I think the argument is going
to as far as that particular mode of degradation is
concerned, it is not going to change.
MR. ROSEN: It is not an EPU specific
problem. All I am trying to get someone to say is
that it won't get down and damage the fuel, and hit
the fuel or the controller out drive, or something
like that.
Can you say that, that it can't get below
the separators and get down to the fuel? Can you say
that?
DR. KRESS: If you have ever seen those
separators, it would have to be a mighty small piece
to get down there.
MR. KOTTENSTETTE: How big a part are you
saying has broken off? Is it something that size or
a piece of something this long?
DR. SCHROCK: Well, your experience with
the crack should tell you something about what a
potential piece may be, and what it's size and origin
might be.
MR. MCGEE: But if it resulted in a piece
being broken off on one end and taking away the
stress.
MR. ROSEN: You have a lot of experience
with cracking of these things to know that they don't
result in pieces, but I don't have that similar
experience. And cracking can be a funny thing, and
you could end up with a crack that proceeds in a way
that a piece comes loose in my world.
Now, I am only asking whether that piece
could go down and cause some real damage in the fuel
or in the control rod drives.
DR. KRESS: It is about the size of a
quarter. It wouldn't hurt the control rod drive.
MR. PAPPONE: This is Dan Pappone. The
region that we are talking about is outside of the
shroud, and the fuel in the control rods are inside
the shrouds. So we have got an area there --
DR. KRESS: Yes, it would never bother the
control rods.
MR. KNECHT: If it went outside the shroud
region, it would drop to the bottom, and where the
recirc pump suction is. So unless it is just the
right size part, and just with the right dimensions
and weight, and all these improbabilities, it is not
going to cause any problem.
I mean, the one thing about the drain
channel cracks is that those have been several inches
long. They are not little flakes of something.
So if something were to break loose, and
it is hard to imagine that when all that stress is
relieved, it is going to be a large part. There has
been no evidence that any part has ever come loose.
CHAIRMAN WALLIS: And all of this is
because of the drains are shaking up above?
MR. KNECHT: Well, that and probably --
DR. KRESS: It is probably residual
stresses like he said.
CHAIRMAN WALLIS: Well, the velocity
through the drains is pretty low isn't it?
MR. KNECHT: I'm sorry?
CHAIRMAN WALLIS: Gravity drain or
something?
DR. KRESS: Oh, yes. There is hardly any
velocity at all.
CHAIRMAN WALLIS: And so there is nothing
there that is going to happen. It is the drains that
are shaking.
MR. KNECHT: And that is creating the high
cycle --
CHAIRMAN WALLIS: And are these drains
being tested? Have they been tested at higher
velocities in a testing facility? Is there a separate
effects test? You take each separator and test it?
MR. KNECHT: Not so much from a flow
induced vibration standpoint, but from a performance
standpoint, we have done extensive testing on the
dryers and separators.
CHAIRMAN WALLIS: So if there were any
kind of residences or anything --
MR. KNECHT: Well, we are well within the
range of experience.
CHAIRMAN WALLIS: And you have run them in
a separate effects test at these flow rates?
MR. KNECHT: Yes, with the uprated flow
rates, we have data that supports that.
CHAIRMAN WALLIS: Yes, you have.
MR. KNECHT: Now, I guess one other point
to make here is that we have had at least three plants
that have operated at an extended power uprate for
several years now, and at least two of them.
And we have had some KKM that have
operated up to not quite the 120 level, but they have
been operating much higher than their original design.
And they have shown virtually no evidence that there
is increased cracking because of the uprate.
DR. KRESS: Is their power level
comparable to --
MR. KNECHT: It is slightly higher than
Duane Arnold. Duane Arnold is one of the smaller
units.
CHAIRMAN WALLIS: And do they use the same
kind of separators?
MR. KNECHT: No. Hatch and KKM are very
similar, and KKL is slightly different. And by way of
conclusion, and I think we have gone through most of
this already --
CHAIRMAN WALLIS: The percentage figures
that you are giving there on the Hatch, and KKL, and
KKM, what are those again?
MR. KNECHT: Those are power updates above
the original power level.
CHAIRMAN WALLIS: So they are the new
power updates compared to the old power?
MR. KNECHT: The current uprating power
versus the original power.
DR. SCHROCK: I guess you said these three
are not the same as each other, but it wasn't clear
that you meant the comparison to Duane Arnold.
MR. KNECHT: Well, Hatch and KKM are both
BWR4 units, and have pretty much the same dryer.
DR. SCHROCK: The same dryer? Okay.
MR. KNECHT: KKLs and BWR6s have slightly
different dryers.
DR. SCHROCK: I thought because they were
foreign that they might have a difference other than
that.
MR. KNECHT: No, other foreign plants have
different dryers, but these are similar to Duane
Arnold. Again, the dryer really has an operational
function, and for testing it and repairing the cracks,
and that sort of thing, is really an investment
protection issue.
There is no loss of margin with the
structural integrity basis of the dryer, because the
main steam line break does not change. And we think
we know where to look for flow induced vibration
cracks based on the experience.
Again, Duane Arnold has not seen any, but
we know pretty much where to look. They are visually
inspected at every outage, and so there is kind of a
confirmation there that can be managed. And they are
also repairable.
So we don't see a safety concern with
these dryers, and the integrity of them and the
performance of them is managed by the utilities.
CHAIRMAN WALLIS: And the visual
inspection, this is with some sort of video device?
MR. KNECHT: It can be. Once it is in the
dryer pool, there is usually a camera that is used to
inspect them. But there is no hard requirement on how
that is done. I think that is up to the utilities.
Any more questions?
DR. POWERS: I noticed in your SAR that
you discuss increases in the vibration levels for your
recirc drives, and that you looked at those by
extrapolating some results from start up testing. Can
you explain more about that to get to the kinds of
recirc close that you are going to have at the power
uprate for that test data that are applicable?
MR. KNECHT: Well, the flow rate in the
recirc system increases just slightly to overcome the
pressure drop.
DR. POWERS: I see.
MR. KNECHT: It is not a very large
increase.
DR. POWERS: Okay. I was thinking it was
proportional.
MR. KNECHT: It is about a one percent
change.
DR. POWERS: That explains it.
MR. BROWNING: This is Tony Browning again
from Duane Arnold, and the next presentation that I am
going to co-give with Dan Pappone from GE is on the
ECCS analysis that was done for the extended power
uprate.
Dan is going to get up and talk about the
methodology side of how the analysis was performed,
and then I will get up and talk about the plant
specific results, and the conclusions.
Again, we are trying to demonstrate that
we have got adequate operational and safety margins
from the LOCA perspective at the extended power uprate
conditions.
MR. PAPPONE: Okay. The methodology that
we are using is the SAFER/GESTR methodology, and it is
kind of an intermediate methodology, where we are
taking advantage of the technology development, and
basing the primary analysis on realistic, a fairly
realistic basis, using nominal models and inputs.
But at the time that the methodology was
approved, we still had to live within the original
50.46 in Appendix K requirements. So we do calculate
a licensing basis PCT that uses the required Appendix
K models, and that is the PCT that is used to compare
against the 2200 degree acceptance criteria in 50.46.
We also, because we are doing a nominal
realistic analysis, we also do an upper bound PCT
calculation to demonstrate that this licensing basis
PCT that we calculate is sufficient --
DR. KRESS: What do you mean by upper
bounds?
MR. PAPPONE: Well, we essentially work
through to what we expect would be a true plant PCT
given the modeling uncertainties, and the
conservatisms that are in the SAFER code would account
for those.
We account for the test uncertainties, and
then there is a set of significant input parameters
that would vary at a two sigma level to come up with
an upper bound level, and so we are doing an
uncertainty analysis.
DR. KRESS: So a two sigma level rather
than an upper bound? It is a continuous distribution.
You are picking out the two key parameters that
determine it and see what you get.
MR. PAPPONE: Right.
CHAIRMAN WALLIS: How much does a two
sigma above mean?
MR. PAPPONE: By the time that we factor
in all of the uncertainties and the two sigma part, we
are usually looking at something like 300 to 400
degrees above the normal temperature.
And then also we do have a restriction
that was placed on the methodology itself in the SAR
that approved the methodology, and we have a
restriction on that upper bound PCT. We are not
allowed to let that go higher than 1600 degrees.
DR. KRESS: Well, that is actually built
into your --
MR. PAPPONE: That was a condition on the
SAR that approved the methodology.
DR. KRESS: How did they arrived at that
limit?
MR. PAPPONE: Two pieces; one is the test
data that was submitted at the time, the actual bundle
heat up test data. Those tests only went up to 1600
degrees because they stopped the test at that point to
protect the test bundle.
And the other part is that the upper bound
PCT evaluations that are in the generic LTR, licensing
topical report, were in the 1600 to 1700 degree range.
CHAIRMAN WALLIS: So you might argue that
on the 600 degree margin to maybe 200?
MR. PAPPONE: Well, I don't want to push
that. That is a nice thing to have, but we are also
looking at relaxing this and bringing it before the
staff.
DR. SCHROCK: I have a question concerning
the decay heat evaluation in this method. My
recollection of the SAFER/GESTR methodology was that
you had used the 1979 ANS standard, with a lot of
evaluations for different fuel conditions, different
points in life and so forth.
But you say then that in the end that you
were required to do an Appendix K evaluation, and so
that would mean that you would have to use the decay
power specification there, which was the older draft
ANS standard, 1971-1973.
MR. PAPPONE: That's right.
DR. SCHROCK: In the SAR, it takes about
the may-witt (phonetic) approach, and that is
confusing to me. I mean, what I just described is
either a best estimate approach, which is the '79
standard, or the conservative approach which is in
Appendix K, which is the '73 standard, draft standard.
So how does may-witt (phonetic) get into this at all?
MR. PAPPONE: May-witt is used in the
containment LOCA analyses, and was originally used in
the containment LOCA analysis. It was never used in
the ECCS performance for the clad heat up.
DR. SCHROCK: Yes, you're right. That's
where it is here. So you are using a different --
MR. PAPPONE: What we were using was --
well, the nominal calculation and the upper bound
calculation is the '79 ANS 5.1 standard with that
uncertainty.
And then the licensing calculation, that's
where we pick up the '71-'73 ANS 5.1 standard.
DR. SCHROCK: And this May-witt is not in
LOCA?
MR. PAPPONE: That is not in the ECCS
LOCA.
DR. SCHROCK: In the ECCS considerations?
MR. PAPPONE: Right. That was in the
containment LOCA.
DR. SCHROCK: And it is just a sort of
fact of history that you -- that you had May-witt
plugged in there, and nobody ever changed it. Do you
think it is better for containment analysis?
How can one be better for LOCA and the
other one be better for --
MR. PAPPONE: I don't know the basis for
using May-witt in the original containment analysis,
but the current power uprate containment analyses we
were using in the '79 ANS 5.1 standard with the two
sigma uncertainty on that.
CHAIRMAN WALLIS: And so May-witt has gone
away completely?
MR. PAPPONE: May-witt has gone away
completely. The only time that we would see that is
if we are comparing back to the original calculations.
Say if we are doing a benchmark calculation. I am not
familiar with the statements in the SAR.
DR. SCHROCK: Well, the statement in the
SAR is that the May-witt decay heat model used in the
current licensing basis.
MR. PAPPONE: Now, is that in the
containment section of the SAR?
DR. SCHROCK: Right.
MR. PAPPONE: Yes. Well, Steve or Tony
may know. But I think that is a case where you redid
the containment analysis a couple of years ago, that
is when we would have moved off of May-witt.
MR. BROWNING: Right. Now, the FSA cases
of record are the original containment evaluations
that were done, and they were done with May-witt. So
we were highlighting to the staff that we had
undergone a change in methodology as we went through
EPU.
CHAIRMAN WALLIS: Well, the staff accepts
the new methodology.
MR. BROWNING: Correct.
CHAIRMAN WALLIS: Well, is there a problem
with this?
MR. PAPPONE: Or is it just a historical
notation in the SAR.
DR. SCHROCK: Well, I was trying to
understand why there would be any use made of May-witt
at this point in time.
CHAIRMAN WALLIS: Well, there isn't. It's
gone.
MR. PAPPONE: It's gone.
CHAIRMAN WALLIS: And so we can forget it.
DR. SCHROCK: It says in the SAR that it
is the current licensing basis.
MR. PAPPONE: And so continuing. It's
Tony's turn.
MR. BROWNING: And on to the plant
specific analysis and results. The analysis was done
for the Duane Arnold specific ECCS configuration, and
what was unique for BWR4 was the fact that we had LPCI
logic, and so we have to look at that in a single
failure evaluation space because we have a
vulnerability there that some of the other designs
don't have.
And which is the failure of the LPCI
inject value to open, which completely starves the
vessel for LPCI flow. So that factors into the single
failure evaluation that is unique to us.
And then we do the full break spectrum
evaluation to confirm that the design basis accident
is the double-ended guillotine break of the suction
line is the worst case, and that we validate that the
large breaks do dominate over the small breaks.
So we look at the small break spectrum as well.
And for the plant specific results, the
licensing basis PCT that we talked about and that we
do the conformance to 50.46, we came up with a
calculation of a bounding value of 1510. So we have
a great deal of margin with the regulatory limit.
CHAIRMAN WALLIS: LB means licensing
basis?
MR. BROWNING: Yes, PCT, and then the
upper-bound PCT.
CHAIRMAN WALLIS: So which is lower bound
and upper bound, and its licensing phase is an upper
bound?
MR. BROWNING: Yes. The jargon. So the
upper-bound PCT is only 1350, which is well below the
1600 limit, and we also see that the upper bound is
below the licensing basis. So we meet both
requirements.
CHAIRMAN WALLIS: This is much like what
you have pre-EPU is it?
MR. PAPPONE: There was only about a 10
degree change in the licensing PCT DBA.
MR. BROWNING: Right. So as you see here,
there is an across the break spectrum of break sizes,
from the small break, all the way up to the DBA case.
You can see the change due to the EPUs,
and the little squares are the pre-EPU cases, and then
the triangles are the EPU cases. So you see the trend
follows, and then when you get to the DBA case, they
are very close. They are within 10 degrees of each
other.
And then you can see where the upper bound
at the DBA case shows up.
DR. POWERS: In fact, doesn't your EPU
temperature, peak clad temperature, go down?
MR. BROWNING: Yes, slightly.
DR. POWERS: And that is because of the
flattening out of the core --
MR. BROWNING: Yes, the same phenomena
that we saw earlier. The peak bundle has a little
more flow because we --
DR. POWERS: I looked at that, and I said
to myself that this has got to be red. I have got to
see this.
MR. BROWNING: I think you will see the
words in the staff safety action are counter-
intuitive.
CHAIRMAN WALLIS: The limit on the right
hand, the three square feet -- what is the limit?
MR. BROWNING: It is 2-1/2 square feet,
yes.
CHAIRMAN WALLIS: And which pipe is it?
MR. BROWNING: That is the recirc suction
line. That is the largest pipe that we have on the
vessel. So, you can see -- well, the trend stays the
same, and the results go up a little bit.
MR. ROSEN: And the solid lines are done
with the Appendix K models. I just noticed that in
the cartoon that you showed before.
MR. BROWNING: Yes.
MR. ROSEN: And that shows that that
number was about 2000 degrees.
MR. BROWNING: Oh, that was just a
representative cartoon. Those were not the plant
specific results. That was just to get across the
jargon.
MR. PAPPONE: That was to show which limit
went with -- or which temperature calculation went
with what limit, and the relative relationships.
MR. BROWNING: Right.
MR. ROSEN: The one on the left looks okay
relative to the numbers on the right.
MR. PAPPONE: Well, the one on the right
is okay, too, because it is the one that is compared
to the 2200.
MR. BROWNING: But it is not the Duane
Arnold result.
MR. ROSEN: So it is not the number, your
number?
MR. BROWNING: It is not our number, no.
And as Dan has explained, the methodologies is where
we try to build in the margin, especially using the
upper bound technique that account for all the
uncertainties, and the licensing basis PCT still has
to apply the conservative Appendix K models for the
regulatory conformance.
And then the acceptance criteria are
conservative as well. So for the plant specific
results, we saw obviously no impact on safety margin
because we had a great deal of margin to 2200.
And then the operating margin is obviously
maintained by that same operating condition.
DR. KRESS: What would you do if those
numbers went all the way up to the 2200 on your
Appendix K?
MR. PAPPONE: On the licensing PCT?
DR. KRESS: Yes.
MR. PAPPONE: That's fine. We have got
plans that are licensed near 2200, and we have --
DR. POWERS: And there is definitely one
at 2183.
DR. KRESS: Yes, that's what I thought.
MR. PAPPONE: We do have a couple of the
early plants that are PCT restricted after 2200.
DR. KRESS: Well, this is in terms of
margin. If you are at 2200, you still have sufficient
margin.
MR. ROSEN: This goes to the question of
whose margin is it.
MR. BROWNING: The 2200 up to the field
cladding failure point, that is the licensing margin
and that is the sacred turf. What we are talking
about down here is the margin to 2200 and this is the
operating latitude. And as long as we maneuver within
here --
MR. ROSEN: I would propose the standard
that if you would license up to the 2200, then the
margin is the licensee and the vendors. And the
answer to the question is whose margin is it. It has
been licensed up to nearly the 2200. So I think that
is QED.
MR. BROWNING: Right. And now we are on
to your favorite topic.
DR. POWERS: We are about to move on to a
topic that I know will go quickly because PRA invokes
a little interest in this committee. I wondered if
the members wanted to take a 10 minute break in order
to build up their strength to get through this.
DR. KRESS: No, let's go on.
DR. POWERS: Apparently they want to
charge ahead. Any acquisitions that I am a slave
driver will not be tolerated. Okay. So, Brad is
going to come up here, and he looks like a brave,
strong young man. He has taken a few slings and
arrows in a checkered career here, huh?
MR. HOPKINS: My name is Brad Hopkins, and
I am a PRA engineer at Duane Arnold. The purpose of
the PRA evaluation for a power uprate was to identify
possible vulnerabilities resulting from power updates.
These may come from potential sources,
such as changes in system criteria possibly, or maybe
from changes in human error probability. I would like
to note at this time that a power uprate is not a risk
informed application.
But nonetheless we are interested in the
question of risk. That is, does power uprate
constitute undo risk in some way or form.
DR. KRESS: How do you identify
vulnerability?
MR. HOPKINS: Well, we will look at, or
what do we use as a criteria for vulnerability, go to
the next slide.
DR. KRESS: I didn't want to say that word
because I get criticized every time I use it.
DR. POWERS: As well you should.
MR. HOPKINS: We will answer that
question. My second bullet here is we have a
guideline that tells us how much of an increase in
core damage frequency or large/early release frequency
constitutes a significant increase.
We used or we compared our results to the
EPRI PRA applications guide. We also -- and I think
the NRC has been using Reg Guide 1.174, and we
compared to that also to make sure that we meet that.
Now, the areas that we looked at are
equipment, reliability, and we look at initiating
event frequencies, and we looked at system success
criteria, such as how many pumps do we need to operate
to have adequate core coverage, or how many SRVs do we
need to open to adequately depressurize.
And finally we looked at human error
probabilities. Now, we didn't have anything too
interesting in the first three bullets.
DR. KRESS: How do you actually look at
the effect of power updates on equipment reliability?
MR. HOPKINS: How do we look at the
effects on equipment reliability?
DR. KRESS: Yes.
MR. HOPKINS: There is -- well, I guess
there is not a hard and fast methodology that we could
find if you take a good look at it, but we tried to
assess what equipment might be seeing higher duty,
such as the feed water pumps.
And we recognize that some equipment does
have higher duty, and failure rates may be higher.
But I think with the maintenance rule in effect now,
we have good programs for monitoring the effectiveness
of our safety related equipment. So we don't really
anticipate --
DR. KRESS: So in your PRA, you just used
the same failure rates for the equipment?
MR. HOPKINS: For this assessment, we
wound up inserting the same failure rates for the
equipment.
DR. KRESS: But you did review because you
went back to see if you thought there was any reason
to change those?
MR. HOPKINS: Yes.
MR. ROSEN: You are talking just about
reliability, and are you also talking about
unavailability as well?
MR. HOPKINS: Well, unavailability as
well.
MR. ROSEN: The slide just says
reliability.
MR. HOPKINS: We identified all basic
events that had a raw value of a certain value, and we
focused in on those pieces of equipment, the equipment
that we felt was significant.
And we asked ourselves is there any reason
that we should increase the failure rate of this
equipment, and I think in all cases that we said no.
So I am going to focus later on in the
presentation on focusing more on the human error
probabilities, since those are the ones that have the
most impact.
Here is a summary of our results, and I
have a column for the base value, or our present PRA
numbers, and a value for extended power uprate, and in
the right-hand column --
CHAIRMAN WALLIS: The PRA predictions are
valid to three significant --
DR. POWERS: At least. That's always.
MR. HOPKINS: We will take a quick look at
the question of uncertainty on the last slide here.
But, no, I don't have uncertainty drawn up here.
But the computer calculates it out, of
course, to --
CHAIRMAN WALLIS: And you are arguing is
the change is what you are looking at, and not
something that you have a better handle on than the
absolute value?
MR. HOPKINS: Right. Here it is the
change that we are interested in.
DR. KRESS: If I look at your base case
CDF and LERF, I get an early conditional failure
probability of .05 and thereabouts just in my head.
For Mark-1s, I am used to .5s and .4s for that. Do
you guys have that good of a containment? It's a
Mark-1 isn't it?
MR. HOPKINS: So you are comparing the
level one to the level two?
DR. KRESS: Yes, as .05 is a pretty good
number, and for Mark-1s, I am used to an order of a
magnitude higher than that in PRAs.
MR. HOPKINS: Yes, and so the level two is
lower than we might expect.
DR. KRESS: Yes.
MR. HOPKINS: I guess I don't have a real
good answer for that.
DR. KRESS: I guess the question would be
why is your particular plant looking so much better
than other Mark-1s?
DR. POWERS: I will make a guess.
DR. KRESS: Okay.
DR. POWERS: A drywall spray.
DR. KRESS: They have a drywall spray.
DR. POWERS: They have it and they are
using it.
DR. KRESS: That certainly could make a
difference.
DR. POWERS: Because the reason that you
get the high failures on the Mark-1s is either a melt
flow across the floor without water, or an overheat at
the seals up at the top. And the spray takes care of
both of those.
DR. KRESS: That is probably a good
explanation, Dana.
DR. POWERS: That's my guess.
MR. HOPKINS: That sounds very good to me.
In the future, I think utilities are seeing some value
in providing PRA results to the public, and making it
publicly available. I think we will see that trend in
the future.
And on the human error probabilities, we
reviewed all human error probabilities with a raw
value of 1.06 or greater, and then we employed a map,
a thermal hydraulic code, to determine whether the --
MR. ROSEN: How did you select 1.06? It
seems so timid.
MR. HOPKINS: Okay. Well, 1.06 --
MR. ROSEN: I would have thought that you
would pick a number like two at least.
MR. HOPKINS: Well, 1.06 corresponds to an
increase in core damage frequency of 1 times 10 to the
minus 6. So any increase at this event, if an event
would cause the core damage frequency to increase by
1 times 10 to the minus 6 or more, then we would
evaluate it. And there were about 20 or
so --
DR. POWERS: I can't help but point out to
the members that this is what we have been asking the
staff to do for the human performance program plan for
a long time.
So these particular evaluations ought to
be very interesting to us; and the question you ask is
are these humans doing as well as we would like them
to do here, and here you have a basis for looking at
this.
DR. KRESS: Why do you call it a MAAP
thermal-hydraulic code? I wouldn't have characterized
it that way.
MR. HOPKINS: As opposed to a probablistic
--
DR. KRESS: I would have characterized it
as a severe accident code, but a relatively poor
thermal hydraulic code.
DR. POWERS: That is not how you
characterize it in private.
MR. HOPKINS: Well, we could call it a
transport code. We will call it a transport code, a
radio nuclide transport code.
CHAIRMAN WALLIS: He is calling it thermal
hydraulic to try to give it respectability.
MR. HOPKINS: We recognize that it has
limitations. Next slide, please.
DR. POWERS: But in fairness wouldn't it
be pretty adequate for this?
DR. KRESS: Yes, I think that would be
perfectly adequate for this. For a BWR, it is
actually pretty good for this sort of stuff.
DR. POWERS: And all it is worried about
is heat and mass here.
DR. KRESS: Yes, this should do fine for
that. I didn't mean to put it down.
DR. POWERS: What do you mean you didn't
mean to put it down.
CHAIRMAN WALLIS: That is a good first
approximation to thermal hydraulics. There is no
energy. It is MAAP.
MR. HOPKINS: I think we maintain a
questioning attitude when we use MAAP, and we try to
compare it with more detailed codes when we can, or
when that is possible.
Now, I would like to go through the five
most important operator actions that we found, and it
is not my point to dwell in great detail on each of
these.
But more to give you a sense of what is
causing the most increase in the core damage
frequency. Most of the increase came from ATWS
events. So four of these operator events apply to
various ATWS scenarios.
So the first one is failure to initiate
standby liquid control. So this is applicable to ATWS
events where the main condenser is not available.
Therefore, all of the energy is going down into the
suppression pool.
DR. KRESS: The spray is not available to
them either? Is the suppression pool spray not
available?
MR. HOPKINS: Well, in many cases, yes, I
think the sprays are available.
DR. KRESS: Yes, there is too much heat
going in there.
DR. POWERS: There is too much heat going
into the containment.
MR. HOPKINS: Now, we look at two
different time frames for injecting standby liquid
control. If we are able to inject early, then later
on in the event we only need one RHR service water
train, and one RHR train to remove the decay heat from
the water.
If we are not able to inject early, we
still have an opportunity to inject standby liquid
control a little bit later. But if we inject later,
then we need to use both trains of RHR service water
and RHR for adequate core cooling.
CHAIRMAN WALLIS: What is the formula that
relates to that? Is there a magic correlation that
says that when you go from 6 to 4 that --
DR. KRESS: It is an EPRI correlation.
CHAIRMAN WALLIS: All right. So that is
based on experience?
MR. HOPKINS: It is an expert opinion is
what it is.
CHAIRMAN WALLIS: Oh, so it is based on
data.
MR. HOPKINS: We used a variety of
methods. That is not my area of expertise. So I am
not able to address it in really good detail.
MR. ROSEN: But fundamentally those
techniques take into account the fact that operating
under stress when you have less time, you have a
higher likelihood of failure?
MR. HOPKINS: That's right. That's right.
But in this case, like Steve was saying earlier, our
operators are well practiced in injecting standby
liquid control. We cover it often in the training.
MR. ROSEN: Would you go back to the prior
slide for a minute. Now, you see, that is the point
that I made earlier, that for early initiation, with
the time reduced from 6 to 4 minutes, but the
deterministic analysis assumes 2 minutes.
MR. HOPKINS: It seems like we are overly
favorable on the deterministic analysis.
MR. ROSEN: Well, you are overly
pessimistic here. But they are not the same, and I
think I understand why. One is a best estimate, which
is this one; and the other one is a conservative, or
is an analysis for deterministic purposes.
MR. HOPKINS: Right.
MR. ROSEN: I wish they were the same
somehow, but I am having trouble reconciling two
different estimates.
MR. HOPKINS: I think we are looking at
two different outcomes possibly.
MR. ROSEN: But we also know in this case
-- and Steve -- I am having trouble with your last
name.
MR. KOTTENSTETTE: Kottenstette.
MR. ROSEN: Kottenstette. He told us that
the four minutes and the two minutes are both
achievable times because everything the operator needs
to do is in front of him in the control room;
information and the mode switch and key.
So it is irrelevant whether it is four or
two minutes. The point is that the operators can take
those actions, and it is in their training program,
and it is in the simulator.
It is a critical task, the training
program, and they can take it in either case within
the four or two minutes. All right. Go on.
MR. HOPKINS: All right. This one is
failure to inhibit ADS. Now, for an ATWS, for most
ATWS scenarios, we want to prevent automatic
depressurization from occurring.
The reason for this is that if you
depressurize, then the low pressure emergency core
cooling systems initiate automatically, and they dump
a lot of water into the vessel.
And we have a concern of a reactivity
excursion when that happens. So we really need to
inhibit ADS.
CHAIRMAN WALLIS: So your ECCS system is
not borated?
MR. HOPKINS: That's correct.
MR. ROSEN: Well, it is starting to borate
it, but very slowly.
MR. HOPKINS: Correct. So here the
available time is reduced from 16 to 10 minutes, and
we have a corresponding increase in the failure
probability for that event.
And failure to reduce power via the
lowering of reactor vessel water level. Another means
of getting our power level down is to lower the water
level down to below the level of the feed water
injection spargers.
By doing this we avoid the need to
depressurize the vessel by keeping the suppression
pool temperature below its heat capacity temperature
limit. The available time is reduced from 15 minutes
to 12 minutes, and we have a corresponding increase in
the failure probability.
DR. KRESS: In your failure to initiate
standby liquid control, you have 14 minutes for late
initiation, the failure probability was about .09, and
on this one you have got 10 minutes for the ADS, and
it is .03 apparently.
How come the failure probability is lower
for a 10 minute than it is for a 14 minute action?
Has it got something to do with the type of complexity
of the action or something?
MR. HOPKINS: Right. We would be
factoring in the complexity of the action.
DR. KRESS: And that is built into the
model somehow?
MR. HOPKINS: Yes. And here we are really
combining two of the previous operator actions for a
little different scenario here. This one is an ATWS,
where the turbine bypass valves are available.
Now, the turbine bypass valves can pass
about 24 percent of reactor power. However, the power
is about 45 to 48 percent. Therefore, we still have
a significant amount of energy going down into the
torus water level.
In this scenario the operator is not able
to get the power level down, either by lowering the
water level, or by injecting standby liquid control.
So we increased the failure rates for both
of these by the same amount as what we saw previously.
DR. KRESS: You uncover the core when you
lower that water level?
MR. HOPKINS: Do we uncover the core?
Yes, I believe the EOPs have us go down to minus --
about minus 30 inches.
MR. POST: This is Jason Post. That is
the collapsed level. There is still a two phase level
swell that is well above the top of the active fuel.
MR. HOPKINS: Thank you, Jason. The last
one -- okay. Now we have looked at all of the ATWS
events. This one is applicable -- this one is a
failure to depressurize the reactor vessel, and it
applies to transients, small LOCAs, and medium LOCA
events.
So if your high pressure systems are not
able to inject, it is very important for the operator
to manually depressurize the vessel so that the low
pressure systems can turn on. So this is a fairly
significant operator action in our PRA.
So I guess for transients and small LOCAs
the available time is reduced from 65 minutes to 55
minutes, and so these probabilities are pretty low
compared to the other ones.
I hope that the operator recognizes that
he is not -- that he doesn't have any water going in
the vessel. I think it is something that is pretty
easy to see, and the action is easy. He should be
able to do it in an hour.
We looked at external events, and here we
are looking at things like high winds, floods,
tornadoes, transportation, chemical hazards, and we
didn't see any effect of power uprate on those events.
However, for fire and seismic, those were
the only external events in which we felt that there
was a measurable effect. Now, for here, we carried
the operator actions through our fault trees for fire
and seismic PRA.
And we found less than a one percent
increase here, and so we didn't find anything too
interesting in external events. No additional unique
hazards were identified.
For shutdown risk, here power uprate is
judged to have a negligible effect on our overall
ability to adequately manage shutdown risk. And since
about 1992, we have employed EPRI's Sentinel model for
monitoring risk during refuel outages.
So here we look at both the defense and
depth in meeting various safety functions, and we are
calculating probability of boiling in the core region.
So we think that we have had a very good
handle on shut down risk. We are experienced with it
by this time, and we think that with a power uprate
that experience will continue.
DR. POWERS: I guess I don't understand
why when you think about it that if you have a power
uprate of 20 percent that you must have roughly a 20
percent increase in decay heat load.
And so your time to boiling must be
roughly 20 percent shorter than it was before. So the
time that you have available to recover from some loss
of cooling capacity must be about 20 percent shorter.
MR. HOPKINS: Yes.
DR. POWERS: So shouldn't that mean that
you have roughly a 20 percent increase in risk being
shut down?
MR. HOPKINS: That's correct.
DR. POWERS: And a window of shutdown, and
I don't mean all of it. But in a window of shutdown
where boiling risk is reasonably high.
MR. HOPKINS: You are correct. The decay
heat values are higher. And we track very carefully
the number of systems that we have available for
removing decay heat, and at any given time during the
outage we would know exactly how many systems we have
to have operating to meet that load.
But in general there is only a few periods
of the outage where the times are very short. That
would be the transition periods when you are cooling
down the vessel, and when the water level in the
vessel is at its normal level.
But for most of the outage the reactor
cavity is flooded all the way to the top to allow for
fuel moving. And therefore the times -- we have on
the order of hours, and sometimes 24 hours for later
periods in the outage for responding to events,
whether it is loss of decay heat removal, or
inadvertent drain down events.
DR. KRESS: Has your PRA been subjected to
the industry peer review process?
MR. HOPKINS: Our PRA went through the
industry certification process four years ago.
MR. ROSEN: It was one of the first, I
think.
MR. HOPKINS: We were one of the first.
We had a very favorable certification, and I think one
of our real strengths is our documentation out there.
We have a living PRA program that was developed within
a qualitative framework.
MR. ROSEN: Now, I thought you were going
to say in response to Dana's question is that you do
get more decay heat as he points out, but
that you end up not getting to shut down temperatures
as quickly as you would now.
So that ultimately the way that you
control shutdown risk is to basically wait a little
longer before you could initiate shutdown operations.
MR. HOPKINS: Right.
MR. MCGEE: But it ends up being an
operational impact where we need to keep the shutdown
for a longer period of time before going into other
phases of an outage.
DR. POWERS: It seems to me that becomes
a time period that bean counters will attack, and the
pressure to shorten that.
MR. HOPKINS: Well, there will be pressure
to shorten that, but the bean counters are already in
this case very happy, because they have been running
at 20 percent.
DR. POWERS: They are only happy quarter
by quarter, and the next quarter, they are going to
want another 20 percent.
MR. ROSEN: But the plant staff should
point out to them that while it is true that it is
going to take a few more hours to get into shutdown
operations, they should be thinking about all the
money they have made while the plant ran at the
extended power uprate.
MR. HOPKINS: Well, that vessel is still
pretty hot when the mechanics are unbolting the head
bolts. They will be doing a dance. Now, uncertainly.
In our original IPE submittal, we addressed
uncertainty with a sensitivity analysis. That is to
say that we don't have a formal rigid uncertainty
analysis for our PRA.
For the present study, we selected
operator actions that were sensitive in the first
place. That is, the first step of this study was to
look at those parameters that are sensitive.
One other thing we did was we looked at
all of the low worth operator actions, and we doubled
their failure rates all at once. We ran a single case
with all of those values doubled.
DR. KRESS: This South Texas guy here is
going to ask you why you didn't increase those by a
factor of 10. That's what they did for their effect
of QA on the reliability of low worth components that
are not safety significant.
MR. HOPKINS: But not for a power uprate.
We would be talking about an exemption request.
DR. KRESS: Yes, you see, an exemption
request.
DR. POWERS: As long as you are harassing
the South Texas guy, I will harass him some more.
Wait as long as you want to. The decay heat load that
you have to deal with is still higher by 20 percent,
and it still shortens down all the times that you have
to boiling by 20 percent.
DR. KRESS: So that doubling didn't have
any significant effect.
MR. HOPKINS: The doubling did not.
CHAIRMAN WALLIS: Are we back to the
beginning?
MR. HOPKINS: We are not, and that is the
last slide.
DR. POWERS: Are there other questions
that people would like to ask about the PRA? Not
seeing any and not looking very hard for any, Ron, did
you have any closing comments that you needed to make?
MR. MCGEE: I just wanted to thank the
committee today for allowing us this time to present,
and by my count we have three things that we need to
follow up on.
And they are Mr. Wallis' question
concerning the stress analysis, and Mr. Powers' had a
question and I believe we will be able to follow up
with something on that.
But then also we have the post-LOCA H202
monitoring question that I think we will be able to
address tomorrow during the staff's presentations.
Other than that, are there any other questions for me
at this time? If not, thank you.
DR. POWERS: If there are no further
questions, I will turn the meeting back over to the
Thermal Hydraulics Subcommittee Chairman.
CHAIRMAN WALLIS: Dr. Powers has done his
usual and hasn't kept us late. So I am very happy to
recess exactly on time at five o'clock, and we will
reconvene at 8:30 tomorrow morning. Thank you very
much.
(Whereupon, the opening meeting was
adjourned at 5:00 p.m, to convene at 8:30 a.m. on
Wednesday, September 27, 2001.)
Page Last Reviewed/Updated Tuesday, August 16, 2016